Null meters still in production? How to replace a Keithley 155 null meter?

[Kleinstein]

There is rather limited need for a null-meter and there are still some of the old null meters around. So the market for new meters is pretty small as most owners of a 752 likely have a suitable old null meter. It is still a bid odd not to have new ones, as it does not have to be expensive and may be usefull for a few other tasks too. Compared to the old ones a new one could offer some 4 digit resolution and better than 1 % accuracy to also use it as a more general battery powered voltmeter for small voltages (though not a full nV meter). The main point would be to have low bias and maybe a trim / test circuit for the bias. For good isolation battery operation is an easy solution as modern meters can be low power.

With the 752 the source impedance (bath sides combined) should be at some 100 - 400 K ohms. So there is no need to have super low noise, as there is always quite some resistor noise. It is more like a need for very low input current noise. So when using an AZ OP to build a null meter it would be more like looking for a type like AD8628 or even max4238  and not LTC2057 or OPA189.
The old null meters are also not really low noise, but good enough for the job.

[TimFox]

When actually adjusting for a null, I find an old-fashioned d"Arsonval meter easier to use.
For a nanovolt meter, a reasonable digital readout may be better.

[Kleinstein]

Anlog and digital read out have there advantages. With a digital signal one can use a more FIR type filter and this way get relatively fast settling, which can be a big plus.

[e61_phil]

The Keithley 182 has a DAC output which gives an analog signal scaled to FS of the range. That DAC is updated with the display. Therefore, it applies the same filtering and you can connect a meter with a needle.

[guenthert]

Null meters are comparatively (say against a longs scale DMM) simple devices (self-calibrating, no need for high linearity or speed).  I think one with the performance of a HP419A is today very well within the reach of  a dedicated amateur.  The Keithley performs quite a bit better (low enough bias current that a bucking voltage supply was not deemed necessary), but here people rose to greater challenges.

HP419A can be considered obsolete and many of them are meanwhile old enough that their neon bulbs / LDR based choppers don't work within specification or at all anymore and hence occasionally available for reasonable prices on the secondary market.  I'd think the (double-walled!) chassis, voltage divider, large MCM and low thermal EMV input connectors would make a good starting point for a self-made one with modern amplifier.

[syau]

Wonder if a nanovoltmeter be used as null meter 

[RoadDog]

I’m surprised no one is making a quality analog null meter anymore. There are other options yet for sub PPM measurements I don’t think anything compares. I’ve used a few different ones and the best was an 845ab that was in great condition. It was fast and extremely accurate. When the capacitors get old they get slow to null and start giving bad readings. The chopper amps get old as mentioned and you lose linearity through the ranges.

The battery powered 845 is much better than the rack mount 845ar in my experience.

[Dr. Frank]

@ martinr33, e61_phil, Kleinstein, syau:

The null-meter requirement for 720a, 752a is to be uncertain to <= 500nV or <= 200nV respectively, (see manuals), which in practice includes resolution and noise.

For sure, nV-DMM fulfill both requirements, but they may be used only, if their bias current is < 1pA.
This is also not specified for those under consideration (34420, 182, 2182, etc), as low noise and low bias are always contradicting parameters.   

Newer, low noise Chopper OpAmps also have high bias on the order of 20pA and more, and therefore are not suitable at all. 2057: 30pA typ, up to 300pA!
Only older ones like the LTC 1052 might do the job, but must be selected to below their typical bias, i.e. 1pA typ., 30pA max.
Here again, lower bias means higher noise.

Those old optical chopper null meters are the appropriate instruments for this job.

Frank

Edit: I just had a look into my recent tests on the 752A. From that I have to emphasize, that a proper shielding configuration, following the 752A manual addendum, is much more important for noise figures, and in turn for zeroing correctly, than having an ultra-low noise null-meter.   

[branadic]

Fortunately, it was shown that K155 F845 can also work with modern (H11F1) component based optical chopper, no need for neons and LDRs or mechanical choppers.
https://xdevs.com/fix/f845ab/#chopi

-branadic-

[syau]

Fortunately, it was shown that K155 F845 can also work with modern (H11F1) component based optical chopper, no need for neons and LDRs or mechanical choppers.
https://xdevs.com/fix/f845ab/#chopi

-branadic-

Any body have the schematic fir the H11F1 chopper ? Want to upgrade my 845.

[alm]

Dave Wise's posts ans responses by others here and further might help.

[branadic]

The idea that a Fluke 8588A can serve as a null detector might come from the marketing claim of the 1281:

Add the fact that it offers exceptional 0.1 ppm (± 2 μV) single-range linearity from zero to 20 V, and you realise that the Model 1281 not only substitutes for Weston cells. It also doubles up as a highly sensitive null detector (with an input impedance >10 G½) and a Kelvin Varley divider.

Source: https://xdevs.com/doc/Datron/1281/doc/1281_spex.pdf

As the Fluke meters (8508A, 8558A and 8588A) are kind of successors of the 1281 this claim might have sustained without further proove? 

-branadic-

[1audio]

I just saw this https://www.analog.com/media/en/technical-documentation/data-sheets/adgm1001-1002-1003.pdf  MEMS spdt switch that could be in effect a mechanical chopper. Could this be the basis of an improved null meter? Its not cheap at $25 but its good for 100 million cycles. . . 3.5 Ohms on resistance.

[Kleinstein]

I just saw this https://www.analog.com/media/en/technical-documentation/data-sheets/adgm1001-1002-1003.pdf  MEMS spdt switch that could be in effect a mechanical chopper. Could this be the basis of an improved null meter? Its not cheap at $25 but its good for 100 million cycles. . . 3.5 Ohms on resistance.

They seem to have some high resistance to ground (likely to allow capacitve actuation of the switches. This may be acceptable, but would need some extra care. Otherwise a leakage current in the 0.5 µA range is a big no go.

I see not real problem in using one of the lower bias AZ OPs (e.g. maxa4238, ltc2055, AD8551, LMP2011, LTC1052, ... maybe up to AD8628) combined with a trim / compensation for the bias current. There is a limited need for a discrete build chopper and if so, I would consider CMOS switches or JFETs (e.g. 2N4117). The zero voltage check could be with a latching relay or mechanical switch if needed.
The LDR choppers are not magic and the H11F1 photo fets are more like poor performance (quite some offset and still possible charge injection) by todays standard.

[ttttrigg3r]

Thanks for doing that. I'm not that savvy with electrical circuits, but I'll analyze it further. I will take your recommendation and follow up with Fluke to see how they're doing it on their end with the zeroing method.

[ttttrigg3r]

The 752A is a current product, it would be amusing if they had some decades-old null meter doing this job during production.

They use the 845 for the 752 and 720.

You're saying Fluke uses the 845 null meter to zero the 752? In that case, we're doing exactly what they're doing...

[wutieru]

I think nullmeter can be replaced by KEYSIGHT 34420A or KEITHLEY 2182A nanovolt meter if using in Voltage measuerment.

[Kleinstein]

I think nullmeter can be replaced by KEYSIGHT 34420A or KEITHLEY 2182A nanovolt meter if using in Voltage measuerment.

It depends on the application:
In some cases with a low source impedance those low noise DMMs are a good solution, way better than the old Fluke 845 or similar. In this case other DMMs can also work as good or better than the old null-meters.
In the case in question at the start of the thread (use with a Fluke745) those nV meter are however a poor choice, as they are not suitable for a high source impedance.

The old Nullmeters like Fluke 845 are not especially low noise - noise wise a Keithley2000 or HP34401 are likely better. The point where the Fluke 845 is better than most DMMs is with the input bias and maybe isolation from mains (though battery operation would be the obvious better solution). They are more like inbetween a DMM and a electrometer: low bias in the <10 pA range, low current noise and a moderately low noise to resolve some 100 nV with enough time (e.g. 10 second average).

The market seems to be too small (and maybe enough of the old ones still alive) to offer a new improved alternative, which would not be very difficult or expensive (BOM wise, but the small volume may add). The AVM2000 is a more modern try in that direction, though in my oppinion not a very good one.

[Dr. Frank]

I have published my findings on the 845AR vs. 3458A using them for the 752A.

https://www.eevblog.com/forum/metrology/influence-of-switch-resistance-in-hamon-dividers/msg4207087/#msg4207087

Frank