Is the quest of replicating a Datron 470x calibrator totally foolish or crazy ?

[DC1MC]

Well somehow getting bored of trying to find a V/I/R calibrator of reasonable cost and weight  , I was browsing the almost unreadable scans of some Datrons 4707/4708 and looked firs at the digital part, that I found to be really naive, even for the time, and also I've had a look at the analogue part and, while complex, it didn't seem so magically unreproducible, hell, even the original transistors used in the original chopper along with the most other components are still available as NOS. Nothing like the Ferrenti unobtaniums form the Datron multimeters.

If I'm not missing something ultra-obvious, I think that at least the low voltage part should be replicable with the current technologies and components for not too much money, and in the worst case I'll make a precision power supply out of it or at least gain some experience   .

Now of course, everybody knows that something is not possible, until some fool doesn't know and tries, what are you saying, did someone already tried it, any experiences, advice and stuff are welcome.

Or else I'll have to make a big dent in my X-mas bonus with this guy:
https://www.ebay.de/itm/Kalibrator-Datron-4000-DC-Kalibrator-Spannung-Stromstärke-Widerstand/401619940787

 Cheers,
 DC1MC

[Echo88]

Assuming the Datron uses a PWM-DAC: Its not the problem to replicate the pcb of such calibrators like the Datron or a Fluke 5700A.
Its just an insane amount of work to replicate them/characterise them and in the end you will spend more time and money on rebuilding something that can be bought used and with known specs.
I still have a Fluke 5700A PWM-DAC PCB which i wanted to use as a high-precision +-10V DAC, but invested my time in other projects. You can buy the PCB from me and use that as a starting point.
Lets take a look at another piece of equipment nobody has yet copied (AFAIK), despite it being relatively simple: the Datron 4910.

[Kleinstein]

There is one important part that is difficult to copy: that is a proven track record of reliability.

The digital part is expected to be relatively simple, it's likely a more or less standard 8 Bit computer of some type. Nothing fancy at that time, as there is no real screen. Still the computer part needs some software - writing this can be tricky, as compilers for the old CPUs where not that common and advanced.

The tricky part is more in the analog side and with the details. It might be a good idea to have a look at the Datron 1281 service manual to see what type of bodges may be needed to get high performance. Much of these point's are the part that is no in the schematics, but in the layout and careful selection of parts and having the tong at the right angle.

There are quite some points, one would do different today (e.g. replace pulse transformers with optical means). PWM generation in an µC.

Some of the transformers and also some dividers can be custom made and difficult to get.

[DC1MC]

Thanks for the answers so far, so AFAIU, it's not a (completely) crazy idea, just hard to be done, but who doesn't love a challenge  , now for some punctual answers, in order:

@Echo88: Well, a lot of things could be bought used, but there are different degrees of used, knowing my luck I'll end with the one that had the elcos vomiting their soul over the irreplaceable parts  .
If I'll ever manage to produce something resembling some stability I'm sure that some forum members with cool toys can be bribed to help me characterize it .
The idea here it's not to get some used parts from existing calibrators and stitch them somehow together, but a new design that somehow it's in spirit of the old one, and the Datron design is free of unobtaniums and doable with what I have, of course if you or other forum member can recommend other interesting calibrator schematics, I'm all ears .
The Datron 4910 seem to be a set of 4 LTZ1000 based 10V references in the same box, the isolation cuts reminds me of something  :



Anyway, we have a flurry of 10V references design, but just one gone calibrator project that died around 2016.


@Kleinstein: The proven track of reliability  has to start somewhere, the digital part is at the level of Arduino/CPLD/ATiny and I will totally not use some obsolete compilers for 6800 . Same with the pulse transformers, there are plenty of them to chose from and they cost almost nothing, I don't think that they are something very special to switch a transistorized flip-flop, they have to be balanced and stable, I've seen a some nice candidates.
As I've said, I want to replicate the concept, not the old schematic, on the selection of parts/layout/orientation of tongue , after the proof of concept, I'll definitely use the forum wisdom (you're on the first place on the list)  and replace them with their best equivalents in order of influence of the overall quality. A resistive divider it's a divider, it can be made of high precision parts or of affordable parts, as long as we know what it influences, then this is it, the rich will get Vishay vacuum packaged dividers, the poor will put what they have/find, if the general schematic supports it, then it's OK.

 So keep the suggestions, advice coming please, I'll try to answer to everybody, while getting to become friendly with Kicad and do a schematic of the main chopper, hopefully someone will help me with a way to simulate it, especially the Bessel filter.

 Cheers,
 DC1MC

[Kleinstein]

Only replicating the general concept makes more sense, as the the computer part is old and could be done much easier with newer parts. Just mentioning NOS transistors made me think of a more or less direct copy.

Also for the analog part there are quite a few old style circuits with discrete JFET pairs, which today would likely use JFET OPs. So things might get easier in some areas.

The AC part also has one relatively odd point - the quasi sine generation. This looks odd by todays standards with DDS chips available.
So given modern parts, I would expect a quite different AC section: more like a DDS type generator and possibly also a different kind to AC detector/ amplitude correction.
The AC part might anyway end up as a second, later project.

There are also schematics for the Fluke 57xx around somewhere.  At least worth a look, if you really plan to build such a system.
There are some DIY PWM DAC circuits around too.

[Echo88]

The magic of Datron 4910 lies in its PWM-divider, which substitute the resistor-divider to get 10V.
I have no experience with the Datron-schematics, but thought about the Fluke 5700A-PWM-DAC a lot and might help if questions arise in this respect.
5700A Service Manual with schematics: https://xdevs.com/doc/Fluke/5700a/5700A_old_sm.pdf Page 542 (540-548)
Page 99 presents an overview of the circuit-blocks used in the DAC Assembly.
LTSpice-Simulation of the 5700A-PWM-DAC, including the PWM-Filter you mentioned: https://www.eevblog.com/forum/metrology/2000-post-teardown-and-study-of-fluke-5700a-calibrator/msg1372679/#msg1372679
To get the extremely good specs of such a calibrator you need to understand the circuit-tricks which are used, such as the sense current cancellation or the avoidance of RDSon and its associated losses in switches Q4-7.

I dont want to talk you out of this project by any means. 

[Johnny10]

I have often wondered if it was possible to replace the reference in a Datron 4000A with the LTZ1000.

You can buy Datrons in US for 400-500 not working.
They use all discrete easy to find parts..

[DC1MC]

One more set of answers:

@Echo88 Many thanks for sharing the simulation, I've discovered that, at least for viewing, LTspice works on Linux under wine, hopefully I'll be able to simulate my own schematic. Also thanks for recommending the Fluke schematic, I've only had the most cursory look so far but I'll study it later, now I have a little schematic to finish.

@Johnny10 Don't mention US anymore or I'll start talking about "American privilege"  , you guys are awash in cool CHEAP stuff, no wonder you lead the innovation in everything, here in the social-democratic world, you have to pay 20% VAT for every derelict crap that you buy for yourself, shipping costs included in the taxable amount  . 4-500$ Datron calibrators, it's kind of a dream.

Anyway, the reference voltage question it's a very good one, that I've intended to ask myself to the local volt nutters:

- Could any of those  modern and available 10V references be put in series and still keep a bit of precision and stability ?

- @Kicad friends: does the netlist expored by the strangely named schematic capture program can be loaded into the LTspice, or even better yet, can LT spice be integrated with Kicat (the absolute latest version, nightly).

 Cheers,
 DC1MC

[DC1MC]

Also my eyes are fairly stranded and tired by the crappy quality of the schematics I've found for Datron 4708 and 4707, if somebody has the real deal on paper I'm willing to buy it or pay for a professional scanning to share it with the forum.
Also if there are such quality scans somewhere and I've missed them, kindly please let me know.

 Cheers,
 DC1MC

[Echo88]

Yeah, ebay.de has way less interesting test equipment compared to the US. Can someone please tell again how to see all ebay.com-articles, not only those who are enabled to be internationally shipped?
 

[DC1MC]

Yeah, ebay.de has way less interesting test equipment compared to the US. Can someone please tell again how to see all ebay.com-articles, not only those who are enabled to be internationally shipped?
 

This is how I do it, open a private browsing window (Ctrl-Shift-P for Mozilla), go to ebay.com, do NOT login and browse the (almost) unobtainable stuff.
Now if somebody can tell me how to turn off this unbelievable shitty automatic translation on ebay.de, I'll definitely  PayPal you a beer.

 DC1MC

[Echo88]

Nope, there isnt a difference in article-count between Private Mode and normal surfing. I think one needs to use a VPN-tunnel, change the ebay-account to a US-based-adress or something like this.
Want to get a voltage reference like a Datron 4910/Fluke 732A, for a reasonable price.

[DC1MC]

Nope, there isnt a difference in article-count between Private Mode and normal surfing. I think one needs to use a VPN-tunnel, change the ebay-account to a US-based-adress or something like this.
Want to get a voltage reference like a Datron 4910/Fluke 732A, for a reasonable price.

Now and then, using the method described, I got "No international shipping", but more worrisome is the lack of good FETs, booth N and P channels, to replace the J1xx used in the chopper, of course there are a lot of NOS around but it seems that Vishay and everybody else is going to quit making nice FETs .

So any suggestion of replacing the following is welcome:
N-FET
Siliconix J112, J108

P-FET
Teledyne J174
Siliconix J175

These were initially sorted according with their parameters, and now only the useless ones are available on the main distributors (I have a feeling that only because were not sold yet, and once gone, they'll be gone 4 ever).

Any suggestions for replacements are most welcome.

 DC1MC

[Vtile]

Nope, there isnt a difference in article-count between Private Mode and normal surfing. I think one needs to use a VPN-tunnel, change the ebay-account to a US-based-adress or something like this.
Want to get a voltage reference like a Datron 4910/Fluke 732A, for a reasonable price.

Now and then, using the method described, I got "No international shipping", but more worrisome is the lack of good FETs, booth N and P channels, to replace the J1xx used in the chopper, of course there are a lot of NOS around but it seems that Vishay and everybody else is going to quit making nice FETs .

So any suggestion of replacing the following is welcome:
N-FET
Siliconix J112, J108

P-FET
Teledyne J174
Siliconix J175

These were initially sorted according with their parameters, and now only the useless ones are available on the main distributors (I have a feeling that only because were not sold yet, and once gone, they'll be gone 4 ever).

Any suggestions for replacements are most welcome.

 DC1MC

J108 and J112 are still in production from ON semi.

[jfphp]

A good starting point for a modern DC V and I calibrator is the Adret/Marconi 103A with its F/V converter. Manual is detailed and the modern circuits allow a tremendous X100 precision and stability (we have allready tried), a dimension downsize... Readout and commands can be made by a PIC or a PC. Another story is the AC.

[Kleinstein]

Nope, there isnt a difference in article-count between Private Mode and normal surfing. I think one needs to use a VPN-tunnel, change the ebay-account to a US-based-adress or something like this.
Want to get a voltage reference like a Datron 4910/Fluke 732A, for a reasonable price.

Now and then, using the method described, I got "No international shipping", but more worrisome is the lack of good FETs, booth N and P channels, to replace the J1xx used in the chopper, of course there are a lot of NOS around but it seems that Vishay and everybody else is going to quit making nice FETs .

So any suggestion of replacing the following is welcome:
N-FET
Siliconix J112, J108

P-FET
Teledyne J174
Siliconix J175

These were initially sorted according with their parameters, and now only the useless ones are available on the main distributors (I have a feeling that only because were not sold yet, and once gone, they'll be gone 4 ever).

Any suggestions for replacements are most welcome.

 DC1MC

Many those fets are still available in SMD (SOT23) case as MMBFJ...,
MMBFJ112 and MMBFJ175 are no problem to get.

If I understand the Datron design correct, it uses very low ohms and matched switches. The Fluke5700 design uses compensation of the resistance by using separate switches to supply most of the current.

[DC1MC]

Now and then, using the method described, I got "No international shipping", but more worrisome is the lack of good FETs, booth N and P channels, to replace the J1xx used in the chopper, of course there are a lot of NOS around but it seems that Vishay and everybody else is going to quit making nice FETs .

So any suggestion of replacing the following is welcome:
N-FET
Siliconix J112, J108

P-FET
Teledyne J174
Siliconix J175

These were initially sorted according with their parameters, and now only the useless ones are available on the main distributors (I have a feeling that only because were not sold yet, and once gone, they'll be gone 4 ever).

Any suggestions for replacements are most welcome.

 DC1MC

Many those fets are still available in SMD (SOT23) case as MMBFJ...,
MMBFJ112 and MMBFJ175 are no problem to get.

If I understand the Datron design correct, it uses very low ohms and matched switches. The Fluke5700 design uses compensation of the resistance by using separate switches to supply most of the current.

Your understanding is, of course, correct, the J108 it's the elephant in the room, and I've seen the MMBFJ guy but seem to be off regarding the specs of the originals.

 Well, we'll see what comes of it, most of the stuff it's ordered, the scary part it's the filter but for a proof of concept I should use what it's around.

 Many thanks for the support, advice from masters like you are always most welcome .

 Cheers,
 DC1MC

[Kleinstein]

The filter part should not be that scary. I don't know the circuit of the 4708, but the 4910 filter looks not that bad, if one uses modern JFET based OP instead of the old  discrete JFET pair + NE5534.

For the MMBFJ... types and the exact specs, I have seen datasheets which are for both J... and MMBFJ... with the only difference in the case. There can be some differences depending on the manufacturer. Also the ultimate leakage and maybe low frequency noise may be effected by the case, but hardly low impedance digital switching parameters. There is quite some scattering in the parameters anyway. So a J109 might well meet the J108 specs if one is lucky.  As they seem to use some compensation, it might need selected fets.

[MK]

A good starting point for a modern DC V and I calibrator is the Adret/Marconi 103A with its F/V converter. Manual is detailed and the modern circuits allow a tremendous X100 precision and stability (we have allready tried), a dimension downsize... Readout and commands can be made by a PIC or a PC. Another story is the AC.

A quick google did not turn up any, do you have a location to share where one could get the manual(s)?

[Echo88]

https://doc.xdevs.com/docs/Adret/103A/ unfortunately in french

[Kleinstein]

The instructions have an English part at the end.

The Adret 103A is using a similar PWM ADC type to the Fluke5700 or Datron 4910,...
However it looks like lower quality, with an LM399 reference and using just CMOS logic chips to drive the PWM. The OPs also look a lot less accurate.

I would consider using CMOS switch ICs like DG419 and the on resistance compensation like the Fluke circuit. This would be considerable easier than the discrete FET switches and could also have less transient load on the reference sources. Those low resistance switches from discrete FETs can have quite some transients seen by the reference - so one would need quite some filtering here.
AFAIR the resistance compensation scheme needs a voltage that is something like 50% higher for the compensation. So with a maximum of some 30 V for the DG419 one could in theory go up to about 20 V reference if really wanted. I would still expect the reference buffer before the PWM stage one of the more tricky parts. Chances are one could use something like a 14 V reference as a starting point, from two 7 V reference in series or for the beginning with a times 2 amplifier.

The fluke circuit includes an optional a times 2 gain. This makes the circuit look rather complicated. It gets much simpler without that extra optional gain.

[DC1MC]

@Kleinstein Well, if you can sketch something that replaces those bloody sorted and "endangered species" JFETs with some COTS circuits, it will be sehr nett . Any picture of a piece of paper will be enough, it can't be worse than the Datron 4708 schematics that I have.
My goal is to produce a reference divider as a module, that depending of the quality of used components could go as precise as possible. The Datron one seem to be the one that I understand the most and seem doable within my means, that mean a bit over average volt passionate, so to say .
Looking at Fluke's diarrhea of relays and complications I've been kind of put off, but if you can somehow simplify it and still keep it precise and doable, then it's wonderful.
At my level
I was just looking at some Ethernet network and ISDN pulse transformers, they are dirt cheap, because another goal is to keep it reasonable cost effective and with parts with a long active life, no some soon to be unobtaniums. I've replaced so many optos that I've kind becoming allergic to them .

Anyway, if nothing will show up, I'll continue with the Datron calibrator implementation, but it's never too late for a good design.

 Cheers,
 DC1MC

[Kleinstein]

In a simplified version the R_on correction and thus linearity improvement Fluke is using in the 5700 is about in the way shown in the attached drawing. I hope I got the resistor rations right. The circuit is wrong with the lower 2 OPs ! The main idea is that essentially all the current is coming from the lower PWM switch with the higher voltage and the upper switch is only correcting small errors. This way the current through the upper switch is very small and thus the on resistance much less important.  In the form shown the lower switch has a differential voltage that is 1.5 times higher than the original reference.

The filter in only simplified with a simple switch. The main buffer should likely be an AZ type, possibly with an extra driver and sense. It might also need an extra buffer for the dividers - but this depends on the rest of the circuit.

With modern µCs I would not use galvanic insulation for the PWM signal, but get the PWM signal directly from an µC (e.g. AVR or similar 8 Bit) and only send control values via UART and opto-coupler. The modern µCs tend to cause not much more noise / spikes than  sychronising flip-flops needed if the PWM signal is send via pulse transformer or similar.

p.s. corrected version of the circuit is at post #32/33

[branadic]

It's been a while ago now that I've worked on the schematic for a PWM-DAC based on the approach presented on EDN and modified by Andreas. Still haven't managed to draw the board, but will go on soon. I put the schematic here for your inspiration.

-branadic-

[Kleinstein]

If I remember right the EDN circuit was not that linear. There is no compensation for the switch resistor difference, and the max4053 switch is relatively high resistance.
The interesting part might be the filter: it is a sampling filter, that could thus result in very good ripple suppression. However sampling a signal with some ripple (after a first and maybe second filter stage) will cause some error and thus nonlinearity unless there is a phase adjustment to sample at zero crossing).  For use with a DMM, i would prefer ripple over nonlinearity.

I just found an old thread about PWM DAC. It looks like someone already kind of copied the datran 4910 circuit.
https://www.eevblog.com/forum/metrology/volt-reference-diyer-ltz-10v-pwm/msg1246656/#msg1246656

The copy is quite close to the original - one could definitely replace the JFETs + NE5534.

[DC1MC]

In a simplified version the R_on correction and thus linearity improvement Fluke is using in the 5700 is about in the way shown in the attached drawing. I hope I got the resistor rations right. The main idea is that essentially all the current is coming from the lower PWM switch with the higher voltage and the upper switch is only correcting small errors. This way the current through the upper switch is very small and thus the on resistance much less important.  In the form shown the lower switch has a differential voltage that is 1.5 times higher than the original reference.

The filter in only simplified with a simple switch. The main buffer should likely be an AZ type, possibly with an extra driver and sense. It might also need an extra buffer for the dividers - but this depends on the rest of the circuit.

With modern µCs I would not use galvanic insulation for the PWM signal, but get the PWM signal directly from an µC (e.g. AVR or similar 8 Bit) and only send control values via UART and opto-coupler. The modern µCs tend to cause not much more noise / spikes than  sychronising flip-flops needed if the PWM signal is send via pulse transformer or similar.

Wow, this design it's really compatible with the Datron extravaganza ?!?!?!
This is crazy !!!
Before I'm looking closely into it, I'm asking some naive questions first:

1) The output voltage it's strictly depending of the 1/0 ratio of the PWM pulse ?

2) Where does the ferocious Bessel or similar filter go ?

3) What it will be the base frequency of the PWM pulse ?

4) What else needs to be done to make it a usable design for a mini-calibrator and replace the Datron design, I'm more than willing to invest some time and materials to produce some prototypes ?

I'm looking to some low power uC to generate a high resolution pwm ab sofort !!!

 Cheers,
 DC1MC

[Andreas]

1) The output voltage it's strictly depending of the 1/0 ratio of the PWM pulse ?


3) What it will be the base frequency of the PWM pulse ?


I'm looking to some low power uC to generate a high resolution pwm ab sofort !!!

1) no, depends on switch resistance over voltage and on/off time difference.

3) I would go for a multiple of mains frequency. (or 244 Hz will also be a good start for 50 + 60 Hz simultaneous).

I would use either PIC24FV32KA302/304 or ATMega1284.

with best regards

Andreas

[DC1MC]

1) The output voltage it's strictly depending of the 1/0 ratio of the PWM pulse ?


3) What it will be the base frequency of the PWM pulse ?


I'm looking to some low power uC to generate a high resolution pwm ab sofort !!!

1) no, depends on switch resistance over voltage and on/off time difference.

3) I would go for a multiple of mains frequency. (or 244 Hz will also be a good start for 50 + 60 Hz simultaneous).

I would use either PIC24FV32KA302/304 or ATMega1284.

with best regards

Andreas

 How is the switch resistance variation with the temperature compensated, Keinstein was mentioning that there is a switch resistance compensation but I cant't see it, dumm Kopf me ?

 DC1MC

[Andreas]

Hello,

In branadics cirquit it is R71 + R19 (carefully adjusted to the switch used).
In EDN cirquit it is the 5R resistor at the same place.
But of course that works only for exactly one temperature.

If you have a processor you could also do compensation in software.
(but you have a hen and egg problem).

For switch linearity (without adjusted compensation) see also here
https://www.eevblog.com/forum/metrology/7-5digit-diy-voltmeter/msg1400157/#msg1400157

with best regards

Andreas

[Kleinstein]

The Circuit based on the EDN idea, that brandic showed needs those resistors to adjust for the ON resistance of the switch. This can work over a limited temperature range.

The circuit I showed is a simplified version to show how the Ron compensation in the Fluke 5700 works. Here the one capacitor shown would be replaced with the filter circuit (e.g. 3 OPs + quite some large caps), like in the Datron 4910 or Fluke.  The circuit is still missing the reference and reference buffer. 

The rather complicated JFET switching part is replaced with the DG419 - at the time the Datron4910 was designed the choice of CMOS switch chips was still rather limited. I don't see a reason the DG419 would be much inferior, especially with R_on compensation. There might be even advantages from less charge pulse during switching.

The output voltage would depend linear on the 1/0 ratio, but there still would still be an offset, e.g. due to leakage current and charge injection from the switches.  There can also be some limitations at the extremes.  So its really linear only from some 2-98 %. So to get close to zero one would need to add an offset (e.g. 5%) and have a zero detection circuit for the output. The output range would than be some -3% to 93% of the reference voltage.

The circuit I showed is also only the coarse part - so it takes some small additions to get the lower 8-10 Bits.

The Ron compensation depends on the resistor ratios. Normally something like a 100 times error reduction sound plausible. With some adjustment one might get an even better improvement.

The PWM frequency is a compromise between settling and filter caps that prefer a higher frequency and switching errors, offsets that would be smaller with lower frequency. So some 100-500 Hz is likely a good compromise. The main part to change would be the capacitors. These can be quite large, as for best performance some should be PP caps with quite some capacitance. There might be a chance that PPS type caps could also work, but the data I have found so far are confusing: some show very good DA performance (even better than PP) and some show poor performance (more like Pet).

For just PWM generation (analog side) a small µC like PIC16 or Mega88 or even a 8-14 Pin version  should be well sufficient.  The control of the rest (PC interface side) would be a different thing. Here it would take more memory etc.

[Pipelie]

I have replicated the Datron 4910 and success at a couple months ago.

I only use the concept of Datron 4910, the digital section is replaced by a CPLD and change the JFETs to the CMOS Analog switch, filter remained the same without JFET. the output stage is similar.

