Electronics > Repair

Fluke 732a repair

(1/8) > >>

I had been repairing Fluke 732a Voltage Reference that I purchased in "for parts or repair" state. This is a bit of a worklog/teardown, covering repair and 12v battery upgrade, for anyone who is interested. Click photos for high resolution versions (hosted by TiN).

It appears that sometime in the past the batteries in this Fluke 732a failed, leaked and caused some cascading failures. The unit came to me without batteries and with some parts of battery compartment missing.

After initial warm up:
10V read 10.00008    (8ppm out)
1.018V read 1.017998  (2ppm out)
1V read 1.000111  (111ppm out!)

1V reading was off beyond what could be adjusted. Resistors responsible for division of 10V reference down to 1V are located inside of the oven.  I attempted to partially disassemble the oven to look for 1V problem. I wanted to see if I can fix it, but half way though the disassembly things got complicated enough that I was not sure I would be able to assemble everything back the way it was. There is a lot is loose individual wires on the inside. The foam over last 30 years hardened and chances of me being able to put loose wires back into hardened foam were about zero.

Here are some teardown photos. Top and bottom covers are off:

Top shield is off, revealing foam. The top piece of foam got fused with side foam making it very difficult to remove it without crashing. I used Stanley Blade to separate foam. Same blade was used (whole width of it) as pry bar of sorts to lift the top foam plank:

Inside you can see another layer of foam along with special card with links to adjust 10V reference in 5ppm steps. I switched the link to lower reading by 5ppm:

Underneath second foam plank there was an oven assembly:

I removed its cover and pulled it out a bit:

At this moment I decided to stop and concentrate on the rest of the unit. After all if resistor/resistors responsible for 1V output had drifted by 111ppm in 30 years, there is no guarantee they stop drifting once I add additional adjustment network trying to bring 1V in range.  By the way, it was actually very difficult to assemble the unit back from this point.

This is regulator PCB Assembly:

Note the charring underneath of two large resistors. I guess it was the failing battery that caused them to overheat. It appeared that resistors were replaced. I made sure they are working and also desoldered a small polystyrene cap next to them to check it. Cap was Ok. I cleaned up the worst of charring from the board. I could not remove all of the compromised material without totally destroying the PCB. The large electrolytic cap appeared physically lumpy and not  even cylindrical. It was of cause replaced. One of two transistors on the bottom of this board had hardware missing that was designed to tie it to bottom shield for heat dissipation. I bough and replaced nylon non-conductive washers and corresponding screws. I kept old thermal sill pads. I hope that is OK.

Battery compartment was missing one metal panel on the side. Thankfully I have a small milling machine and I was able to make an aluminum cross plank that is now holding batteries in place and together:

 also decided to convert the unit to using two 12V batteries instead of four 6V batteries. I ordered two of: "EnerSys Genuine NP5-12 Genesis NP Series 12V 5Ah" from Amazon. They seem like the same brand and kind that Fluke is using, just much cheaper from Amazon. I fully charged and load tested both batteries @ 250mA and one was 2% less than nominal capacity and another 5% more. Close enough for me. Switching wires and connectors for using 2 batteries instead of four was completely trivial. Now terminal location for new batteries came out to be different than existing holes. Milling machine to the rescue again. Since I made new holes a bit smaller than old, I used kapton tape around them to prevent me from accidentally shoring the batteries while installing/removing them. See yellow insulation in photo:

Two 12V battaries are a bit smaller than four 6V, so I had to create few spacers to keep batteries from rattling.

Battery compartment includes a small lamp that apparently is not used as a lamp, but as some sort of regulator.  :-+   The PCB next to the lamp showed discoloration, but lamp itself appeared ok when I removed and checked it:

Power supply circuit appeared fine. I just changed two electrolytic capacitors and replaced Schaffner time delay bomb of an EMI filter, and replaced it by Delta 06GEEG3E. I hope it is sufficient:

Now Fluke 732a had been powered and running for few weeks. It would not stabilize for first two weeks, but now looks stable. I will be monitoring it for months to come and still want to do a study of thermal behavior of internal boards that appeared charred to make sure the problem is not continuing.

Is here is a member in a derivable distance from NJ with a calibrated Fluke 732 or something similar? I would really want to synchronize my 732 to theirs. Please let me know by a PM or something.

