Ultra fancy resistor divider

[garyk]

Hello everyone, hope this qualifies for the beginner category.    I've got lots of experience in circuit design, but virtually none involving high voltages or negative voltages.

I'm trying to design a dynode control circuit for a Hamamatsu photo multiplier tube (PMT).  My objective is to be able to adjust the PMT gain over a wide range, by increasing or decreasing the number of enabled dynodes.
You'll see by referring to p.98 (figure 5-20) of http://www.hamamatsu.com/resources/pdf/etd/PMT_handbook_v3aE-Chapter5.pdf there is a suggestion to short a certain number of dynodes.

My quandary is how to handle this. My supply voltage to the PMT (aka -HV) is -500V and I want to be able to enable/disable each dynode remotely, either via a 330K resistor on the divider segment, or a 0 ohm to short it to the adjacent dynode..  I'm guessing some BJTs or FETS, coupled with microcontroller GPIOs may be a good start here, but I've got lots of uncertainty.

Any suggestions or ideas?

Thanks!
Gary

[T3sl4co1l]

Shorting out legs of a multiple resistor divider results in a shift in the voltage per (remaining) taps.  It would seem you need a constant current source, so that the voltage per tap is constant (assuming the current draw of the PMT itself is negligible compared to said current).

Tim

[garyk]

I believe modifying the voltage at the dynode taps is acceptable, since this is one of the mechanisms for adjusting the PMT gain.  The key, however, is that the taps you still have in circuit have a consistent voltage drop  between all of them; the resistors themselves should take care of that. 

But, if I throw a FET or BJT or some other switch into the equation, things may go awry.

[Seekonk]

I designed a high speed version of exactly this 20 years ago.  Wish I could remember.

[Alex Nikitin]

Hello everyone, hope this qualifies for the beginner category.    I've got lots of experience in circuit design, but virtually none involving high voltages or negative voltages.

I'm trying to design a dynode control circuit for a Hamamatsu photo multiplier tube (PMT).  My objective is to be able to adjust the PMT gain over a wide range, by increasing or decreasing the number of enabled dynodes.
You'll see by referring to p.98 (figure 5-20) of http://www.hamamatsu.com/resources/pdf/etd/PMT_handbook_v3aE-Chapter5.pdf there is a suggestion to short a certain number of dynodes.

My quandary is how to handle this. My supply voltage to the PMT (aka -HV) is -500V and I want to be able to enable/disable each dynode remotely, either via a 330K resistor on the divider segment, or a 0 ohm to short it to the adjacent dynode..  I'm guessing some BJTs or FETS, coupled with microcontroller GPIOs may be a good start here, but I've got lots of uncertainty.

Any suggestions or ideas?

Thanks!
Gary

If you "disable" a dynode you will just kill the gain of a PMT more or less completely. What is shown in the book is the way of restricting the gain by using one of the dynodes at the anode, essentially ending the multiplication chain at that dynode. Do you need to change the gain more than 10-20 times? If you need to adjust the gain within that range,  change the supply voltage to the divider chain, that is all. If you need to "gate" a multiplier, quickly switching it off to prevent an overload, you can do in by reversing the voltage between two dynodes. If your input signal is very strong you may consider using the circuit shown in the book, however you may get better results just by using an ND filter in front of the PMT, otherwise the PMT life could be very short indeed.

Cheers

Alex

[garyk]

I'll need a substantial gain variation; target is to be able to switch between half gain (~5 or 6 dynodes in the divider chain) and full (all 11 dynodes active).  After one of these two gains is selected, I would fine-tune the gain within that range by adjusting the PMT supply voltage.

Alex, what is this ND filter you mention?

[PTR_1275]

A ND filter is used in photography (And other fields I am sure). Stands for Neutral Density and is basically a shaded piece of glass that is uniform and doesn't add any colour change. Imagine sunglasses for a camera lens.

A ND2 will be 1/2 the original light, ND3 is 1/4 the original light, ND4 is 1/8th the original light. Each ND number increase is 1 stop which is half the light of the previous step.

