Electron gun- what's wrong here?

For the design and construction details of ion guns, necessary for more advanced designs and lower vacuums.
User avatar
Steven Sesselmann
Posts: 2060
Joined: Thu Aug 11, 2005 1:50 am
Real name: Steven Sesselmann
Location: Sydney - Australia
Contact:

Re: Electron gun- what's wrong here?

Post by Steven Sesselmann » Sat Aug 06, 2016 6:27 am

David, you should be measuring lower than -1000V after its rectified, and after the load resistor. Your DVM may not like voltages over 1000V and I wouldn't try measuring the 30 kHz AC side.

For this kind of measuring it's a good idea to make yourself a 1/1000 divider with a 1G/1M resistor, so 1V = 1000V

PS: Can you draw the circuit please.

Steven
http://www.gammaspectacular.com - Gamma Spectrometry Systems
https://www.researchgate.net/profile/Steven_Sesselmann - Various papers and patents on RG

David Kunkle
Posts: 284
Joined: Thu Apr 17, 2014 4:43 pm
Real name: David Kunkle

Re: Electron gun- what's wrong here?

Post by David Kunkle » Sun Aug 07, 2016 2:25 pm

Apart from occasionally going over 600V, I see now why the DMM wouldn't like it- 60 Hz vs 30kHz. My HV probe output is divided by 1000 already.
Attachments
circ 001.JPG

User avatar
Dennis P Brown
Posts: 1288
Joined: Sun May 20, 2012 2:46 pm
Real name: Dennis P Brown
Location: Glen Arm, MD

Re: Electron gun- what's wrong here?

Post by Dennis P Brown » Sun Aug 07, 2016 6:48 pm

HV probes work terribly in that low end range; while hand held multi-meters just tend to fail in that same range since it (due to various factors) tend to read lower voltages than the actual voltage, which quickly exceeds the rate level of the meter causing it to fail.

Why not just make your own divider as Steve said (two cheap, low power resistors and that's it - his example works fine) for the multi-meter? So very simple - just design the range for the the high end being 2000 volt range and you are both safe and handle this grey area that doesn't have a good easy to buy solution relative to hardware.

Rex Allers
Posts: 171
Joined: Sun Dec 30, 2012 8:39 am
Real name:
Location: San Jose CA

Re: Electron gun- what's wrong here?

Post by Rex Allers » Mon Aug 08, 2016 2:13 am

[Arrgh. I guess I flubbed submit. Second try on whole post.]

David,
I'm not sure what is going wrong but I do see one problem. For the output voltage this TDK CCFL supply can make, a single 1 KV diode is too low a voltage rating.

Rather than doing a lot of speculation, I decided to test. As I mentioned in an earlier post, I have a TDK CCFL module that looks like yours. A while ago I was doing some testing on a different supply and had the prototype I used with a simple rectifier and a divider for measurement. I reconfigured to attach that output to the TDK. Here are pic of what I used.
TDK Test pic.jpg
TDK CCFL with rectifier. Original test supply greyed out for clarity.
TDK Test 1.png
TDK test circuit. Simple rectifier and measurement voltage divider.
In the circuit diagram, on the left, I have shown the TDK module with its basic output circuit. I have bypassed the two caps by connecting to point 'A' to take the output directly from the transformer. On the right is my voltage divider for measurement. The main dropping resistance is 300 M ohm (three 100 M ohm resistors in series). The output is measured by a DMM across ~300 K ohm. Earlier, I had calibrated this ~300K value to give near exact 1 V out for 1 KV HV, or divide by 1000.

Testing the circuit with +5 VDC into the TDK module, I measured 1.26 V, so 1.26 KV output. I got about 1 KV out with about 4 V in. Therefore the 1 KV rating on your diode is probably exceeded a good bit. The diode I used is rated for 10 KV which is overkill in this circuit and probably drops the HV out a bit below what it could be.

The circuit you used is a voltage doubler which uses the caps on the TDK board. To verify it works the way I expected, I reconfigured the circuit and added a second diode to match your circuit. Here it is.
TDK Test 2.png
TDK test with voltage doubler.
It did work as expected and with 5V in, I measured 2.5 V or 2.5 KV out.

