Circuit design of branded electrometers

[Kirill V.]

[..]
Current about 1.4-1.5 pA, therefore, insulation resistance more 4 POhm.
[..]

On a breadboard?  That is, er, quite remarkable.

The non-inverting input hangs in the air, the signal is received via a coax cable with polyethylene insulation, the external shield is connected to the buffer output and guarding forms. Nothing remarkable, the classic construction of measuring instruments
Update:
I made a mistake when working with decimals, sorry. There's about 14-15 pA, not POhms, of course. But the breadboard has nothing to do with it

[Kirill V.]

Correct.

First I was a bit confused what you meant by drift related to the input, to be clear, this is not drift - it is interaction of input bias current with input impedance.
This impedance consists of the resistive part and the capacitive part, the capacitive part will be charged by the bias current, so the output "drifts" - or better said settles - to the final value given by the resistive part.

As you already noticed, if the output hits the high voltage rails, the NFB does not persist.
This leads to a voltage difference of the op amp inputs and finally will be clamped either by protection circuit between the inputs or to the bootstrapped power supply rails.
When the protection kicks in, the input impedance will drop dramatically.
To protect the protection circuits from to high currents there is the need for current limiting the inputs, this is simply achieved by putting a suitable resistor at each input (most op amps are rated for 10mA absolute max).

If a specific resistor with specific properties is needed, I would recommend to buy one - up to 10G are awailable for ~1EUR with not to shabby specs.

Ok, thanks. In fact, the parasitic leakage resistance and the bias current compensation resistor form a voltage divider. This can be represented by an equivalent circuit. Please note that this is not bootstrapped supply:

We can achieve zero output with two independent adjustments. But the bias current, offset voltage and all resistances drift with time and temperature. Here any drifts referred to the input are multiplied by the noise gain equal to one.
Now equivalent circuit with BS supply

Here, a positive feedback is formed through the parasitic leakage resistance and the bias current compensation resistor. This is what confuses me
Update:
The diagram I drew exactly corresponds to the ohmmeter mode. This means that if there is the slightest mismatch in the bias current compensation, the difference current will flow through the insulation of the input circuit. This means that with significant input circuit resistances, the adjustment must be very fine and painstaking. And the difference current will still appear over time and temperature changes. Correct me if I'm wrong

[Kirill V.]

Кирилл, посмотри здесь.
Схема похожа на то, что ты ищешь.

Cyril, look here.
The circuit is similar to what you are looking for.

http://bbs.1ppm.cn/topic/276/%E8%87%AA%E5%88%B6%E6%8C%87%E9%9B%B6%E4%BB%AA%E7%9A%84%E6%80%9D%E8%80%83-some-thoughts-on-diy-null-meter/16

Thank you, Sergey, I will study this. А вы нарочно пишете по-русски на международном форуме?!

Friends, is there any difference between styrene and polystyrene capacitors?

[Kleinstein]

Friends, is there any difference between styrene and polystyrene capacitors?

No there is no difference, just a shortened name. Just simple styrene is a liquid. It is the polymer that is used to make styrofoam, capacitors and some cheap plastics.

Keep in mind they don't like higher temperature and most solvents.

[serg-el]

PTFE wire test.  Useful.
https://misrv.com/teflon-wire-test/

[Kirill V.]

Do not expect that the voltage stays near 0V when the input is floating...

Okay, now tell me, friends, what does the carefully adjusted and long-warmed K617 with float input show on the display? As well as other electrometers that you are dealing with. Is there a significant difference with the connected cable and without it?

[Kleinstein]

Like with other very high input impedance meters, the expected behavior with an open input is drift of the displayed voltage. How much depends on the individual instrument. The slower the drift the better - as the drift rate is given by input current divided by input capacitance  Adding a cable chances are the drift would slow down as the cable adds capacitance. If the cable is poor quality (poor isolation) the voltage may stabilize.

[Kirill V.]

Thanks, Kleinstein. That's what I meant - drift. So it's inevitable.

[Kirill V.]

How does the K617 electrometer implement the SUPPRESSION function-at the analog or digital level?

[MiDi]

It is just as Nulling on DMMs, so only digital math.

[Kirill V.]

Then if the digital part subtracts from the ADC result ,say, 10 Volts, and the input voltage is 200 Volts, then the display will show 190 Volts, but the analog part is already working almost at the limit, right?
Nulling/suppression at the analog level has some important advantages?

[Kleinstein]

There is not much advantage to get nulling at the analog side. There could be a small advantage to have a crude offset adjustment at the analog side, as the offset tends to be temperature dependent and often, but not always with  FET circuits the temperature effect is larger when the offset is large.  So the analog "offset" trim is more like a TC trim.  The digital zero also includes later effects all the way to the ADC. The big advantage there is that is can be fast and does not need a trimmer that may be drifty.

The offset compensated by the digital part should be more like < 500 mV,  not 10 V.  A main part would be a difference in FET gate threshold.

[Kirill V.]

Why often do not use the offset adj. pins of the operational amplifier? The question is not related to the suppression function

[MiDi]

If adj. Pins are used, most of these op amps suffer from elevated noise and drift introduced from adj. circuit.
Nowadays in most situations just choose an op amp that meets offset voltage requirements.
For volume production it is simply a cost factor to manually adj. something.
It is much cheaper to do a full digital and automated factory calibration.

[Kirill V.]

But the external offset adjustment circuit also has its own noise and drift. What are the advantages?

