Zj-27 ionisation gauge controller for the people

[Robert Clarke]

The ZJ-27 is an ionisation gauge tube which is readily and cheaply available from Chinese suppliers. This conversation is an attempt to gather information and ultimately to establish a basic design for an easily reproducible, open source controller.

[Robert Clarke]

I would love if any folk with experience in this subject could give an overview of what it takes to drive this kind of gauge tube.

[Bob Reite]

Look at the spec sheet., you will find the filament requirements and other parameters there.

[Robert Clarke]

Hi Bob. I have those details to hand. I am just attempting to get a narrative going in the thread. Starting with an overview and perhaps some general comments form someone who has built a similar bit of hardware.
If there are no replies in a couple of days I'll copy the full specs here and hopefully get some good ideas going

[Aleksa Djokc]

First, the basics, this is a good start: https://avs.scitation.org/doi/pdf/10.1116/1.5025060

Seems like not that many people know what it takes to drive and read the ion gauge tube.

When I first started with BAG tubes I had no clue how they works or how to drive them. Unfortunately, there is zero information on the web about how to drive them or read the current output. I was lucky enough to find some very old BAG controllers and take a look how they did it.

The first and most simple task is to make a circuit that controls the constant emission of 0.1 to 10mA, so lets start from this point.

In my first design, I've used a simple constant current circuit with LM358 and a LED diode for the output, that LED was a part of TRIAC/DIAC/Photo resistor network that drove the filament supply. The set point was adjusted with a trim pot and a TL432 reference, the current was read from a point on the filament (shown in the provided link). By using a Triac to drive the transformer, I had some issues that were solved in a not so good way, first, there were a lot of oscillation because there was no hysteresis, so the filament output would start jumping all over the place until it settled down on the setpoint, after some time, it would break again, the solution was simple, add a capacitor in the feedback loop to slow down the reaction time, this worked, but the output, when first energized, overshoots and undershoots several times before it settles and stayed on the setpoint. A better way is to control the filament with DC, the output of the OPamp should go to a transistor that controls the filament. If done this way, the circuit must be isolated from the rest of the tube, it will sit at 30v bias voltage.

As I've posted the schematic in the previous thread, here is another one that is the same: https://i.stack.imgur.com/BtnzA.png

This circuit is for a constant current source for laser diodes. Instead of the diode, use the filament. The supply voltage (V3) should be set with an external power supply to the maximum voltage that the filament can be run at, just as a safety, in case the circuit doesn't regulate the output. The reference diode should be tied to the supply rail of the OPamp. The + input on the op amp is the setpoint input, this will be adjusted to give a constant emission current of 1mA, the (-) input of the op amp should be tied to the shunt resistor on the filament. The shunt resistor should be sized to give a voltage in the range of 1V/1mA, or there about(meaning the setpoint at the + input should also be in that range).

It may sound confusing at first, but when you get the hang of it, it will not be so complicated.

[Gustavo Cervantes]

I was alerted by Aleksa on this topic, there is a spanish forum which has a lot of good content regarding vacuum work, not sure how good the translation on google part would be, but here is a hot ionization gauge made y one of its members:

http://www.cientificosaficionados.com/f ... 5y5yNreK-g

Also someone made a cold cathode gauge:

http://www.cientificosaficionados.com/f ... hp?t=19438

There is a few missing images cause the threads are a bit old and the forum moved not long ago

[Robert Clarke]

Hi all,
Below is a quick high level comment on a matter which confused me in the past.

I began my vacuum journey with John Strong's Procedures in Experimental Physics. In this awesome book the ionisation gauge tube is described as a modified triode tube in which the glass bulb is connect to the apparatus to be measured. Easy right?

So what's all this talk of BA tubes? Well, there comes a point at very low pressures where the physical design of a modified triode tube causes a problem. Around 1x 10-5 the tube begins to create X-rays. The new charged particles created by the x-rays look to the 'collectrode' as the ionised gas current it expects to see; therefore at and below these levels of vacuum the readings are erroneous.

