Joe Ballantyne Fusor v1.

[Richard Hull]

Joe, I have experience with that very metering system and it is why I pleaded with you to investigate. Simple is good and bad. (There is that @#^%*! double edged sword again). Most folks do not realize that 100% of all TC gauge systems have near zero components in them. I am going to add to the tube chart FAQ the simplified facts about manufactured TC gauge systems. I now supply that full expanded explanation at the URL below to go with the chart of TC gauge tubes.

viewtopic.php?p=94827#p94827

Richard Hull

[JoeBallantyne]

So I had a post I wrote last night - an update on yesterday's efforts - and it took me a couple of hours to write. I did a cut and paste to a backup notepad .txt file halfway through, so I wouldn't lose it. But forgot do do another backup of the post before I clicked submit. And because of either my TP-Link router, or Comcast, the post was lost, because for some reason I keep getting disconnected from sites repeatedly. So when I clicked submit, fusor.net kicked me to the "you must be logged in to post page" and AFAIK dumped my post completlely on the floor. Sometimes I can't get connected to mainline sites either because of DNS issues. Comast + TPLink in the Redmond WA area is certainly not a good combination.

I was so angry about losing an hour+ of work, that I punted. Still haven't summoned the will to rewrite the damn thing.

Coalman, does the server keep http POST requests that it rejects around anywhere. If so, and you can dig up a post submission from about 1am PST last night, I would certainly love to get my work back.

It is really pathetic that neither chrome nor edge nor any other web browser I know of, doesn't make a browsable backup of every web form you click submit on. Or are even just working on. It would save users so much frustration and anger, and wasted hours of life. The browser had all the data, but there is no easy way to get it back. Unfortunately I closed the tab, so there is now really no way of getting it back. I mean browsers have been around for basically 30 years now, and they STILL don't save work you are doing on forms on websites so that you can recover your entries if something goes wrong. LAME!

I guess now I have to be hard core about always writing every post in Word, so I don't risk losing it due to typing it directly in the site. Pretty sad how technology still SUCKS in so many ways.

Joe.

[Richard Hull]

Joe you are not alone. I tend to compose in word and save frequently. Over my 23 years here I have lost many posts by composing in this editor at fusor.net. If I choose to not go to word, some here might have noted that I hit submit before I am done and leave a little note..."still working on this".
I just keep submitting until I am done. There is no real fix other than composing in Word and saving as you go and then transfer the finished text.

If I write a "teaching paper", of some length, I do it in word and save it as a PDF. I then do a 3 line intro in a post in the forums to tell what the attached PDF is all about.

I hope you finish your build on your post and share your thoughts with us.

Richard Hull

[Paul_Schatzkin]

Cue Billy Crystal: "I hate it when that happens!"

I couldn't quite follow you off the cliff their Joe. It sounds like you had the right idea when you started copying to a .txt file. What I don't quite follow is: once the post got past a few sentences and you were pasting into a different window (the .txt file) why didn't you just stay in that window until you finished the post. Does Winders Text Editor auto save or is it still you-have-to-remember-to-save? (all the numerous text-editing applications I run on a Mac auto save by default).

Anyway, my usual MO is, once I get into the second or third paragraph I jump to a text editor, finish there, then paste back into the post window. And if it's a short one like this, I'll select/copy to the buffer before I hit 'submit.'

Sorry for the lost effort. Damn, Billy Crystal again.

--P

[JoeBallantyne]

So I'm going to resurrect this thread and continue posting about my efforts. (Yeah, I REALLY hate losing work to stupid computer glitches. 5+ months is a long time to cool off.) I have a half written post about where I was, which I will complete and then I will make a few posts about different areas of effort, as the day by day, blow by blow account is no longer possible as too much time has gone by, and although I can remember the highlights and the take aways from the work, I cannot recall the exact timelines.

I have a couple of large (1x24 inch) He3 tubes I bought from Richard Hull at HEAS a few years back, and have been looking into building a moderator for them. I found what I think is really quite a good deal on HDPE on Amazon. It is $129 plus whatever sales tax is in your jurisdiction for 3 25x450x600mm slabs of HDPE sold as cutting boards. That works out to approximately 1x17.7x23.6 inch slabs, or rounded off, 1x18x24 which of course is what they are marketed as being. The metric dimensions are actual however, so you don't get the full 18x24 inches. This was MUCH cheaper than the equivalent from TAPPlastics which wanted $284 plus $78 shipping for 3 1x18x24 inch HDPE cutting boards.

Of course I'm assuming if you use Amazon you have Prime, and so the shipping is free. (It was free for me here in Redmond, WA. YMMV. The item is NOT a Prime item, but shipping was free anyway.) There are currently 19 of these left in stock.

I figure I can make 2 5x5x24 inch moderators for my He3 tubes with that plastic. It is about 50lbs of HDPE. The HDPE is natural - not dyed - and they are REALLY nice cutting boards... but pretty soon for me they will be in pieces.

Here is a link: https://www.amazon.com/Hakka-3-Board-Wh ... ast_sto_dp

Joe.

[Richard Hull]

I lost a lot of of file material I put up in the last terrible undesired backup purge this year. I found my original super moderator for the 1X24 tube pdf file missing in one of my FAQs for those interested here is my precise build.

Richard Hull

[JoeBallantyne]

Richard, thanks for the .pdf describing your moderator setup.

My plan is similar, but will have 4 1x5x24 inch pieces, and 2 1x2x24 inch pieces. That should use up one of the 18x24 slabs completely, and use a little more than 7 inches of the 18 inch width of the second slab.

So looking at the 5x5 inch end straight on, the He3 tube will go in the center where there is no X.

XXXXX (1x5)
XXXXX (1x5)
XX XX (1x2) (1x1 hole) (1x2)
XXXXX (1x5)
XXXXX (1x5)

That means there is a minimum distance of 2 inches of moderator which is 5.08cm A little less than the optimal 5.5cm for 2.5MeV neutrons, but should be OK. We shall see. I am going to build it and try it out. I will actually probably make each piece slightly bigger than 5 inches so that I get 3 equal sized large pieces ~5.1 inches and 1 smaller piece of ~2.05 inches from each of the first 2 18x24 slabs. The exact widths will depend on the kerf of the saw blade I use to cut the material. But the goal is to make 3 24 inch long cuts and end up with 4 pieces that are all exactly the right width.

Joe.

[Richard Hull]

I am sure your design will perform flawlessly. Anywhere from 2.4 to 3-inches of total containment all around with HDPE will do a fine job.

Richard Hull

[JoeBallantyne]

So here follows the first half of the post that I lost back on the early morning (probably about 3AM or so) of March 10th. This is the part that was backed up in a notepad doc.

Today, Wednesday 3/9/2022, I got my deuterium system connected to the fusor, in a vacuum tight configuration. I of course immediately tried to see if I could get any neutrons out, but it was not to be, yet.

To make the deuterium delivery system vacuum tight, and get it done today, I ordered a couple of unions from Grainger online, drove over to Seattle, forked over the $50+ for 14 tiny pieces of 316 stainless steel (7 pieces per union). Swagelok uses a nut, and a double ferrule to clamp down on the circumference of the tubing as the nut is tighted on the male thread/seat, and make a high pressure seal which is normally stainless steel on stainless steel, and can support pressures into the thousands of PSI. Lots of manufacturers now make Swagelok compatible parts. The unions I got were A-LOK made by Parker, but they worked just fine with the Swagelok male connectors I had on my equipment.

I needed to be able easily move and form the tubing, and I only need this to be vacuum tight, so the highest pressure it will see is 14PSI or so. (1 atm) I therefore chose to use 1/4 copper tubing, acquired from HD (Home Depot) in a coil of 20 ft. The copper tubing was very malleable, so it was easy to unbend it from the coil, and make some big loops to get from the deuterium cylinder to the top of the KF16 cross that currently has the 2 TC tubes attached. I moved the TC tubes to the side, so that I could run the tubing to the top, since that minimized the amount of bending required, and also puts the least stress on the connection, as the weight of the tubing is coming straight down instead of hanging off to one side.

I had to cut the end off the tube as it came from HD, because it was dented, and I couldn't get the nut, or the ferrules over the end. The dent made it out of round, and there was no way the nut or ferrules would go on. So I cut an inch off the end using a copper pipe cutter (a must, because the tubing must stay perfectly circular). This is one of those things where you have a circular cutting wheel that you clamp down gently on the location you want to cut, then spin it once around the tube, tighten it a little, spin around the tube again, rinse and repeat until the tube or copper pipe is cut. You can't hope to cut the tubing with wire cutters or a hack saw, and think you will get the swagelock parts over the end. You won't.

After cutting the tubing, you put the nut, then the back ferrule, then the front ferrule on the tubing, insert the tube all the way into the male seat, slide the nut and ferrules up to the connector, and first hand tighten the nut, and then tighten it with a crescent wrench. Because I was using copper tubing with stainless (not ideal, it is better to use brass, but I just got stainless because that is typically these parts come in), and because I only needed this to be vacuum tight, I was VERY gentle when tightening with the wrench. Especially because I want to be able to reuse the connection if possible, and if I crunch it all the way on the first use, I won't be able to do that. I probably tightened the nut to 2-4 ft lbs. Maybe less. Again I just need vacuum tight not 5000 PSI. Plus I'm tightening into copper which is SOFT.

