V7-45 electrometer made in Belarus (welcome in attoamps world)

[bsw_m]

Further investigation of my instrument revealed 2 problems.

One - long settling after integrator reset, but this is not very big issue.

The second is the change in the measured current readings over time, which in fact makes the last discharge insignificant. Despite the gradual increase in the instrument readings from the beginning of the cycle, after the establishment and until the end of the cycle, the instrument error is within the specification.

I think this effect may be due to the DA of the integrating capacitors.
I will investigate this effect in more detail and may be able to improve it.

On first graph, big spikes is integrator reset. And start new measurement cycle.
Second graph - more detailed one measurement cycle.

[Vgkid]

That rust is not good.

[TiN]

Little rusty, makes project 10 times more fun to repair 

[bsw_m]

This is real pain.
Time will tell if this DRC-6 can work again.
According to some insider information, tuning the DRK-6 was very difficult and could only be done by one person. The design of the DRK-6 looks quite simple, but alas, there are a lot of nuances. And alas, they are no longer recognizable, the documentation was destroyed, and the developer of this capacitor is no longer alive. And there is no one to ask.
As far as I know, 6 original measuring devices were developed for tuning and checking the characteristics of this capacitor. But what they measured, what they were, and how the DRK-6 was tuning / checked - there is no information.
I will also try to restore my second killed DRK-6, but I have no illusion of success.

And our sellers are really not good people, under the guise of a normal product they try to deceive and sell complete trash. (
It is very difficult to find such a device in an adequate condition. And into which they did not climb with crooked hands and totally did not kill him.

[Kleinstein]

There are probably a few tricky points with the vibrating capacitor part. Target should be getting a stable amplitude and about the same amplitude on both sides. If everything is symmetric there should be 2 modes of vibration, the wanted one with the opposite phase on both sides and one at a slightly lower frequency with same phase on both sides. The magnetic excitation should mainly excite the wanted mode, but if not fully centered, the other mode could be excited too. The 2 modes are probably very close in frequency(e.g. ~ 1 Hz) , so it may be not so easy to separate them. This can interfere with amplitude regulation.
Best chance would probably be to make sure the excitation is symmetric, likely coarse tune mechanical and than maybe fine tune with a resistor in parallel to one of the coils.

Another tricky point could be that an external magnetic field would lower the frequency of the wanted mode and possibly bring them even closer together. However I am not sure how strong this effect is. In the extreme case one may be able to run it as a parametric oscillator. The force would be proportional to the square of the field.
The detection part may also couple magnetically and the square law would require some bias field.

The next higher modes would than be at some 6.6 times the frequency, so high enough that they should not interfere.

I have some experience with smaller similar resonators.

I don't like the resonator construction very much, I would have used some extra common part to make it more like a classical tuning fork and this way get more frequency separation and less coupling to the fixed end.

[doktor pyta]

Interesting article about MEMS reed electrometer: 'Room-Temperature Sensing of Single Electrons Using Vibrating-Reed Electrometer in Silicon-on-Glass Technology'
DOI: 10.1109/LED.2018.2876506
+ paste DOI number here: https://sci-hub.st/

Contains coparision with Keysight B2987A.

[RoGeorge]

Interesting, I didn't expect to see sharp metal edges, including a sharp tip left from the wire snipping (can be seen in the pic, the sharp tip at the end of the horizontal gold wire).   

Seems interesting to me because I always thought electrons tend to fly away through sharp tips/edges, and leave the boundary.

Shouldn't be better (in terms of charge leakage) with all the parts rounded?

[bsw_m]

Interesting, I didn't expect to see sharp metal edges, including a sharp tip left from the wire snipping (can be seen in the pic, the sharp tip at the end of the horizontal gold wire).   

Seems interesting to me because I always thought electrons tend to fly away through sharp tips/edges, and leave the boundary.

Shouldn't be better (in terms of charge leakage) with all the parts rounded?

If you can read in Russian. There is patent for that electrometric unit.

[RoGeorge]

Sorry, I don't speak Russian, I can at most recognize a few letters, simply because some decades ago I (and many others here, in Romania) happened to use Russian IC's for TTL equivalent chips, memories, EEPROMS, etc.  I was into Z80 Spectrum "business" back then.   

Not saying that the manufacturing with those sharp edges is bad, or that the sharp terminations is impacting the performances.  Just genuinely asking why there is no rounding of the sharp edges, and trying to learn more.

As a non Russian speaking is hard for me to translate those pics, is there any particular key passage or thing I should know in order to learn more, please?

[bsw_m]

In fact, these points do not affect the noise level of the instrument. Moreover, taking into account the size of these elements, it is very difficult to make rounding on them.
For reference, the wire diameter of the contact block is 0.16mm.

[bsw_m]

Commutator in this instrument have very clean and nice construction!

[exe]

Production of new Sapphire insulators.... negotiations are underway with the plant[/li][/list]

What's the price approximately?