I'm also building a PWM DAC that can source +/- 10V, the circuit is base on Datron 4000 and combines with the idea of ACAL from fluke 5440. it's working now, but I'm still having problems that related to the ACAL function.

[DC1MC]

@Pipelle - Very interesting project, could you share (some of) the schematic with us ?

 Many thanks,
 DC1MC

[Kleinstein]

The plan for the PWM DAC I showed earlier here has a mistake with the voltages for the lower switch. The corrected version is not that much different, even slightly simpler.  One might want to adjust the resistors shown as 5 K a little, a slightly higher value to first order compensate for the R_on of the lower switch.  The resistors could also be higher value.

The place where the capacitor is shown, there should be a filter section like Fluke, / Datron use them.

[DC1MC]

The plan for the PWM DAC I showed earlier here has a mistake with the voltages for the lower switch. The corrected version is not that much different, even slightly simpler.  One might want to adjust the resistors shown as 5 K a little, a slightly higher value to first order compensate for the R_on of the lower switch.  The resistors could also be higher value.

The place where the capacitor is shown, there should be a filter section like Fluke, / Datron use them.

Thanks for the refinement of the schematic and sorry for the naive questions, but:
Assuming that the 12V on the schematic it's the actual Vref, and the upper switch switches from from Vref to Zero, the small negative offset to be added is onstead of  the GNDA part of the upper switch or on the not yet described fine tuner part ?
The 7-pole Bessel filter from the Datron could be used here as well, or should I implement the Fluke specific one ?
Finally, an interesting project, it prompted me to start sorting and start adding what parts I have into the PartKeepr stock management program, very nice application, too bad it was abandoned, but is extraordinary useful as it is. If out of this comes the organizing of my stuff, and is good.
If the schematic presented here stabilizes, I'll order the precision OP and the CMOS switches, I have some 1% metal resistors around the values shown to sort from, so next WE I should be able to do the first tests, finally finding some use for the FeelTech 6600 AWG as an initial PWM generator, so I need to fix its power supply, output OP and other stuff that was idling and of the recapped Datron 1082 as a reference voltmeter. 

But if out of this crowdsourcing comes out a (relatively) simple, cheap and reproducible reference divider, that, depending of the parts used, can go from adjustable power supply to calibrator grade divider, I will consider it the achievement of 2018 and you'll be one of the people to thank.

 Cheers,
 DC1MC

[TiN]

I wish all the best to TC on this huge project. I had similar ideas and bought many parts for such a project some years ago (you still can find shameless thread at EEVBlog about it), but gave up on this and build Fluke 5720 from scrap dead parts (all PCBAs are originals, though, just some missing chassis parts are DIY).

[DC1MC]

I wish all the best to TC on this huge project. I had similar ideas and bought many parts for such a project some years ago (you still can find shameless thread at EEVBlog about it), but gave up on this and build Fluke 5720 from scrap dead parts (all PCBAs are originals, though, just some missing chassis parts are DIY).

We are talking here of a different level, at least financially speaking , but your expertise is anytime welcome.

 Cheers,
 DC1MC

[Kleinstein]

The offset to add would be a negative offset. This could be either a shift in the negative side of the reference, or just adding a negative current through a high resistor to before the filter "cap". The fine DAC part could also be added that way, possibly together.  The fine DAC part often is also a PWM circuit, but could in principle also be just a 12 Bit ready made DAC chip. The resolution would normally be split in a way to get some 16 bits from the coarse PWM DAC part shown so far and than some 8-12 extra bits from the fine DAC. For adjustment of the fine to coarse scale it might help of the fine part could help if the fine DAC could cover a little more than 2 coarse steps, maybe 4 steps. As the filter is already there, the obvious way is to use PWM for the fine part too.

With the offset added, the DAC would cover a voltage range of something like -100 mV to maybe - 12 V (with a 14 V reference Level).
Due to the higher voltage at the correcting switch it would get tricky to get a much higher reference level - some 20 V might still just work.
I have not seen it in the Fluke / Datron schematics, but It might help to have some feedback path at around zero output to adjust the Zero setting. This would be something like an AZ OP with quite some gain (e.g. 100-1000 times) going to an ADC (could even be the µC internal one). This could also help to bring the fine an coarse scale together, kind of a build in null meter.

PWM from the generator would be only for a very crude test, to check for ripple.

The 2 resistors at the switches should better be relatively low TC, as transient self heating could have an linearity effect.  So I would consider here something like 15-25 ppm/K thin film resistors (not too tiny form factor), that usually come in 0.1 % tolerance. They are still not that expensive. The resistor would also need to be rather stable as it sets the part from the fine DAC.

The Filter part of the Fluke and Datron circuit is rather similar, and both should work. Here it might need at least 2 relatively bulky and possibly expensive high capacity (e.g. 1 µF)  PP-film caps. So one might be tempted to use a PWM frequency that is not too low as this would allow for smaller caps. Chances are one could get away with MKS type for the other caps in the filter, if one accepts a little longer settling time.

[DC1MC]

@Kleinstein, so to summarize:

- 1 Coarse 16-bit PWM setter to split the Vref in 64K coarse parts.
- 1 Fine 8(16)-bit PWM setter to cover just -/+2 coarse divisions on full range (how to implement it analogically, the Datron uses for the fine part another precision divider of the reference, can we use something else ?!?! )
- 1 Zero amplifier with some AZ (what is this, sorry for not knowing it) OP with an arbitrary, but stable, amplification factor, going to either uC DAC or some kind of zero detector comparator going to an gpio pin.
- 1 SPI 12bit DAC like Maxim 4921 (suggestions welcome) mounted on the negative rail to inject some zero compensation, during signaled idle time  or  when the output is set to zero/calibration requested, it should do some calibration to cancel the drift caused by temperature, charge injection, whatever.

 Is my Zusammenfassung  correct ?

 Cheers,
 DC1MC

[Echo88]

I thought about using a High Resolution PWM-Generator when i wanted to avoid the 2. precision divider for the lower voltage resolution part a few years ago.
This post shows a few ways to do this without resorting to GHz-clocks to produce the PWM: https://www.mikrocontroller.net/topic/386557
I dont know anymore if any of the proposed solutions like the PICOLLO-series or the STM32F334 is cabale of the necessary resolution at about 190Hz (5700A-DAC PWM Frequency AFAIR).

[DC1MC]

I thought about using a High Resolution PWM-Generator when i wanted to avoid the 2. precision divider for the lower voltage resolution part a few years ago.
This post shows a few ways to do this without resorting to GHz-clocks to produce the PWM: https://www.mikrocontroller.net/topic/386557
I dont know anymore if any of the proposed solutions like the PICOLLO-series or the STM32F334 is capable of the necessary resolution at about 190Hz (5700A-DAC PWM Frequency AFAIR).

Using just one PWM channel for the full resolution it very enticing indeed, and I think MC56F8455X can actually do it, while having even a nice ADC and other useful goodies, the price: 6,70EUR at Mouser for the 48pin variant that is still solderable.

https://www.nxp.com/docs/en/data-sheet/MC56F8455X.pdf

Are there any downsides of this one high precision PWM solution, I don't know, like the analogue switches switching time, and so on ?

 Cheers,
 DC1MC

[branadic]

The suggested PIC24FV32KA302/304 or ATMega1284 do have multiple 16bit PWMs, which are appropriate for the task of a PWM controlled voltage divider.

http://bbs.38hot.net/thread-15839-1-1.html reflects the theory behind it.

A little bit of search on www.mikrocontroller.net turns out all the knowledge that went into the EDN 32bit PWM DAC modified by Andreas (2 bit of overlapp so a total of 30bit). https://www.mikrocontroller.net/topic/148472#2309830

-branadic-

Edit: Another approach to consider is PWM modulation. This approach is used in the "LM399 PWM DAC reference" based on the "ION" MickleT. presented here as a representitive of the russian forum (radiokot.ru). The original design comes from adver, but sources can be found on Sergei's google drive.
You can find the source code for two versions:
- 8bit + 14bit fractional
- 8bit + 16bit fractional
and lot's of pictures.

[Kleinstein]

To add the fine ADC there are 2 options. One is using precision resistors,  the one already at the DAC and another to add the fine part. With some 0.1% accuracy one could get some 8 bit accuracy for the fine part. So the fine part could offer an extra 8 Bit (maybe 9 or 10 at best ) bit resolution. Here it would not even need overlap. The divider only needs to be accurate to the full scale of the fine DAC, not to the full resolution.
The resistors need to be good quality, but no absolute need for a extra precision divider.
Especially with a little more resolution from the coarse part and less from the fine part, the resistors get less critical. Modern µCs can be a bit faster than the old days discrete logic and thus get a slightly high PWM base clock.

The alternative might be to not assume an accurate resistor ratio, but do a kind of internal cal cycle. In this case one would need some overlap, so that the fine ADC would cover some +-2 coarse steps. For the adjustment one would use different combinations to get the zero. From the measured near zero cases one would get a high resolution (e.g. 0.01% range) for the scale ratio between the coarse and fine steps. For the adjustment it helps to have a few overlapping steps, as one would get the fine steps corresponding to some 2,3 or 4 coarse steps and thus some extra resolution. I would expect the adjustment only to be needed rather infrequently, so it does not matter if it takes quite some times.

However the step ratio would likely not be a nice integer ratio like 1 coarse step corresponds to 1024 fine steps, but an arbitray number like 5123.6 fine steps to a coarse one. So when setting a DAC value, there will be rounding errors, though very small. When setting a fixed voltage, one would have that rounding problem anyway, as the steps would not correspond exactly µVs. With the measured ratio it is two scales with some odd size (e.g. 120.6 µV and 0.068 µV). The rounding problem is similar, maybe slight less as the steps can be smaller.

For the offset, I dont think that one would need an analog adjustment. It would be more like a rather crude but fixed offset to make sure one can reach the zero and a little below. The exact numerical zero reading would be corrected numerically by adding to the DAC setting.
A ready made DAC chip like the MCP4921 could in theory be an alternative to a PWM DAC for the fine part. But with the filter already there PWM is easy for the fine part and 12 Bit is easy, even if a simple circuit. In the old days one might have used an DAC to change the fine scale setting to get exactly 1024 fine steps to a coarse step, but today I would prefer the numerical way.

To speed the adjustment procedure and maybe add some resolution it helps if the zero detection it not just a comparator, but a high gain amplifier and an ADC, even if only some 8 bit. So the µC internal ADC should be good enough. It's a little like a Null-meter - low gain accuracy, but high amplification, low noise to detect even small deviations.

A single super high resolution PWM channel is possible too and simplifies the math. The delays from switching are not a special problem here - these only limit the use of the very extreme ends (like < 0.1% or > 99.8%), but this also applies to lower resolution. The coarse DAC needs to be precise to the full resolution anyway, so the demands are the same. There are 2 small downsides to those high clocked µCs: one is they tend to use a PLL clock an this can produce extra jitter that might be visible as extra noise. The other is that those higher speed µCs tend to produce more EMI problems, that might end up as extra offsets / noise somewhere. To avoid interactions, e.g. via ground currents the µC creating the PWM should  not do much else during use.  So even using the ADCs might already be too much.

[DC1MC]

The two PWM generators solution is understood (and most likely will be used in the design that I want to implement), I was just being curious about a hypothetical high resolution PWM channel, if it can be used at least theoretically or are other factors that forces the use of two channels. There are a lot of ns in 100Hz, if one can make a precise PWM signal, will it be OK ?

