Electron gun power supply issues: Stumped by a current meter

[ChristofferB]

Hi!

I'm working a bit on electron/ion sources and there's one aspect I can't get my head around.

Let's say one wants a simple electron gun:

 - A tungsten filament with 6V DC and then a plate with a pinhole infront of it.

To extract electrons, either:
 - the plate needs positive bias (let's say 200V), and the filament grounded,

or:
- the filament at -200V and the plate grounded. This is what's usually done.


So how does one go about designing a PSU where a high current low DC voltage can be offset to a high voltage?

Aside from using batteries and letting them be at HV potential I can't figure it out.

Ideally I want to have a bias box where the filament DC current and voltage can be controlled, and then the filament output can be offset by an external HV PSU.

[Siwastaja]

Aside from batteries you mention, isolated transformers are a trivial way to produce isolated voltage sources that can be arbitrarily connected in series or whatever.

Some capacitive coupling happens in transformer windings though between primary and secondary, sometimes this is an issue.

Such transformers can run at mains 50/60Hz directly, or be part of higher frequency switch-mode converter, in which case physically smaller transformer does the job.

[ChristofferB]

Ok so one could use an isolation transformer to power say, a small OEM switchmode psu and then have that insulated from the world by ceramic standoffs or similar?

The isolation transformer would need to be capable of dealing with as high voltage as one wants to bias, i guess?

[ChristofferB]

..or get an isolation transformer that steps down to the voltage one wants, or have a variac infront.

Actually, these small potted PCB transformers has primary/secondary isolation of up to around 5kV, so I guess having 1kV DC on the secondary wouldn't really be an issue. Perhaps I'm overthinking this.

[Siwastaja]

HV is difficult, and it's dangerous, too. You need to really understand the insulation ratings and capacitive coupling of the transformers and power supplies.

Variacs especially are usually completely non-isolated.

[ChristofferB]

Agreed, this is why I want to use standard commercial HV PSU's that are current limited in the 100's uA range instead of just ripping the entire E-gun and PSU assembly out of an old oscilloscope.

The variac I was considering to put before the isolation transformer for a variable filament voltage.

Another issue is whether I can just assume the HV PSU doesn't mind having 3-12V AC fed into it. The HV PSU's I'm going to use are basically just inverter/transformer/CW-multiplier to output, fairly standard.

[TimFox]

"Filament transformers" were so named since they were used to power filaments or heaters in vacuum tubes, including rectifiers where the cathodes were at high potential.
As a typical modern example, The Triad F6-12 (rated at 12.6 V CT and 2.5 A with 115 V primary) is rated for 2500 VAC insulation primary-to-secondary and 1500 VAC insulation windings-to-core.
The same series includes lower secondary voltages.
Either heat your filament with AC, or use a simple rectifier (carefully insulated) from the secondary of such a transformer.  Adjustment can be done with a Variac in the primary. 

[ChristofferB]

Filament transformers, nice, that's the buzzword I was looking for, it seems. Perfect, thanks!

The F6-12 might be ideal.

I have drawn up a quick concept for a circuit, I'd really like some feedback!

[NiHaoMike]

You shouldn't need a variac and a current limiting resistor, just the variac by itself will do. An adjustable end stop for the variac will let you use the variac itself for soft starting while maintaining previous settings. To enforce soft starting using the variac, you could have power switched using a 3D printed cam on the variac shaft operating a microswitch.

If you want to be able to use higher voltages than what commonly available transformers are rated to, you can make your own transformers with insulation able to withstand the higher voltage (obviously, you have to know how to properly design it, it would be much easier to do it at a somewhat high frequency on the order of kHz to tens of kHz) or have a motor turning a generator with a plastic rod.

[TimFox]

It's not a bad idea (but not necessary) to use a series resistor approximately equal to the hot resistance of the filament (it does waste power) to moderate the inrush current at power-on.
However, a larger value can be placed instead on the primary side of the transformer (note that the resistance there is N² higher there due to the transformer turns ratio N:1) for improved safety.
AC on the filament is usually considered to be better than DC for various obscure reasons, so long as the application is not hum-sensitive.
Back in grad school, where we needed to power a control dome sitting at about 100 kV, we made our own isolation transformers from two Variac cores and a few turns of RG-34/U high-voltage coax (grounding the coax shield at only one end) to give something like 10 or 20 VAC for distribution within the terminal dome.

[ChristofferB]

Ah, that makes sense. The filament current is proportional to the variac voltage anyway.

The filaments will mostly be broken bulbs. maybe up until 2A would be nice.

A variac preset would actually be nice for repeatability.

I'm not sure winding my own transformer would be worth the hassle - there are quite a few canned mil-spec filament transformers on Ebay. If one wants higher voltage one could just do a two-stage acceleration system where maybe -1000V extracts the electrons from the filament, and then a completely isolated electrode with +10kV accelerates them.

[David Hess]

So how does one go about designing a PSU where a high current low DC voltage can be offset to a high voltage?

Take a look at how later analog oscilloscopes did it.

