Author Topic: Stable High Voltage (3kV) Photomultiplier Supply  (Read 3422 times)

0 Members and 1 Guest are viewing this topic.

Offline sixtimessevenTopic starter

  • Frequent Contributor
  • **
  • Posts: 330
  • Country: ch
    • Flickr
Stable High Voltage (3kV) Photomultiplier Supply
« on: February 01, 2019, 10:48:52 am »
Hi

I need a stable 3kv supply for my xray florescence setup to get the best energy resolution from the tube. Unfortunately my current supply is a simple switching converter which has a lot of ripple and is generally unstable.

My first idea was to just add a floating series regulator but then I saw the Agilent handbook on power supplies

http://literature.cdn.keysight.com/litweb/pdf/5952-4020.pdf

Basically they take a unregulated hv supply from a 60Hz transformer, rectifier it via voltage doubler and connect it in series with a series regulated supply.
The supplies which employ that scheme (and I found so far) are the HP6110A and the HP6516A. The 6110A in particular has some rather nice specs with stability in the 0.001% range and ripple of only a few millivolts.

http://apapes.physics.uoc.gr/equipment/documentation/HP_06516-90001.pdf
https://nscainc.com/wp-content/uploads/pdf/A_6110A.pdf

Now the 6110A comes up on ebay from time to time but I'm doubtful about the capacitors in there. Plus I do not really need adjustable. Fixed voltage and current limit would be just fine. So I won't need the high voltage control winding, circuit, low voltage offset and just minimal adjustment circuits. Also, I wondered if I should replace the "voltage input"section with a opamp based design and leave the driver circuit (Q10, Q9) as it is.

I had a look on the transistor they used for the 200V series regulator and there seems to be a recommended replacement, the NTE124

http://www.nteinc.com/specs/100to199/pdf/nte124.pdf

However I had a look at some datasheets and if I happen to interpret the "save operating area" graph correctly this one
https://media.digikey.com/pdf/Data%20Sheets/ST%20Microelectronics%20PDFS/BU208A_508A_508AFI.pdf

Should give me about 30mA dc current @1000V VCE? To bad it's obsolete  :-DD


2000V and 4000V 1uF film capacitors are probably the priciest parts.
https://www.digikey.ch/products/en/capacitors/film-capacitors/62?k=2000v%201uF
https://www.digikey.ch/products/en/capacitors/film-capacitors/62?k=&pkeyword=&sv=0&umin2049=0.1&umax2049=2&rfu2049=%C2%B5F&pv1293=89&pv1293=29&pv1293=171&pv1293=49&pv1293=139&sf=0&FV=ffe0003e&quantity=&ColumnSort=0&page=1&pageSize=25


I would love some feedback, pointers or tips on other HV supplies from HP (didn't find too many) to have a look at.








 

Offline Yansi

  • Super Contributor
  • ***
  • Posts: 3930
  • Country: 00
  • STM32, STM8, AVR, 8051
Re: Stable High Voltage (3kV) Photomultiplier Supply
« Reply #1 on: February 01, 2019, 11:31:01 am »
Why not using a CFPR (Current Fed Parallel Resonant) converter and a voltage multiplier? The CFPR can be also easily regulated for stable voltage. Output will also be fairly clean (less switching mess, as the diodes will operate in soft recovery, due to the sinewave voltages).

I mean, that two-transistor self-oscillating job is more than enough to supply a CRT or photo-multiplier I guess.
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 15148
  • Country: de
Re: Stable High Voltage (3kV) Photomultiplier Supply
« Reply #2 on: February 01, 2019, 11:55:07 am »
Do you really need 3 KV for a photo-multiplier ? The ones I know need more like a little below 1000 V to  get pulse hight resolution.

Usually the current is pretty low (e.g.  maybe 1 mA, mainly for the divider).

I would avoid using large capacitor and better use a higher frequency high voltage generation. So more like a 20 kHz resonant converter and than filter caps more in the 20-100 nF range. With voltage doubling rectifier the caps will only see something like half or even less of the voltage.

If one really needs linear regulation this only has to deal with a smaller adjustment range, like 100 V or so. There should be plenty of transistors available, e.g. those made for video CRT drivers in old TVs (e.g. BF869, BF469 - but many similar ones to choose from).
 