Once I got the Swagelok connectors attached, I gently worked the KF25 connector that was on the other end of the copper tube so it was sitting nice and square on the top of the KF16 cross + KF25 adapter, without exerting any sideways forces on the KF25 connector. This was easy to do because the copper tubing is soft. I then clamped the KF25 in place, and opened the butterfly valve on the pump, and the chamber and tubing pumped right down to ~35 microns. Just to check if things were tight, I closed the valve to the pump, and watched the TC gauges to see how fast the chamber would rise in pressure. Normally it rises about a micron or two a minute. It was a little faster than that, so I put a little bit more torque on each of the Swagelok connections. Repumped the chamber down, closed off the pump, and this time the vacuum stayed put at 35 microns like it normally does.

With the copper tubing supply line under vacuum, I also opened up the valve on the low side of the regulator, to pump out everything all the way up to the regulator. This regulator has a gauge that shows the vacuum level on that side of the gauge, and it was off course pegged at -30 inHG. The gauge is interesting in that it is graduated in PSI for pressures above 0, and in inches of mercury for vacuum. The regulator appears to also regulate into vacuum which is very nice.

I had previously fully closed (turned it CC until it was clearly not regulating and the handle would turn easily) the regulator so that it should not allow any gas through. Then I very carefully cracked open the valve on the cylinder just a bit - it actually took more turns than I thought it would to start to open up - and there was a little hiss, and the gauge on the high pressure side ramped up to 500psi, and held steady. I immediately shut the valve on the cylinder off again fully, since I have read on fusor.net from multiple folks, that just that charge on the regulator is enough to run things for a while. I realize now, that BEFORE doing that I should have also partially closed the regulator while the low side was under vacuum to also evacuate the other side of the regulator all the way to the valve of the tank. Then reclose the regulator, and crack open the main valve on the tank.
Unfortunately I did not do that, which means my charge of deuterium is on the high pressure side of the regulator is actually mixed with air. But its not that bad, because 1 atm is 14 psi, and I added 500 psi, so (500/500+14) = 97.27% of the gas on the high side is D2. Not too bad. But it would have been better if I had evacuated that side too.

At this point the chamber was being pumped, and was at ~30 microns as was the gas feed tube. I closed the low side regulator valve, tightened the regulator down until it just started to admit deuterium into the low side of the regulator, it got up to about -15 inHG, and stopped, I figured that was likely good enough. I then cracked open the low side regulator valve and watched the TC gauges on the chamber. They rose steadily all the way up to ATM - they don't have any resolution up there, it was probably more like what the regulator pressure was showing which was -15inHG or about 1/2 ATM. At this point I had basically filled the evacuated chamber with D2. I shut off the regulator ou

And there you have it - yes the backed up text in the notepad .txt file did cut off mid word for some reason.

Joe.

[JoeBallantyne]

Everything that follows is now what I can recall 5+ months later of what I did. I will do my best to recall things accurately, but it is not the same as when documented the night following the work. Some things I do have notes on, but much I do not, of necessity there will be significantly less detail.

What I did that evening was to fill the fusor with D2 gas up to the max pressure allowed by the regulator. Then closed off the D2 regulator output valve and opened up the vacuum butterfly valve to pump out the fusor. I repeated this probably 3 times or so. The intent being to flush the fusor chamber with D2 repeatedly, so that on the last fill, the D2 concentration would be maximized. After closing of the D2 valve, I brought the voltage up as high as I could to light off a plasma, and try to make neutrons. As I recall, I wasn't able to get the voltage up to the max of the power supply because the pressure in the chamber was too high - the lowest I could get it on my TC vacuum gauges at the time was what looked like about 35 microns or so on the gauge. If I raised the voltage too high, the current would spike, and the fuse on the input to the high voltage supply would blow.

So the next task was to improve my vacuum system so that I could get a lower pressure in the chamber so I could raise the voltage higher than 15kV or so.

Joe.

[Dennis P Brown]

The first step for any fusor is to get the chamber below 10^-3 and at least mid 10^-4 torr; of course, lower is better. Flushing a fusor at low pressure (10 of microns or higher) is not a very effective methodology. This does a very poor job of removing water vapor (a serious contaminate) and other contaminates.

Far better to get down to a few microns and add a little D2 and strike a plasma to attempt to clean the chamber; however, if their is small leakage, unlikely you will get a system clean enough for measurable fusion.

First, can your pump (no load) get to a few microns? If so, next the chamber must have little leakage (like a micron a minute - not out gassing; that will improve with cleaning.) After getting the chamber/system well sealed, then use a plasma of D2 to clean it. After that an attempt to obtain measurible fusion can be attempted.

Best if you can get in the low 10^-4 torr range.

I will warn you - lack of instrumentation (being blind to vacuum parameters) makes doing fusion extremely difficult. Not knowing your actual pressure readings is a serious problem that (if you haven't already) must be corrected. Of courser, reaching only 15 kV or so means your pressure is far too high for fusion; and flushing with hD2 won't fix that issue. Unable to get into the 10^-4 torr is not making your life any easier, either.

[JoeBallantyne]

So one of the most awesome things about this site is how people share and help each other.

Last March when I was posting about my first efforts building a fusor, and the problems I was having accurately measuring vacuum, Bruce Meagher messaged me out of the blue, and offered to loan me an MKS901p with a KF16 connector to which he had attached a board he had assembled that was the design Finn Hammer had made for a digital display of the MKS901p pressure.

Of course, I immediately took him up on his offer, and a few days later I got a package in the mail with a beautiful MKS901p with the display board and power supply for running it.

Amazing! Everything I have done since with my fusor was possible because of the help Bruce offered me, and because Finn Hammer designed a great little board, and sent some of the bare boards to Bruce - who added all the components and got them working. Thank you Bruce, and Finn.

Comparing the MKS901p with the digital display to the Varian 531 TC and Varian 801 gauge, is like comparing a Ferrari with a Yugo. OK, they both may be able to move you from one spot to another, but MAN, the driving experience is NOT the same.

The other thing that was great about the MKS901 Bruce loaned me, was that he zeroed it for me before he sent it out, as he had a vacuum setup that allowed him to do so. So from mid March on, I finally had an accurate way to measure my vacuum pressure levels, and to better calibrate the 531 TC gauges I had. This allowed me to get my pumping system properly setup and to find and fix leaks in the fusor setup. I’m a little embarrassed to admit that I am still using his loaned digital display board with my own MKS901p as I have not yet finished building my own digital display setup although I have gotten most of the parts.

One of the great things about the 901p is that a digital readout of pressure that can go well under 1 micron is so much easier to read and use, than trying to squint at a Varian 801 gauge and decide which tiny narrow line, the tiny narrow needle is sitting on. When the single digit micron pressure lines on the gauge display are about half a millimeter apart, and shifting your head even a little from side to side gives you different readings from parallax. (Yes I know that is why good meters have a reflective bar behind the needle on them, so you can eliminate parallax by lining the needle up directly in front of its reflection. That is still a PITA compared to immediately reading a number off a very bright little screen.)

The MKS901p with a digital display is great. Thanks Bruce!

Joe.

[Paul_Schatzkin]

So one of the most awesome things about this site is how people share and help each other.

That is really cool to read, Joe. Heartwarming in a very real way. Like by setting all this up I've done something worthwhile after all (something my ex-wife would take issue with...).

While the world is focused on a (not really) net-gain reaction that lasted a billionth of a second, the people here are learning how to build and run sustainable fusion systems.

Yeah yeah, I know. Countless orders of magnitude from Q...

but,

still...

... this is a thing of unimaginable – and promising, goddammit – beauty.
.

(Fusor 'star mode' photo by Brian McDermott)

I look forward to seeing yours when it happens, Joe. Send pictures!

--P

[JoeBallantyne]

Here are some pictures of the deuterium gas setup I was previously discussing. This is the state as it was back on March 9th 2022.

The nice Matheson SS regulator that regulates into vacuum. Note the high pressure side is charged with D2 at 500psi as discussed previously. It is really comforting to be able to keep the actual cylinder always valved off so the max possible lost D2 at any time is just the amount sitting in the tubing on the high pressure side of the regulator.

The D2 cylinder itself.

The D2 supply plumbed with Swagelok connectors and a 1/4 inch copper line to the fusor.

One thing to keep in mind when dealing with the CGA350 connectors that come on most D2 cylinders is that they use a LEFT HAND THREAD!
Critical to remember especially when you are first trying to unscrew the nice pretty cap that covers the 350 connector on the high pressure valve that ships with the cylinder. You must turn it CLOCKWISE to take it off. That bit me when I first tried unscrewing it. Fortunately after I first tried to get it off (by unscrewing normally - in the wrong direction) without using very much torque, I thought, this is brand new, it is just a cap, and it should be pretty easy to get off, maybe it is LHT, and indeed it was. Turning in the direction that would tighten normal connections (clockwise), it loosened immediately and came off. It is a very strange experience the first time you "tighten" something hoping it will indeed loosen instead.

I labeled the nut on the regulator to remind myself in the future if I ever need to swap in a new cylinder.

Joe.

[Richard Hull]

I had a 59 De Soto and took it to a June car show some years ago. I had one of my lady friends along for the ride. It blew a tire on I-64. Now, like any guy, I know to merely loosen all the lug nuts before jacking the car. I struggled in the hot sun to loosen one lug nut after another. Not one of them would cooperate. The girl I was with said, "let me try". What!? Silly girl thing! Ha! I was bushed and fetched a cool drink from our ice chest and let her have at it. With my back turned, slogging down the cold soda, I suddenly heard the unmistakable screech of a lug nut breaking loose. I turned around and she was now working on a second nut!

