How to minimize backstreaming from diffusion pump

Every fusor and fusion system seems to need a vacuum. This area is for detailed discussion of vacuum systems, materials, gauging, etc. related to fusor or fusion research.
Post Reply
Jeff Robertson
Posts: 58
Joined: Tue Jul 19, 2011 11:58 am
Real name:

How to minimize backstreaming from diffusion pump

Post by Jeff Robertson »

Apologies if this has already been addressed, I've been abusing the search function but haven't found a direct answer to this question.

What's the best way to connect a diffusion pump to vacuum chamber? From what I've read/heard from others, connecting the diffusion pump directly to the chamber is bad news and will definitely result in some oil backstreaming. What's the most practical way to minimize this oil backstream? I've seen a lot of talk about cold traps, but I don't know if this is the most viable way to go, or even how exactly to make a cold trap. A couple articles I was reading mentioned using liquid nitrogen as a cooling agent in cold traps, which seems way more complicated than anything I've heard anyone do.

Thanks as always.
Jeff Robertson

Edit - Don't know if it's worth mentioning, but I'm using DC-704 diffusion pump oil.
SteveHansen
Posts: 294
Joined: Sun Aug 05, 2001 8:37 pm
Real name:

Re: How to minimize backstreaming from diffusion pump

Post by SteveHansen »


Most diffusion pumps have some sort of optically dense (can't see straight line through it) baffle which helps considerably. You may also see a water cooled cold cap over the upper jet. A baffle will slow the pump down a bit but that's actually good as it reduces the amount of gas needed to maintain pressure in the fusor and it creates a pressure drop that helps to keep the pump inlet in molecular flow while the chamber is in transition flow.

DC-704 is a good, low vapor pressure fluid. Santovac is better but at many times the price.

Finally, when you are flowing gas, the outlet of the chamber should be in transition flow as opposed to full molecular flow. That has a ballasting effect and will also help to minimize any backstreaming.

A lot of Varian's pre-turbo leak detectors didn't use a cold trap. Just baffling and low vapor pressure fluid.
User avatar
Carl Willis
Posts: 2841
Joined: Thu Jul 26, 2001 7:33 pm
Real name: Carl Willis
Location: Albuquerque, New Mexico, USA
Contact:

Re: How to minimize backstreaming from diffusion pump

Post by Carl Willis »

Hi Jeff,

I have successfully used DC-704 oil in diff pumps with no traps, down into the 1E-7-torr range.

A lot of older directives mention the need for traps, because older hydrocarbon pump oils continuously broke down into volatile cracking products and most uses benefited from chilled baffles or traps. That is no longer the case. Polyphenyl ether oils and silicone oils are more common now, and are generally heat- and even oxygen-stable.

Some vacuum processes are sensitive to trace contaminants from modern pump boilers. Any proper diff-pumped UHV (< 1E-8 torr) system will still include baffles or traps. The purpose of the diffusion pump in a continuously-pumped fusor is to improve speed over mechanical pumps at operating pressures in the millitorr range. This is a very modest requirement, well-suited to smaller diffusion pumps using cheaper oils and no baffles or traps.

-Carl
Carl Willis
http://carlwillis.wordpress.com/
TEL: +1-505-412-3277
Jeff Robertson
Posts: 58
Joined: Tue Jul 19, 2011 11:58 am
Real name:

Re: How to minimize backstreaming from diffusion pump

Post by Jeff Robertson »

As always, thanks for the informative responses guys.

Carl, your success without cold traps is relieving to hear, as I was dreading having to learn/implement yet another additional piece of vacuum technology. I have some time now so I want to clarify how my vacuum system will operate. I am working with a glass bell jar, with the diffusion pump originally planned to be mated directly to the base. If one was to sit on top of the inner grid, they would be able to see into the diffusion pump (but not very far). The pump is currently locked away in a lab which I won't have access to until monday, so unfortunately I can't inspect it and give you any information on its optical density. I am fairly certain, however, that vision directly down into the pump is impeded by a baffle.