Offline sixtimessevenTopic starter

  • Frequent Contributor
  • **
  • Posts: 330
  • Country: ch
    • Flickr
Re: Stable High Voltage (3kV) Photomultiplier Supply
« Reply #3 on: February 01, 2019, 12:34:34 pm »
Do you really need 3 KV for a photo-multiplier ? The ones I know need more like a little below 1000 V to  get pulse hight resolution.

Usually the current is pretty low (e.g.  maybe 1 mA, mainly for the divider).

I would avoid using large capacitor and better use a higher frequency high voltage generation. So more like a 20 kHz resonant converter and than filter caps more in the 20-100 nF range. With voltage doubling rectifier the caps will only see something like half or even less of the voltage.

If one really needs linear regulation this only has to deal with a smaller adjustment range, like 100 V or so. There should be plenty of transistors available, e.g. those made for video CRT drivers in old TVs (e.g. BF869, BF469 - but many similar ones to choose from).

I currently use a small 10-Stage Russian built tube with mixed results.

First idea was a XP2020 which would have taken a max voltage of 3000V. But after some searching today I found a better option.

So next step will probably be a 10-Stage XP2102. It only requires 1000-2000V - Yay. It's affordable and seemed reasonable with the 9.5% resolution.
https://wwwusers.ts.infn.it/~rui/univ/Acquisizione_Dati/Manuals/Philips%20XP2020.pdf
They go on Ebay for about 30 Euros, but under the old phillips brand which now is photonics:
https://www.ebay.ch/itm/Photomultiplier-Tube-XP2412-SQA-Phillips-PH21-pC-PHR-9-20/253087238554?hash=item3aed2cc99a:g:jTUAAOSwRYBZi4xz:rk:1:pf:0


Quote
Why not using a CFPR (Current Fed Parallel Resonant) converter and a voltage multiplier? The CFPR can be also easily regulated for stable voltage. Output will also be fairly clean (less switching mess, as the diodes will operate in soft recovery, due to the sinewave voltages).

CFPR seems like an interesting option. Thank you!
 

Online David Hess

  • Super Contributor
  • ***
  • Posts: 17427
  • Country: us
  • DavidH
Re: Stable High Voltage (3kV) Photomultiplier Supply
« Reply #4 on: February 01, 2019, 03:28:10 pm »
Basically they take a unregulated hv supply from a 60Hz transformer, rectifier it via voltage doubler and connect it in series with a series regulated supply.
The supplies which employ that scheme (and I found so far) are the HP6110A and the HP6516A. The 6110A in particular has some rather nice specs with stability in the 0.001% range and ripple of only a few millivolts.

Many Tektronix oscilloscopes use that method to tightly regulate the 1 to 3 kilovolt cathode supply to the CRT but the high voltage part comes from a high frequency inverter which makes filtering easier.

At low currents I would consider using a flyback transformer based boost converter to drive a diode-capacitor voltage multiplier and at higher currents a regulated Royer converter which has the advantage of a clean high frequency sine wave output which is easy to filter and a lot of commonality with CCFL making it economical.

Linear Technology application note 118 discusses high voltage low noise DC to DC converters.
« Last Edit: February 02, 2019, 02:11:58 am by David Hess »
 

Offline Yansi

  • Super Contributor
  • ***
  • Posts: 3930
  • Country: 00
  • STM32, STM8, AVR, 8051
Re: Stable High Voltage (3kV) Photomultiplier Supply
« Reply #5 on: February 01, 2019, 03:38:58 pm »
Putting a non-regulated supply in series with a regulated one does not create a regulated high voltage supply. Or what did I miss?

High voltage secondaries at mains frequency are pan in the ass to work with.  Even the Russians knew not to do that in their CROs. They only had the filament winding (with the reinforced insulation) on the mains transformer. Cathode (and anode)  voltages were supplied by a tiny DC-DC converter + multiplier.
There is just too many hassle making a mains frequency high voltage transformer. There is little gain in doing that.

« Last Edit: February 01, 2019, 06:04:11 pm by Yansi »
 

Offline sixtimessevenTopic starter

  • Frequent Contributor
  • **
  • Posts: 330
  • Country: ch
    • Flickr
Re: Stable High Voltage (3kV) Photomultiplier Supply
« Reply #6 on: February 01, 2019, 04:44:49 pm »
Putting a non-regulated supply in series with a regulated one does not create a regulated high voltage supply. Or what did I miss David?