The solution? Turn the triode inside out! This is what Bayard and Alpert did circa 1950. The problem with the Olde world Triode was that it had a large surface area cylindrical collectrode. While this was great for collecting the ionised charges we deliberately create, it is also a huge target for the more random x-ray induced charged particles.
So Bayard and Alpert made the collectrode like a pin. This pin still had the bias charge which made it attractive to the ionised gas molecules being measured but was a much much smaller target for the x-ray induced charges. And so the gauge could continue to be useful at lower pressures.

So for clarity the ZJ-27 is in the older and now surpassed modified Triode construction. It is not a Bayard Alpert Gauge tube strictly speaking. However for our controller project I don't believe this makes any practical difference.

[Robert Clarke]

Hi all,
May I start by thanking you all for your responses so far.
In an attempt to keep my solutions as simple as possible (off the shelf) may I ask if the following is feasible.

Dealing only with adjustment of Filament power supply for a moment. The old literature has a variable power supply to filament and the mA gauge on the grid/anode. No active feedback at all!

Can one simply pump down to 10-3 or whatever is safe, and while watching the grid current gauge, just trim the voltage and so current through the filament?
Example: I have a 3.3v 6A SMPS. These are very common and easy to find. It has a small degree of adjustment possible with a trim pot on the front. Can I use a power resistor to get close to the 1.3v across the filament and tweek the SMPS that until I see 1mA on my grid gauge?

I realise this is horribly wasteful in terms of Wattage but these parts come from China and any waste I should be worried about has already happened flying them halfway around the world!

[Aleksa Djokc]

Safe vacuum is in the range 10-4mBar and below, these tubes probably use pure tungsten filaments, if they were yttria coated iridium then you would be able to go up to 10-1mBar.

You could control emission current like that, but you will immediately see why you need a circuit to do it for you. Just be careful with the voltage, try not to go above what's specified in the datasheet.

[Robert Clarke]

I accept I may have to crawl before I run!
The Zj-27 is described as having a Nobel metal oxide cathode so hopefully I have not already burnt it out.
Aleksa, I do not doubt your advice is correct. And I'm sure I will ultimately use active feedback control but my learning is slow and takes much energy. Every little success pushes me on to try a better solution.
Tomorrow I will check anode current with my 'voltage divider' plan. See how it varies with changes to the SMPS trim pot. I will report back.

[Robert Clarke]

I just tested with my Zj-27 still sealed in it's glass tube. I was under the impression they were shipped at a very high vacuum but it seems this may not be so.
Regardless I pumped full wack through the voltage divider as described above. I saw 30mA on my analogue gauge attached in the 200v anode/grid circuit. EEk!
I trimmed it and at 1.03V across filament I was seeing 1.5 mA on my analogue gauge. At this voltage the filament was dull red.

Official specs of the ZJ-27 are as follows:

Filament 1.3V @2.5A
Filament to ground 25V
Gate/Grid/Anode to ground 225V/200V (depending on the source of info')
'Collectrode' to Ground 0V

What I assume is de-gas 10V/3A

More to follow...

[Robert Clarke]

Moving on to the 'Collectrode' end of the gauge, previously Aleksa suggested the LMC622 as a suitable OpAmp to sense the ion current.
This OpAmp has an Input Bias Current of 2-4 pA, if I understand the datasheet correctly. Aleksa tells me this is the critical parameter to look for in this situation.
Also a high quality resistor 100m to 1Gohm is required I hope to find more details on this requirement shortly. I will order am OpAmp or two and see what I can see.

[Alex Aitken]

Have you seen https://github.com/m-labs/ionpak btw?