So I'm not sure if exceeding the rating of 1 KV for your diodes by 26% is enough to kill them or not. For each of the diodes in your circuit you should be using two 1KV diodes in series.

In your last post you mentioned frequency. Some notes I have from earlier say this TDK supply probably runs the transformer at about 50 kHz, but that shouldn't matter. The caps in either of these circuits should filter out most of that to effective DC on the output.

It wouldn't surprise me if there is someting going wrong with the probe you are using to measure, too. In the last posts you haven't said much about what it is and I didn't go back to look if you covered that earlier. Do you know what its input resistance is? For not overloading this small supply, it should probably be at least 100 M and higher is better. My divider in this test is about 300 M.

You didn't say anything about how you constructed the circuit, either. I built this test by soldering "flying" components above a ground plane. This isn't the kind of voltage and frequency to try with one of those white plastic pluggable prototyping boards.

So as I have shown, it should work. I hope something here gives you a hint why it isn't working for you.
Rex Allers

David Kunkle
Posts: 284
Joined: Thu Apr 17, 2014 4:43 pm
Real name: David Kunkle

Re: Electron gun- what's wrong here?

Post by David Kunkle » Tue Aug 09, 2016 2:32 am

Thanks for all the info everyone. Thanks for going to all that trouble, Rex.

Probe info: 1-40 kV DC, divider ratio 1000:1, input resistance 1000 M Ohms- nominal. Also checked it against my main HV PS and its meter. In good agreement from 200 to 3 kV. (Hard to adjust it below 200V.)

Seemed the easiest route to try first was what Rex suggested. I re-wired it with 4 unused diodes. (Happen to have bought 6 total.) 2 diodes in series in place of each one. This time I got a max of 1800V, but it drops steadily until about 3-400V in a matter of seconds when I'm measuring it with the probe, and then stays there even after I remove the probe, "let it rest", and test it again- until I cycle the power to the inverter.

David Kunkle
Posts: 284
Joined: Thu Apr 17, 2014 4:43 pm
Real name: David Kunkle

Re: Electron gun- what's wrong here?

Post by David Kunkle » Tue Aug 09, 2016 3:03 am

Thought I 'd check it over one last time. Found one questionable connection. Also, I was setting the DMM to V so I could mentally multiply the result by 1000 and not have to wonder what units it's in. Turns out this screws it up somehow. (It automatically goes to mV when powered on.) Between the connection and leaving the DMM alone when I turn it on, I easily and very consistently get a reading of exactly 2,300V DC! Looks like I'm out of the woods and learned a few things along the way.

I don't have an oscilloscope. This should be a fairly smooth, continuous DC output- not pulsed- right?
If I understand this right, each output from the inverter is 5 mA max. So 10mA if wired together like I have them? Will the max mA drop proportionately to the increase in V with the doubler circuit? And also, the V will drop if much of a load is placed on it? I suppose I could measure the V again when it's hooked into the filament circuit and running on the chamber.

Rex Allers
Posts: 171
Joined: Sun Dec 30, 2012 8:39 am
Real name:
Location: San Jose CA

Re: Electron gun- what's wrong here?

Post by Rex Allers » Wed Aug 10, 2016 6:40 am

Did it again. Typed for about half an hour but didn't save locally and it went poof when I submitted. Maybe I wasn't really logged in?

Shorter version...

Good to know you got some output results that look good.

I checked numbers on the module here and it is CXA-L10A. Same as you and Steven.

Found basic data here:
http://www.tdk-lambda.com/products/sps/ ... bodye.html
and you can download full datasheet from there. Datasheet has a configuration they call Connection A, which is the two output caps in parallel as for the doubler circuit. There is a section of data specific for that.

Datasheet says 900 Vrms into open load. With simple rectifier and light load we will measure peak voltage or 900 * 1.414 = 1273V. Very close to what I measured before.

Then data sheet then says typ. 10 mArms, but that is into about 27 kohm load. Ohms law says the output voltage at 10 mA is only 270 Vrms! Not uncommon: take your pick, voltage or current. Don't ask for both at same time.