I did not know that in the 50's, MOPAR (Chrysler cars), had normal CCW loosening nuts on one side of their cars but CW loosening nuts on the other side.
She didn't know either, but when it wouldn't loosen one way she tried the other way, (tightening to we normal He-Men) and it broke loose.

We replaced the tire and with a bit of egg on my face, went on to win "best unrestored car" in the show. A photo of that day with me at the wheel driving to the awards table.

Nuts can be nuts and make you nuts. They can, on rare occasions, be very backward things.

Richard Hull

[JoeBallantyne]

Fusor Vacuum Pump Saga

This post will detail my efforts at getting a vacuum pump setup that is sufficient for fusor work. One of my goals was to do this without using a 2 pump setup. (Yes, I DO know that the vast majority of folks building fusors, use a 2 pump solution for their vacuum system.) I wanted to see whether I could dispense with using either a diffusion pump or a turbo pump. Primarily because having a second pump doubles the probability of failure of the pump station – 2 things that can fail instead of one. Furthermore turbos require a somewhat complicated electronic controller – which means if you are using a turbo, you now have 3 things that can fail – mechanical backing pump, turbo pump itself, and the turbo controller. Furthermore, using two pumps requires more vacuum plumbing than one, as you must plumb the pumps together in addition to plumbing them to the fusor.

It turns out that you CAN get a pump setup that only uses one pump, which will get you down to fusor level vacuums ie: ~1 micron, but you either have to get lucky, or be willing to spend a pretty penny.

Fortunately, I got lucky.

My very first vacuum pump setup on the fusor was using an Edwards E2M8 dual stage direct drive pump, but with that I could not get down to less than about 5 microns, and typically it liked to run after getting hot, at more like 12-14 microns. This was good enough to let me make a plasma, but not good enough to properly evacuate the fusor for D2 runs.

I also had an old Welch 1405, but that would only pump down to about 35 microns. Much worse and not even close to what is needed for a fusor.

However, I also had bought from a University of Washington (UW) surplus auction at some point about 10 years ago what looked to be a brand new old stock, still in the box Welch 1402 pump head, that appeared to be at least 40 years old, but never used. It still had the factory tags on it. It turns out that both the Welch 1405 and Welch 1402 use a 1/2HP electric motor, and the diameter of the drive pulley on both pumps is 10 inches. So, I decided to replace the 1405 pump head that only pulled down to 35 microns with the 1402. This entailed properly locating and drilling 4 new holes in the pump base plate for the 1402 (the hole layout IS different between the 1405 and 1402), and then bolting the 1402 head down, and then reattaching and adjusting the tension of the drive belt. One thing I discovered about these old Welch pumps is that they have plates that go across the bottom of the pump base, underneath the drive motor that the motor mount bolts screw into, which function just like conflat plate nuts, so you can just loosen the motor mount bolts and retighten them to remove and adjust the drive belt tension without ever having to access any nuts underneath the pump base. Very nice touch, IMO. Especially since the complete pump assembly is NOT light. The 1402 with motor and base weighs in at 120+ pounds. Unfortunately, the pump bolts do not screw into similar plates, so you are stuck tilting the base up and getting 2 wrenches on bolt head and nut to attach the pump head to the base. But even if there had been plates on the 1405, they wouldn’t have worked for the 1402 anyway because the bolt spacing is different.

The new old stock 1402 pump head came filled with oil that looked brand new – I drained it to take a look, and it was a clear pale yellow clean oil. I had purchased some Inland 19 Ultra vacuum pump oil, which I decided to use instead. So I saved the original drained oil and loaded in a batch of brand new completely clear 19 Ultra. Then since the pump had been sitting for decades, I manually turned the 10” pully drive forward, and backed it up a bit if it got stuck, and then moved it forward again. There is an arrow on the pulley indicating the normal direction of rotation. It has been several months since I first brought this pump up, so I don’t remember how much time I spent manually turning the pump. But it is very important that you are sure that the pump can turn manually in its normal direction without getting hard stuck, before you switch on the motor to run it. If the vanes are sticking, they will loosen up the more the pump is turned. Everytime it sticks in the forward direction, just move it back and forth a bit until it unsticks and you can turn it more in the forward direction. The pump will loosen up the more you turn it manually until it stops getting stuck. You may need to spend and hour or more working the pump manually to make sure it won’t get stuck.

Once the pump would turn in the forward direction normally without getting stuck, I plugged in the pump motor, and it ran just fine. For a while. Then I heard the start of what sounded like a clanking noise. Unfortunately the clanking noise got worse as the pump heated up. I wasn’t sure exactly how a new old stock Welch mechanical pump was supposed to sound, but I was pretty sure that a loud metallic clank was not good, so I powered it off. After letting it sit a while, I fired it up again, and at first again the pump sounded fine, but eventually it again started to clank. I thought perhaps that oil wasn’t circulating properly in the pump, so I pulled the 00 rubber stopper plugging the pump hose barb intake, and added a bit more Ultra 19 oil down the pump throat as it was running. Didn’t seem to make much difference to the clanks. Of course doing this caused the pump to gurgle loudly and spew oil mist “smoke” out its exhaust while the stopper was out of the intake.

Basically, whenever the pump started to heat up, the clanking would start.

I thought maybe it was the new oil I was using. The new oil looked a bit less viscous than the old oil, perhaps the pump required a heavier weight oil. So, I swapped in the original oil.

No difference. Still clanked when it got hot. Plus I noticed that the ultimate vacuum that I could get with the old oil (5+ microns) was significantly worse than the vacuum I could pull with the Island Ultra 19 (~2 microns). So, I swapped back in the Inland Ultra 19. It wasn’t the oil viscosity or type that was causing the clanks.

After getting frustrated with the repeated clanking, I mused that maybe the clank wasn’t really that bad after all, and perhaps the pump just clanked normally. So, I decided to just let it run and clank. (Really BAD decision… DO NOT do that if you have a Welch vacuum pump that clanks. They are NOT supposed to clank.) Pump ran for a while, clanking, and the clanking got louder and more frequent as the pump got hotter, and the pump kept getting hotter, and hotter, and started to slow down… and the motor was getting really hot. And the pump kept slowing down and slowing down until it stopped. At which point I realized I was being a total moron, and that I might have frozen up my brand new old stock pump head, and that I was probably going to burn out the motor as well! So, I immediately switched off the pump, and let it cool down for a several hours. (I’m telling you about my stupidity, so that you don’t make the same mistake.)

After the pump cooled off, I then tried carefully moving the 10” diameter pump pulley back and forth by hand to make sure it wasn’t actually frozen, and fortunately it was not. Once it would again rotate fully in the normal drive direction by hand, I drained the oil, and put in a brand new batch of Inland 19 Ultra. (1402 pumps actually hold quite a lot of oil – a little more than ½ a gallon.) The drained oil was full of tiny particulates, and was no longer clear, but instead slightly grey – likely due to the presence of both metallic and vane particulates in it. It looked much like the very first oil change made on brand new lawn mower engines – assuming that first oil change is made after 10 hours or so of use and the oil is still clear enough to see through.

After this epic fail, I decided to try a new tack, which was to run the pump until it just first started to clank and then immediately switch it off and wait for it to cool down. Then then turn it on again until it clanked, switch it off, cool, rinse and repeat. My theory was that most likely the pump vanes which are typically phenolic (the original label states they were fiber vanes) had absorbed moisture during the several decades the pump had been sitting at atmosphere in the Pacific northwest humidity. So, when the pump heated up, the vanes would get too constricted to slide freely up and down in their slots in the rotors, and would cause the clanking. I reasoned that if I ran the pump enough up to the point where the vanes started to be constricted, but not let it actually damage itself by clanking, I could slowly wear the sides of the vanes down (and dry them out by keeping them at vacuum pressures for significant time), and eventually the pump might run continuously with no clanking.

This is exactly what happened. At the beginning the clanking would happen when the pump got to about 95 to 100 degrees Fahrenheit. As the runs continued, the temperature at which the clanking would start slowly rose. Eventually, after many cycles of running until clanks started, and then stopping to cool off, the pump could get fully up to its running temperature of 120-125 degrees Fahrenheit, with no clanking! (An inexpensive Chinese made handheld electronic thermometer with a laser pointer was very useful in keeping track of the temperature of both the pump head and the motor during this process.)

Once the pump finally appeared to be able to run continously with no clanking, I decided to just let it run for 24 hours straight, which it did without any clanking. After that, I measured the vacuum the pump was pulling at the end of the hose it was connected to, and measured 1.5 microns.

Success!

Morals of the story:

1) Welch makes really nice mechanical pumps. The 1402 pumps are especially nice.

2) If you have a brand new (or brand new old stock) Welch mechanical pump, it can pull down enough by itself to run a fusor – down to 1 micron or less.

3) Never let a Welch pump clank. That is not how they are supposed to sound. When running they make a quiet gurgle, with perhaps some random aperiodic clicking that is not very loud. While they are pumping down from atmosphere, the gurgling sound is much louder. The manual specifically warns against letting the pumps pump against too high a pressure continuously. Pumping at 1 torr or higher for extended periods of time can damage the pump. (It also spews oil mist out its exhaust vent, or into the exhaust filter.)