[Robert Clarke]

Alex,
Thanks for that link I had not seen it.
They state that the ZJ-27 is shipped at 10-5 to 10-6mBar. If that is consistent with all the tubes manufactured then I think testing with the tube still sealed should be safe.
The pdfs of their circuits also have great information. I will prototype with the Lmc662 TIA style circuit for now to assist my learning. Once I have some clearer understanding of how the straight forward circuit works I will revisit the pdf and see how they implement it.
I intend to aim for somewhere 5v to 0v output from the opamp to hopefully interface with an Arduino ADC or even a moving coil meter.
I think the Lmc662s will not arrive until early next week so in the meantime I will see if I can see ripple in my high voltage smps.
I will also rebuild my power resistor to a slightly higher resistance
Thank you all again for assisting my project

[Robert Clarke]

Aleksa may I ask you some questions regarding your description of the laser diode CC supply circuit?
Is this adjustability necessary because once the tube is installed conditions might change and emission current will be different?

I have redrawn the circuit of the laser diode. if you have time could you tell me if it looks correct to you?

Many thanks
Rob

[Aleksa Djokc]

The first thing you learn when starting with high vacuum is how much everything is unstable and outgasses like you wouldn't believe. The stability of the emission current will never be stable on its own, it will always move around depending on many factors. The only thing, for now, what you need to worry about is the filament temperature variation and gas load. The schematic you have drawn is indeed just a constant current source that takes feedback from the shunt resistor. The feedback point is not from that resistor but from a resistor between the anode and cathode. What the BAG tube needs is a constant emission current with a varying filament current that adjusts it self for stable emission. So the feedback point needs to be between the cathode and anode. Take a closer look at the schematic of the BAG tube.

[Robert Clarke]

Aah! I did not realise there was should be a shunt resistor between Anode and Cathode on the BAG tube. TBH the only thing I have seen in my research between the Cathode and Anode is 180v-200v! I'm sorry for asking silly questions, for you this must be very obvious.
I have been looking at the Github repository and the Ionpak circuit diagram but I can not see where the anode/cathode shunt feedback comes from. I am trying to piece together information from various places to learn.
Perhaps this project is just beyond my understanding at the moment. Tomorrow is a new day. Thank you Aleksa.

[Aleksa Djokc]

For a beginner, the IonPak is very complicated, personally, I don't like how they did the layout, very confusing and hard to understand. If you have the power supplies on hand, and you have made the CC circuit, just don't connect the feedback point to the shunt on the schematic, instead, connect it to a shunt resistor between the anode and ground. Keep in mind that you must make isolated power supplies, they cannot share the same ground reference. Filament power should be provided from a high quality transformer(good isolation), 180-200v can be supplied from a SMPS, 30v filament bias is not critical. If I saved my files from proteus, I would have gladly posted the schematic...

[Aleksa Djokc]

Here is a quick sketch, the component values are not real, don't use them as a reference. The concept should be ok, this is the most simplest way I could think of. If you get any oscillation, you will need to use a capacitor in the feedback (integrate).

[Robert Clarke]

Many, many thanks Aleksa. I'm so grateful for the time you put into explaining this for me and doing the sketch. I will take some time and study it. I think part of my problem in understanding this situation is that I have not reconciled in my mind how the voltages and their respective currents exist in the circuit and tube.
I have been reading so much which just refers to current without mention of the voltages involved, which is OK until I need to figure out values for components!
Is it safe to think of the filament-grid electron flow as an equivalent resistor? like 180/0.001=180,000Ohm?

[Robert Clarke]

Also I must think of this in the terms of the isolated power supplies. I have been basing my understanding on very simplistic power supply diagrams. TBH isolated power supplies within the one circuit is new to me! So I will think about this information for a day or two and breadboard something.
BTW my god the LMC662 is sensitive! I did a quick circuit and I could not move in my seat but it would register! Fantastic!

[Richard Hull]

I have used the LMC 6041 for many years. 2 femptoamp bias.
My cat walking across the room 8 feet away with a 7 inch antenna on the input makes it go rail-to-rail.

https://www.ti.com/lit/ds/symlink/lmc6041.pdf

Richard Hull

[Aleksa Djokc]

I would strongly recommend checking out some of these YT channels in order to get some deeper understanding of how the electrons flow in a vacuum environment.

Applied science, glasslinger, the thought emporium, mr carlson lab, Eric Dalgetty, jdflyback, Signal Ditch...