[Alex Nikitin]

I'll need a substantial gain variation; target is to be able to switch between half gain (~5 or 6 dynodes in the divider chain) and full (all 11 dynodes active).  After one of these two gains is selected, I would fine-tune the gain within that range by adjusting the PMT supply voltage.

Alex, what is this ND filter you mention?

PMT = Photo Multiplier. A full gain would be about 10^6, with only half of dynodes used - about 10^3. There is also a limit of what a photo-cathode can produce and it's life is essentially equals to the maximum charge you can get from it. With the gain only about 1000 the output current would be in nA area for maximum light input without shortening the life of the tube. A PMT has a rather large dynamic range from single photons to about 100MHz-1GHz in photon counting equivalent, more for short pulses. All you need to do is to adapt your light level to the PMT range so it would be covered completely. The PMT gain adjustments are the last resort, usually for accurate gain calibration.

Cheers

Alex

[plesa]

For controlling gain of PMT will much better to control only voltage and not playing with dynodes.
Everything depends on you PMT tube, -500V is quite low for PMT (many PMT will start amplification between 250-300V).
If you want to decrease the gain of PMT do not forget that photocathode can be damaged easily when exposed to light with applied potential. Maximal photocathode current in is datasheet as well.

Another dirty alternative can be measure only current from photocathode for measuring higher light level or use limited ammount of dynodes and last dynode use as anode.

[vindoline]

For controlling gain of PMT will much better to control only voltage and not playing with dynodes.
Everything depends on you PMT tube, -500V is quite low for PMT (many PMT will start amplification between 250-300V).
If you want to decrease the gain of PMT do not forget that photocathode can be damaged easily when exposed to light with applied potential. Maximal photocathode current in is datasheet as well.

Another dirty alternative can be measure only current from photocathode for measuring higher light level or use limited ammount of dynodes and last dynode use as anode.

+ to what pleas says. The gain of a PMT is very sensitive to dynode voltage. I believe it varies proportional to the square of the voltage. i.e. small changes in voltage give large changes in gain. So you wouldn't "fine tune" your gain with voltage adjustments. What is your application? Perhaps we could offer better suggestions if we knew what you were trying to do.

[Marco]

So I assume the PMT works best at higher inter-node voltages, such that it makes sense to disable dynodes rather than do everything by voltage control. Method 2 from page 98 seems to me to be a bit deceptive though. Lets say Kathode voltage is 800 Volt in normal operation, when you short D5/D6/D7 with the Anode the Kathode voltage actually has to decrease to 500 Volt at the same time to maintain the same internode voltages.

This makes handling everything with switches/relays a bit complex with a constant voltage power supply. A constant current power supply would make it easier.

[garyk]

Thank you guys for the feedback thus far.  I wanted to clarify that I have manually changed the dynode count by adding and removing resistors to the circuit, and changing from 330K to 0 ohms on some of them - the design behaves as anticipated.  What I am looking for is a way to enable remote control of the dynode count.  I want to be able to modify it via our control panel.

FWIW, this PMT will be used as the receiver in a CD (circular dichroism) spectrometer.

[Alex Nikitin]

What I am looking for is a way to enable remote control of the dynode count.  I want to be able to modify it via our control panel.

FWIW, this PMT will be used as the receiver in a CD (circular dichroism) spectrometer.

If the application permits (i.e. the switching speed is not an issue etc), I would just use relays for switching.

Cheers

Alex

[Marco]

Don't switch out too many stages in this way though. You are increasing the voltage across each remaining stage and the power consumption with each stage you switch out. It works, until it blows up.

[garyk]

I was mistakenly assuming relays were still larger mechanical devices.  I found that solid-state relays are very common, and have small enough footprints that we can handle quite a few of them in a limited space. 
Solid state relays is my new POR. 

Marco -- We do have a controllable power supply, so we can readily adjust the voltage to limit the inter-dynode voltage drops.

Thanks again to the folks who chimed in here.

-gary