4) If your pump has been stored a long time and clanks when it starts to warm up, bring it into service by repeatedly running it until it first starts to clank. Then shut it off, let it cool, and run it again. Do this repeatedly, until it can run continuously without clanking. Once it runs continuously, you should probably consider replacing the oil used during the bring up process with new clean oil. You will also probably have more success eliminating the clanking, if you replace any decades old oil right away before starting the bring up process, since the old oil is probably full of moisture as well, and will therefore slow down the process of getting the pump vanes to slide properly when the pump gets hot. (Since oil already loaded with moisture will take a lot longer to pull moisture out of the vanes, than oil with no moisture.)

5) The oil you use in your mechanical pump really does matter. Especially if you are trying to do a single mechanical pump solution for a fusor. Modern mechanical pump oil that has been doubly refined – like Inland 19 Ultra – is going to perform better than oil that has been absorbing moisture while sitting at atmospheric pressure for decades, and likely was never as refined as the new oil is anyway.


The original tags that came attached to the Welch 1402 pump head. Pretty clearly 60's or early 70's era tags.

The pump after swapping on the 1402.

The initial pressure measured at the end of the vacuum hose using the MKS901p and display loaned to me by Bruce Meagher.

An early pressure pulled on the fusor by the 1402. Later the pressure would drop significantly as the chamber outgassed its adsorbed moisture.

Joe.

[Richard Hull]

I have two 1402's sitting around that pump deep like yours. In addition, I have two 1396 Pumps by Welch. I have tried to sell all of them at HEAS for years, but their simple mass of hundreds of pounds, each, keeps them anchored fast to my lab floors. I would use them, but the Precision 5 CFM I have used since 2000 matches them. Thus, I let a sleeping dog lie.

I used a 5 CFM yellow jacket direct drive from 1997-2000 and did my first fusion in 1999 using only the yellow jacket( 5 micron low end) before switching to the belt drive precision in 2000 and continued to do fusion without any secondary pump until 2003!

I wasted a bit of deuterium gas purging what little air remained via the normal differential pumping of the single pump, but I did fusion! It can be done! All you need is a good deep pumping ( below 10 micron) mechanical pump. (note: I did have the famous and unobtainable micro sieve sold by Lesker that helped a lot.) I think it left the market as it fouled the oil in the mechanical pump as you used the electric heater in the thing to clean it up. This forced you to gas ballast for an hour or two to clean the oil as you heated the micro-sieve. A tedious process to be sure and chased me to a diff pump in 2004.

My system got a diff pump in 2004 and a Turbo in 2018. Both made fusion a lot easier, saved a lot of Deuterium gas and made startup to fusion a lot faster.

Richard Hull

[JoeBallantyne]

Additional Pump “Improvements”

Once my 1402 pump was running, and connected to the fusor with a vacuum hose. One of the first things I decided I wanted to do, was get rid of the ugly 2” diameter red rubber vacuum hose, and replace the hose barb connector on the pump with a KF25 connector. That way I could plumb from the pump directly to the fusor using standard KF fittings and stainless steel bellows. The Welch 1402 has a 1-20 threaded connection for the intake connection. One inch diameter – 20 threads per inch. It is NOT an NPT (national pipe thread), or a BPT (british pipe thread) connection. Straight 20 tpi all the way through – which means you HAVE to buy an adapter made specifically for that threaded hole. I purchased a KF25 Welch pump adaptor for 1-20 for a good price on Amazon from Brownian Motion Technology, and then unfortunately had to immediately return it once I got (with quite great difficulty – more on this later) the hose barb connection off.

Why you ask did I have to return my brand new adapter? Because it turns out that Welch has “old style” and “new style” pump heads, and the adapter I bought initially from Amazon was for “new style” Welch pumps. I had no idea there were two types of pumps and corresponding KF25 adapters, until the one I bought didn’t work. My pump head is probably about 50 years old, and requires an “old style” adapter. The new style adapters have an o-ring that comes in them and which gets clamped around the circumference of the threaded input hole, and you simply screw them in, and the o-ring seals directly to the outside of the pump body. Old style pumps however, have a recessed seat machined into them about a quarter of an inch deep, and then a smaller threaded 1 inch hole inside of that. The machined recess is too small for the outside of the o-ring holder of a “new style” KF25 adapter to fit through, and too large to allow the o-ring to seal properly. So, you HAVE to get an “old style” KF pump adapter that fits properly in the recessed seat, and which uses a metal gasket like the hose barb which comes standard on “old style” Welch pumps.

You can find these “old style” KF25 1-20 adapters from a couple different suppliers, but they are about 50% more expensive than the new style ones which are available from a bigger pool of suppliers. Straightforward capitalistic competition (or lack thereof) at work.

Getting the original hose barb off my Welch 1402 pump was surprisingly difficult. They are installed VERY tight. I had to use an oversized ~20 inch long crescent wrench to get mine off, and I had to put a LOT of torque on it. The thread is a standard thread (not left hand), but I would estimate those hose barbs are put on with at least 125 ft-lbs of torque and probably closer to 175-200 ft-lbs. I had to tip the pump up on one end of its platform so that the hose barb was horizontal with the floor, and the motor was directly above the pump, and then put the crescent wrench on the hose barb so that the wrench was at right angles to the hose barb, but also parallel to the floor, and then essentially put all of my 200+ pounds on the end of that 20 inch wrench with an impulse to get the hose barb to start to loosen. You can’t use a standard socket because the hose barb is too long, so it is important to get the crescent wrench very tight on the barb so there is no way it will slip and bugger up the machined in hex nut at the base of the barb. Bugger it up, and you will NEVER get it off. (Unless of course you are then OK with using a 3 ft or longer pipe wrench on your buggered up hose barb.) Fortunately I got mine broken loose without messing up the machined hex nut.

Once you break it loose, it comes right out, but the 1-20 threads are very fine, and getting my newly purchased KF25 “old style” Welch pump adapter to start cleanly in the threads was also tricky. Be careful to fully clean out the machined seat recess in the pump (without having anything fall into the pump). Make absolutely sure you start threading the adapter in by hand, and be sure it is not cross threaded, before you tighten it with a wrench. The trick I used was to put the adapter on the threaded input, then “unscrew” it until I felt the spot where the initial thread starts (you can usually feel it click down a little, but it is really tough with 20 tpi), and then gently screw it in by hand. It took a few tries for me to get it, and unfortunately even when I had it right, the adapter screws in pretty tight. I could not screw the adapter in by hand all the way down to the metal washer, not even close, but I could get it to go enough turns by hand to be sure it was not cross threaded.

Another thing I would suggest – and which I regret not doing myself, is to place the metal washer in the recessed seat and get it perfectly centered in the seat and around the 1-20 threaded hole BEFORE you start screwing in the adapter. That way the washer will hopefully stay centered until the adapter clamps it to the pump body when tightened fully. The metal washer that came with my KF25 adapter was sized in a way that makes it difficult to get it correctly positioned – especially if you don’t do it beforehand with the adapter out of the seat so you can see how it is aligned. Ideally the washer would have an outer diameter that is almost as large as the diameter of the machined recess it fits into, and an inner diameter just a bit larger than the threads of the adapter. Neither was the case, the washer was significantly smaller in OD than the recess into which it seated, and the ID was significantly larger than the threads of the adapter, which means it had a significant amount of play in it, and I suspect that in my case it ended up completely off center when finally clamped down, because in the end, after the new KF25 was fully installed, and the pump ran for a while, the new adapter seal leaked.

Of course it didn’t leak immediately. That would have made the cause of the leak obvious. At first the pump worked great. Pumped down to about 1.5 microns on the input immediately. So I then connected it to the fusor, and it pumped that down to the mid single digit microns also. But after about an hour, all of a sudden the pressure in the fusor started to rise until it hit 18-20 microns. When nothing had changed. It didn’t make any sense. I checked all the KF clamps, tightened ones a bit that needed it, but no change. Pressure in the fusor stayed at about 20 microns.

I started to get worried that I would have to try to track down a leak somewhere in the fusor, and was feeling a little bit sick about it – as it would probably take a long time, when I sat back and thought, wait a minute. I just changed the hose barb on the pump to KF, that is what has changed most recently, and is most likely the cause of the problem. What if there was some residual oil (because when I cleaned the machined recess where the KF adapter screws in, I used clean Inland 19 Ultra oil) that had been preventing a leak, but it got sucked through the leak, and now the pump is pulling on atmosphere though a tiny orifice. (Likely where the metal seal was misaligned to the adapter seat itself.)

I thought if that were the case, then I could drip some new pump oil in the recessed seat around the base of the KF25 adapter, and fill it with pump oil all the way around, and see if that resolved the higher pressure the fusor was at.

BINGO. A few seconds after I dripped some pump oil off the end of a screwdriver into the crack between the base of KF25 adapter and the recessed seat, and filled it up with new pump oil, the pressure in the fusor dropped right back down to single digit microns, like a switch had been thrown. It was almost instantaneous.

So, now I had a nice KF25 adapter on my great Welch pump head that worked fine as long kept the space between the side of the adapter and the machined recess in which it was seated filled with high quality pump oil. But as soon as all of that oil got sucked through the leak, which typically takes a couple of hours or so, my KF25 adapter leaks, when the hose barb it replaced never did.

Sigh.