@Kleinstein: for the practical implementation, I was thinking of these capacitors for the filter, are they OK, (the guy has a lot of parts that may be used) ?
https://www.ebay-kleinanzeigen.de/s-anzeige/10x-ero-mkt-1825-kondensator-1-f-10-100-v-rm-15/979765578-172-18783

 Cheers,
 DC1MC

[Kleinstein]

In the filter at least the first 2 capacitors towards the signal should be polypropylene. The caps from the link are MKH type and thus not really suitable here, because of high dielectric absorption and thus sluggish settling. They might be still good for the later stages. The PP caps would be something like Wima MKP4 or MKP10 series.  To my surprise one can get them for about $0,50 to $1. Still they are large, e.g. 7x16.5x27 mm³
So not too bad, but still large.

The single super high resolution PWM is possible - not just hypothetical. Some of the ARM based (and comparable) µCs offer some PWM with additional delay stages to get a time resolution of a few 100 ps. Thus 24 Bit resolution at still more than 100 Hz. One can also push the resolution a little (e.g. 1-3 extra bits) by using additional sigma delta like modulation.  The main possible problem here is jitter from the PLL clock. The specs I found so far did not tell how much jitter over longer (e.g. ms scale) time. Chances are it can work, but the 2 PWM solution with a 8 Bit µC might offer slightly lower noise.

[Andreas]

Hello,

Why using MKT (polyester) when you can have polypropylene for nearly the same price

https://www.digikey.de/product-detail/de/kemet/PHE426HB7100JR06/399-5960-ND/2704614

with best regards

Andreas

[DC1MC]

@Kleinstein - OK then, no more of the DSP/MCU combo with high-res PWM, 2xPWM will be used, was anyway more of a FMI question.
Now, what is an AZ OpAmp (examples will be nice, really curious, this thread has awaken some stuff that I thought to be lost in the Lethe) ?
I'm now the proud owner of 40 PRÄZISIONS WIDERSTAND 10K / 0.01% CAR5Y-10KLI from WELWYN, because Vishay won't send me any vacuumed foil samples , so if we can keep the needed resistors as combinations of series parallel of these ones, this will keep the costs of the prototype in check.

- In the end, what is the consensus, how do we fit the two PWM together ?
- The zero thing I believe you've said to be something like a small fixed negative thingie, connected where, coarse switch ground ?
- The OP used are OP07, are the the most price-performance suitable devices ?
- To minimize the zero ADC should I get some i2c/spi device and isolate it form the MCU and then eventually use some dreaded optos and fully isolate it ?

@Andreas - My link was for 10pcs !!!, but if the polypropylene guys are so much better, so be it, the price difference it's not that big.

 Cheers,
 DC1MC

 

[DC1MC]

So, here's the Datron implementation of their 7-pole Bessel filter @125Hz, any ideas how recalculate for the  suggested 1uF caps, and what shold be used for the OP, if possible without those weird JFETs.

 Cheers,
 DC1MC

[Kleinstein]

The polyester caps are still considerably smaller (e.g. something like a factor of 5-10) and cheaper at low voltage:
The low cost PP can be at around 0,5 EUR, while the cheap MKS is more like 0,2 EUR.
The ebay link is for a pack of 10 - still why buy from an obscure source when you pay not much more from the normal parts source, where you get the other parts anyway.

The Datron 4910 filter uses 7 caps of 1 µF - so thus can get a bit bulky with all PP, but still not too bad.
The Fluke 5700 filter uses only 5 large caps, but 3 of them are 66 µF - this will get large in PP.

I have not found design formulas for this type of filter. It's quite an interesting circuit, but could not find it's name.
So it would be a question of simulation and try and error to change more than just the frequency.
Currently I would more prefer the Datron circuit with smaller caps and JFET OPs instead of the discrete JFETs+OP.


An AZ-OP is an auto zero or chopper stabilized OP. This could be something like an ADA4522 or AD8551 (with bootstrapped supply). Such an OP should also be used for the main buffer.

The 2 OPs to drive the correcting switch could be OP07 or maybe OPA172. Here it depends a little how the circuit continues to the right, as there may be an extra power stage and external sense inputs.

Connecting the PWM signal could be with 2 resistor to the input of the filter. One from the coarse PWM and one larger (e.g. 5 M or a little more) from the the offset and fine PWM combined (with extra resistors).  The 5 M resistor to bring in the fine PWM might be slightly tricky, as large resistors tend to be less stable. So one would still use some kind of divider to reduce the voltage of the fine PWM first. The 100 K resistance was only a crude first assumption, a slightly lower value might be better.

The high quality 10 K resistors would be more something for an extra divider, not directly related to the DAC. A few could also be used for Ohms or current reference if needed. 2 of these could be used for an initial gain of 2 before the PWM part. This depends one reference planed to use. With LM399 or 1N829 I would more consider 2 references in series to start with.

For the resistors I would consider something like 15 or 25 ppm/K thin film ones, maybe using 2 in series for the critical resistor. These should be good enough here. The high value resistor for the fine PWM could be a few in series if needed.
I think I would prefer the measured ratio way over a precisely set ratio, especially with the high resistor value that is often less accurate.

The Datron filter does not need to use those discrete FETs - there are now JFET based OPs with better performance. Something like OPA172, OPA134 or at the high end OPA140 would be good.

[DC1MC]

I have a number of new and original TL072CP and TL074CN, cold they be used in any of the design parts, or should I ignore them and refill with better OP ?

 DC1MC

[Kleinstein]

A very low cost version would be rather unusual. I don't think it makes sense to really go down so far as using TL072 or similar. They might be still ok for the OPs for the on resistance compensation.

For the filter I would prefer lower noise OPs like OPA172 or TLV171 at least. One might get away with one OP less.
The output buffer should still be an AZ OP - about the lowest cost is MCP6V27 or similar. It would need supply bootstrapping because of the low maximum supply voltage.

The zero detection could be the same type MCP6V27.
If really needed one might get away with 74HC4053 switches at 7 V and with modified resistor values to limit the voltage.

So it is possible to build a low cost version, but there are limits where it does not very much sense.
Normally such a circuit would be more in the no expense spared category. For a hobby build single sample there is not much sense in looking for the lowest cost parts - one does not need to use $20 resistors, but should not discuss much about a few $1 OPs.
Getting down to the cents is something if build is large numbers.

[Echo88]

Indeed, too really get the specs of such a circuit one needs to use suiting OPs with good specs. I dont think bootstrapping low voltage AZ-OP Amps is the suitable way in this particular case and would like to propose the OPA189 or the ADA4522.
If youre (or anyone else) ordering the OPA189, i would also want to take a few. Dont want to buy at Mouser/Digikey just to test those, but they look mighty fine spec-wise. 

[DC1MC]

@Kleinstein Sorry, I was not formulating correctly my previous message, I don't want to go to counting cents, is that I've already got the TL072/074.

So, to summarize again:
Ripple filter: OPA172,
Output buffer: ADA4522 or AD8551
Switch compensation: also OPA172, OP7 or... ?
What about LTC1151CSW from Andreas design ?


Big switch selection (I will order them): DG419 or MAX4053A or...

I'm now in a selection block, I would like to place an order for the OP and the capacitors, I'll go with 3 sets of 10: 0.22, 0.47 and 1uF, to be able to mix and match the filter.

Whatever you gentlemen are choosing will be ordered tomorrow, no it's too late, I want to sleep on it a bit.

 Cheers,
 DC1MC

[DC1MC]

Indeed, too really get the specs of such a circuit one needs to use suiting OPs with good specs. I dont think bootstrapping low voltage AZ-OP Amps is the suitable way in this particular case and would like to propose the OPA189 or the ADA4522.
If youre (or anyone else) ordering the OPA189, i would also want to take a few. Dont want to buy at Mouser/Digikey just to test those, but they look mighty fine spec-wise. 

I would sooo much want to finish the BOM, let's do the draft, and of course if any of you gentlemen would like to play with some OP or other parts, it will be my pleasure to add them to my order.

 Cheers,
 DC1MC

[branadic]

I suggest drawing some schematic first to finalize the BOM needed. Next step would be to layout some board. A board is then ordered very quick at low cost on Elecrow (10x boards of size 10x10cm² for $4,90 plus shipping).
An almost final schematic makes discussion easier and I'm sure there will be some combatants following the project by parts and by boards.

-branadic-

[Kleinstein]

Before selecting the OPs, one would need to make clear which reference to use, especially the voltage level.
The MAX4053 is only low voltage, even with only a 7 V reference one could reach the limits.

The Fluke 5700 uses a AZ-OP with bootstrapped supply. So this is not such an unusual solution. The choice of higher voltage (e.g. > 10 V) AZ OPs is not that large. The main contenders are  ADA4522, OPA189/188 ?, LTC2057 and some old LT1050 and similar.
The first 3 are low noise, but also high bias. Especially at the low cost level the AZ OPs are 5 V supply.
Not every supplier can deliver each type - so the final choice also depends on the source.
My low cost choice for an AZ OP would be the MCP6V27, which would need bootstrapping the supply for the buffer.

For the other OPs one should know the circuit first. And the same board could be used for low cost or good quality as well.

The output should likely have a kind of power stage and separate sense lines. At the negative side this might need quite some effort. This might mean moving the negative side of the reference circuit. So it might make sense to include the reference circuit too.
If a 7 V only reference is used, one might have to add a gain stage somewhere to get at least 10 V. One might get a circuit that could use either amplification or 2 refs in series.

[Echo88]

I meant that the circuit of course can benefit from bootstrapped OPs, but that this feature in this case unnecessarily complicates the design which DC1MC tries to realize, while similar high spec OP-Amps like the ADA4522/OPA189 with suitably high supply voltage are available.

[Kleinstein]

The part looked at so far is mainly the PWM DAC and thus the DC-voltage part of a calibrator.  The AC part is likely a completely different and quite difficult part. For the DC part, one might include something to also alow some current calibration.s

I don't think this would be like a lab supply that turns over from voltage control to current control, but more like 2 separate modes and a crude current limit for protection. At least for the higher currents, it might be relatively easy to add the current mode:  This could be a series of shunt resistors and a switch to change over the feedback from the "Buffer" to get a kind of constant current source. To be on the safe side it may need some protection to at least protect the shunts as long as the internal current source is used.

Depending on the current ranges, it might need a slightly more powerful output stage, maybe up to a little over 1A or at least 100 mA. For 1 A one might consider to have this only at low voltage (e.g. 6 V) to limit the power dissipation.

The other point could be the maximum voltage: the DG419 switches and many other CMOS ones are working well up to a little over 30 V supply. The second switch used for the compensation sees twice the reference voltage as the swing  (could be reduces to 1.5 times with a bootstrapped supply and maybe 1.2 times with different resistor values). So a 14 V reference (like two LM399 in series) would be about the maximum for a 30 V supply ( actually more like -8 V and +22 V, maybe -10 V and + 24 V with 34 V to the chip).
Higher output voltage would need some kind of amplification (like in the Fluke 5700). When using a higher supply AZ-OP like ADA4522 this would not be that difficult. The use of a low voltage OP with bootstrapped supply is what makes the 5700 circuit quite complicated. With gain (e.g. 2 times) active the accuracy would be lower, but would extend the voltage range to some 22 V.
For the polarity I would guess the DAC could be positive only and turn over just at the outputs if needed.

The DG419 looks like a reasonable good choice with some 20-30 Ohms at 30 V supply and some 20-50 ns dead time.  So I don't think we would need to look for a better one here.

[alm]

Does it make sense to go through the trouble of using an extremely linear DAC for current sourcing? Wouldn't the quality of the shunt resistors limit linearity? Would a traditional transconductance amplifier perform much worse?

[Kleinstein]

It's true that the shunts can limit the current linearity. I don't see how a trans-conductance amplifier would make things better in this aspect. I know TIAs mainly for measuring small currents, like below 1 mA, where self heating is less of a problem. Self heating would be the same or worse with an TIA.

The point with the current source part would be that it would be relatively low extra effort. The main part would be the shunts and an output stage that can deliver more than a few mA. I think it would at least make sense to include it as an option in the discussion.

The higher current shunts would be used with a low voltage and thus only some 0.1 - 2 % of full scale of the DAC anyway. So accuracy would not be that great from the DAC side anyway. The smaller currents, like 1 mA might not be that bad, if the resistors are good.  For the small currents self heating is much less of a problem and one can even use a higher voltage, like 1 V.

Especially if build without an amplified output, one might be able to use the sense inputs in an unconventional way and than use an external 4 wire shunt. The DUT (e.g. DMM to measure the current) would than be between the high side drive and one side of the external shunt, and the sense inputs go to the sense lines of the external shunt. So it may not need to be part of the circuit, maybe just consider that case when designing the protection at the sense input(s).

As the filter takes quite some time to settle, does it make sense to have some kind of output disable ?  So the filter could settle with the output still off.

[alm]

I don't expect a TIA to perform any better, but it might be simpler than adding even more switching to the DAC, and more as a separate module, simplifying design even further. In the past they were even separate instruments that would be connected to a voltage source (AC or DC). Another advantage would be that the same amplifier could be used for ACI.

If it can be simply added to the DAC circuit, then sure.

I think an output disable makes sense. I wouldn't want a calibrator output some any signal that is outside the specified tolerance.

[Echo88]

Thinking about the ADA4522 and the OPA189 i made a spreadsheet to compare their parameters which are critical for DC-applications.
The duo/quard-versions werent considered in the spreadsheet, since their specs arent as good as the single-amp-packages. Also they increase they chip-temperature.

Questions:
It would be nice if someone could explain why the PSRR/CMRR/Avol-diagrams arent showing the typical/maximum-parameters, which are listed in the Specifications.
Also the Bias/Offsetcurrent-Histograms in the OPA189 are way better than the listed typical/max-specs?
I assume that the varying Offsetvoltage vs. Common-Mode-Voltage in the OPA189 is a separate effect to the normal Common-Mode-Rejection?

The ADA4522 looks quite good compared to the OPA189, which i favourited at first, especially since it doesnt show the varying Offsetvoltage vs. Common-Mode-Voltage.

Thoughts/Opinions/Did i state something wrong? 

[DC1MC]

I thank you all for the effor, needles to say that I'm getting more and more confused  , but hopefully we'll have a schematic/BOM by Saturday for Digikey.
Also getting requests from Germany if someone needs parts on this order, please PM let me know.

 Cheers,
 DC1MC

[Kleinstein]

I don't see any so special parts that would need an order from the US (e.g. digikey): the ADA4522 would also be available from Farnel/HBE or Arrow in Europe. The other parts are even more common.

@echo88:
The DC CMRR is essentially the same as an offset voltage that depends on the common mode voltage.
The diagram may be from a few samples tested in detail. These are sometimes slightly better than typical grade ones.
At the very high level the CMRR and PSRR values could also be limited by the measurement system.

I don't think this circuit would need the absolute highers performance amplifiers. The voltage is still relatively high (e.g. usually > 100 mV) and there are other sources of noise (especially the resistors) and drift (e.g. from leakage)

The circuit may not even need a 2 nd high supply voltage AZ OP,  as the main buffer / amplifier behind the filter. The amplifier for the Zero adjust could use an amplifier with 5 V supply (this would help with an lower cost ADC like the µC internal or maybe an MCP3421). One may not even need an AZ OP here, if one has some MUX before the input to select the point to test and include a zero here. At least the zero check signal would be low impedance, so no need for a high impedance source.
Anyway the exact choice of OPs is not yet the important point - they can be changed even with a ready made board. It's only the question if something like a bootstrapped supply is needed. Given that the ADA4522 is not that expensive I would keep the circuit simple and use an OP with high supply range. This would be especially an advantage if the output would be with gain. I would consider an gain option worth to include.

The relatively high bias should not be a big problem for the buffer after the filter, as the impedance is constant and there would be an zero adjustment anyway. AFAIK the AZ OPs have quite some input current, but this current is rather constant.

So far the circuit I have in mind would start with 2x1N829A in series, so that the PWM DAC would give some -0.2 V to +11 or 12 V. The output amplifier (e.g. with ADA4522) would be switchable between gain 1 and about 2. The maximum output could be something like up to 24 V and maybe 100 mA, possibly also 1 A up to some 5 V.  An extra, crude hardware voltage limit would probably be a good idea. The output would be the usual 4 wire output with extra sense inputs. To keep the circuit simple, the sense input may not be very high impedance though, especially the low side sense.

With an extra 4 wire resistor one could also use this as a current source, e.g. for current up to some 100 mA or 1 A and down to maybe some 100 µA, depending on the resistors used. Precision for the current may not be that high, especially for the higher currents, as the DAC would operate at the low end. If the resistors would be internal, I would stay with manual switching to keep the circuit simple - the switch would be not critical. The tricky part could be some protection (shunts are potentially rather expensive).

The supply would likely need some mains transformer.  However the voltage needed are a little odd, like  -9 V, + 5 V, +15 V ?, + 22 V and maybe +28 V.
So it one might either have quite some loss at linear regulators or would need more than 2 secondary voltages.
Unless high current is needed, the total power should still be moderate (like < 10 W).

So far this could provide voltages up to about 22 V and down to maybe 100 mV with reasonable fine settings.

I don't yet have a good (simple) idea on how to allow an test of the output gain stage.  It could be done with a second amplifier made especially for adjustment. Instead of towards  0 V this would be against some 10 V. So it can be done, though not very elegant. The check/adjustment would be a fixed resistor divider against 10 V from the PWM-DAC in x1 mode and x 2 mode. This would also give a value to the 10 V resistive divider that might be used for an auxiliary output with possibly lower noise. The resistive divider does not need to be long term stable - so no super high grade resistors need, but likely still good ones.

At least for the beginning I would have the digital control via an isolated UART (likely with USB-uart bridge) to a PC. So that the display / keys would be a later step. The PC side would still need to do quite some controls / math. 

[DC1MC]

Some things, unfortunately my day job is extracting most of the energy out of me, but still:

- In the end, is there a type of OP that can be used in ALL stages of the project, I would really want to do the BOM by WE ?
- The 1N829 are bloody expensive: $18.01, the LM399AH are $13.45, can they be connected in series as well ?
- What else should I put on the order, in @branadic design I've seen this integrated rezistive divider, LT5400-x seem to be nice precision dividers, should  add a couple for testing, what values do you recommend (maybe the 1M ones for PWM injection).
- I'm doing a Digikey order because I want to also get a 50MHz TCXO for my FeelTech cadaver, and there is no  other place to order it, what other components, besides the 1uF polypropylene caps should be there ?


 Best regards,
 DC1MC

EDIT: The power supply needs to be asymmetric, like +30/-8V ?

[splin]

Thinking about the ADA4522 and the OPA189 i made a spreadsheet to compare their parameters which are critical for DC-applications.

Questions:
It would be nice if someone could explain why the PSRR/CMRR/Avol-diagrams arent showing the typical/maximum-parameters, which are listed in the Specifications.

Odd isn't it. I asked AD the same question, but I don't think the answers were entirely satisfactory:

https://ez.analog.com/amplifiers/operational-amplifiers/f/q-a/16967/ada4522-cmrr-spec/143590

I accept that measuring AC CMRR is difficult above 80dB but other datasheets seem to be consistent. Perhaps 50dB reduction in CMRR from 0 to 10Hz is in fact perfectly reasonable and perhaps a characteristic of auto-zero amps? My problem being that I'm used to seeing CMRR graphs that start flat from DC a low frequency then start to roll off at some frequency which varies depending on the GBW.

So beware of diagrams and graphs in datasheets -  I've found many cases where they don't match the typical specs, are questionable or downright wrong.

[EDIT] Ok CMRR v frequency graphs don't actually start at DC.

Looking at the zero drift OPA2187; DC CMRR spec is 140dB. Figure 11 shows CMRR of 140dB flat from 1 Hz up to 4kHz, which is consistent. But presumably TI's graph extrapolates from CMRR measurements at DC and those at higher frequencies where the measurements aren't limited by the test setup.

So perhaps the AD4522 CMRR graph below 1kHz actually shows a 110dB measurement limit and not that of the amp in which case it shouldn't have been included in the datasheet. If it was the real amp CMRR, then at what point does the 160dB DC CMRR drop to only 110dB?

[splin]

So far the circuit I have in mind would start with 2x1N829A in series, so that the PWM DAC would give some -0.2 V to +11 or 12 V. T.

Why zeners? Those are very expensive by virtue of their rarity rather more than performance. They have a good TC but still need selecting for best performance and especially low noise. See my reply #835 on the LTZ thread.

https://www.eevblog.com/forum/metrology/ultra-precision-reference-ltz1000/msg614899/#msg614899

[splin]

- The 1N829 are bloody expensive: $18.01

They are from Digikey, but Farnell (UK) sell them for £5.46 ($7.04) one off. Same company, Newark want $7.03 in the US but only $4.04 from element14 APAC (Singapore) - don't know what the delivery costs would be though.

Allied Electronics and Automation (US) $5.33

Still not cheap for a zener though.


[EDIT] looks like element14 APAC don't deliver to Europe or the USA

[DC1MC]

Somehow related  :
https://www.ebay-kleinanzeigen.de/s-anzeige/fluke-5700a-kalibrator-transkonduktiver-verstaerker-fluke-5220a/997008335-239-7973

If somebody has 5.5K EUR this couple of almost working devices are available in my town, we can even meet an have a beer 

What about the asymetric OP supply, does anybody have some answers for my previuos stuff, andbody from DE want something on a shared buyfrom DIGIKE ?

 Cheers,
 DC1MC

[Echo88]

Bargain, but sadly i dont have this kind of money. 
Id like 10x ADA4522-1 "ADA4522-1ARZ-ND" -> 25€ if possible. PM me for address/Paypal please, thank you.
There are quite a few schematics mentioned in this thread, im a bit lost which exactly do you want to replicate? Otherwise i cant give component-suggestions.

[Kleinstein]

The 1N829 zeners seem to be rather variable in price. Depending on the source they can be considerably cheaper than LM399, but also more expensive. The advantage of the 1N829 zener is there possibly lower noise compared to an LM399, but it would take quite some effort to chose the right current to also get a low TC, at least over a limited temperature range (e.g. 20-30 C environment). The LM399 is more like simple to use, but not very low noise - still about standard for 6 digit DMMs. It's possible to have 2 LM399 in series too.

There is no one OP to fits all.

So the ADA4522 is good for the output buffer / amplifier.
For the filter circuit (2 or 3 OPs)  I would think about  OPA141, OPA134 or OPA172  depending on the required quality (mainly noise).
For the two OPs for the auxiliary voltages for the R_on compensation part the OP07 or  OPA172/171 could be good enough.
For the buffer between the reference and the PWM part, one could use an AZ OP (e.g. ADA4522) or a good BJT based like OP27/ADA4075/OPA209/ ... .
The reference itself would likely need 2 more lesser grade OPs, like OP07 or even LM358 (with LM399).
The resistors should not be that critical - so maybe some 25 ppm/K or similar thin film resistor for the main PWM path. I had 100K in my plan, but the ADA4522 might want a slightly lower value, like 50 K.

For the zero (and possibly also 10 V for gain check see below) detection lower cost AZ OPs like MCP6V71 (or 81) should be sufficient. Alternatively one could to the chopping by hand with an more normal low noise OP (e.g. OP27) and CMOS MUX.

For the input protection one could consider depletion mode MOSFETs like BSP135 and likely some low leakage diodes (e.g. BAV199).

For the OPs a non symmetric supply is not a problem. The OPs don't get and thus don't know about GND.

For having a possible gain (e.g. x 2)  at the output one would need some extra parts:
2 good resistors (e.g. some 20 K, <10 ppm/K) for the Gain.
A good precision low bias OP as a buffer for the R_On compensation and possible guards etc.
Another buffer for an auxiliary output at some 10 V for gain adjustment. Here an OP07 would be sufficient - for also external use one could consider a higher grade (e.g. ADA4077 or AD4522).
For the switching between gain of 1 or 2 one might get away with just CMOS DG413 or similar. JFET switching would likely be better.

I am still not sure about required protection. It would probably be hard to make the source as foolproof as an DMM. With a known circuit however repairs would be not that difficult just in case.

[DC1MC]

@Echoo88 The PM will follow for the Digikey parts.
The schematic that I've wanted to reproduce is the Reference Divider and the non-AC (for the moment) parts of the Datron 4707, in the mean time, @Kleinstein showed some issues with the proposed schematic and came with a an concept based on CMOS integrates switches, then other forum members chimed in with different other schematics, the the discussion slided over the the advantages and disadvantages of AZ OP versus normal ones then I'm totally confused, so I'll get a few of each OP mentioned, some polypropylene capacitors for integrator and filter, here the consensus is clear, they are the best, a 50MHz TCXO for an unrelated project and some other small parts.
To make it for the shipping costs I'm offering a shuttle service for the German members.
Basically this is it, until someone comes with a ready to test schematic, I'll have to do my own testing.

In the meantime I'm looking at my hometown calibrator and wondering where can I find 5K EUR )

 Cheers,
 DC1MC
 

[DC1MC]

One quick question: where can I find some of these nicely isolated posts for soldering some puns together for hi impedance path ?

 Cheers,
 DC1MC

[Echo88]

https://www.digikey.com/product-detail/en/keystone-electronics/11301/36-11301-ND/2170335
https://www.buerklin.com/de/loetstuetzpunkt-fuer-leiterplatten/p/12h2712

[DC1MC]

https://www.digikey.com/product-detail/en/keystone-electronics/11301/36-11301-ND/2170335
https://www.buerklin.com/de/loetstuetzpunkt-fuer-leiterplatten/p/12h2712

Today I Learned

 Thanks,
 DC1MC

[DC1MC]

Last call for the Digikey shuttle, anything that would be nice to add for a novice experimenter to the list of parts ?

 Somebody snagged already the Fluke I've posted above, maybe it will post a teardown and a thanu-you, or it will appear later for sale at 10K+ 

In any case, I heed to proceed with my little mini-calibrator project, final canditate for schematic are welcome.

 Cheers,
 DC1MC

[branadic]

One week later... any news on your project? Still haven't seen a schematic your are planning to build up.

-branadic-

[DC1MC]

True, I'm still mulling over my options, read and scratch my head, I've started to actually understand the difficulty of the project and why many of the "simple" versions with MCUs were really never taking off, and remains just concepts.
Even the original Datron TTL-diarrhea it's actually more than meets the eye at the first look, not to mention the Fluke schematic that I've barely had a look yet.
I kind of see why there is no agreed upon stable schematic that is reproducible and a newbie could be pointed to it, like for example the multitude of excellent voltage references projects.

It seem that there are two categories of people involved in this:

- Enthusiast dilettantes that didn't know/look into the problem enough, and eventually produces some "simple, quick" schematics, that of course are never implemented and like many simple obvious solutions mostly don't work .

- People that have actually tried to do it and said "screw this", tighten the belt and bought a profi instrument.

I'm momentarily in-between, I'm still waiting for a breakthrough but with not too much time at the end of the year, the progress it's not so great .

Whatever the final income will be, I really thanks to all of the members that contributed ideas and insights, they were really valuable to me (and probably for others) to better understand the actual difficulty of the task.

 Cheers,
 DC1MC

 P.S> TiN probably laughs like > !!!

 P.S2: If still a (semi-)polished schematic shows up and it's deemed acceptable by the metrology gurus I will try to implement it if I can afford and my efforts are continuing.

[Kleinstein]

I have a reasonable idea on how the circuit should look like. However there are a few points, where one has to decide what is really aimed for.  The main points to choose are:

1) Reference. This might be something like  1 x LM399, 2 x LM399, 2 x 1N825, ... 2 x LTZ1000 ?
 Here the main point is starting from about 7 V or 14 V. Starting from 14 V would be more stable as there would be no amplification set by resistors and still affordable with LM399 or 1N825

2) Output stage with gain, or without.  Without gain would be easier, but the maximum voltage would be some 11 or 12 V.
    With gain one could get something like twice the voltage as a second range.

3) How much output power is needed.  A higher current output stage might want some extra protection, like a variable current / power limit.

4) Especially with gain, there is the question on how to realize gain switching and maybe polarity reversal would be done.
   Switching could be with something like max327 (expensive CMOS switches), JFETs or relays or maybe mechanical switches.

5) For the version with gain, the question is if gain check needs an external reference or not.

6) Which µC to choose to create the PWM and how to enter the values. Not really critical - more like hard to get consensus because of personal preferences.

7) provisions for smaller voltages, especially below 1 V.  Should be something like a divider and buffer - so this could be separate.

how to do the negative side sensing: This could use an extra amplifier, which would be the circuit that is easier to understand. Alternatively one might use a floating reference circuit, which slightly depends on the reference used: The circuit would be a little more tricky, but could be less critical parts.

9) how much ADC resolution for zero sensing, maybe other uses. Here it's mainly µC internal (e.g. 10 Bit) or maybe an SD-ADC chip like MCP3421.

10) how much input / output protection: Especially the output is hard to protect very well against external higher voltage

My preference so far would be:
 - Starting from 2xLM399 or 2x1N825A (optional footprint, slightly lower voltage)
 - Output stage for some 100 mA max with limits at about 1 mA , 10 mA , 100 mA via mechanical switch, limited sink capability.
 - Output stage likely with optional gain (e.g. x 2), but no polarity reversal
 - Gain switching with JFETs  ( 2 Settings)
 - Sense lines with not very high input impedance - so kind of a slight compromise here.
 - Floating reference for negative side sense
 - PWM from AVR (e.g. Mega88), directly coupled. Control via isolated UART from PC. Only minimal function in the AVR , CAL etc. on the PC side.
 - PWM switching with 2x DG419 for the main part and 74HC4053 for the fine part.
 - external supply with some +25-30 V (if with gain), + 15 V, + 5 V, -10 V
 - not sure about zero sense ADC

The point however is that I currently don't have much time and an other project to go first.

[Kleinstein]

Here is a still crude circuit.  There are some resistor values given, but they still need some final adjustments / check.

The reference part is still missing and one might want to consider a reference + buffer - that is the main still open point.
The reference would be relative to AGND = sense low.
Another part still not included is the zero detection. This might well need another JFET switch to separate it from the sense input, if not needed.

For the filter one could consider a 3rd OP for higher order filtering. It kind of depends on the PWM frequency used.
The OP IC8 is there for current compensation for the negative sense pin = analog ground. This can include the reference current.
The OP IC9 is compensating most if the input current for the sense input.

IC3 should be an ADA4522 or similar AZ OP. The other OPs are less critical.
The JFETs for switching should be a different type, more like MMBFJ202 or similar.

For protection one might use fuses and high power clamps for the sense inputs.

[DC1MC]

Thanks a lot @Kleinstein  , you're keeping my hope alive.
I see that the reference buffering seem to be a sensitive topic, but it also seem to be unavoidable, as long as we're "sucking" some electrons from it, that is drawing some current.

Could you kindly explain to this neophyte why it's such a big deal, even with the modern OP, do they really degrade so much the quality of the reference (noise/drift/tempco/??) that people are reluctant do buffer their voltage references ?

 Cheers,
 DC!MC

[Kleinstein]

Even a good amplifier is less accurate than a piece of wire. So one avoids amplifiers not really needed.  For a BJT base Buffer, noise is usually less of a problem, but long term drift can be a downside. With an AZ buffer, noise often is a problem and so are the extra current spikes from the AZ part of the OP. It is mainly the rather old circuits where we get schematics and in these days amplifiers where still rather expensive and not as good.

The difficulty is that die CMOS switches give some extra charge injection current peaks, and these peak could upset the reference. In addition the settling time of the reference might be relatively long and this could cause INL or require more minimum time for the PWM signal. It depends on the reference circuit too: With 2 times LM399 one could use the bridge type circuit with one OP to drive the output and the reference currents. This would probably be OK as the output is already buffered. With a rather fast OP, like OP27, OPA209 the OP should also settle relatively fast after the pulses. The drift and noise of an OP27 is lower than that of most references at 14 V so it is not such a big deal to have a buffer. The higher current 1N825 might prefer the circuit usually used with 7 V refs : an amplified voltage to provide the current source and than a separate buffer. For the2x LM399 one might  even get away without a buffer and use just some capacitance.

There is also the "buffering" problem at the ground side. I think one can get away with just sufficient capacitance (e.g. 1 µF+10 µF) here - so there would be quite some capacitance from the sense low to drive low. The voltage would be limited by protective diodes anyway. I think this would be acceptable for normal use, where there is just a wire connection.

For the zero detection amplifier one could consider a differential input (e.g. AD620 or similar) and extra MUX (e.g. DG409) so that one could use this also for a gain check / measurement. The 4 possible inputs could be sense against zero, sense against an auxiliary 10 V (e.g. divider) and one amplifier zero or reversed polarity. There could be one optional external input.

[Vtile]

Isn't that 1n82x series be pretty similar in specs, only that 1n829A is most suitable for production as the zero TC point is closer to the nominal 7.5mA.

Ie. I had a few 1n823 zeners where the zTC point were closer to 4.5 mA than nominal 7.5 mA, with wild variations between individual parts what comes to higher order fluctuations from temperature. The cost is friendlier for a hobbyist though, you can get 10 1n823 for a cost of one 1n829A to sort and pick. A null bridge is a good way to pick the best part(s) for the reference, in that way you can use another zener as a ref to get in the mV differential range, even when starting without stable reference, like I did.

As going to derail more the topic, is my dilettante thinking in somewhat right track that with higher current reference the part is less susceptible to noise from load and stray voltages around the circuit, but slower to stabilize because of self heating?

[Kleinstein]

With 1N823 the scattering is quite high - one might not find a suitable current in all cases. So there is a certain risk to get a batch of diodes that can not get zero TC. How expensive the 1N829A is depends on the source. For 1N825 chances are good to find zero TC at an acceptable current.
The price also varies quite a lot between sources. So it might still be ok to get 1N829A.

The higher current zener diodes tend to have lower noise. The zener diodes are rather low impedance, so noise pickup or stray (leakage) current is not a problem at all. It may still need some sorting to get low noise units, as there can be quite some scattering in the noise level.
Self heating is higher, but the time constant for stabilization is about the same.

[Echo88]

How much noise does such a Zener produce compared to a fast and easy solution like the LM399 or the more expensive LTZ1000?

[Kleinstein]

The zener noise depends on the type and sample. The 1N82x are supposed to be relatively low noise, especially if selected. So noise wise they may come close to the LTZ1000. Except for the bad samples they should be lower noise than LM399.

The downside of the 1N82x zeners is that they need an selected current to adjust the TC and still get a low TC only for a limited temperature range (like +-5 K).  Noise performance can also vary and might need a check. Especially with 1N823 the current for low TC might end up rather high.

The LM399 is kind of easy to use and low TC even over a large range -  the downside is quite some noise.

One could design the board to alternatively accept LM399 or 1N82x , with just a few resistor values to change because of the different current. So the choice of reference is not that critical yet. 

[Andreas]

The 1N82x are supposed to be relatively low noise, especially if selected.

The 1N829 are normal temperature compensated zeners, and not low noise buried zeners.
Due to the high current (7.5 mA) against the LM399 zener current (0.4 mA) there is some compensation.

On volt-nuts you find a longer thread of WarrenS who had experience with selecting them.
But as far as I understood the selection was mostly for low popcorn-noise.
And you need a good pre-ageing with high current (50mA for one year) before selection.
https://volt-nuts.febo.narkive.com/9QIiUfpN/some-questions-to-zeners-1n823-1n829

My own measurements 0.1-10 Hz 1/f noise on some samples (some NOS + some fresh devices)
showed that there is a large dependancy regarding the manufacturer and also a certain stray from diode to diode.

Motorola 4 samples (NOS) 2 - 2.5 uVpp
ST 3 samples (NOS) 11.5 - 22.5 uVpp
APD 5 samples (new from Farnell) 2 - 2.8 uVpp

LM399 typical 4 uVpp (but partly with a large stray).
LTZ1000 typical 1.2 uVpp (2uVpp max).

And of course in every case you have to do a screening for popcorn noise.

With best regards

Andreas

[Vtile]

My dilettante ideas again. Have anyone tested (definedly, but which were the results.  ) what happens if you bake these TC zeners with close to 180 deg. C for a few tens of hours and then cool them back with slow slope ie. 10..20 hours.

To partly answer to myself yes relaxation heat treatments seems to be used but at lower temperatures and longer times at least for whole units.

[z01z]

There is a paper on how Datron did the aging and selection, here's the link from the 2dw23x topic.

REDUCING OF THE TIME DRIFT OF ZENER STABILIZATION VOLTAGES, famous DATRON, Thanks zlymex !
https://www.eevblog.com/forum/metrology/t-c-hysteresis-measurements-on-brand-new-lt1027dcls8-5-voltage-reference/?action=dlattach;attach=239750

[DC1MC]

There is a paper on how Datron did the aging and selection, here's the link from the 2dw23x topic.

REDUCING OF THE TIME DRIFT OF ZENER STABILIZATION VOLTAGES, famous DATRON, Thanks zlymex !
https://www.eevblog.com/forum/metrology/t-c-hysteresis-measurements-on-brand-new-lt1027dcls8-5-voltage-reference/?action=dlattach;attach=239750

So now I can buy 50 of these guys:
https://www.ebay.de/itm/5x-1N823-Voltage-Reference-Diode-Zenerdiode-6-2-V-0-5-W/332779932314
And start the burn-out procedure and in a month I'll have some cool references  ?
I can do a simple rezistive oven, do the diodes really need to be under power, and at what current ? If yes it's some kind of ion/doppin migration happening or what ?
Is it worth doing it ?

 Cheers,
 DC1MC

EDIT:
Actually is it something technologically special for the 1N82X series, I've seen on evilbuy 6.2V Zeners form Vishay that costs 6EUR/100
https://www.ebay.de/itm/Zener-Diode-VISHAY-500mW-10-25-50-100-Stuck-Spannung-2-4V-75V-Dioden/192070809110

Are there any other similar diodes that can be bought in bulk and aged ?

[Kleinstein]

The 1N82x are somewhat special: it is not just a 6.3 V zener, but a zener and normal diode in series. So more like a 5.6 V zener and 0.7 V drop from the forward biased diode. A normal 6.3 V zener would have a considerable positive TC (e.g. 2-3 mV/K).  Within the class of normal surface zeners the 1N82x are also relatively low noise.

There are similar diodes as ZTK6.8 - though about as expensive and I don't know about noise or drift.
Then there are Chinese 2DW232 reference diodes. Form some sources they are very low noise. However there seem to be quite some scattering - so more like 1N823 not 1N829. The voltage is a little lower and it's a transistor like case (TO-39) with access to the point between the 2 diodes.

For aging there are sereral effects going on: dopant / impurity diffusion, relaxation of stress from the case, charge build up on isolated traps in the surface oxide (a little like an EPROM). So it might help to do heating by the diode current itself.

[DC1MC]

The 1N82x are somewhat special: it is not just a 6.3 V zener, but a zener and normal diode in series. So more like a 5.6 V zener and 0.7 V drop from the forward biased diode. A normal 6.3 V zener would have a considerable positive TC (e.g. 2-3 mV/K).  Within the class of normal surface zeners the 1N82x are also relatively low noise.

There are similar diodes as ZTK6.8 - though about as expensive and I don't know about noise or drift.
Then there are Chinese 2DW232 reference diodes. Form some sources they are very low noise. However there seem to be quite some scattering - so more like 1N823 not 1N829. The voltage is a little lower and it's a transistor like case (TO-39) with access to the point between the 2 diodes.

For aging there are sereral effects going on: dopant / impurity diffusion, relaxation of stress from the case, charge build up on isolated traps in the surface oxide (a little like an EPROM). So it might help to do heating by the diode current itself.

 -  So you say that technologically speaking, the 1N82X are having this extra junction and similar models, at least from this poin of view,  are ZTX6,8 and 2DW232, others are not known ?
 -  Also is it really possible to raise the junction temperature to around 100C just by injecting current, without destroying  the diode due to side effects ? That would be cool, because will make the life of "Zenerzüchter" so much easier than using an  external heater ?


 Thanks for sharing this cool information 

 Cheers,
 DC1MC

[Andreas]

And start the burn-out procedure and in a month I'll have some cool references  ?

Perhaps in your dreams:
- did you read about the yield of WarrenS for zero TC current for 1N823 against 1N825 in the linked thread?
- did you read about my results for noise especially for "ST" parts that I have measured? Maybe you have more luck but I fear: no.

with best regards

Andreas

[DC1MC]

And start the burn-out procedure and in a month I'll have some cool references  ?

Perhaps in your dreams:
- did you read about the yield of WarrenS for zero TC current for 1N823 against 1N825 in the linked thread?
- did you read about my results for noise especially for "ST" parts that I have measured? Maybe you have more luck but I fear: no.

with best regards

Andreas

Well, we all have to dream, don't we, it's Christmas soon ?

Now, joking aside, of course I did read your post, ST were having the largest noise, BUT only 3 !!! samples (were they pre-selected or random), what about 200 pieces, I can seralize them in sets of 10 and wipe out a 72V PS and let them burn with a limiting resistor, either cyclic or continuously.
I'm still curious if the burn-in cycles (heating of the junction) can b made with just the current or one needs an external heater ?
In the end now and then we should keep an eye on this diode guys, who knows when a cool lot may show up, of course there no lack of expensive stuff, but the proletariat needs also precision ,

 Cheers,
 DC1MC
I

[Kleinstein]

With zeners from one batch chances are good most of them will behave similar and a few are bad. So if 3 units from one source are high noise one has to expect something like 199 noisy ones and maybe 1 that is worse if you test 200.

If low noise without the price-tag of the LTZ1000 is the priority, the Chinese 2DW232 maybe the best bet. Here starting from something like 10 units might make sense to find some 4 good ones. Still it may be a batch thing so one could get 10 good ones or none at all. Here selecting would be for suitable current at zero TC and checking for not too much noise and maybe also check for drift after a few 100 hours. The big unknown with these is the long term drift.

If it is just about getting a good TC, the LM399 is a good bet. The big advantage is that it's easy to use: no current to adjust and essentially always a low TC.

[Vtile]

Don't get my ramblings get your way of getting LTZ1000 etc. proven units. My measuring capabilities ends to 10ppm resolution, but it just interesting to follow and wonder in thoughts in this area, more than regular electronics. 

The discrete zeners are definitely satisfying to work with though, when no walkers are breathing your neck.

The heat threatment idea just popped to my mind from working with class, metals and gemstones (not this metaphysic jimba-jamba superstition) where all areas the heat treatments do give extreme results, extending the qualities far beyond regular. So why not a device which is constructed from metal, class and gemstone. 

[z01z]

-  So you say that technologically speaking, the 1N82X are having this extra junction and similar models, at least from this poin of view,  are ZTX6,8 and 2DW232, others are not known ?

My understanding is that 1N82X is "just" a regular Zener and a regular diode in one package. However the Zener voltage was selected such that is has the same temperature coefficient as a diode, but with opposite sign. This makes it possible to find a current that gives zero TC.

Do they worth the hassle? I suppose it's a telling sign that Datron started using LTZs instead. If they had smaller noise than an LTZ, then maybe, but it looks like they don't.

[Echo88]

Since the invention of the LM399 and LTZ1000 basically nobody used Discrete Zeners as references for high accuracy DMMs/Calibrators, since the time spent on selecting suitable zeners is not worth it. I still have a testboard with two populated LM399 and a few populated KX-boards which only needs precision resistors and LTZ1000. You can have them for free if you want.

[DC1MC]

I do , how should we do it ?

 Cheers,
 DC1MC

[Echo88]

Sent you a PM.

[Kleinstein]

For the first board I would go with the LM399: they are easy to use and still affordable, even when using 2 of them.  At least the last circuit I showed has the reference directly at the low side sense terminal and thus could be a little tricky with an external reference board. At least one would need a special board made for a separate negative side supply. So one would probably have the reference as part of the PWM-DAC circuit. One could still have the option to choose between 1 Reference and amplifier x 2 or 2 references in series and a buffer. This would be just 2 precision resistors or 1 simple one. So one could still decide later.  For the LM399 it might be cheaper to use 2xLM399 instead of 1xLM399 and 2 precision resistors. A first test could even use just LM329 to seen if the DAC works.

One point still open is the choice of µC. In principle something like an AVR should be sufficient to create the PWM signals.

Another point might be the idea of not using PWM for the fine part, but a ready made (e.g. 12 Bit) DAC chip. The advantage would be to avoid possible interactions between the 2 PWM signals, that might cause DNL errors as some specific codes, if the isolation between the 2 PWM signals is not very good.
Depending on the µC and frequency the coarse part could provide some 15-16 Bit of resolution. Something like 1 bit might be lost for overlap to simplify the calibration. So even 15 bit coarse - 1 bit lost and 10 Bit fine resolution would give 24 bits of total resolution.

[DC1MC]

Sent you a PM.

 Me as well

 Cheers,
 DC1MC

[branadic]

Any progress on this topic?

-branadic-

[DC1MC]

Yes, I've got the capacitors and poor-man OP, collected the precision resistors, some P and N JFETs, and now I want to place a big order on Arrow to assemble a collection of OP suitable for testing (Kindly please contribute to my fresh topic in this forum regarding the OP), also Kelvin cables for my little multimeters and I want to come with a schematic this month after some experiments.

Also looking for one of these small CPLD/FPGA boards to implement the PWM generator, I've looked at your experiments with ATtiny and friends, this stuff really has jitter and bundle of other problems, I'll take my chances first with the original Datron schematic with JFET switches, with a bit of improvements from Kleinstein and you and many others.
In the end I'll learn something at least or even produce some usefull results.


It's clear from your results that the MCUs with embedded 16/8 PWM are not the answer as far as I can see, if the FPGA/CPLD implementation with external TCXO it's not absolutely jitter free, I'll do a discrete implementation, there is not very much that I can do for the analogue part with my resources, but I totally don't accept any BS on the digital part, it has to be perfectly synchronized and jitter free.

So yeah, if the couch that I must carry tomorrow (along with other house chores) will leave me some energy, I'll test the  summer/filter/buffer thing with humble TL081, as the couplings are mostly AC and try to see how much is the noise/drift using a fixed set of 2 values: zero and the output of the LM399AH reference. Also a good time to test the dual 18V power supply.

If I can't maintain these two values and this stage drifts or has noise already, then it's no need to bother with PWM dividers and worry about charge injection, the whole base has to be redone with better components until it's statically stable.

 Cheers,
 DC1MC

[branadic]

In my opinion there is no need for CPLD/FPGA. And there is no problem with jitter or other stuff using ATtiny. What you read about is the goal of Andreas to reduce current for use of the circuit as a buffer stage for a portable LTZ1000 reference. Hence he tries different approaches such as sleep mode and smaller clock source to reduce current, with the result of severeal drawbacks.
There's no need to reduce current as much as possible on a DAC device here. I think the circuit could be modified to work as a 8.16 bit PWM DAC directly powered from a 10V reference voltage.

-branadic-

[perdrix]

You'll find a good scan of the 4808 manuals here: https://groups.io/g/Datron-Wavetek/files/Datron%204808

[DC1MC]

You'll find a good scan of the 4808 manuals here: https://groups.io/g/Datron-Wavetek/files/Datron%204808

Thanks, I was missing the AC guy, I hope someone will upload schematics for the DC ones in a bit of a higher resolution, even if one member send me a bit better one it still strain your eyes to look on the schematics details.

EDIT: Holly crap  , it's worth reading the sticky post of the forum, I've got a newer version of the PWM schematic with EXCELLENT quality in the aptly named  file:

https://doc.xdevs.com/docs/Datron/4910_4911/4910%20c20090120%20%5B8%5D.pdf


 Cheers,
 DC1MC

[Kleinstein]

I see no problem with a PWM signal from an µC. It would help to have hardware to support at least 16 bit PWM. I see no real advantage in using an CPLD or FPGA - it's more of a problem in having low level logic signals that might need extra level shifter. In both cases (µC or FPGA) one could consider external flipflops for synchronization. However chances are they are not really needed of the supply to the µC/FPGA is clean.
Other than one simple sync stage discrete logic would be more of a problem. I see no need for an TCXO - a normal crystal clock is well good enough for PWM. It's more about a clean supply.

To get a very high resolution, like more than 20 Bits, there are mainly 3 options, that could be combined in some areas:
1)  2 PWM stages for coarse and fine part to get the extra resolution, after adjustment of the fine stage. This is what the Datron 4910/Fluke5700 do.
2)  Super high resolution PWM from some special µC with kind of analog timing interpolation (timing resolution better than ns)
     Some µC offer this, to directly get PWM with an effective base frequency in the GHz range, though the real clock lower. 
3)  good resolution PWM (e.g. 16 Bit) and additional SD like modulation to keep the contend at the lowest frequencies small.
    This still needs a reasonably high clock.
 
The PWM stages from the Datron 4910 and the Fluke5700 like circuit discussed before are different. The Datron one goes for very low R_on and good matching. The Fluke solution goes for compensation of different R_on with additional circuit. In my few the Fluke way may be easier. The difficulty in the Datron circuit could be hidden in keeping the supply clean.
Just the simple PWM with an DG419 would probably have limited linearity, due to R_on differences.

[branadic]

Just the simple PWM with an DG419 would probably have limited linearity, due to R_on differences.

Chances are good that the difference between both switches are not that hugh. However, we can check that on a few samples I have on my desk.

-branadic-

[Kleinstein]

Linearity can be quite sensitive to the R_on values. One has to compare R_on difference with the resistor for the RC filter. So it only takes something like 1 Ohms to get a few ppm of INL error with a 50-100 K resistor at the filter. Specs for the DG419 are less than about 5 Ohms difference - with a suitable supply it can be better if the extreme ends are avoided.

[DC1MC]

Linearity can be quite sensitive to the R_on values. One has to compare R_on difference with the resistor for the RC filter. So it only takes something like 1 Ohms to get a few ppm of INL error with a 50-100 K resistor at the filter. Specs for the DG419 are less than about 5 Ohms difference - with a suitable supply it can be better if the extreme ends are avoided.

Is this source of INL possible to be compensated with some look-up table or formula, as much as it's beautiful to have 10uV/bit all the time, the MCU power is quite a bit more nowadays.
Also I did some anemic tests and indeed on the very low values the linearity it's really bad. Is there any way to get over this, do we need to have some kind of offset or use different polarity references per switch, one injecting positve voltage and one injecting negative. Sorry for the naive stuff, but I'm really curious what it's actually done for small values around 10-50mV.

 Cheers,
 DC1MC

 

[2N3055]

AOIP SN 8310 (Omega CL8300) actually goes from -10% to 100%. So on 10V range it goes from -1V to 10V, with zero and full scale calibrated in.

[DC1MC]

AOIP SN 8310 (Omega CL8300) actually goes from -10% to 100%. So on 10V range it goes from -1V to 10V, with zero and full scale calibrated in.

Wow, that is really an obscure device, no schematics that I can find.

[Kleinstein]

One may be able to do some numerical corrections, if the INL is accurately measured. I would consider the Fluke type analog compensation of R_on differences more practical. This reduces the current through the switches and thus may reduce the expected INL effect or R_on differences by something like a factor or 100 maybe a little more.

The Datron circuit uses JFETs (J106) with < 6 Ohms resistance and thus likely better matching, especially if selected pairs are used.

[2N3055]

AOIP SN 8310 (Omega CL8300) actually goes from -10% to 100%. So on 10V range it goes from -1V to 10V, with zero and full scale calibrated in.

Wow, that is really an obscure device, no schematics that I can find.

No schematics..
It uses pretty much this principle, it has 2 DG419 doing PWM (lo res and high res), coming from MCU buffered through HCT logic . Input ref is -1 to +9V, output from filter is -0,7 to +7V.  That goes to output amplifier that does voltage scaling and in current mode switches in CC source using same reference. All points are calibrated from software.
Basically same principle.

[DC1MC]

How did they done the shifting ?

[branadic]

AOIP SN 8310 (Omega CL8300)

It's "only" 20ppm accuracy.
Brief Teardown: OMEGA CL8300 (AOIP SN 8310) DC voltage / current standard



-branadic-

[macaba]

I've been working on a PWM DAC design with software correction of linearity, it now has a better linearity than I can measure (my DMM6500 has about 0.7PPM non linearity in the worse region).

It might be of interest to readers here to know that 0.7 ohms of switch mismatch causes 8 PPM of non linearity in my design.

I'm now designing hardware correction of non linearity in an attempt to reduce a 3 point calibration procedure (being performed frequently to accommodate the sensitive-to-temperature variation in the analog switches) to just 2 point calibration (performed less frequently).

Question: Are the Datron 470x, and similar, ~20 million or ~200 million count devices? (i.e. 1uV vs. 100nV LSB)

[Kleinstein]

According to the service manual, the 4708 has some 30 million counts: some 0.6 µV steps with a 0-20 V range.
Anyway the difficult part is the accuracy of the coarse part - The actual final steps size from the fine part is more like a fine detail. Especially now with modern digital electronics and often higher clock speeds, the step size is more like the easy part.

[macaba]

Anyway the difficult part is the accuracy of the coarse part

Makes sense. I've got it easier (maybe harder in some ways!) - I'm generating a single PWM signal for the whole range from a counter with 6GHz effective clock, that is giving me 300Hz PWM with 1uV steps on 20V and is working extremely well.

[Kleinstein]

Doing at the resolution with a single high resolution PWM (e.g. fine timing in some µCs) is possible. However it depends on the accuracy of the chips. Inside a µC there can be nasty interactions, via supply variations and so on that can shift transitions a little. I have seen some odd interactions in my AVR based ADC circuit (e.g. the µC internal ADC clock likely effecting the GPIO pin timing. A more moderate resolution would allow for a separate sync stage based on a flip flop.

I would today get more resolution from the coarse PWM (like 16 bits) and than maybe use a conventional DAC chip for the little fine part.

[perdrix]

You can find the 4808 schematics here: <https://groups.io/g/Datron-Wavetek/files/Datron%204808>

The 4808 is pretty much a better 4708.

David

[Chris56000]

Hi!

I hope it isn't because I might want to build something similar as well!

I'm not bothered about using old 68000 architecture as long as I can filch an EPROM dump for it!

Can anybody offer me a suggested replacement circuit idea for replacing the gas–discharge displays with an LED solution?

Chris Williams

[TheDefpom]

Hi!

I hope it isn't because I might want to build something similar as well!

I'm not bothered about using old 68000 architecture as long as I can filch an EPROM dump for it!

Can anybody offer me a suggested replacement circuit idea for replacing the gas–discharge displays with an LED solution?

Chris Williams

I designed some led display modules to replace the broken displays of a 1062 (which is the same display), these modules run from standard LED drive voltages as I modified my display driver board to be sourced from 5V instead of 180V.

I did a 4 part video series on a 1062 (I also covered the reverse engineering and design requirements in a live stream just prior to those videos)

[dmderev]

Hi,

I know that the discussion is old, but is there anybody who could share the experience on the following:

1. What are the advantages or drawbacks of using inherently linear sigma-delta DAC like 20 bit DAC1220 vs PWM if I were to design the calibrator today? Is it larger voltage levels (10V vs 2V) and drifts which can make a difference?

2. Is the advantage of PWM that it does not need to be calibrated in midpoints and therefore acts like a infinitely accurate voltage divider?

3. Any notes on long-term stability of sigma-delta DACs and the need to keep both reference and DAC in thermally controlled enclosure (oven)?

4. Isn't it more feasible to use an integrated ~12bit DAC as a fine tuning controls for MSB instead of PWM?

5. I could use thin film R2R resistor array as first 10-12 bits DAC, this should probably get good long-term stability, and then followed by fine-resolution DAC like DAC7880 (it is old fashioned laser trimmed 10V MDAC). Could I achieve same or better accuracy as the mentioned calibrators (assuming that I measure the output value of each MSB bit and digitally compensate for mismatch)?

6. Any experience with long-term stability of integrated DACs as sources of calibrated programmable voltage? I.e. long term drift in case of laser trimmed resistive DACs vs modern current-steering DACs?

Thanks!
 

[Kleinstein]

The Sigma delta DACs have limitations to the linearity. It's similar effect as with a PWM DAC and ADCs.  One part is the switch resistance that can be nonlinear and voltage dependent. Another class of problems is from dynamic effects on switching (e.g. voltage dependent charge injection). The SD DACs use very fast switching and is thus more sensitive to those switching effects. Some also use a variable switching frequency, depending on the input code. So the charge injection directly contributes to the output.

The switching effects are temperature dependent and may also change over time. So a stable temperature may be a good idea, but it depends on the chips if really needed.

It is very feasible and in my oppinion a good idea to have a more normal DAC for the fine part together with PWM for the coarse part.  What I don't like on using 2 x PWM is the possibility to interference between the 2 parts, if they use the same clock and maybe even the same frequency. As filter is already there, PWM is of cause simple. So using a DAC chip instead is not a simplification, but more extra caution.

The PWM DAC is especially linear in the central part, if the switch resistors are matched or compensated. If not, one has a square law contribution and needs an extra adjustment at the mid point ( correction term or use the mid point to trim the switch resistance).
A big plus for the PWM DAC is that the long term stability is expected to be relatively good, especially with compensation of the switch resistance.

R2R resistor arrays alone don't make a very stable DAC. Part of the problem are the switches. There are some discrete implematations of rather stable R2R type DACs for a calibrator, but this needs quite some effort, e.g. with low resistance FET switches and wire wound / BMF resistors and trimmers.

[dietert1]

I think one important aspect is bandwidth. While many commercial ADC and DAC products are meant to perform at least at audio bandwidth or more, for a precision calibrator one doesn't need that. For utmost accuracy one can use time division schemes that are more like those used in high precision meters, thus reducing the switching by large factors, e.g. 1000. Also opamps work much better at low bandwidth, like 100 Hz or so.
I would not recommend reengineering nor replicating an ancient design but rather implement something similar using modern parts, e.g. those zero-drift amplifiers we now have.
Recently i studied a little a Prema 6048 8.5 digit meter. At the time they made proprietary microcircuits to implement their time division ADC scheme at 3.6 MHz: a multibit delta-sigma converter, where a 100 Hz PWM serves as multibit DAC. Meanwhile i got something similar working with a STM32L432 MCU, yet at 64 MHz. Don't need that proprietary part anymore and get almost 20x more resolution. And the STM32G4 MCU family provides high resolution PWM with an equivalent clock of 5.44 GHz or 184 pico seconds. Other brands offer similar capabilities.

Regards, Dieter