Also don't forget to do some reading: http://www.tubebooks.org/

The thing about isolation, as you can see from the sketch, there is a return path from a separate power supply that will introduce 30v bias in to the feedback of the op-amp, I have avoided this problem by running the filament from a transformer(AC), that gave me the needed isolation. So, in theory, the schematic I drew may not work correct, instead, there should be a SMPS driving the filament and feedback would just adjust the PWM, the transformer would provide the isolation and allow for biasing. Try it anyhow, just limit the input voltage to the filament...

Electrometers are a lot of fun, they can be used for many interesting experiments and measurements. Keep in mind that when dealing with such low currents, there is a whole science behind shielding and guarding the path that needs to be measured, surface leakage currents, material leakage, loading impedance, filtering, etc....

[Robert Clarke]

Thank you both!
I just read the first half of 'Inside the Vacuum Tube' by John F. Rider. It was very entertaining. I got the book for this very project. I got as far as Pentodes and thought that's as much as I can handle for now!
I keep thinking back to what is the simplest solution for me to get this working. So I might start with an analogue meter for electron current with manual trimming of the filament power supply and an LMC662 collector circuit.
I realise this is very clumsy but small wins help keep the momentum for me.
How would one estimate the ion current in order to determine the correct components?
Am I correct in assuming the voltage of the ion current is the bias voltage?

[Aleksa Djokc]

The output of the collector is measured in current only, voltage is not measured. If you are using the tube as a it arrived sealed, if it is sealed under high vacuum, expect the current to be in the range of nA to pA. For example, 10-4mbar is 1uA, 10-5mbar is 100nA, 10-6mbar is 1nA etc...

Here are some more good reading materials...

https://sci-hub.se/https://doi.org/10.1063/1.4869873

https://physics.mff.cuni.cz/kfpp/skript ... ratura.pdf

[Robert Clarke]

Thank you for those links. I will read them as soon as I am at my desktop.
In the meantime I revisited the basics of the opamp circuit for sensing. And I now understand what you said about the measurement being in current only.
As the circuit is a transimpedance amplifier, the voltage I was so concerned about develops across the feedback resistor. Silly me.

So the feedback resistor is chosen based on expected current and desired output voltage.

Does this sound correct?

[Aleksa Djokc]

Yes, that is correct.

Start with something in the range of 10 to 100M ohm, that will give you some room for experimentation.

Just to mention a little situation I had with measuring low currents, when I was cleaning my leak detector, it had a metal-ceramic feedthrough that carried the collector current to the outside world, directly above sits 5889 electrometer tube with a glass sealed 1G ohm resistor. After degreasing, acid clean, water rinsing in lab grade demineralized water for half hour, acetone and IPA, I dried the whole detector cell in a dry nitrogen environment at 60c. When I reassembled the whole unit, of course it had problems, the output from the electrometer was always pegging the needle, this was the result of not cleaning the ceramic feedthrough enough(imagine not being "good enough" after all that I mentioned above). The solution was to again heat up the whole cell, dry it under vacuum and in the chamber where the tube lives, purge it with freon R114 that eliminated all the moisture that would eventually come back and affect the measurement. This story should give you some idea of how cleanliness is very important, and that low levels of current flow in very delicate way. Keep in mind this is picoAmps range, so not that dramatic, imagine going femto..

[Robert Clarke]

Aleksa,
I could imagine going femto must be quite the trip!
Last night I salvaged a filthy 10 Mega Ohm bleed resistor from a microwave oven capacitor,

It was covered in some mysterious oil which I scrubbed off with IPA. When I installed this in my LMC662 TIA circuit it immediately became very sensitive to me touching the croc clip jumper wire. This is how I am currently connecting the 'collectrode' to the opamp. I realise this is the most horrible situation in these circumstances. Especially that it is part of a rats nest of wires in my prototype. Induced currents a go-go, I'm sure.
At 10Meg my results have been variable. The clearest indication I am seeing is a difference of 50mV on my output. This being most obvious when I disconnect power to the tube. It jumps back up to the 'off' voltage.
So, my thoughts are: This is induced current in my jumper wire OR it is working and the tube is at high vacuum and I should be using 100meg or Gig Ohm resistors.
Pesky work got in the way over the last few days. I will order some high value resistors and see what I get.