One step forward, two steps back.

Does the “old style” KF25 adapter come with multiple metal gaskets? Of course not. Is the metal gasket easy to acquire on its own. Of course not. Is the metal gasket made out of the same stuff as say a typical washer – of course not. I didn’t examine it that closely but from the quick glance I gave it before installing the adapter, it looked like it was probably made out of aluminum. It was definitely a softer metal so that it would make a nice metal to metal seal between the stainless steel of the KF25 adapter, and the cast iron of the pump body. The only problem with it IMO was that it was sized so that if you didn’t line it up perfectly beforehand, it is not big enough to ensure that the offset seal will actually make a vacuum tight seal all the way around the base of the KF25 adapter. It certainly did not in my case.

Now, I did buy more than one of those KF25 “old style” adapters, because I have more than one old style Welch 1402, but I wasn’t keen on effectively wasting one of them just for its metal washer.

So, for now, since I don’t typically run the pump for more than a couple of hours at a time anyway, I simply make sure to drip some clean oil around the base of the KF25 adapter in the recessed seat, so that it will pull a good vacuum for a couple of hours. And I have to remember if the pressure suddenly starts to rise – seemingly inexplicably – that I need to drip a bit more oil into the recess in the pump head around the adapter.

Kind of a pain, but I do still prefer to have everything plumbed in with KF fittings and metal bellows than using a big ugly long old red rubber hose – that of course never leaked at all. The price of vanity I suppose… Someday, I will bite the bullet and attempt to reinstall the KF25 adapter with a metal seal that is very carefully and perfectly centered in advance. But not yet.

When I installed the adapter I did torque it a lot, but not quite as much as the original hose barb was torqued. I did not want it to be quite so difficult to take off, if I ever actually needed to take it off. Based on how it felt when I was torquing it at the end, I am quite sure that the washer did get clamped tight, and that it was off center and one side might have possibly been extruded so the inner edge of the washer was on the outer edge of the adapter. I won’t know for sure until I take it apart and examine it carefully before installing it again with a new metal washer/seal.


Takeaways:

1) Make sure you get the metal seal lined up and centered perfectly before installing any of the “old style” KF25 Welch 1402 pump intake adapters. Do NOT just slip in on the threaded end of the adapter, and screw in the adapter, and hope for the best. That’s what I did. It was a fail. Should the metal seal be sized so that slipping it on, would just always work. Of course it SHOULD be. But it is NOT.
2) Although KF is wonderful, and looks really nice, if you have a working hose barb, and hose, and you know it doesn’t leak, think long and hard about whether you really need to swap out the ugly hose, for pretty stainless KF. It really might not be worth it.
3) If you have a mechanical pump, and it doesn’t pull a great vacuum when you put a vacuum gauge on the inlet, BEFORE you decide to tear the pump apart, try dripping some clean new high quality mechanical pump oil around the base of the pump inlet where it screws into the pump body (be it hose barb, KF, or otherwise) and see if all of a sudden your pump pulls down better. If so, you have a leak right at the base of the inlet where it seals to the pump body. It will only take a few seconds to check, and may save days of unnecessary work. On my current Welch 1402, it means the difference between a pump that pulls down to 1 micron or less, and a pump that pulls down to 5-18 microns or so. (The longer I let the pump leak, the worse the leak gets, as more and more of the oil that was blocking the leak gets sucked through the leak, and the effective leak aperture gets larger.)

[Richard Hull]

Let all sleeping dogs lie! The KF25 idea is great, but.......... With straight threads, Welch, I am sure, used some sealant that would seize like the hand of god when they installed the barb years ago. I would solve your problem with a thick vacuum grease around the base.

I would have used only a short length, of the thick red India rubber hose to the factory, welch installed, barb. (sleeping dog left to lie) Then stuff into the hose a "barb to KF 25" adapter. (cheap, Duniway Stockroom)

I my case I use a 6" length of red rubber hose to the barb to KF 25 adapter to my 4 way KF25 SS cross and then to all SS plumbing thereafter.

The thick rubber hose is totally vacuum tight and seals to any barb over 100% of all imperfections with a good SS clamp at both ends. There would have been nothing wrong with a one inch length of hose Welch barb and clamp to a barb to KF25 with a clamp where the metal to metal distance was near zero. I found the old fashion huge diameter red rubber hose to be fantastic! I purchased my 1 foot length new from Duniway back in 1999. I still have the 6" cutoff from that original length. The rubber is far more supple and gives great vibrational separation from the pump to the hard plumbing.

You went to a lot of trouble to be a purist, which is laudable, but you woke up the sleeping dog and got bit!

[JoeBallantyne]

Neutron Moderator Construction

This post will detail the construction efforts for the neutron moderator design that I mentioned in an earlier post.

Last summer, the best price I could find on 1 inch thick HDPE was a set of 3 18”x24”x1” HAKKA cutting boards from Amazon. I bought a couple of sets. In August 2022 they ran $129 for the set of 3, now in January 2023 the same set is $150. I’m sure as the Federal Reserve (FED) continues their eternal crusade to debase the dollar the price will continue to rise. Somehow the FED thinks its rai·son d'ê·tre is to steal 2% a year from everyone on the planet who holds any amount of dollars. I guess if you steal just a little bit at a time from EVERYONE, on a permanent ongoing basis, it is somehow OK. But I digress…

I used one set of these 3 1x18x24 inch cutting boards, to build 2 5x5x24 inch moderators, the design of which I outlined in an earlier post. There was also enough left over to build the core of a small neutron oven for activation purposes.

It turns out that cutting HDPE is non trivial. I read online that the best way to do so is to use a triple chip grind (TCG) blade. So, I bought a Diablo D1084L TCG blade from Amazon https://www.amazon.com/gp/product/B001TH8HK8?th=1 that was highly rated (a LOT of people really like Diablo saw blades, and they make a 10 inch TCG carbide tipped blade). Swapped it onto my table saw, and then made a partial sacrificial 1” cut in a board so I could measure the kerf (width of the cut made by the blade). Turns out this blade had about a 0.1” kerf so 3 cuts would burn 0.3” of the 17.7” wide boards. I calculated that I should be able to make 3 pieces 5.1” wide with 2.05” left over. (Just FYI, cutting wood with that blade was like cutting through butter with a hot knife. It was a lot easier to cut through wood, than to cut through the HDPE.)

I very carefully adjusted the rip fence on my table saw, and made another sacrificial cut (or two) into a piece of wood to be sure that I would have exactly 5.1” of material cut off, and then I went for it. You can’t cut HDPE too slowly, because the plastic will heat up too much and start to melt, but you can’t go too quickly either. The feed rate I used on the 1” boards was about ½ to 1 inch per second. I made 3 cuts down the 24 inch length of each of two of the boards, and ended up with six 5.1x24 inch pieces, plus two 2.05x24” pieces. I then made two more 5.1” wide cuts in the 3rd cutting board, to get a total of eight 5.1x24” pieces, and then I had to adjust the rip fence to cut two more 2.05x24 inch pieces off. After that I had a piece left over that was a little over 2.9 inches wide, and I had 2 complete 5x5x24 inch moderators for my 24” long He3 tubes.

I decided to use that extra strip to make a neutron oven in conjunction with another ½ x 12 x 18 inch HDPE cutting board. I cut that 2.9x23.7 inch long strip into 4 equal lengths, and then cut the ½ inch thick cutting board into 9 ~4x6 inch pieces. It turns out that it was MUCH harder to cut the ½ inch HDPE than the 1” HDPE as it tended to start melting quickly. The feed rate on the ½ inch thick HDPE had to be significantly faster than the 1”. Even with a faster feed rate, I could not prevent the ½ inch HDPE from starting to melt by the end of the cut. I think the HDPE material itself for the ½ inch cutting board was different from the 1”. It seemed to be less dense, and it certainly appeared to have a lower melting point as well. The other issue, is that since the board was thinner, it was less able to sink heat away from the cut. I managed to get the ½ inch board cuts completed, but had to do significant cleanup of the cuts after they were done because of the melting. The 4x6 inch squares cleaned up reasonably well by scraping the edges with the edge of the blade of a flat tipped screwdriver, but I was happy that the 1” thick boards cut beautifully and needed no cleanup at all. (I didn’t realize how nicely they had cut until I started trying to cut the ½ inch thick cutting board.)

Since the 1” thick HDPE cutting boards were actually 25mm thick, not 25.4, they were just a tad under 1” thick. This meant that I could not slide my 1” diameter neutron tubes easily in and out of the central slot in the moderator. Side to side was not an issue, as I could just adjust how far apart the two 2.05” wide central strips of HDPE were, but top to bottom was just a tad too small. To fix that, I used another Amazon HDPE 18x24 inch cutting board I had purchased that was just 1/16th of an inch thick. I used a box cutter razor to cut off two 2.15” wide strips (intentionally a little wider than the central 1x2.05x24 pieces of HDPE, and I put them directly underneath those 2 central strips. This gave me enough headroom to be able to easily slip the He3 tube in and out of the moderator. Here is a link to those boards: https://www.amazon.com/gp/product/B07B89ZLTC/ it turns out that the price differential from end of August to now on this set of four cutting mats is even more dramatic. End of August 2022 one set of these cost $10. Now one set is $31. Ya gotta love the FED!

One of the original 25x450x600mm cutting boards.

Three of them in the ~50lb box they shipped in.