I tend to suck at describing things with words, so I included a picture (taken from a paper I'm writing) of the vacuum system I plan to build. Additionally, I included a picture of my diffusion pump, which was kindly sold to me by a fellow member of these forums. I've searched long and hard for more information on the pump, but sadly I don't think Edwards supports or makes this specific type of pump anymore.

So, based on the images I attached, would you guys still say it's safe to push forward without a cold trap?

Cheers,
Jeff Robertson

Edit - The image of my vacuum system wasn't showing up, so I re-uploaded it as a jpg to see if that would fix it.
Attachments
vacuumsystem.JPG
diffusionpup.JPG
SteveHansen
Posts: 294
Joined: Sun Aug 05, 2001 8:37 pm
Real name:

Re: How to minimize backstreaming from diffusion pump

Post by SteveHansen »

I don't see any kind of baffle.

I would recommend putting in a simple optical baffle. The quickie figure below shows one that works. Make 3 aluminum disks that just fit into the pump's inlet. You can cut them with a hole saw. Then use the saw to make two "mousebites" in each disk Cut these so that when the three pieces are stacked on top of each other at 90 degrees you can't see through the stack. Drill a hole in the center of each for a 1/4-20 bolt. Stack them so that there's about 1/4-3/8 inch between them and they sit just above the top jet when inserted into the pump.

This is the type of baffle used in the first DP I ever owned - one made by NRC. I've never used a chilled trap but have used some kind of optical baffle on each pump.

Steve
Attachments
baffle_plate.jpg
baffle_plate.jpg (4.84 KiB) Viewed 3128 times
SteveHansen
Posts: 294
Joined: Sun Aug 05, 2001 8:37 pm
Real name:

Re: How to minimize backstreaming from diffusion pump

Post by SteveHansen »

One other note, Jeff. I don't see any kind of isolation/throttle valve at the diffusion pump's inlet.
User avatar
Richard Hull
Moderator
Posts: 14992
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: How to minimize backstreaming from diffusion pump

Post by Richard Hull »

You will definitely need a throttle valve between the diff pump and your chamber, as Steve noted.

I am currently unaware of any successful fusioneer who has ever used a cold trap and no one has announced using even an optical baffle, though these are sometimes part of a surplus diff pump's assembly. The key is to use the best fluid you can afford and not to make any mistakes during operation and shutdown of the diff pump. Check this forum's FAQs for proper operation in a fusor environment and a lot more data on diff pumps in general. (relative to the fusor of course)

viewtopic.php?f=25&t=4142#p21550

I see you have a water cooled pump. This more or less lashes you to the yard arm of having a continuous flow of water from the mains with associated drain or developing a cooled, closed circuit, recirculating water system. Most fusioneers, (I think 100%), opt for an air cooled diff pump system which is a time and effort saver.

We are not vacuumists, (vacuum heads), which has a concomitant purity clause associated with the rite of passage. We can have contaminants and other garbage in our vacuums that would frighten the vacuum purist.

We use the diff pump to get somewhere in the warranted 10e-5 torr range. To a deep vacuum guy this is "filth city". Since we work in a flowing gas, (deuterium), environment in the 10e-2 range, the filth is only 1 part in 1000 or, the fusion fuel out numbers the filth 1000 to 1. This is a highly acceptable level of operation.

So, bottom line, start with as clean a system as possible, use the best diff pump fluid you can buy and try for an air cooled pump and you are pretty much past any vacuum issues in fusor operation. We are here to do fusion and not spend a lot of gear grinding time and anguished moments over an ancillary vacuum system. Give your vacuum system a lot of attention up front and you won't be fiddling with it unduly during operation.

Some folks think I am in charge here. Far from it. What knowledge I impart in this amateur fusion quest comes from 14 years of being "lashed to the wheel", having personally committed all the sins and made all the mistakes, poking around in the fog of amateur fusion. We all have fashioned this site through a good deal of trial and error, learning and sharing as we went along the path.


Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
SteveHansen
Posts: 294
Joined: Sun Aug 05, 2001 8:37 pm
Real name:

Re: How to minimize backstreaming from diffusion pump

Post by SteveHansen »

I'll agree and disagree here. True, the vacuum requirements for a fusor are pretty trivial in terms of pressure and contamination control. Far more critical are the niceties of operation: it's very easy to really mess up a diffusion pumped system if you stall the pump. Proper throttling and control of the gas feed are very important.

