Looking for HV schematic 400v, easy to build from scrap parts

[Vancata]

Hey there again folks.
One friend hobbyist is building Geiger counter and asked me for some cooperation. We are looking for some easy (scrap-parts) schematic that we can do for the power supply. We are going to use 6v (4xAA) or 9v battery, and we are looking for ~400v AC to "attack" the Geiger–Müller tube.

We have tried the 2nd schematic from here http://www.techlib.com/science/geiger.html but we wasn't able to get anything close to 400v. Even the Orcad simulation needed like 45mH coil to start giving some voltages, so we are not sure if this thing is going to run properly in the real world.

If anyone can think of other stable schematic that he seen somewhere and that don't need specific transformer (this is the rod in the wheel in our case/the main problem at the moment) will be glad if he share it with us.

We already did the counter part  but we can't properly test it with out the tube output

[David Hess]

I would use a charge pump inductive boost converter like that shown on page 93 of Linear Technology application note 47:

http://www.linear.com/docs/4138

[Vancata]

I would use a charge pump inductive boost converter like that shown on page 93 of Linear Technology application note 47:

http://www.linear.com/docs/4138

Cannot find the LT1073 in our local supplier here. The cheapest other DC/DC regulators start from like 5$ here. Might go for something like this in the end, or try to wind a freaking transformer on one ferrite ring.

[David Hess]

I would use a charge pump inductive boost converter like that shown on page 93 of Linear Technology application note 47:

http://www.linear.com/docs/4138

Cannot find the LT1073 in our local supplier here. The cheapest other DC/DC regulators start from like 5$ here. Might go for something like this in the end, or try to wind a freaking transformer on one ferrite ring.

Any step-up switching regulator will work in this configuration.  What you should look for is one intended for micropower operation which usually means some type of burst mode.

Usually the output voltage is limited by the maximum switch voltage but in this case, multiple stages of diode-capacitor voltage multiplication allow the use of a relatively low voltage switch while generating high voltages at low currents.

[Nermash]

Check this out:
https://www.sparkfun.com/tutorials/132

I have also built one using the pwm output of a PIC18 mcu driving mosfet and IIRC 100uH inductor. No feedback, no isolation, bit crude but it worked fine for SBM20 tube, 5V/380V boost converter.
You can also look at tl494 or ka7500 app notes, they are easily found in atx psus.
Experiment a bit

[miguelvp]

Do it at your own risk:

A CFLs electronic ballast is a AC to DC then DC to AC convertor outputting 1000V to the Fluorescent Lamp.
Voltages and Amps in there can kill you. So that is 8 times of the input.

For 400VAC you will need 50 DC for this ballast and you will have to tap after the rectifier and remove the transformer and circuitry before the rectifier.

Here is an instructable that goes even beyond 1000V and using AC instead of lightbulb using a flyback transformer in the output of the ballast.

http://www.instructables.com/id/MAKE-A-HIGH-VOLTAGE-SUPPLY-IN-5-MINUTES/?ALLSTEPS

VERY dangerous if you touch the wrong thing.

Safer way cheap but no spare parts:
http://www.xoxide.com/coldcathodes.html

Still hundreds of volts output from just 12V input, so no touchy the output

Cold Cathode lighting is the most commonly used form of case lighting on the market. Advancing on the standard 12V neon, cold cathode lamps operate by use on a power inverter, which converts the standard 12V input to hundreds of volts of power. This is done by minimizing the amperage, via the inverter unit (included with the kit). The end effect is an intensely bright and thin tube, the most popular form of lighting today.

Our kit is fine tuned to push out as much brightness as possible, while still keeping the best lifespan to brightness ratios. We have tuned our kits to operate for up to 30,000 hours of burn time, while still competing with the brightest tubes on the market.

This kit includes everything you need for a complete lighting solution. We include a Molex connector with power splitter, boxed inverter, and 2 acrylic encased cold cathodes.

The cold cathode mounts easily via adhesive square attached to the end of the tubes.

[David Hess]

A high voltage power supply intended for a cold cathode fluorescent tube is way too powerful for a Geiger tube.  It will work but the extra complexity for a high power design is not needed.

Another place to look is photographic strobe power supplies but they are intended for relatively high power loads as well.

[ajc]

Why not have a look at Jeff Keyzer's Geiger counter design - http://mightyohm.com/blog/products/geiger-counter/design-files/.  His is based on a 555 timer for the DC/DC drive signal (which is also a well known nixie tube power supply design approach).

BTW, my understanding is that you want a DC HV supply not AC, but I haven't played with GM tubes myself, so I could be wrong here.

[LukeW]

You want DC, not AC. Typically anywhere from about 400V to 1500V DC, depending on your GM tube. 900V is probably the most common defacto standard for US-made Cold War era Geiger counters and GM tubes.

400VAC driving a single Cockroft-Walton style voltage doubler stage might make sense, though.