Older oscilloscopes used a big line voltage transformer with many isolated 6 or 12 volt windings for various filaments.  Newer ones did the same thing, but with a high frequency transformer as part of an inverter.  Regulation is accomplished by monitoring either the primary side peak voltage, or the peak voltage from one of the secondaries, commonly the high negative voltage secondary since this one was critical for deflection accuracy in an oscilloscope, but it could be any.

For a simple modern implementation, I would use a standard pulse or inverter transformer which are available with various winding ratios.  If an AC output is acceptable, then that is all that is needed.  For DC, a bridge rectifier on the output makes DC, and then diodes are added to the feedback network on the primary side to compensate for their voltage drop.

[james_s]

The vast majority of electron guns I've dealt with run the cathode at ground potential and use HV on the anode, I mostly worked on CRT displays.

Xray tubes typically have a split HV supply with a grounded center tap and the filament is powered via a special transformer that has very high isolation between the primary and secondary.

[David Hess]

The vast majority of electron guns I've dealt with run the cathode at ground potential and use HV on the anode, I mostly worked on CRT displays.

Oscilloscope CRTs use a positive anode and a negative cathode because this allows the electrostatic deflection plates to operate at ground potential.

[james_s]

Oscilloscope CRTs use a positive anode and a negative cathode because this allows the electrostatic deflection plates to operate at ground potential.

Yes but I've only rarely worked on those and never had to mess with that area of the circuit. The vast majority of the CRT displays I have worked on and am familiar with are magnetic deflection.

[jonpaul]

Bonjour, CRT filaments have been run at high potential with an HV isolated transformer wdg on the FB,  t 1-5 turns 20 kV wire, also specifically insulated mains transformers.

A few portable scopes (EG Mason Engineering)  even had batteries to run the filaments as well as the 1960s Scientific American Amateur Scientist column on a home built Van De Graff atom smasher!

Suggest the excellent 1960s Tektronix Circuit Concepts book "Cathode Ray Tube circuits"

https://w140.com/tekwiki/wiki/File:062-0852-01.pdf

Bon Chance,

Jon

[ChristofferB]

Sorry for adding onto a 2 month old thread but I'd just like to share the results!

I bought a 220V-> 2x 6.3V 3A filament transformer (from a HAM who was nice enough to insulation test the secondary to 3kV!)

Wired the filaments in series, and connected to each center pin of a BNC, added another BNC for high voltage bias in. The transformer is housed in a NIM module box and controlled by a small external variac.

Below it's seen running a light bulb at ~6VAC with +500V bias. Disregard awful connection of bulb, the HV supply is very puny (300µA at 500V)

Did I overspec the dimensions of the box slightly?  At some point I'd like to put a filament current meter with a current transducer in there.

[NiHaoMike]

Looks like plenty of space to add a built in variac and a built in high voltage supply.

[CaptDon]

I assume you intend to run this 'gun' in a high vacuum chamber so it operates like the electron gun in a C.R.T.? If you intend to shoot electrons/ions in open air the filament will burn up and with air as the insulator you won't be firing anything off the smoldering filament.

[strawberry]

someone wants floating supply then grounds it together all system bypassing isolation transformer back to power outlet, classics

[ChristofferB]

Yes, the filament and electrodes will be in a 10^-7 mbar vacuum chamber, the filaments being different kinds of crushed incandescent and halogen bulbs.

I'll work on adding a simple passive current meter to the box, with the ranges of 3A, 1A, 100mA and maybe 10 mA. I've never worked with current sense transformers before but i'm thinking something like this: A load for the current sense, rectifier bridge, and a simple volt-meter with some ranges.

[strawberry]

could be fun adding gases, photonicinduction showed high power neon bulb with filaments
passive current sense transformer can measure AC component only for DC component need hall current sensor

[NiHaoMike]

Yes, the filament and electrodes will be in a 10^-7 mbar vacuum chamber, the filaments being different kinds of crushed incandescent and halogen bulbs.

I'll work on adding a simple passive current meter to the box, with the ranges of 3A, 1A, 100mA and maybe 10 mA. I've never worked with current sense transformers before but i'm thinking something like this: A load for the current sense, rectifier bridge, and a simple volt-meter with some ranges.

Move the resistor to the output of the bridge rectifier, that'll make linearity a lot better especially at lower currents.

[ChristofferB]

Ok, current sense transformer / meter added but doesn't behave correctly. I may have messed up the circuit, I'd love some feedback!

-Should the "parallel" resistor just be the internal resistance in the coil, and then the range switch be dividing resistors to ground instead?

Having the 100 ohm resistor across the sense transformer is what the datasheet recommends:
https://el-supply.dk/shoppix/5692.pdf

but maybe that's only applicaple for a high impedance voltmeter.

[ChristofferB]

Okay my brain is hazy from christmas food and my 3rd jab, I don't think the previous circuits makes sense. It makes much more sense to me to just connect the meter directly and then put resistors in parallel to divide the current. This brings us to this schematic.

This has an 1A range that's just straight across the meter, just using the coil resistance.
When on this scale, using a light bulb as a load, with a DMM in series to check primary current, the analog meter shows 0.25 mA with 500 mA going through the bulb.

What the?

It's very clearly half off, but I cannot fathom why. I'm clearly missing somthing.

The simplest solution would of course be to make another loop through the transformer...