Wanted - Fluke 732B DC Standard

[KJ-90]

I'm wondering if anyone out there has a Fluke 732B DC Voltage Standard for sale. For some reason, they never seem to show up on eBay, except every few years. The Fluke 732A, its predecessor, also is now very seldom seen on eBay as well. In the past, just a few years ago, you'd see far more volt-nut items on eBay. Other similar Fluke items which appeared often were the 731A and the 510A AC "Standard Cell". But for some strange reason, these items now seldom appear, or if they do, they're not in good condition. Does anyone know what happened? eBay used to be such a great place to find all this stuff.

The Fluke 732B that I am looking for needs to be in good physical condition and work, but it's okay if it has minor issues such as: needs a new battery, the cal is expired, it needs to be cal'd, broken cal labels\seals, and similar issues.

[plesa]

About year ago were 2 on ebay for $300 and more than moth ago one for $4k. Check also the dove-bid and similar surplus marketers.
In general it is quite rare.

[Conrad Hoffman]

The 732 is still a go-to standard and commands a decent price. I do remember more of them coming up a few years ago as cal labs were shutting down or being bought out, but not at any price I could afford.

FWIW, having a single standard isn't sufficient to maintain the volt. You really need three of them so you can inter-compare to be sure all are good, plus have the ability to take one out for certification and do another comparison when it gets back so you know nothing's changed.

No doubt the 731 A or B isn't as good, but they can be tweaked to be quite a bit better on tempco than expected. I use three of those to maintain the volt, plenty good enough for 6.5 digit meters, but probably no more than that.

[ManateeMafia]

The 732B is hard to come by at almost any price. The most recently sold one went at a price higher than I was expecting which tells me there are people willing to risk $4K for an $8K reference.

These references can also have issues. I have seen a 732B with an unusual drift rate of 6ppm/yr. It would take some time to prove a used reference has issues and probably not returnable through PayPal.

I own both 732A's and 732B's. The 732B is great for shipping and battery life. The 732A is still a good reference, but it sucks down batteries during an outage or traveling. If you live close to a cal lab, the 732A is still a good choice.
If you can get either of them fully functional, they will be just as good a reference.

Conrad is correct about having multiple standards. I am not sure you can measure drift of a 732x reference with even a 3458A. Also, it will be a slippery slope since having one will only lead to wanting more.

[HighVoltage]

I also have been looking to find a good 732B for a long time now and finally gave up. (for now)

Instead I got 4 x Fluke 731B' two in really good shape and two need repair.
Eventually I will have 4 of them working, so I hope, to maintain my volt standard this way, until a 732B will show up.

[VintageNut]

I also have been looking to find a good 732B for a long time now and finally gave up. (for now)

Instead I got 4 x Fluke 731B' two in really good shape and two need repair.
Eventually I will have 4 of them working, so I hope, to maintain my volt standard this way, until a 732B will show up.

Similar for me. I bought 2x 731B and they both work properly. One has an issue with the front panel meter which makes no difference in how it functions.

[zlymex]

I got five 732B, and I tested other two.
Believe me some 732Bs are good but some are not. Those on sales are especially prone to out of spec for annual drift.

[Theboel]

I got five 732B, and I tested other two.
Believe me some 732Bs are good but some are not. Those on sales are especially prone to out of spec for annual drift.

Zlymex can You share What the caused for out of spec drift

[zlymex]

I got five 732B, and I tested other two.
Believe me some 732Bs are good but some are not. Those on sales are especially prone to out of spec for annual drift.

Zlymex can You share What the caused for out of spec drift

The cause is the 7V to 10V step-up resistors(R401=6.2k, R402=11.42k).
I modified all my 732Bs removing the 1.018V output connections, and replaced them with internal 7V(from the connection point of R401 and R402) so that I can measure the 7V outputs directly. Those 7V outputs are all drifting less than 1ppm per year compared with my 4910, but some of the 10Vs were drift much more than 7V. I replaced R401/R402 for one of my 'bad' 732Bs with similar hermetic PWW resistors taken out from other fluke equipment but the improvement is small.
Photo below shows the resistors taken out(left) and to be placed(right).

[HighVoltage]

Very interesting, zlymex, thanks

You have five .... just amazing!
Are the schematics and or service manual available for the 732B?

[zlymex]

Very interesting, zlymex, thanks

You have five .... just amazing!
Are the schematics and or service manual available for the 732B?

Yes, instruction manual is available on Fluke site below, in which there are maintenance information and schematics.
However, there is no schematic for the A3 PCB(containing the refanp) in the oven assembly, only a block diagram.
http://us.flukecal.com/products/electrical-calibration/electrical-standards/732b-direct-voltage-standard-0?quicktabs_product_details=4

[VintageNut]

For the 731B, the reference device and the associated resistors are factory selected as a complete group. This leads me to believe that replacing the resistors on a 732B requires characterizing the reference device and the resistors all as a group. Maybe build a large oven that you can mount the ovenized PCB into that  allows easily exchanging and tweaking resistors

My opinion is that there is some art with the science of building a stable 1ppm box for 10V. Sometimes the art is beautiful. Sometimes the art is ordinary. Sometimes the art is ugly.

It may require some years for the ugly to be exposed.

Just my opinion.

I got five 732B, and I tested other two.
Believe me some 732Bs are good but some are not. Those on sales are especially prone to out of spec for annual drift.

Zlymex can You share What the caused for out of spec drift

The cause is the 7V to 10V step-up resistors(R401=6.2k, R402=11.42k).
I modified all my 732Bs removing the 1.018V output connections, and replaced them with internal 7V(from the connection point of R401 and R402) so that I can measure the 7V outputs directly. Those 7V outputs are all drifting less than 1ppm per year compared with my 4910, but some of the 10Vs were drift much more than 7V. I replaced R401/R402 for one of my 'bad' 732Bs with similar hermetic PWW resistors taken out from other fluke equipment but the improvement is small.
Photo below shows the resistors taken out(left) and to be placed(right).

[dr.diesel]

A bit OT, but what's the difference, advantage/disadvantage of the black vs white front panel 731Bs?

[Theboel]

I really like to hear what the theory behind the drift of high quality fluke resistor

[Conrad Hoffman]

For the 731B, the reference device and the associated resistors are factory selected as a complete group. This leads me to believe that replacing the resistors on a 732B requires characterizing the reference device and the resistors all as a group. Maybe build a large oven that you can mount the ovenized PCB into that  allows easily exchanging and tweaking resistors

My opinion is that there is some art with the science of building a stable 1ppm box for 10V. Sometimes the art is beautiful. Sometimes the art is ordinary. Sometimes the art is ugly.

It may require some years for the ugly to be exposed.

Just my opinion.

I tuned up one of my 731A units pretty much that way. Cardboard box and a small light bulb for an "oven". If I remember right, I adjusted the current through the reference by changing a resistor value, and then (I think) had to change another to bring the voltage to the correct value. It was well worth it, as the unit now has a tempco near zero at room temperature. It wasn't that great previously. The long term drift seems to be minimal- my last cal was 12 years ago and an intercomparison between my three units of different vintage is probably less than 10 ppm.

[saturnin]

The cause is the 7V to 10V step-up resistors(R401=6.2k, R402=11.42k).

Correct me if I am wrong, but I think R401 and R402 are not 7V to 10V step-up resistors. 10/7 gain of U1 opamp is set by R1=50k and R2=118k (located in Z1 hybrid), so exact gain of U1 is 1+50/118=1.4237xx. Vref=7.024V (U4, refamp), then exact value of 10V output is: 7.024*1.4237=10.0000xx V

IMHO, R401 and R402 are used to set the operation point of U4. If 7V ref voltage drifts less then 10V, then I would suspect R1 and R2 in Z1 could be culprits.

Does it make sense? 

[chinapp]

The cause is the 7V to 10V step-up resistors(R401=6.2k, R402=11.42k).

Correct me if I am wrong, but I think R401 and R402 are not 7V to 10V step-up resistors. 10/7 gain of U1 opamp is set by R1=50k and R2=118k (located in Z1 hybrid), so exact gain of U1 is 1+50/118=1.4237xx. Vref=7.024V (U4, refamp), then exact value of 10V output is: 7.024*1.4237=10.0000xx V

IMHO, R401 and R402 are used to set the operation point of U4. If 7V ref voltage drifts less then 10V, then I would suspect R1 and R2 in Z1 could be culprits.

Does it make sense? 

i agree with

[Mickle T.]

I disagree with.
732 working well without R1 and R2.

[saturnin]

Mickle T.,

You used a complete different circuit topology...

I simulated (LTspice) the circuit used in 732B and the simulation does confirm that R1/R2 ratio affects output voltage. I admit that R1/R2 change has to be huge to cause a measurable output voltage difference (according the simulation, 4% change in R1/R2 causes approx. 2ppm change in the output voltage). The circuit is remarkably stable as concerns R1/R2 drift...

[guenthert]

Mickle T.,

You used a complete different circuit topology...

I simulated (LTspice) the circuit used in 732B and the simulation does confirm that R1/R2 ratio affects output voltage. I admit that R1/R2 change has to be huge to cause a measurable output voltage difference (according the simulation, 4% change in R1/R2 causes approx. 2ppm change in the output voltage). The circuit is remarkably stable as concerns R1/R2 drift...

He sure did and your circuit is closer, but doesn't quite match the original either.

In [1] a kind, brave^Wskilled, experienced fellow disassembled a (working!) 732B and shared pictures.  There one can see an AD OP97 being used (there might be variations and more recent units might use a different OpAmp -- meanwhile I found that the OP97, as well as the LT1007A you picked, are both low power successors of the OP7) close to the RefAmp.  I suppose that's the OpAmp used to scale the output voltage.  This is significant in as much as the manual [2] states, that a line regulation unit is used to buffer the output (and ensure that not more than 12mA are drawn).  Now the OP97 is not a rail-to-rail amplifier and cannot drive it's output higher than 1.5V below Vcc, which makes driving the series-pass elements directly a bit of a challenge (Q202 is to be taken symbolically here, methinks).  I guess there must be some indirect approach taken, which then causes further delays / phase shifts and might risk oscillations.

Then there is the obscure current cancellation stuff ...

FWIW, in the manual it is stated:
"The next major component that is used to generate 10V is the 10V resistor string. This
string (R401 and R402) is located on the Oven assembly. These are the feedback
elements that set the gain to multiply the reference voltage up to 10V. In its simplest
form, the voltage of the 10V output is expressed by the equation Vo = (Vref)(1 + R401/R402).
The value of R401 is always 6.2 k , and given the value of Vref, R402 is
wound to the value so that the 10V output equals 10V within ±1 mV"
I just wished, I could see it that way. 


[1] https://www.eevblog.com/forum/testgear/fluke-732b-dc-standard-teardown/
[2] https://xdevs.com/doc/Fluke/732B/FLUKE_732B_734A_INST.pdf
(well, originally at http://us.flukecal.com/products/electrical-calibration/electrical-standards/734a-voltage-reference-dc-voltage-standard?quicktabs_product_details=4)