Tektronix Technical Forums are maintained by community involvement. Feel free to post questions or respond to questions by other members. Should you require a time-sensitive answer, please contact your local Tektronix support center here.

Electrometer damage due to fast voltage changes

6514, 6517 Support
Post Reply
cvtITLza
Posts: 5
Joined: October 8th, 2021, 12:33 am
Country: South Africa

Electrometer damage due to fast voltage changes

Post by cvtITLza » October 8th, 2021, 6:25 am

Hi
We have a Keithley model 6517b electrometer in a measurement setup as depicted in the attached diagram. The setup shown is duplicated to implement two separate but identical systems.

Quick description of attached diagram:
1. The body of the ionization chamber, the chassis of the electrometer, and one side of the CAEN high voltage power supply (HV PSU) are all connected to the safety ground of the system.
2. ‘Input low‘ of the electrometer "floats" on the 500V provided by the CAEN HV PSU. So all the electronics inside the electrometer floats on this voltage. See footnote [1] below.
3. The voltage on ‘input high’ of the electrometer will differ slightly from that on the ‘input low’ terminal. Only due to the fact that the op-amp operation is not perfect and the virtual ground on ‘input high’ cannot be perfectly maintained.
4. The connection to the ionization chamber is through a 50-meter long triaxial cable. The ionization chamber is located in a distant high radiation area. So the high voltage is fed to this chamber through a long cable that presents a big parasitic capacitance. The CAEN HV PSU ramps the 500V output up and down at a user-programmable rate to prevent high currents during charging and discharging of the parasitic capacitances.
We have been using this configuration very successfully for a long period of time now without any problems.

Problem description:
Recently we had a scenario where the utility power (220Vac) fell away on all of the electronics systems at the same time. After switching on the systems both electrometers were found to be damaged as can be seen in this video. https://www.youtube.com/watch?v=ARAsAR-AisY

We monitored the output voltage of the CAEN HV PSU when utility power is removed. It was determined that the voltage ramp down function is not active when primary power is removed abruptly. The high voltage just falls away immediately.

This led to the suspicion that the electrometers might have been damaged due to the uncontrolled ramp down of the voltage during the power failure. This uncontrolled discharge of the 50 meter long cable attached to the electrometer probably causes high voltage and/or current transients on the sensitive inputs of the electrometers.

All the electronics systems are now on an uninterruptable power supplies (UPS) system that executes a controlled shutdown when utility power falls away. This is however not enough to prevent damage to equipment because the CAEN power supplies can still be (accidentally) powered off by unplugging the unit or flipping the power switch. There are several places in the power chain where this could be done.

The way to ensure the prevention of equipment damage is to implement the voltage ramp down function separate from the CAEN unit. This should be external to the CAEN power supply. Initial thoughts are to implement a protection circuit that the electrometer, HV PSU, and long cable can be connected to. When power falls away the circuit should discharge the cable in a controlled fashion to ramp down the voltage seen on the electrometer input at a rate that is within the specifications (i.e. dv/dt rates) of the electrometer.

Questions:
1. I have searched for the dv/dt specifications of the electrometer’s input in all related Keithley documentation and could not find a thing. Are such dv/dt specifications available?
2. The only related specification that I could find is “The maximum common-mode input voltage (the voltage between input low and chassis ground) is 500 Vpeak. Exceeding this value may create a shock hazard.”
The warning says shock hazard and not equipment damage.
In light of this, does it seem feasible that the uncontrolled ramp down of the high voltage in this setup could have caused damage to the electrometers? Especially since the common-mode voltage provided by the HV PSU is at the 500V limit, i.e. no margin.

Possibly related posts on this forum:
1. viewtopic.php?f=39&t=142732
2. viewtopic.php?f=39&t=142717
3. viewtopic.php?f=39&t=129321
I will link these posts back to this post.

Footnotes.
[1] A less important/relevant observation is that the V/F converter also "floats" on the 500V of the CAEN HV PSU to allow it to be connected to the electrometer's 2V output.
Attachments
Electrometer setup.pdf
(508.47 KiB) Downloaded 18 times

Andrea C
Keithley Applications
Keithley Applications
Posts: 1625
Joined: October 15th, 2010, 10:35 am
Country: United States
Contact:

Re: Electrometer damage due to fast voltage changes

Post by Andrea C » October 9th, 2021, 8:19 am

Hello,

The dV/dt is not something we would specify. But it is most likely playing a role together with the capacitance of your 50 meter cable and other system aspects and is generating a rather large displacement current thru the 6517B in your case.