[Rich Feldman]

Nice resistor recovery, Robert! A resistor with an honest kilovolt rating. Does your picture also show if MO capacitors are designed to break connection between terminals and the foil roll, if lid bulges from overpressure?

The talk about femtoamps, and Richards' comment:
>>My cat walking across the room 8 feet away with a 7 inch antenna on the input makes it go rail-to-rail.
reminded me of Wikipedia page about "Electrostatics".

Cat_demonstrating_static_cling_with_styrofoam_peanuts.jpg (53.3 KiB) Viewed 2823 times

[Richard Hull]

My transimpedance amp in image below. It is built in a tiny Pamona BNC component box using the LMC 6041. Data sheet URL in my post above. This is the item and its stub antenna mentioned that is hyper sensitive to the outside world of charge exchange between it and my cat at range and the tapping of my toe on the carpeted floor 6 feet away. Charge exchange in motion in a world filled with electric fields we can't sense.

I built this little guy back in 1995. I have been on this for years as just another one of the HEAS experimental demos and sidelines in electronics. It makes one think long and deep. Nothing beats the doing, and then some deep thinking. All to weigh and consider. It can be considered a cool toy by those with little depth of thought not use to stopping to smell the roses.

Open up a new world for $2.64 from digi-key

https://www.digikey.com/en/products/det ... OPB/364306

Hook the thing up with 4 coin cells +/- 6 volts to ground as a simple unity gain amp use no other components, no IC socket no pc board, float in air with antenna hooked to the positive op amp terminal. Amazing!

Note: a zeroing attempt is a futile gesture if ultimate sensitivity and a true study of the outside world is the goal. If you must, install a highly insulated antenna input to ground shorting button to temporarily zero. (this effort on your part will be a miracle in itself). Needless to say never touch the antenna unless you first touch the circuit grounding first...(battery supply center point....usually case metal shell, if you enclose this thing like my little Pamona box.)

Place this hooked to a scope via a long enough cable to isolate it from large grounded objects. Use a lunch box type O'scope with an LCD display that retains it trace image as a fading trial on .2-.5sec/division and marvel in experiment.

Richard Hull

[Robert Clarke]

Rich and Richard you can rest assured my cat will be tested with the electrometer once I have it constructed. No more secrets Alberta.I await 100meg and 1 gig resistors.

Would I be correct in assuming that a single rail power supply will be ok for Ion current sensing as it can't go negative...or is that positive...I'm starting to mix up conventional current, electron current and Ion current and photoelectric current!

[Robert Clarke]

Rich,
Regarding the MO capacitor. The lugs in the photo have a tiny spot welds to the tabs of the foil in the dielectric roll. The lid came away without any pulling or distortion on the tabs so I assume they are 'fuses' of sorts. Is it even possible spot weld aluminium foil to iron? Maybe there is more going on here than I can see!

[Richard Hull]

Current is current regardless of any classy names we ascribe to it as regards to the mechanism from which it springs into measurable existence. Polarity and direction of current flow is also just a handy way to try and understand a special case. Current flows whenever an extant electric field is disturbed via some form of motion either in air or a conductor. All charge in this marco world be it in a battery or between fields is strictly due to the electrons establishing a field condition either by an over abundance at a point or a forced lack thereof at another point. Chemistry is 100% electronic in nature.

Richard Hull

[Robert Clarke]

Hi Folks!
Here is a little update of progress. My high Ohm resistors arrived so here is my circuit in progress...

This will eventually be housed in a metal surround, probably a copper pipe. The glass feedthrough in the picture is from a bulb that was over the bathroom sink. I hope this will isolate the signal enough.
In the meantime I have built a manual adjustment for electron current.

This is just the existing trimmer VR from the 3.3V power supply brought out to the front od the enclosure. There is a nichrome wire resistor in series with the filament of the ionisation gauge tube.