That said, if better practice can be achieved by simple measures, then incorporate them. Most of my vacuum needs do not significantly exceed what's needed of a fusor. But, I don't leave finger prints on the wetted surfaces of components, I don't own any vacuum grease and I always make sure that I can't see the top jet of a diff pump from the chamber.

But then, I'm a vacuumhead.

Steve
User avatar
Richard Hull
Moderator
Posts: 14992
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: How to minimize backstreaming from diffusion pump

Post by Richard Hull »

My admonition about starting with a clean system will cover the neatness and the throttle valve will cover the optical exposure of the jet stack to the chamber while in operation.

Note: The fusor is always operated and gas flow only attempted when the throttle valve is barely cracked open, (jet stack totally invisible.)

I have my right angle throttle valve mounted directly to the Diff pump and then take off at a right angle a 6 inch long 1.5 inch diameter SS flex hose in another turn onto the chamber. No one working a fusor should go from pump to valve to chamber in a straight stacked fashion unless they are looking for trouble.

Again, at the higher vacuum levels we are looking for some naturally installed reduced conductance. After all, in operation, we will have about a 1/32 inch or less opening diff pump-to- fusor chamber as we strangle the diff pump to where the micro slow gas leak into the fusor is maintained at about 8-20 microns; about where most mechanical pumps are bottoming out. This is why you can certainly operate a successful fusor with just a very good mechanical pump coupled to a zeolite trap, provided you are willing to waste a little deuterium at start up.

Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
SteveHansen
Posts: 294
Joined: Sun Aug 05, 2001 8:37 pm
Real name:

Re: How to minimize backstreaming from diffusion pump

Post by SteveHansen »

Violent agreement - all good approaches to proper practice.

In the industrial world a lot of people question why throttle valves, even for large pumps, are only spec'd from just cracked to about 10% open. Max open for control purposes may be well under 10% open or less than 5% of the pump's rated speed. This is simply because you only need full pumping speed to reach base pressure and then you just want to sip gas during the process cycle.

Same with a fusor, you need to throttle the pump way down so that you keep the pump inlet in molecular flow, the chamber is at the right pressure for the plasma and you don't eat up your D2 supply.

Q = S x P rules.

Steve
Jeff Robertson
Posts: 58
Joined: Tue Jul 19, 2011 11:58 am
Real name:

Re: How to minimize backstreaming from diffusion pump

Post by Jeff Robertson »

Steve,

I was under the impression that I could just have a valve between the roughing pump and the diffusion pump, and I just wouldn't turn on the diffusion pump (or open up the valve) until pressure levels were low enough. Should I remove the valve between my roughing and diffusion pumps (as seen in my diagram) and just put it between the diffusion pump and chamber, or should I have valves in both places (making 3 valves total)?

Additionally, thanks for the aluminum baffle idea. The next time I have access to my university's metalworking shop I'll see what I can whip up.

Richard - Yeah ideally I wanted an air cooled diffusion pump, but I got a hold of this pump so I've decided to make do. I already have a water pumping system installed and it does it's job, so it'll suffice for now. Additionally, if I understood you correctly, you said that a right angle valve in series with a hose will impede the oil vapor enough to not require a baffle at all? As long as you keep the valve barely cracked open.

Based on the ideas you guys have been discussing, I attached a modified image of my vacuum system. Does this look more suited for the type of vacuum in a fusor? Also, is the valve between the two pumps (I marked it in red) still needed?

Cheers,
Jeff Robertson
Attachments
vacuumsystem2.JPG
SteveHansen
Posts: 294
Joined: Sun Aug 05, 2001 8:37 pm
Real name:

Re: How to minimize backstreaming from diffusion pump

Post by SteveHansen »

Hi Jeff,

You absolutely need a variable conductance between the pump and chamber for pressure control purposes. Use of the other valves depends on how you want to operate the system.

If you just want to pump straight through (no roughing or foreline valves) then you will just have to wait for the pump to cool down before opening the chamber. Adding a foreline valve (mech pump to DP) and a roughing valve (mech pump to chamber) will permit you to isolate the DP while hot so you can open the chamber.

You need a gauge to monitor the chamber pressure so that you can do the crossover at the right time.

There is a downside to having a 3 valve system - it's simply easier to screw up and open the wrong valve at the wrong time. The proper initiation to manual sequenced DP operation is to have to clean up the resulting mess. I favor the single valve approach as 1) being over 60 I get distracted easily 2) I'm not a production shop and 3) I can always find something to do while the pump cools.

The isolation/throttling valve I have on my little Varian pump is an inline KF40 bellows valve and there's no line of sight through it. I could easily get away without any additional baffling but the pump has a built in baffle and there's no incentive to remove it.

There's no need for a hairpin turn in your pump's inlet line. Put the valve there.
User avatar
Richard Hull
Moderator
Posts: 14992
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: How to minimize backstreaming from diffusion pump

Post by Richard Hull »

Make do with a water cooled pump?!! A water cooled pump is ideal if you have the flowing water on hand, and you do! Make use of it.

If you want a fusor, you absolutely, positively have to have a throttle valve between the fusor and the top of the diff pump.

Finally, you have absolutely no need whatsoever for the line, "tee" fitting and valve directly to the chamber from the fore pump. Replace all that in your diagram with just the vacuum gauge. For fusor use, you only should use two valves at most.

I usually let the foreline and fore pump up to air after all valves are closed and all pumps are turned off by undoing and lifting my KF clamp and fitting on the forepump for a second and then re-tightening it immediately.

Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
DaveC
Posts: 2346
Joined: Sat Jun 30, 2001 1:13 am
Real name:

Re: How to minimize backstreaming from diffusion pump

Post by DaveC »

I hope this doesn't confuse things, but Jeff's last sketch seems to me to be almost the ideal arrangement of plumbing.

From my experience with a variety of vacuum systems, it is just about mandatory to have a valved "roughing line" to the device being evacuated. Rough pumping through the diffusion could set you up for indavertent contamination of your very pricey diffusion pump oil. Granted water will eventually exit because of the high temps in the diff pump boiler, but I prefer to coax as much water out as possible via the roughing line and gas ballasting - even N2 flushing, if needed.

As has been already mentioned, you only need to run with a wide open high vac valve position until the chamber has become rid of the adsorbed gases and such.
Then by reducing the high vac valve opening until the pressure just stabilizes, one gets a good idea how much cleaning remains to be done.

Although it's definitely overkill, I often used my storage oscilloscope to monitor the ion gage output current. The line slope, on a 50 sec . per trace sweep time, gives a good visual of how well the system is cleaning. (I find that watching a vacuum system pump down is one of those high tech. time wasters that is also quite relaxing.....)

For extended operations, (like 6-8 hrs straight) I typically back off the High Vac valve as just stated and also partially close the fore line valve (between the diff pump and the roughing pump). It doesn't seem to affect the pumping rate at that point, and it gives a psychic comfort in appearing to aid in minimizing backstreaming.

My system is a vintage (read that as old) bell jar based unit that has a water cooled diff pump with a water line traced inlet pipe between the High Vac valve and the diff. pump body. System came that way and it seems to keep the back streaming to a very low value. I get very nice metallizing results, which seems to corroborate that little backstreaming is occurring.

I have done some demo fusion-like operations, but use extreme caution to avoid any beam impingement on the jar walls. The system has a separately valved inlet to the base plate for gas inlet, venting or whatever. This allows N2 flushing, or D2 inlet...

Between the throttle action of the High Vac valve and the inlet leak valve, a reasonably good pressure control is possible. But as we've all discussed before and elsewhere, this is a delicate balance, whose success depends mostly on leak valve quality and having the properly low enough gas pressure for the gas inlet. The use of a fine capillary greatly aids the control process.


My water cooling system uses a simple backyard fountain pump in a 5 gal bucket, with an air cooled recycled auto heater as the radiator. Works very nice... only maintenance issue is replacing water from time to time due to evaporation and the dog drinking it (!).... that and to flush out the bugs that wander in....

FWIW.....

Dave Cooper
Post Reply

Return to “Vacuum Technology (& FAQs)”