Look at these schematics for some ideas:

http://mightyohm.com/files/geiger/geiger_sch_fixedR5R6.png

http://dlnmh9ip6v2uc.cloudfront.net/datasheets/Sensors/Radiation/Geiger_Counter-v30_2.pdf

You can either go for a boost converter sort of approach using an inductor, or a transformer-based approach. You can use common, cheap transformers that you can find cheaply or recycle, such as 8-ohm speaker drive transformers, or 600-ohm telephone line isolation transformers from modems for example.

Another approach that is sometimes used is to use a small CFL inverter board, which is a cheap surplus thing such as you'd find in a scanner or laser printer or monitor or perspex-window case-modded PC to drive the CFL lamp.

This might typically output, say, 400-500V AC. You'd need to remove the series output capacitor after the transformer, add a single-stage voltage doubler to the output, and regulate the input voltage to get the output voltage that you need.

[IanB]

Disposable flash cameras have a circuit that produces ~300 V DC from a 1.5 V battery. Put two or three in series and you can have 900 V.

[miguelvp]

Another approach that is sometimes used is to use a small CFL inverter board, which is a cheap surplus thing such as you'd find in a scanner or laser printer or monitor or perspex-window case-modded PC to drive the CFL lamp.

This might typically output, say, 400-500V AC. You'd need to remove the series output capacitor after the transformer, add a single-stage voltage doubler to the output, and regulate the input voltage to get the output voltage that you need.

This is exactly what I recommended as a safer way. But just wanted to note that since they are planning to operate it with just 9V the output will be 3/4 of what it was designed for, so it should be around 400V already.

But they should measure the output (carefully and not with a scope unless you know how to measure this by removing the grounds on your probes and using the scope math to check it). Actually don't know about other scopes but mine has a max voltage input of only 300V RMS

And don't use a cheap DMM without enough protection either.

Edit: just remembered that even I said it was battery operated, still be careful measuring the outputs.

[miguelvp]

I just tested one of those cold cathode inverters for lighting your computer and I was getting 518VAC from a fresh 9 VDC battery (9.236V)

35KHz and it shoots to over 600V on power up for a little bit, didn't use my scope buy my DMM, since my scope can only deal with 300V

[David Hess]

I just tested one of those cold cathode inverters for lighting your computer and I was getting 518VAC from a fresh 9 VDC battery (9.236V)

35KHz and it shoots to over 600V on power up for a little bit, didn't use my scope buy my DMM, since my scope can only deal with 300V

Usually you would use a x100 oscilloscope probe and the inexpensive ones are good to 2000 volts.

[miguelvp]

I just tested one of those cold cathode inverters for lighting your computer and I was getting 518VAC from a fresh 9 VDC battery (9.236V)

35KHz and it shoots to over 600V on power up for a little bit, didn't use my scope buy my DMM, since my scope can only deal with 300V

Usually you would use a x100 oscilloscope probe and the inexpensive ones are good to 2000 volts.

Thanks.

Mine are 10x but they are rated for 300V Cat II max input as well. Might pick up some at a later time. Rigol branded ones are a bit pricey for my taste at $260 a pop from tequipment, although they are 300 MHz probes so well worth it if I needed them.

I guess for that high voltage 60MHz 100:1 will do and those are fairly cheap around $20 each.

But my newly acquired UT71B was enough for checking this out. I could datalog the inputs and post the result since it has an opto-coupled USB interface but. I love this meter! (I know not the D but only paid $50 for it)

[David Hess]

One problem when working with low current high voltage power supplies is that even if your favorite multimeter can measure the output voltage safely, sometimes its 10 megohm input resistance is too low.  High voltage probes for multimeters can help with this though even if the higher voltage range is not needed.  Most start have an input resistance of 75 to 1000 megohms which helps a lot.  This problem applies to x10 oscilloscope probes as well which makes x100 probes useful.

[miguelvp]

Would a 100KHz True RMS DMM be accurate to measure AC? That's what I used.

Edit, I meant 100KHz not 10

[Vancata]

Thanks a LOT folks! Really refreshed me with a lot of ideas.

Just measured one ballast for neon lights, it gave more then 1kV AC. My tube is soviet one:  STS-5 -5 we found it in one local supplier http://www.gstube.com/data/2398/#circuit5
For voltage meter i use one low class Extech 430  to do high voltage measurements, and especially when i measure ac driven current, because of the trms that it suppose to have.
For the scope measurements I have only one cheap usb hantek scope. Anyway I know the drill, but really paying attention to not short my usb with high current sources with common GND, and really I'm not going to use on the main outputs. Anyway in the upcoming months will get one old analog one and one 1:1 mains transformer (could be found easily 2nd hand here).

So you really filled me with some ideas, will go hunt some old printers/scanners/cameras with flashlight et.c and tinker with them 

[JulietMikeBravo]

I have built several geiger counters, the most practical circuits I have found are Charles Wenzel's and derivatives, like:



My own complete implementation:



The HV generator in these circuits shuts off when the desired voltage is reached, and tops off the storage capacitor as needed, perfect for the low current draw of geiger tubes. When the circuit is idle it usually draws less than 500 microamps.