[Kleinstein]

Most modern OPs often don't have the Offset trim pins anymore. They are rarely used as the factory adjustment / precision got better - 100 µV offset specs are no longer that special. In addition the offset pins can add noise, pic up hum and effect the frequency response.

There is still some use to analog offset trim in OPs. Especially with BJT based OPs the offset trim also gets a low temperature drift So something like the old 741 can get low drift (e.g. 0.5 µV/K range) if the offset is trimmed to zero.
There are a few, rare examples where the offset pins can be used for alternative uses  not originally intended  (e.g. as a FB to get a kind of instrumentation amp, or to add a correction signal from another circuit).

[Kirill V.]

Let's discuss the compensation of the bias current. The K617 implements temperature-dependent correction using two thermistors, right? I propose for this purpose an IC temperature sensor with a known dependence of the output voltage. It is possible to enter a scale factor with the desired sign using the op-amp.
What do you say?

[Kirill V.]

I would also like to learn about the noise aspects when using bias compensation similar to K617

[Kirill V.]

Two thermistors in the bias current compensation circuit - I'm interested in the type and specifications

[MiDi]

I would also like to learn about the noise aspects when using bias compensation similar to K617

There are mainly two noise sources from bias compensation, thermal current noise from impedance (essentially the 250G resistor) and voltage noise from supply that transforms to current noise through the resistor.
You want to keep the current noise contribution of bias compensation low.
Maybe it is not that important for electrometer application as there is large contribution to current noise from charged particles picked up by the input.

Two thermistors in the bias current compensation circuit - I'm interested in the type and specifications

I guess there is nothing special on these 100k thermistors, they are for (partially) compensation of change of bias current of input JFET over temperature.
The glass encapsulation helps to reduce long term drift and effect of humidity and maybe pressure.

[Kirill V.]

Thanks MiDi. Yes, it looks quite clear and logical. I just compare my view of this with the opinion of professionals.

Maybe it is not that important for electrometer application as there is large contribution to current noise from charged particles picked up by the input.

Current noise flowing through the high internal resistance of the signal source can cause noise voltage, but parasitic capacitances that create a low-pass filter come into play. This is a nice bonus from parasitic parameters:)
I want to develop a temperature-dependent system similar to these two thermistors but using a temperature sensor. While I am at the stage of brainstorming. It is quite enough if it will work in a narrow temperature range near the room (like the K617, which has a working temperature range of only 18-28 degrees).

[e61_phil]

Sorry for a slightly offtopic question, but I got a broken Keithley 617 recently and before I start to repair it: Does anyone know if they used cadmium solder inside? I also got a triax cable for the 617 (quite new, from 2007) with a sheet of paper with declares, that the cables and wires contains Cd.

[MiDi]

Thanks MiDi. Yes, it looks quite clear and logical. I just compare my view of this with the opinion of professionals.

Maybe it is not that important for electrometer application as there is large contribution to current noise from charged particles picked up by the input.

Current noise flowing through the high internal resistance of the signal source can cause noise voltage, but parasitic capacitances that create a low-pass filter come into play. This is a nice bonus from parasitic parameters:)
I want to develop a temperature-dependent system similar to these two thermistors but using a temperature sensor. While I am at the stage of brainstorming. It is quite enough if it will work in a narrow temperature range near the room (like the K617, which has a working temperature range of only 18-28 degrees).

I have to regret, I am not a pro, just hobbyist.

As stated earlier, the K617 is a mid 80s design, today one would use appropriate jfet op amp, which were not awailable back in the days - e.g. LMC662, LMC6001 and if you want to get serious: ADA4530-1.
You should ask yourself what your requirements are and if it is necessary to develop a temperature compensation or just use a better op amp.
Do not forget that appropriate high value resistors are expensive and cost more than even the pricey ADA4530-1, so if you can omit bias compensation completely, this will save you cost.

Sorry for a slightly offtopic question, but I got a broken Keithley 617 recently and before I start to repair it: Does anyone know if they used cadmium solder inside? I also got a triax cable for the 617 (quite new, from 2007) with a sheet of paper with declares, that the cables and wires contains Cd.

Philipp,
I would be afraid if there is Cd-solder used, I hope that it is only common solder 
Cd-solder would not make any sense, because there is no need for low T-EMF - and with respect to the design and layout: it makes even lesser sense.
Maybe your question is better suited in the K617-Thread.
Which cable are you talking about?

[David Hess]

There is another way to make a high voltage electrometer input stage *without* bootstrapping.  A varactor diode input stage is limited only by the breakdown of the isolation transformers:

https://www.electronicdesign.com/technologies/analog/article/21801819/whats-all-this-varactorinput-amplifier-stuff-anyway

AC coupling also reduces flicker noise.

[Kirill V.]

I have not viewed this topic for a long time and it turned out that you wrote a lot of interesting things, thank you!
Yes, I know about these old methods. I will tell you in confidence that the USSR has written excellent books on this issue at a very high quality scientific level. It describes a large number of original solutions. I think that even experienced engineers on this forum do not know the things that are available to me in these books Varactors, ceramic converters, vibrating capacitors, air capacitors with amber insulators, methods for measuring the triboelectric effect, electrometer-grade vacuum tubes , and so on... Small-current radioactive sources , ionization chambers for generating small currents...

I have another question about K617. Why do some verification procedures require a jumper between the COM terminal and the chassis?