High voltage secondaries at mains frequency are pan in the ass to work with.  Even the Russians knew not to do that in their CROs. They only had the filament winding (with the reinforced insulation) on the mains transformer. Cathode (and anode)  voltages were supplied by a tiny DC-DC converter + multiplier.
There is just too many hassle making a mains frequency high voltage transformer. There is little gain in doing that.

Well "series" might be a bit confusing. Sorry. What I meant to say is that the HV and the regulated voltage source are coupled in series. Then the regulator takes the feedback from the HV + regulated supply and smoothing out the unregulated HV supply in the process.

Quote
    Basically they take a unregulated hv supply from a 60Hz transformer, rectifier it via voltage doubler and connect it in series with a series regulated supply.
    The supplies which employ that scheme (and I found so far) are the HP6110A and the HP6516A. The 6110A in particular has some rather nice specs with stability in the 0.001% range and ripple of only a few millivolts.


Many Tektronix oscilloscopes use that method to tightly regulate the 1 to 3 kilovolt cathode supply to the CRT but the high voltage part comes from a high frequency inverter which makes filtering easier.

At low currents I would consider using a flyback transformer based boost converter to drive a diode-capacity voltage multiplier and at higher currents a regulated Royer converter which has the advantage of a clean high frequency sine wave output which is easy to filter and a lot of commonality with CCFL making it economical.

Linear Technology application note 118 discusses high voltage low noise DC to DC converters.

I love LT application notes. They are such a joy to read.
My concern in the beginning was that a DC-DC converter introduces a lot of high frequency noise into the output which the bandwidth of the series regulator might not be able to remove. However I guess with additional ferrites (like it is done in for low voltage regulators) and filter caps one can get rid of them.

As to transformers, for 2400V a standard micro wave transformer would have done the job. But the caps get expensive (and scary as hell).




 

Offline Yansi

  • Super Contributor
  • ***
  • Posts: 3930
  • Country: 00
  • STM32, STM8, AVR, 8051
Re: Stable High Voltage (3kV) Photomultiplier Supply
« Reply #7 on: February 01, 2019, 06:05:35 pm »
Jeezus crist...  a MOT for a photomultiplier supply?  You need a mA, not a whole amp at 2kV.
 

Offline Mark

  • Frequent Contributor
  • **
  • Posts: 272
  • Country: gb
Re: Stable High Voltage (3kV) Photomultiplier Supply
« Reply #8 on: February 01, 2019, 06:14:22 pm »
Take a look at electrophoresis supplies, I have a couple that will put out 3kVDC at 400W. 

Load regulation: PowerPac 3000: ±1% at 3000 V for a 50% change in output load. 
 

Offline jmelson

  • Super Contributor
  • ***
  • Posts: 2851
  • Country: us
Re: Stable High Voltage (3kV) Photomultiplier Supply
« Reply #9 on: February 01, 2019, 08:21:56 pm »
Several companies make modular DC-DC converters for jobs like this.  You might be able to puck one up on eBay or similar.  They take something like 12 V DC and put out 1000 - 2000 V, unregulated.  Then, with some high-value resistors and a FET-input op-amp, you make a voltage measuring circuit.  Then, with another op-amp and a Darlington power booster, you make a regulator loop that drives the DC-DC converter.  You can put a couple pole R-C filter after the DC-DC converter to remove any noise.  It is a little tricky to get the regulator loop to close around the R-C filter, but with some care, you can make it stable.  I've built a couple of these, many years ago.

Jon
 

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15794
  • Country: fr
Re: Stable High Voltage (3kV) Photomultiplier Supply
« Reply #10 on: February 01, 2019, 08:43:25 pm »
Have you read "The boy electrician"? ;D
 

Online David Hess

  • Super Contributor
  • ***
  • Posts: 17427
  • Country: us
  • DavidH
Re: Stable High Voltage (3kV) Photomultiplier Supply
« Reply #11 on: February 02, 2019, 02:15:03 am »
Putting a non-regulated supply in series with a regulated one does not create a regulated high voltage supply. Or what did I miss?

Feedback from the combined high voltage is used to control the regulated supply.  In the designs I am thinking of, the regulated low voltage output was combined with the unregulated high voltage through a circuit much like a high voltage DC restorer.
 