I know this is very ham fisted but until I learn more this will have to do for now!

[Robert Clarke]

A small progress report. This is my electrometer circuit as of this evening

Electrometer Faraday cage

The casing was cast by myself and friends over the weekend. The model was a 3d print.

Let's see if it works! More to follow

[Aleksa Djokc]

The glass feed through is a bit of an overkill, a typical BNC or other similar connector would suffice. A typical RG58 coax and alligator clips at the and are good for single digits pA. I would expect the ZJ27 is sealed somewhere between 10-5 or maybe 10-6mbar, in that case, the range is in nA/pA, so 100M might be too much, it may overload the input and cause the output to swing the power rail, so keep that in mind if you see such behavior.

[Richard Hull]

I wonder...........Is this latest posting about building an electrometer to look at the electrical world around you? If so that is fabulous.

You original post was about a vacuum gauge issue and a replacement for a formal controller.

Work for a purely informative electrometer of extreme sensitivity, (Fempto ampere work), would best be posted in the neutron/radiation detection forum as a very sensitive electrometer is the basis for an amateur construction of a very sensitive ion chamber, radiation detection system where 10e-14 ampere currents are the norm.

Richard Hull

[Robert Clarke]

Thanks both for the replies,

Aleksa, thank you for the guidance on what to expect from the output. I have yet to power it up and connect the ZJ27. I have 12dc available with the enclosure so I will use that. TBH I'm not sure what I'm doing with respect to voltages. Its a little frustrating that the only 'variation' I have is when the tube heats up or I fiddle with the filament power. I will report back.
The glass feedthrough is partly a philosophical choice. We are surrounded by such wonders in our everyday objects. These Dumet wire seals in bulbs are amazing. I suppose I wanted to celebrate that a little. Glass to metal seals are very curious to me.

Richard, I am certainly inspired by your stand alone electrometer. And my enclosure is a homage no doubt. But primarily its just to try ensure isolation. The one in the image above is currently mounted within the plastic enclosure seen furthur above. I'll post an image when I have it tidied up. I intend to reuse the pattern and cast another enclosure for a hand held version. Seems like a fascinating device to play with in the wild. It is deliberately scaled to allow for CR2032s or so as a battery pack like the one you describe above.

[Richard Hull]

All the best in this project. You will be amazed if the amp you are using has a bias in the fempto-ampere range. Pico amps are fine, but nothing shows the electrical world as being alive like a hyper low bias amp. I like the glass insulator, make sure to wipe it off with alcohol to remove grease from hands. Do not use an input resistor. Just let the input lead float to the insulator. You are looking for sensitivity not a calibrated response offered by a teraohm resistor effectively shorting out the 5-10 teraohm natural input of the FET gate.

Richard Hull

[Robert Clarke]

In the interim and for the enrichment of the community I post a photo I took of the electrometer part of the Varian 880 gauge controller. If work stops getting in the way, I will progress my build some more.

That transistor is a 2n3904 of all things. Curiously with the base and collector tied to that post. (you can't see from that angle but there are two legs there)
If there are any relevant observations, feel free!

[Aleksa Djokc]

Yes, that indeed looks like a typical ion gauge electrometer design. The transistor is there just as a protection, its used as a diode, the input of the OpAmp is very sensitive to ESD. AD545 is a very good device with only 1pA input bias current, impressive for the time...

[Aleksa Djokc]

Typical voltage that gets out of the ion gauge collector is up to 10v DC, so the 2n3904 is used as a zener and reverse protection, in zener direction it clamps somewhere between 8 to 9V and in the other direction a typical 0.6V,

[Robert Clarke]

*Another aside*
This zener effect is very interesting. A circuit I have used a few times in work has been this:
http://cappels.org/dproj/simplest_LED_f ... rcuit.html

Am I right in saying it uses the same effect in the transistor?

Aesthetically this flasher is very pleasant, analogue looking. Giving a nice fast dimming, with no PWM, which is good for camera.
I will be back on topic next post, I realise this is meandering.