Actual dimensions of the cutting boards.

First completed moderator with central pieces sitting on top.

Both moderators and neutron oven.

Moderators, neutron oven, and large pile of shredded HDPE created by the cutting.

The very excellent blade that did all the work. (This was removed from my table saw as soon as this project was complete, as this blade is only for cutting HDPE.)

Joe.

[Richard Hull]

Great post on the moderator build. Others will benefit from this post, I am sure. While certainly the above post was part of Joe's fusor construction post, it related directly to neutron detection. I think Joe should also post a copy of this in the radiation forum so that it might be accessed in relation to that aspect of radiation detection. I just like to see a great post like the one immediately above posted where it won't be lost in the deep clutter of construction posts. Don't move it, just reproduce or copy it to the radiation forum. It is too good to get lost here in construction.

I am so glad I bought my moderator pre-cut back in 2020 from the local Piedmont Plastics facility here and picked it up in person. Joe's is a virtual same cut as we have the same 3He tubes. I put out a post on my work back then with a PDF. The total cost for it all 100% pre-cut HDPE with tax was $97.00

viewtopic.php?t=13553

If you are in possession of such a wonderful long tube, regardless of price for the HDPE, the finished project results in one of the most sensitive neutron detectors around, once all the electronics are hummed in to perfection. It creates a roar of counts from a whisper of neutrons.

I felt the pain of the recent price increases when I went back to buy pre-cut strips, 6 feet long to create a Lego/Lincoln Log-like set of small blocks of HDPE to assemble any number or forms of small HDPE moderators for activation work around the intimate contact rhodium/Russian STS-5 GM tube.

I got group of long strips of 1-inch thick X 1-inch and 1-inch X 1.5-inch HDPE all 6 feet long. This required my 10-inch chop saw to get a similar blade to what Joe purchased. I was fortunate to have to be able to cut only 1-inch and 2-inch cuts in a series of rapid, near instantaneous chops.

I find that for activation, a "fusor hugging" activation moderator, a "small block" assemblage is just the ticket, especially for fast decay isotopes where a GM tube can be placed for instantaneous counting at shutdown of the fusor. I still have a couple of those long, HDPE strips left, yet to be cut up and a box full of those "toy-like" HDPE blocks. This group of precut strips cost me about $125.00 one year later in 2021. (this was a plank of HDPE 1-inch thick, 6-inches wide and 6-feet long with several flawless, length-wise cuts by Piedmont.)

For the block moderator post, I turned into a FAQ found at......

viewtopic.php?t=14107

Richard Hull

[JoeBallantyne]

Neutron Detection

Of course the vast majority of folks who embark on building a fusor, do so because they ultimately want to do fusion and prove that they have done so. That proof is that they can detect the neutrons which are produced by half of the D-D fusion reactions.

I was no different, and so last spring, soon after I got my deuterium supply system properly connected to the fusor in a vacuum tight way, I tried to make neutrons.

I failed.

Or at least I failed to successfully detect any neutrons that I might possibly have made.

In my experience, getting properly setup to detect neutrons, and getting to a state where you KNOW that your setup will detect neutrons is the hardest part of building a fusor and proving that it does fusion. Why? Because unless you just go fork over $4000 - $6000 to Ludlum to buy a brand new Model 15, or Model 12-4, or Model 2363, you can’t be confident that your neutron detectors will actually work properly. Without doing a lot of work.

I purchased a few used complete neutron detectors off Ebay. Given the overall cost of those used Ebay detectors, I would have been better off just buying a single new one from Ludlum. Especially since none of the used neutron detectors I bought were able to successfully detect the low levels of neutrons I was initially producing with my fusor. In the end it was a 1x24 inch Texlium He3 tube that I bought a few years back from Richard Hull at HEAS in the HDPE moderator I described in an earlier post, that I coupled to a Ludlum 2200 scaler for neutron detection. That 24 inch long He3 tube was by far the most sensitive neutron detection device of all the different tubes and meters that I had.

Last spring (March 2022), I tried using a couple of Ludlum 2363 counters with the “Thors hammer” probe. I never got any counts that were statistically different from background. I tried using a couple different Ludlum Model 15s that I bought used off Ebay as well. Ditto on the failure to detect neutrons at statistically significant levels. (I think there may be issues with the tubes in both of those Model 15s. One of them has a tiny He3 tube, and the other has a larger BF3 tube.) Of course when I bought them, I had no way of really testing them, so I couldn’t know for sure if they worked before the window for returning them closed.

Bottom line is that you don’t know if the neutron detection equipment you are buying off Ebay actually works. The only way you can know for sure, is if you have a neutron source. If you don’t KNOW that your fusor makes neutrons, and you don’t have access to a neutron source, then you really are flying blind. If you don’t get any statistically significant counts, you don’t know if it is because your fusor is actually NOT making neutrons (certainly a very distinct possibility), or if your neutron detection equipment just doesn’t work, or isn’t calibrated properly, or the bias voltage level on the tube is too low, or the threshold on the meter is too high. Now I suppose I could have forked over some dough to Ludlum to calibrate one or more of the meters I bought, but I decided to go a different route. The do everything yourself route. It costs more, takes the longest, and is harder, but that is how you maximize what you learn.

I wanted to be ABSOLUTELY SURE that I could detect neutrons. The only way to do that, was to either purchase, build, or borrow a neutron source. I didn’t know anyone in my area with access to a source, so borrowing was not an option. Purchasing a source, I assumed would be hard. Now, I could be wrong about this, but I suspect that most entities with neutron sources, especially if they are reasonably strong sources, are probably required to have an RSO (Radiation Safety Officer), or a Health Physicist, or both. Certainly you can’t really find neutron sources on Ebay. At least I could not. I must confess I did try. Neutron sources are part of some kinds of moisture meters, and some kinds of density meters, but the kinds with neutron sources do not show up on Ebay. I’m pretty sure they are not supposed to ever show up on Ebay. In my experience, they don’t.

That leaves building a source. So, I thought I would just follow Carl Willis example, go the easy route, and buy an Antistatic Nuclespot 5mCi Polonium source for $289, and use it with a $50 Beryllium disk bought from Ebay to make a simple source. The Nuclespot would only last a year or so, but that would be sufficient time for me to test and calibrate my neutron tubes and meters. So, I tried. I ordered one back towards the end of August 2022. They told me it would be a 12 week wait. Well, it turned out that was wrong, because I finally called them towards the end of January 2023 and cancelled my order and asked for a refund. Nuclespots are in my experience, at least right now, impossible to get. Evidently the manufacturer made some changes to the manufacturing process, and due to that is now dealing with getting NRC approval for the modifications. At least that is the story they told me.

Siemens Pyrotronics F3/5A to the rescue… these are a LOT more expensive now on Ebay than they were just a couple of years ago. I bought a couple from a fellow member of this forum post HEAS, had another one I had purchased off Ebay a couple of years ago, and so was able to make a temporary, very weak ~200uCi Americium 241 Beryllium (AmBe) neutron source. This is only 1/25th as strong as the Nuclespot source would have been, but the advantage is that the half life of Americium is 1143 times longer than the half life of the Polonium in the Nuclespot. The Americium won’t be mostly decayed in a year. (In fact it still retains more than 90% of its original activity levels, even though it is on the order of 60 years old! Yes, it has been a LONG TIME since they put 80uCi of Am241 in a smoke detector… Modern ones have only 0.8uCi or thereabouts.) If I need a neutron source in a year, I can just put the Beryllium foil back in close proximity to the sealed Americium foil, and BINGO!, neutrons. Not very many neutrons mind you, but neutrons. Americium 241 is an alpha source, and Beryllium has the property that if you hit it with an energetic alpha particle, between 30 and 60 times out of a million hits, it will spit out a neutron. If we go with the low end of that estimate, 200uCi is 7,400,000 Becquerel (decays per second), which means we should be getting 7.4*30 = 222 n/sec. In reality the most counts I got with the source next to a neutron tube inside a moderator, was 100-200 cpm depending on the tube I was testing. But the point is, I could get counts! So I could be sure which neutron detection setups were working, and which were not.

None of the Ebay acquired all in one neutron detectors (the model 15’s and the model 2363’s) would register statistically significant count rates with my very weak source. This of course was quite disappointing.

Many of the folks on fusor.net who use stand alone neutron tubes, do so with specialized NIMBIN equipment. They use either homegrown, or specialized commercial preamps, coupled to shaping amplifiers, discriminators, and MCAs etc. I didn’t want to deal with any of that. I wanted to see if I could get some of these stand alone tubes to work directly with stock Ludlum meters – ala Mark Rowley with his Ludlum 3 and Russian CHM11 tube combination. (Although I guess in his case, he did put a bias/ballast resistor in series with the CHM11 tube.)

Well, it turns out that you CAN get standard He3, BF3 and B10 tubes to work directly with Ludlum 2200 and Ludlum 1000 scalers. No bias resistors, no preamps, no shaping amps, no discriminators required. BUT, not using a preamp, means you have to run at least the He3 tubes with a higher bias voltage, to get the tube gain up to a point where the stock meters will register the counts with their threshold adjusted very low.