[Vgkid]

Thanks for those, looking forward to the restoration.

[doktor pyta]

It seems that there are two kinds of sapphire used by the manufacturer: both are transparent but one is colorless and one is violet.
Am I right?
Does anyone know what are differences between them in terms of resistance, other properties?

[Kleinstein]

The colors are from minute impurities. Likely no big effect on the conductivity, though chances are it does not get better.

Sapphire wafers are actually no very expensive - I found them as low as $15 for a 1 in (25 mm) diameter, though only thin. Also optical window are not all that expensive. With a laser they can also make precise holes and get you odd Mickey mouse shape parts.  So I don't think the Sapphire is the difficult or very expensive part any more.

[TimFox]

Is this parasitic current due to chemical reactions, or piezo-electric effects in the stressed Al₂O₃ crystalline sapphire?

[Kleinstein]

Pure sapphire should not be piezoelectric . I would more expect things like dielectric absorption and maybe surface effects, where the surface layer changes to a new equilibrium. There is also a chance to have surface charge on isolators that can persist for a long time. The slow decay can cause a background current.

[TimFox]

Bulk crystalline sapphire is not piezoelectric, but the broken symmetry at the surface can have piezoelectric effects.

[exe]

Looks simple . I wonder how difficult it would be to reproduce this device. I guess, the devil is in details, such as calibration.

[Kleinstein]

Looks simple . I wonder how difficult it would be to reproduce this device. I guess, the devil is in details, such as calibration.

There is not much calibration to go on with the resonator - essential one gain factor related to the amplitude. If used just as a zeroing detector one may not even need that factor to be accurate or even very stable. The setup however seem to have the vibrating part at ground - so the amplitude may matter.
One of the tricky parts is getting the amplitude stable and large enough. Very likely the mechanical resonator is setting the frequency. So no forced external drive but more like feedback from the sense part to the drive part and an extra loop to regulate the amplitude. Regulating the amplitude with a high Q resonator can be tricky, as the system reacts slow (seconds) and the driving signal is not just about amplitude but also the phase. It may be still doable in the analog domain, but it can be tricky, especially with the magnetic drive, that is nonlinear with more force when closer to the magnets. So without regulation the amplitude has a good chance to be unstable.

I personally would consider a piezoelectric drive (or at least detection) instead of the magnetic. For the amplitude, I would expect that one may get quite some strain at some points, so a careful mechanical design and a quite hard material may be needed to avoid fatigue. If possible to machine I would consider to have the reeds thinner at the top to have less moving mass and higher frequency.
If one calculates the accelerations seen at the tips of the reeds, one gets quite impressive numbers.

The magnet cores taken apart look interesting with a local magnetic return - so the main carrier part could be non magnetic stainless. I was wondering how they kept the direct magnetic coupling low.

[TimFox]

How much amplitude of the vibrating plate is required?  It may be difficult to obtain with a piezo element.

[Kleinstein]

To get a good conversion factor the capacitance should change significantly. So the plate distance should change significant by something like a factor of 2. So I would expect some 0.5-1  mm amplitude at the tip as a target.
A good mechanical alignment could reduce the required amplitude - the plates should be quite well in parallel.

The mechanical resonators have a high quality factor (could be at some 100-1000 - in extremes also more) so the amplitude can get quite large compared to the static displacement. So it only needs a static displacement in the 1-10 µm range. Piezos can produce quite some force - however mounting them can be tricky and glue / solder would add damping.

I have seen / used at the university lab a piezo driven vibrating optical chopper that got some 5 mm amplitude, though with a much lower frequency and maybe 3 times the length.

[MadTux]

I have seen / used at the university lab a piezo driven vibrating optical chopper that got some 5 mm amplitude, though with a much lower frequency and maybe 3 times the length.

A thin, long piezoelectric element and vacuum inside the vibrating reed capacitor should do the trick. Air resistance of vibrating reed likely degrades Q a lot, especially at high frequency/large amplitude.
 Without air, it's only the mechanical power dissipation of piezo element, that degrades Q.

[Kleinstein]

The air gives some damping, but one can still get a reasonable high Q. In vacuum I got a Q of slightly over 1 million with a silicon resonator - that is competing with quartz crystals. With metals one usually gets a Q factor more like around 1000-10000 with low damping material. It can be around 100 with high damping materials like cast iron or magnesium. There is also some intrinsic loss in the materials. The usual PZT piezos would likely be glues / soldered to the main metal part - a little like the cheap beepers. For a first test one could do some experiments with such an element and an attached reed for extension.

[TimFox]

To get larger motion, PZ elements are used as "bi-morphs", where opposite polarity voltage is applied to parallel strips so that one expands while the other contracts (like a bimetal strip in a thermostat), exchanging increased deflection for reduced force.  An interesting problem would be shielding the sensitive node of this circuit from the PZ drive voltage.