The use of some simple back to back diodes mounted in light tight box should be protective of the electrometer.

When you pass more and more current thru electrometer, the voltage burden becomes larger. If a diode is present, it will begin to conduct if the v_burden becomes significant.
Use back to back diodes so that you have a bipolar protection device.
In normal operation, the v_burden during amps measurement should be only few micro volts and presence of diode is not zero leakage, but can be quite small.
It is a cheap insurance for the sensitive electrometer.
Section 3 in the 6517B Reference manual has a bit on the input protection using diodes.

cvtITLza
Posts: 5
Joined: October 8th, 2021, 12:33 am
Country: South Africa

Re: Electrometer damage due to fast voltage changes

Post by cvtITLza » October 22nd, 2021, 5:39 am

Hello Andrea,

The suggested protection circuit will provide protection in measurement configurations such as the ones listed below.
Measurement of:
1. leakage current of capacitors
2. reverse diode leakage
3. insulation resistance of cables,

In these configurations, it could happen that the HV source can be connected directly over the ammeter’s ‘INPUT HI’ and ‘INPUT LOW’ terminals due to component failure in the DUT test circuit. Such as a diode or capacitor that shorts.

This scenario above is not possible in my circuit since I do not have component failures that could cause such unintended connections.
I have included a simplified version of my original diagram. It can be seen that only one side of the high voltage power supply is connected to the ammeter terminals. So the ammeter circuitry floats on the high voltage, relative to the chassis ground.

The only way to apply high voltage over the ammeter is to short ‘INPUT HI’ to chassis ground. Sure, this is possible but highly unlikely in my setup. I do not want the diode protection circuit over my ammeter, degrade my measurement performance, just to protect for something that will not easily happen.

I think I have a bigger problem. I have 500V between ‘INPUT LO’ i.e. ‘COMMON’ and chassis ground. This is the maximum common-mode input voltage allowed between these 2 nodes. We can lower this voltage to 350V.

Even with this lower 350V there will still be problems. Transients will be superimposed on the 350V when the high voltage power supply is switched off abruptly. This can exceed the 500v maximum common-mode input voltage specification. I need to protect against this.

I think I should add a Zener diode voltage clipping circuit as shown in the attached simplified diagram. The link below shows 300V to 390V Zener diodes available from Digikey.

Questions:

1. Am I wrong in my analysis above?
I.e. that the ammeter doesn't really need protection but rather protection must be provided to prevent the "500v maximum common-mode input voltage" specification from being violated when the HV abruptly falls away. See attached pdf for the specification detail.

2. If the diode clipping circuit between ‘INPUT LOW’ and chassis ground is the best solution should I use Zener diodes or would TVS or any of the other overvoltage protection devices be better?

I have also attached relevant pages from the Model 6517B Electrometer Reference Manual

https://www.digikey.co.za/en/products/f ... Xvhn5tQ-UA
Attachments
Simplified for post.pdf
(51.66 KiB) Downloaded 14 times
Model 6517B Electrometer Reference Manual -- Pages 55 - 58.pdf
(374.86 KiB) Downloaded 8 times

Andrea C
Keithley Applications
Keithley Applications
Posts: 1625
Joined: October 15th, 2010, 10:35 am
Country: United States
Contact:

Re: Electrometer damage due to fast voltage changes

Post by Andrea C » October 22nd, 2021, 12:32 pm

The use of diode is also protective of an over-current situation.
As more and more current is passed through the electrometer, the voltage burden will increase in value.
If this forward biases the diode a little, it provides a parallel path bypass of the electrometer.

Your original post offers the theory that the sudden removal of the LO side bias voltage was causal to the damaged instrument.
Knowing that I= C *dV/dt, it did seem plausible that a large displacement current might have been pulled through the electrometer especially since you have capacitance of 50 meters of cable.

Use of the zener to enforce a common mode voltage limit is good idea too.

Post Reply

Return to “Electrometers”

Who is online

Users browsing this forum: No registered users and 1 guest