I decided to measure again. Configured back to simple half-wave rectifier (see my last post). Found some resistors: 68K 5w and two 10K 3W. Put them in series; measured 86.6 kohm. Used that as load on supply. I got it cranked it up as far as 674 V out, but that was at input of about 5.2V at 1.5 A. That's pushed beyond rated input. Ohms law says output current was 7.8 mA ; about 5.24 W. The 68K 5W resistor got pretty warm in minute or so at this level.

So to get into the ~1 KV range at mA current, will require the doubler. Lets say input of 5V at 1A might be a safe steady state level. In my measurement the supply efficiency was about 67%. If that held, and we can get 1100 V out of the doubler, the current would be about 3 mA. But the doubler will have losses too. I didn't measure more, but I would think if you can get around 1100 V out, you'd be doing good if you can see 1 or 2 mA.

At no load I would expect very low 30 kHz ripple on the HV DC. With these mA current levels through the doubler, I would expect a lot of ripple.

Fun measuring. I'm actually pretty impressed at what these little switchers can do.
Rex Allers

David Kunkle
Posts: 284
Joined: Thu Apr 17, 2014 4:43 pm
Real name: David Kunkle

Re: Electron gun- what's wrong here?

Post by David Kunkle » Thu Aug 11, 2016 12:52 am

Thanks again for going to the trouble. 1100V and 1 to 2 mA is hopefully enough for what I'm doing. Floating the HV output on the filament shouldn't draw much current- maybe even in the uA's range? The only energy being used is to accelerate the e-'s away from the filament. I'll find out as soon as I find the time to get it hooked up on the chamber. Then hopefully start seeing e-'s fly and measure the final HV the circuit can put out with the filament on.

When I was getting frustrated with getting this circuit to work, I found some Universal Voltronic PS's and thought about getting one and giving up on this circuit. About the same vintage and make as my main HV PS (about 1/2 century old), but not quite enough V or A for a fusor. Could probably get one to my door for about $250 and I got something I could always resell. I'll keep this as my fall back strategy if the inverter comes up short of what I need.

User avatar
Richard Hull
Site Admin
Posts: 10470
Joined: Fri Jun 15, 2001 1:44 pm
Real name: Richard Hull

Re: Electron gun- what's wrong here?

Post by Richard Hull » Thu Aug 11, 2016 5:44 am

As an old vacuum tube guy and a guy who has worked in technical vacuums of 30 to 5 microns .......A good emitting filament close to a plate of opposite polarity in anything less than a 10e-6 torr vacuum might tend to draw a gang o' current (townsend avalanche) and drop a high impedance supply to its knees. I am sure you will report to us on this. Your vacuum level will be the factor here.

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
Retired now...Doing only what I want and not what I should...every day is a saturday.

David Kunkle
Posts: 284
Joined: Thu Apr 17, 2014 4:43 pm
Real name: David Kunkle

Re: Electron gun- what's wrong here?

Post by David Kunkle » Tue Aug 23, 2016 12:43 am

Everything has been thrown together on the chamber and tested (10e-6 Torr range). Best way I could figure to test it is with a central anode (main HV PS switched to positive output). The filament about 10 inches away. With + HV at about 3kV, and the 12V filament switched on, it doesn't draw any e-'s (Nothing on the main HV ammeter). Same conditions and turn on the inverter, and it shows several mA. If the + HV goes much higher, looks like a townsend avalanche- the mA's start going sky high fast, and the main HV PS trips off pretty quick.

FWIW, the accelerator ring that I built into the conflat is just sitting there- not even connected to ground.

With the HV probe, the inverter still shows 2300V with the filament off, but drops to a measly 100V with the filament on!
I get the same results (100V) without the central anode switched on. Apparently I'm getting the same number of e-'s ejected- they're just going to ground when they hit the opposite side of the chamber wall?

Something new also occurred to me lately. I have the filament transformer on the isolated transformer, but wondering about higher V's. Apparently 2300V isn't a concern. Say10kV was being put on the filament. The entire circuit the filament is supplied by is now floating at 10kV. Is there some point where there is danger of arcing to anything conductive nearby? Does that whole circuit have to be isolated/insulated to withstand the kV?

Post Reply