In order to be sure that I knew exactly what the bias voltages and thresholds of my meters were, I bought a used Ludlum Model 500 pulser from a vendor in the calibration business, and used that to calibrate both the HV meter as well as the threshold potentiometer on the Ludlum 2200. (To double check the HV meters, I also bought a couple Sensitive Research electrostatic voltmeters off of Ebay. Which BTW are very cool pieces of hardware that require no batteries or external power source to function and to which you can directly connect the high impedance HV sources on radiation meters because the electrostatic voltmeter input impedance is on the order of 1e15 ohms.) The Ludlum documentation on the 2200 says that it will work with proportional tubes, but the threshold needs to be set to about 2mV for them. So, I increased the amplifier gain (screwdriver adjusting screw with a lock nut that is labeled DISC on the front of the 2200) on the 2200 by turning the set screw clockwise, so that 1.0 on the threshold pot was a 3.3mV threshold as measured on the Model 500, and 10.0 on the pot was a 33mV threshold. This allowed me to set thresholds anywhere from 2mV up to 30mV or so with better resolution on the pot than the initial lower gain amp setting allowed. After locking in the DISC amplifier setting with the lock nut, I adjusted the threshold pot down to 0.6 which corresponds to a ~2mV threshold.

The Texlium He3 tube has an HN RF connector on the end (many of the Reuter Stokes neutron tubes also have an HN connector on them), so I bought an HN to C adapter (UG-702/U) and screwed it onto the HN Texlium connector. Then used a stock Ludlum C to C cable to connect the Texlium tube directly to the Ludlum 2200. No bias resistor or preamp involved at all. (I intentionally did NOT use BNC connectors anywhere in the setup because I wanted to be able to push the bias voltage up to the 2500V limit of the 2200 if needed, without having to worry about significantly exceeding the HV capacity of the BNC connectors.) Put the tube in the moderator with the weak AmBe source and slowly started raising the bias voltage. Nothing at 1000V, nothing at 1500V, nothing at 1600V, nothing at 1650V. Finally when I hit a little above 1700V I got a single count. At 1800V I got a couple more counts. When I got up to about 1850V the count rate started to go up. Turns out, that if you run my Texlium He3 tube at 1900V with a 2mV threshold, it will work perfectly with a Ludlum 2200 no additional hardware required. Which makes it really easy to use.

Now granted, with a preamp, I could probably run the tube at a much “cooler” voltage of 1500V or so. So for the purpose of extending tube life as long as possible, using a preamp and a lower bias voltage is better. But this tube is not going to be running anything close to 24x7. More like a couple hours a week on average, absolute max. So for now, I am perfectly content to run it the simple way, directly connected to my Ludlum 2200.

I had a number of other tubes that I tested to determine what kind of bias and threshold settings were needed to run them directly connected to a Ludlum 2200. This is what I found along with typical counts per minute inside an HDPE moderator with my weak AmBe source, as well as typical background cpm:

Tube/Meter:.....................................Bias and Threshold:.........AmBe............Bkgnd:
Texlium stainless He3 tube 1x24.............1900V bias 2mV threshold..95cpm...........8cpm
Reuter Stokes Aluminum He3 tube 1x22....1900V bias 2mV threshold..101cpm.........12cpm
Russian SI19 1.125x8 He3 Corona............2050V bias 35mV threshold.70cpm..........18cpm
GE B10 1.5x17..................................715V bias 10mV threshold...40cpm..........1cpm
Nancy Woods BF3 1x6.........................1975V bias 2mV threshold...12cpm ..........5cpm
Ludlum Model 15 BF3.........................1750V bias ?? threshold.......2cpm............0cpm
Ludlum Model 2363 hammer.................?? bias ?? threshold............6cpm............6cpm

For the Ludlum model 2363 I just ran it as a unit as it was configured when I purchased it. I left the probe attached to the meter, and put the probe a few inches away from the unmoderated AmBe source.

Here is a picture of the tubes listed above: from bottom to top Texlium, RS, S19, GE B10, Nancy Woods, and the Model 15 BF3

Bottom Line:

1) Large He3 tubes are the most sensitive. But they are hard to come by and cost $$$.

2) You can directly connect many tubes to Ludlum 2200 Scalers and they will work perfectly – with an appropriately low threshold and a hotter bias than would be required with a preamp.

3) Ludlum 2200 scalers are great instruments that can be used to drive many different radiation measuring devices.

4) To be SURE you can measure neutrons with your instrumentation, you need a neutron source.

5) Ludlum 500 Pulsers are nice for making sure your instruments are properly calibrated and for accurately determining bias voltage and threshold voltage levels.

6) If you don’t buy a ready to go, calibrated, guaranteed functional, reasonably sensitive neutron measuring system, getting one setup and being sure it really works, is non trivial.

7) Ebay purchased neutron detection instruments like the Model 15, Model 12-4 and Model 2363 are most certainly not guaranteed to work in all cases. Sometimes (maybe more often than not) you will get a dud. Make sure you can figure that out before the return window closes. (ie: that neutron source I keep talking about, is kind of important…)

Joe.

[Richard Hull]

A great report on neutron tubes and the use of a Ludlum 2200. The 2200 is a fabulous all in one instrument. I reported on mine here and Jim Kovalchick also gave his thumbs up to the 2200. It is the closest thing to a full NIM bin in a box.

As with a NIM Bin, you must learn about the internal electronics an understand how to use them at a level beyond the casual amateur dabbler. it just takes a bit of use, and successes to hone one's skills with the device, as Joe notes.

As to the tubes. I think Joe bought both his 3He tubes from me at HEAS and he paid dearly for them, but you see how sensitive they are in his tests. Nothing touches a large 3 of 4 ATM 3He neutron detector tube.

The Nancy woods tubes are deceptive they are vastly smaller than their external dimensions with very short sensitive regions. There is a smaller tube within the nice larger brass body with the HV BNC on it. BF3 tubes even with a good preamp don't work at all until you really pour in the juice (HV). If you have a good preamp and the voltage to make them sing, size for size they are every bit equal to a good 3He tube. Problem is the amateur is most likely not going to acquire a larger BF3 tube.

I have spoken to the nice ancient GE boron lined tubes of the 50's and 60's in a number of posts before. They are great tubes and use to be cheap on e-bay. Not any more. They will not respond to the ham-fisted amateur as the bias voltage is hyper critical. Most idiots will chase them into their Geiger region and think they are counting neutrons. They are proportional tubes and require a preamp and a critically set voltage to detect neutrons. The 2200 is the ideal instrument to play "let's find the proper bias voltage" when seeking to bring a boron lined tube to heel and count neutrons.

The use of a pulser and digital o'scope aids not only in setup and calibration, but also as a big learning experience.

I can tell that coming out of his trials (expensive ones), Joe is now a self-trained near expert in the use, and more importantly, the understanding of just what it takes to make neutron tubes actually count neutrons! He understand the tubes and the electronics at a core level.

Again a great teaching post...! Too bad it is deeply buried in a post thread in the wrong forum. This needs to be in the Radiation detection forum as a first post


Richard Hull

[Paul_Schatzkin]

Jeez, Joe, I sure hope you're using voice dictation for these posts. I cannot imagine typing all that!

I sure hope others appreciate and benefit from your painstaking attention to the details.

--PS

[Richard Hull]

I have been typing my brains out for years in long posts, especially in the FAQs. Many folks arriving here need to be taken by the hand and "spoon fed" materials in long form. This often demands a completeness that folks in the know can assume and arrive at through logic, but not many noobies are so endowed. Even then, for all my efforts, they do not read for comprehension. One would like to think we attract a higher order of folks, but it ain't always the case.

Long typed FAQs are also all about being explicit enough to keep people safe as we do venture into so areas where safety and good procedure is a must.

As I noted, Joe did a great job. The Ludlum 2200 is a great all in one instrument. I realized his great post related to the 2500 would forever be lost and buried deep in this long construction thread. The effect of this and Joe's post kicked me in the butt to write a FAQ around it and its use in the radiation forum FAQs. It, too, is somewhat of an epistle.

viewtopic.php?t=14785

Richard Hull

[JoeBallantyne]

Neutron Club Entry Request

So on 11 Jan 2023, I was finally able to get my fusor to do provable fusion, by making neutrons that I could detect at statistically significant levels with my neutron detection setup. In addition to working on neutron dectection in the months since March 2022, I also made several changes to the configuration of Fusor v1. I moved the fusor onto a utility cart, along with the D2 cylinder gas supply system and Spellman DXM 70N600 power supply. I moved the Welch 1402 vacuum pump to a different utility cart. Ultimately the plan is to move the gas supply from the fusor cart to the other cart, so that I have a gas supply and vacuum system that is independent and isolated from the fusor itself. That way I can just attach it to whatever fusor I am currently working with. I plan to keep Fusor v1 in permanent operational condition and NOT cannibalize it for parts for future fusors.

I moved the D2 gas inlet for the fusor from its original location on the top of the KF16 4 way cross, to the KF50 flange on the far end of the KF50 tee containing the cathode. This way the D2 gas enters the fusor on one end, and has to flow through the cathode and all the rest of the KF50 tubing of the fusor before it gets to the KF40 butterfly leading to the KF40 bellows that goes to the vacuum pump. In the original setup, the inlet gas could flow directly out to the vacuum pump without ever flowing through the cathode. In the new configuration this is not possible. In order for the gas to get out of the fusor, it MUST flow through the cathode location. (Realistically this probably doesn’t make much difference at all, but it is nice from a conceptual standpoint.) I also swapped out the original hard to use KF50 ball valve, for a KF50 to KF40 conical adapter and a proper MDC KF40 butterfly valve that is MUCH easier to use than the ball valve, and is what I use for throttling the vacuum and controlling the pressure in the fusor. MDC makes some really nice butterfly valves.