Offline ycui7

  • Contributor
  • Posts: 19
  • Country: us
    • SpecialtyCircuits LLC
Re: Stable High Voltage (3kV) Photomultiplier Supply
« Reply #12 on: February 02, 2019, 06:59:39 am »
https://www.spellmanhv.com/en/Products/PMT

very quite high voltage power supply. you can hardly measure the noise with a 1000:1 probe, as the noise is almost on the noise floor of typical digital oscilloscopes. That puts the noise upper limit to less than 1V.

Alternatively, you can go with Matsusada, or certain XP-EMCO module. Do not go with UltraVolt.

Offline Yansi

  • Super Contributor
  • ***
  • Posts: 3930
  • Country: 00
  • STM32, STM8, AVR, 8051
Re: Stable High Voltage (3kV) Photomultiplier Supply
« Reply #13 on: February 02, 2019, 08:16:22 am »
Why would you measure NOISE (AC) using a 1000:1 DC probe? Shouldn't that be measured with an AC COUPLED 1:1 probe instead?
« Last Edit: February 02, 2019, 08:18:37 am by Yansi »
 

Offline Yansi

  • Super Contributor
  • ***
  • Posts: 3930
  • Country: 00
  • STM32, STM8, AVR, 8051
Re: Stable High Voltage (3kV) Photomultiplier Supply
« Reply #14 on: February 02, 2019, 08:18:06 am »
Putting a non-regulated supply in series with a regulated one does not create a regulated high voltage supply. Or what did I miss?

Feedback from the combined high voltage is used to control the regulated supply.  In the designs I am thinking of, the regulated low voltage output was combined with the unregulated high voltage through a circuit much like a high voltage DC restorer.

Do you have any schematic on hand to give an example? I can't still think how one would regulate two supply voltages with just one series pass element. That seems impossible.
 

Online David Hess

  • Super Contributor
  • ***
  • Posts: 17427
  • Country: us
  • DavidH
Re: Stable High Voltage (3kV) Photomultiplier Supply
« Reply #15 on: February 03, 2019, 06:31:53 am »
Putting a non-regulated supply in series with a regulated one does not create a regulated high voltage supply. Or what did I miss?

Feedback from the combined high voltage is used to control the regulated supply.  In the designs I am thinking of, the regulated low voltage output was combined with the unregulated high voltage through a circuit much like a high voltage DC restorer.

Do you have any schematic on hand to give an example? I can't still think how one would regulate two supply voltages with just one series pass element. That seems impossible.

The examples which come to mind are the Tektronix 7704A and 7904 oscilloscope mainframes.  These were their first to use a switching power supply to drive a separate high voltage transformer producing a semi-regulated high voltage output; previous designs regulated the high voltage about by adjusting the high voltage inverter drive.  Since the high voltage negative cathode supply had to be tightly regulated for good deflection accuracy, a separate low voltage circuit with feedback from the rectified high voltage injects DC into a DC restorer which rectifies the high voltage AC to high voltage DC.

As shown below, R4240, R4241, and adjustment R4123 are the feedback divider which uses a +50 volt reference.  Operational amplifier U4110 is the low voltage error amplifier.  Q4115 and Q4105 are a level shifting stage to produce 0 to 150 volts DC.  CR4205 and CR4206 with associated capacitors and resistors are the combined DC restorer, rectifier, and filter.  The output voltage in this case is a very closely regulated -3000 volts which tracks the +50 volt reference.  I measured stability on my 7904 as better than 100 millivolts.

An interesting feature of this circuit at least to me is that U4110 is a plain Jane C-grade 741 despite the high input impedances but they used capacitive bypassing to limit noise and a bias error cancellation resistor.  I would have used a low input bias current 308 or similar but Tektronix was very parsimonious about using premium parts or changing designs once they had something working.
 

Offline ycui7

  • Contributor
  • Posts: 19
  • Country: us
    • SpecialtyCircuits LLC
Re: Stable High Voltage (3kV) Photomultiplier Supply
« Reply #16 on: February 03, 2019, 10:11:26 pm »
That was in my grad college time. There was no kilovolt-withstanding AC coupled probe in academic environment, at least for my lab. You cannot just put oscilloscope to AC coupling and connect 3kV to a GHz oscilloscope input. If you do, boss would be super mad, and you may never graduate. So, I did what I could, although no ideal.

Measuring sub-volt noise on a kilovolt signal is always difficult, particularly due to availability of probe.


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf