Fluke 752A repair

[jfphp]

Hello to all,
I have just bought a Fluke 752A reference divider but I am facing a problem with the 100:1 divider long term drift correction. The long term drift corrections on the self calibration bridge and the 10:1 divider went smoothly (I am not at an end point correction and the variable pot is about in the middle) but on the 100:1 correction the calculation gives a value of + 74 Ohm. The jumpers are allready set for 120 Ohm and the maximum of the correction is 150 Ohm, far from the 194 I need (and the 150 ohm correction leaves a large positive offset I can not correct with the pot). Any idea ? Can leakage (trace of flux on the boards with the résistors) introduce such a drift (traces of repair) ? Thank you.

[doktor pyta]

Hi,
do You use Fluke 845 null detector?
As far as I remebmer it was better to physically disconnect the null detector before using divided output.
Maybe a photo of Your setup ?

[Conrad Hoffman]

It's been a long time but I believe there were leakage tests in the manual. Does it pass those? You might have to do some careful cleaning.

[jfphp]

Thank you but there is no leakage test in the manual I have (Rev 1 4/84).

[Conrad Hoffman]

You're right, they talk about cleanliness, but not about specific tests or values. There's a supplement that suggests value changes aren't unheard of, and module replacement can be necessary. They say the switches can't be cleaned by "normal" means, but IMO if there's a will, there's a way. No doubt that replacements of the modules would be beyond a normal person's means. When I got mine it was a ways out, but it was possible to bring it in by changing the jumpers. I'd still look for contamination and maybe do some cleanup with pure ethanol if necessary, then dry well.

[marcos_bier]

Hello, We have a Fluke 752 with a similar problem. Could you find a solution?
We believe the compensation board have some leakage problems. We could measure isolation resistance between some circuit points obtaining 40 Gohm, which seems to be small to this application.
Now we are trying to "clean" the board.
Regards,
Marcos

[Edwin G. Pettis]

It would take quite a bit of 'leakage' to change such low ohm values as much as you have noted, more than likely you have a resistor/resistors that may have drifted out of spec with time.  I had similar problems in my ESI 240C ratio bridge, I ended up having to correct a number of resistors which had mostly gone low in value, very few resistors above a few hundred ohms had drifted out.  It would take a rather dirty board to produce enough leakage to lower the overall value of the circuits, 40G on a 100 ohm resistor is quite negligible.  I do not recommend trying to adjust the resistors unless you have to needed skills, you could easily end up causing permanent damage and requiring replacement.

[jfphp]

Finaly the 752A went to the Fluke service center (drift in the resistors of the divider). Apparently it is a very common problem...

[Magnificent Bastard]

Finaly the 752A went to the Fluke service center (drift in the resistors of the divider). Apparently it is a very common problem...

Please post back with the repair details (and if you can, how much it cost).

[Echo88]

"Why are there dried tears on the bill....ah, yes. The price..."

[jfphp]

Fixed amount for 752A : about $1400. Included pick up, delivery and CalNet ISO9001 cal certificate validy one year. All module résistors changed and all jumpers resetted to 0. To compare with the EBay price of a 752A : no warranty, no cal certificate...

[TiN]

Very reasonable, considering performance of 752A. 

[Dr. Frank]

... Included pick up, delivery and CalNet ISO9001 cal certificate validy one year....

Was there a description, how that calibration was done, and which equipment had been used?

That would imply a different thing than the regular adjustment (by 845A and stable 20V supply), which you would have to do every time before usage..

So a real calibration would verify the adjustment, just made before..

Frank

PS: There is also no parameter in the specification at all, which might be valid for one year.
There's one statement only:

'1-3.    The 752 is self-calibrated before each use.'

[jfphp]

Cal certificate CalNet is the expression on the bill. In fact you receive from Everett Service Center a "Performance test certificate" : instrument is "fully operational" and is "within manufacturers specifications" (I am in Europe : repaired by Fluke Nederland and send to USA for the certification !). As previously noted, there is no real calibration like a resistor.

[marcos_bier]

Hi and I sorry for my late answer.

We repair our FLUKE 752!

The problem was some leakage in the compensation board, the isolation resistance of the board was in the order of 10^10 ohm between different compensation networks. So it was no leakage in parallel with a low-value resistor, instead, it was in parallel with the main divider resistors. It is not clear why the board presents so bad isolation resistance, in addition, it was an internal problem because measuring the superficial resistance  I easily obtained values of 1 Tohm.

The solution was to make a new board and to transfer all the components.

After that, the divider is close to the balance position for more than one week and it shows short term stability compared with the HP3458.

The price was a new board and my time!
Regards

[Henrik_V]

Fixed amount for 752A : about $1400. Included pick up, delivery and CalNet ISO9001 cal certificate validy one year. All module résistors changed and all jumpers resetted to 0. To compare with the EBay price of a 752A : no warranty, no cal certificate...

Just curious: Shouldn't it be a ISO 17025 calibration (ISO 9001 is just production AFAIK)  and isn't the 'validy time' of a calibration certificate just instanious for the moment of calibration and further acceptance/trust time only in the hand of the owner (and documented in the QMH, if needed)?