Since high voltage is the most dangerous thing about typical fusors, I built a custom connection for my stock MPF KF40 30kV feedthrough with various PVC and CPVC pipe fittings and pipe, so that I can use the proper DXM x-ray high voltage cabling with the correct connectors on both ends (Claymount CA11), and plug it directly into the power supply, and into my custom HV feedthrough connector. I bought the proper CA11 socket to use on this connector and it also has a 225W 150kohm ballast resistor built in. This connector is designed so it can be filled with mineral oil, and hopefully will allow me to get to the full rated 70kV of the power supply without any arcing on the 30kV feedthrough. The connector is a modified and improved version of the dual o-ring design used by Joe Gayo and Jon Rosenstiel. My design does not require using a lathe to mill out the inside of the PVC pipe. It uses stock PVC and CPVC parts available at Home Depot, and a slightly larger o-ring than the one used by Joe and Jon. I plan on making a separate post describing the design and build in detail. This HV connection system ensures that there is no exposed HV anywhere on the fusor or the HV power supply system. The cable is a commercial HV cable and is designed to be used with the Spellman DXM supplies. This dramatically reduces the risk of and worries about getting zapped! (I still make a reasonable effort to stay away from the high voltage cabling and connector, but at least there is no exposed HV anywhere in the system.)

Vacuum System:
Welch 1402 mechanical pump with an ultimate vacuum of ~1 micron.
KF25 and KF40 plumbing to the fusor with 2 KF25 butterfly valves – one just above pump inlet, and one after a typically blanked off KF25/KF16 reducing T. (This normally blanked off KF16 inlet is very useful for doing vacuum testing of any vacuum items I purchase from Ebay or elsewhere. I simply remove the blankoff, and attach whatever I want to test to the KF16 port, and can then pump it down, and determine if it is functional and whether it leaks or not.)

Deuterium Supply System:
100L cylinder of 99.8% D2 gas from Cambridge Isotope Labs.
Matheson SS CGA350 dual stage regulator
¼ inch copper refrigeration tubing with SS Swagelok connectors
¼ VCR SS Needle valve

HV Power Supply System:
Spellman DXM 70N600X3547 70KV 8.5mA power supply
75KV HV cable with proper Claymount CA11 connectors on both ends (proper cable for DXM supply)
Custom made oil fillable HV feedthrough connector with CA11 socket and 150Kohm ballast resistor
Windows laptop to control the DXM using Spellman software and an Ethernet connection to the supply.

Fusor vacuum chamber:
¼ VCR to KF50 adapter (connected to D2 gas supply needle valve)
KF50 Tee
KF50 to KF40 conical adapter
KF40 30KV HV feedthrough (standard MPF 30kV feedthrough see http://mpfpi.com)
KF50 Tee
KF50 viewport
KF50 to KF16 reducing Tee
KF16 4 way cross
KF16 MKS901p
2 KF16 blanks
KF50 to KF40 conical adapter
KF40 MDC butterfly vacuum throttling valve
KF40 bellows – to the vacuum system

Neutron detection system:
Ludlum 2200 Scaler
1x24 inch Texlium SS He3 tube
Bias voltage: 1750V
Threshold: 2mV

I did 3 separate 10 minute neutron background measuring runs with the Texlium tube in the moderator, but the fusor turned off, and got the following counts for each 10 minute run: 72, 51 and 84. Which results in an average of 6.9cpm – which we will just call an even 7cpm.

I started the vacuum pump. Pumped the chamber down to 1.2 microns. Then I closed the gas needle valve, and opened up the valve on the low pressure side of the regulator just a bit. The pressure rose quickly in the fusor up to about 5 microns. (The needle valve does not fully close off the flow of the D2, and I didn’t want to damage it by trying to close it further.) At this point all 3 of the butterfly valves in the fusor and vacuum plumbing were fully open to the pump.

On my Windows laptop I fired up the Spellman DXM control software – which you can download from the Spellman website. Selected the port type as Ethernet, and then set the network port to 50000 and the IP address to the default address used by the DXM of 192.168.1.4. (I had previously turned off the WIFI adapter in my Windows laptop, and set the Ethernet address of the wired port on my Windows box, to have a fixed IP of 192.168.1.2, with a mask of 255.255.255.0 and a gateway address of 192.168.1.1. I plugged both my Windows box and the Spellman into a wired Ethernet 8 port gigabit switch. There was no actual gateway on this local Ethernet, just my laptop and the DXM and the switch itself. Two CAT5 Ethernet cables, one between my laptop and the switch, and one between the DXM and the switch.)

After clicking OK in the DXM software network config dialog box, I got the software control panel for the DXM. Switched the control from local to remote. Set the DXM current mA limit to 8mA. Set the voltage to 16kV. This Spellman model voltage output is negative, so all of the actual HV values on the cathode are negative even though I will not make that explicit in this report – when interacting with the GUI you do not use negative values and it does not report the measured values as negative either. Then clicked enable HV. The monitoring voltage jumped up to ~16kV, but the current stayed really low, and there was no plasma, as the pressure in the chamber at ~5.5 microns was too low for the plasma to light off.

I then very slowly closed the KF40 butterfly valve on the end of the fusor to throttle the vacuum, and the pressure on the MKS901p steadily rose. I was focused primarily on the Ludlum 2200 counts and the viewport – to see when the plasma lit up. All of a sudden the Ludlum 2200 started counting! It was going slowly, but definitely faster than the background count rate. I checked the viewport, and the plasma was lit – it wasn’t very bright, but there was definitely visible plasma. I started making a video of the run as I was so excited about getting neutrons. The pressure was 24.4 microns, the voltage was at 15.91kV and the current was ~0.36mA. That first minute run generated a neutron count of 95 cpm!

It turns out that with those particular settings, 16kV on the cathode, ~24.5 microns pressure, and ~0.4mA, my fusor is very stable. The pressure is steady (rising very slowly), as is the current(also rising very slowly). So I decided to just let it keep running and not mess with any of the settings. I did another 1 minute run. At the end of that second run, the voltage was 15.91kV, the current was 0.36mA, and the pressure was 24.5 microns. The neutron count for that second minute was 142. I then reset the counter for a third minute run with the same 15.91kV, 0.36mA and 24.5 microns and got a count of 108 neutrons. Since the fusor was stable, lit, and making neutrons, I decided to try for a 10 minute run, and switched the Ludlum 2200 to count for a 10 minute interval. The fusor ran beautifully for the full 10 minutes! At the beginning of the run the voltage was 15.91kV, the current was 0.38mA, and the pressure was 24.5 microns. At the end of ten minutes the voltage was 15.91kV, the current was 0.43mA, and the pressure was 24.7 microns. The neutron count for that first 10 minute run was 1479.

I decided to do a second 10 minute run immediately after the first. So the initial voltage current and pressure at the start of that run are the same as the settings at the end of the first run, or 15.91kV, 0.43mA, and 24.7 microns. The voltage, current and pressure at the end of the second 10 minute run were 15.91kV, 0.46mA, and 25.0 microns. That run generated a neutron count of 1716. Interestingly, even the slight rise in the pressure and current during the run made a decent increase in the neutron count over the 10 minute span.

I realize that the voltage and current are both pretty low as typical fusor runs go, but the 24 inch Texlium tube is sensitive, and the difference between the background count of 7cpm, and the average fusor running counts of 148cpm and 172cpm for the two 10 minute runs, is a factor of 21x and 24x the background rate which is statistically significant in my book.

I did pull the neutron tube from the moderator with the fusor running to check what happened to the neutron counts, and with the tube pulled, the counts went down to background level (I got 1 or 2 counts during a 20 second interval or so), and immediately rose again when the tube was reinserted in the moderator.

A week or so later, I did additional runs at significantly higher voltages, and got much higher neutron counts, but this is the data from the first day I made neutrons, and I consciously decided to not push the fusor, as it was running in such a stable way. When running at higher voltages, it is much more difficult to keep the current under control even for a full minute run, and the DXM will shut down when the current hits its upper bound of 8.5mA.

A few pictures of the setup on 11 Jan 2023.

The Ludlum 2200, Texlium neutron tube and moderator.

The Spellman DXM control software running on a Windows laptop.

The MKS901p with Bruce Meagher’s digital display unit.

Fusor V1 with commercial HV power cord feeding custom feedthrough connector.

Closer view of the custom feedthrough connector with ballast resistor.

Joe.

[Richard Hull]

Joe this is a very full report and you are now inducted into the neutron club.. We look forward to more HV runs.

Suggestion: limit your pressure to 6 microns and bring up the voltage. you should over time be able to add more pressure.

The auto shutdown of the supply is a great disadvantage, but over time you will get to learn how to jockey the system.

Richard Hull

[Paul_Schatzkin]

Congratulations, Joe.

And thanks for all the very detailed reports.

--P

[JoeBallantyne]

Thanks Richard and Paul!

Its nice to finally be able to make neutrons.

Joe.

[Paul_Schatzkin]

And I'm sure it will make for some intriguing cocktail party banter.

"And what do you do, Joe?"

"Well, I'm retired now, so I make neutrons.."

Some guys go fishing. Some guys make wood furniture or paint.

Joe makes neutrons.

I just hope your wife approves.

--P