Questions, comments, and suggestions are welcome.

Hi Nikonoid, great pictures !

I've just started yesterday to revive a 732A purchased that summer (it remained unpowered for at least 10 years, but the lab who sold it to me is in Italy, hence few chances that it suffered frost). I've taken some pictures that I may share. Mine is dated 1983 and I have few differences with yours:

* Main and stangest difference is regulator PCB: I do not have Q14, and R24 is replaced by a short, meaning that heater is switched on by a 3906 (Q13) That would not work if the heater conforms to schematics (i.e. 44 \$\Omega\$). My heater measures ... 1.12k \$\Omega\$. Strange, but explains why power transistor is not fitted ...
* All tantalums are orange. My board is Rev.F
* On power supply, your diodes are bigger than mines (1N4817)I'm waiting for parts ordered from RS. Things that will be replaced:

* Batteries, of course. Mine are dated 1987. I will use 4x LC-R064R5P from Panasonic. They have a life expectancy of 6...9 years when used at trickle (that seems a bit optimistic...).
* Chemical capacitors. I will use Kemet PEG124, rated for 27500 hours @ 105°C. Both 330µF/80V will be replaced by 470µF/100V, and the 220µF/100V will be replaced by same value.
* Line filter. Perfectly agree with you, these are time bombs  >:( - I had once smoke from my PM3350 oscilloscope due to filter.Question: did you performed modifications of your 732A, according to manual errata, i.e. mainly addition of R19 (15k) on power supply and +58°C temperature protection on reference case, since you dismantled upper foam?

I did not take the risk to take that foam apart on mine  ;) What I didn't do too is to replace tantalum capacitors by OSCONs - let's hope that these will not short with time. I plan perhaps to add R19.

Any suggestions (especially on oven) are very welcome!

Nice tear down / repair, thanks for sharing.

Does the resistor divider for the 1V may have an oven that is not working right.
May be it will recover over time / change back in the right direction?


--- Quote from: HighVoltage on December 08, 2017, 04:22:25 pm ---Does the resistor divider for the 1V may have an oven that is not working right.
May be it will recover over time / change back in the right direction?

--- End quote ---

The dividers for the 1.018V and 1V outputs are on the same PC Board with the reference IC and all of temperature sensitive components of the 732A including the adjustment pots.  This PC Board is then inside of the oven assembly.
What is the resistance of the Thermistor measured at the Oven Temp Thermistor terminals?  It should be between around 3,500 to 5,000 ohms.  This will tell you if the oven is operating at 45 degrees C.
I have several 732As that the 1V and or 1.018V outputs are beyond the adjustment range of the pots.  Since these are card wound resistors, low values and hand selected you may be hard put to adjust these.  You will have to get inside the oven assembly, which you have already found out is a difficult task to begin with, and then wait until the oven warms up again to find out if you made a correct guess on how much change you needed to make.  This whole process could take many tries before you get it correct.  If you don't need the 1V output then don't worry about it.  Just the voice of experience speaking.

My thermistor is at 3.656k. It is fairly stable. I did a logging of it before and from my memory it corresponded to about 0.02C fluctuation.

My thinking was that since 1V drifted so much but others remained relatively stable it must to do something with one of resistors. I saw that on multiple Dial-a-Source standards I previously repaired, when some precision resistors drifted too much over year. Trimming resistor that is already drifting might be pointless as it will just keep drifting. If I was to replace it, than new resistor will be subject to ageing.

Combined with difficulty of opening the oven I decided to leave things as is.

HighVoltage: did you possibly mean that there are several heater resistors in the oven and one that is the closest to 1V divider is not performing as should?

Is there some specific reason to chance 330µF capacitors to 470µF? My 732a still have some potential noise of +/-1uV. Maybe additional filtering can help that.

I did not know about manual errata and suggested additional modifications. Do you have references to where to get them?

I also did not adjust the regulator, as I did not have a variac. I have one now, so I may do that and other small changes before getting 732a calibrated.

I have a cal lab 10 minutes from me, but they are horrible. Can someone recommend a reputable lab in a derivable vicinity of New Jersey?


[0] Message Index

[#] Next page

There was an error while thanking
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod