I *think* my 720A is sick

[perdrix]

I couldn't do the Self-calibration for the 0.1 and 1.0 settings of the A decade - null meter way out and trim pot having no effect.

I set Decade A to Cal and measured the oil bath resistors.  Most are about as expected with reading of about 9.9K.  But R302 reads 9.793K (pins 3-4) and R311 reads 9.711k  - both reading much lower than expected 9.896k or so.

Your thoughts?

Dave

[The Soulman]

I wouldn't suspect those oil-bath resistors just jet, more likely be a mechanical issue (switch).
Have you've seen the repair on xDevs?
If not, here's a link, it also have the schematics.

https://xdevs.com/fix/f720a/

[perdrix]

I have got the schematics and manual.   With the A decade switch (S1) set to CAL and the main switch (S8) set to CK B, I am fairly sure that the 2nd and subsequent decades are disconnected from the A decade as are R313 and R314 which begin the B decade and are also in the oil bath.

Yes, it's possible that there's some "miswiring" in there but it doesn't look as if it has been touched internally.  I checked that R1008 connects directly to R302 which then connects to R1009, and R1044 connects to R311 which then connects to  R1045.

Dave

[Dr. Frank]

I have got the schematics and manual.   With the A decade switch (S1) set to CAL and the main switch (S8) set to CK B, I am fairly sure that the 2nd and subsequent decades are disconnected from the A decade as are R313 and R314 which begin the B decade and are also in the oil bath.

Yes, it's possible that there's some "miswiring" in there but it doesn't look as if it has been touched internally.  I checked that R1008 connects directly to R302 which then connects to R1009, and R1044 connects to R311 which then connects to  R1045.

Dave

Deck A switch is connected in parallel to the 9k896 resistors, anyhow.
To make sure, hat there is no defect / contamination in the switch, or in the other cables, I propose to de-solder one cable at the switch, coming from the oil bath, for each of both resistor in doubt (so not to heat the oil bath).
That should completely isolate both resistors for determining their values.

Good luck, that it's 'only' the switch.
Frank

[perdrix]

Hi Frank,

I'll do that, but I'm not hopeful as with decade A set to CAL then S1A should be open, S1B should connect to R1001, and S1C to R301/R1005.

So R302 and R311 should not be shunted by anything at all.

I do wish the schematic had the colour coding of the wires as well ...  I know they follow the usual conventions, but ...

I disconnected the black wires from position one of the A decade switch which should isolate R311 from the switch.   Sadly I still get a reading of 9.794k\$\Omega\$ between connectors 3-4 on the oil bath.

To save time I cut the white wire leading to position 21 on the oil bath and measured R302.  Sadly it still reads  9.710k\$\Omega\$

I shudder to think how many Mars Bars Fluke would demand to fix this.

Thanks
Dave

[RobK_NL]

From the manual:

Resistor Can Assembly
  If a resistor in this assembly requires
  replacement, the entire instrument
  must be returned to the factory for
  repair and calibration.

In technical parlance that means you're screwed. 

Even if you could get hold of the individual resistors (matched to the rest!), they are in an oil-filled can that has been soldered shut. Good luck trying to open it without upsetting all other resistors.

[MK]

How would just one resistor in the chain get damaged and not the others at the same time? isnt it more likely that one of the switches has failed internally?
How clean are all the solder joints etc?

[perdrix]

How would just one resistor in the chain get damaged and not the others at the same time? isnt it more likely that one of the switches has failed internally?
How clean are all the solder joints etc?

The solder joints are all clean - as to HOW just two of the resistors ended up FUBARed I have no idea.  Remember I'm measuring them with all switches etc. disconnected direct across the terminals on the oil bath.    Bad solder joints would make them read high, not low

Dave

[perdrix]

In technical parlance that means you're screwed. 

Sadly I fear you are right.

Dave

[ManateeMafia]

Fluke charges big $$$ for repairs and might price it high enough to force someone to buy a new replacement.

If the resistors drifted and are not internally shorted (assumption) why not trim up the values with good quality resistors inline? It may not be the best solution but it is better than sitting on a shelf collecting dust unless you plan on reselling it as a parts unit. The trim values are a fraction of the total resistance so the TCR effects would also be minimized.

[perdrix]

If the resistors drifted and are not internally shorted (assumption) why not trim up the values with good quality resistors inline?

I've no clue whether they've developed internal shorts or not and no way to find out without "opening the box" which I'm in no hurry to do.

I'll admit that heretical thought had occurred to me. 

So I'd need a 185 ohm resistor (9896-9711) and a 102 ohm resistor (9896-9794).  That's not exactly a huge delta so I'll give it a try.

Dave

[The Soulman]

Maybe could be (heavy) contamination between the contacts of the tank, would be 500K and 1M respectively.
Worth a check, if it's not that: 

[Vgkid]

Not sure what the resistors look like in the oil bath. Possibly 2(or more)windings shorted together.

[perdrix]

Well I added series resistors (tricky as you need to get a pretty exact value to get near a balance) only to discover that even then the 0.1 decade balance was awfully twitchy and hard to set, and kept moving around.  To add insult to injury the 0.3 position (which worked the other day) is now way out of balance. 

At this point I think that the 1st decade switch is flakey as well as the resistors.

Sadly no gross contamination to be found on exterior the tank.

I may return to it at a later date, but right now I'm a bit fed up of the sight of it!

Dave

[MK]

Could it be contamination of the oil, if moisture did get in then acids could get to the windings of the resistors in there? Is there a fill hole that you could use to drain and refill? even perhaps use a vacuum pump on the oil you recover to see if it does have any moisture in it?

[perdrix]

OK I went back to it today with a fresh look at the problem and rapidly found an open connection to the A decade switch.

I messed around for a  while adding a resistor (196.4 ohms) in series with R1008 (first two attempts were not quite on the money)

With the series padding resistor hack applied to both out of spec sections, I was able to calibrate the A decade.  Sadly two of the positions on the B decade wouldn't trim up, and nor would the S2 shunt adjustment.  The 0.1 calibration was still massively fussy and kept moving around.  I tried running the trimmer back and forth, but to little avail, so I think I need to replace R1041 (though this instability may be in the tank).

Do I need to use WW 20 turn timmers or can I use Cermet?  If I need WW, any suggestions on where I can get them.

Thanks
Dave

[perdrix]

Some while later I'm back to this 720A.    I've added resistors in series with R1008 and R1044 to bring the resistance for those parts of the A divider to a value that it would calibrate OK.  I also had to remove five turns (about 1.3 Ohms)  from R1051 as the reading across R313 and R314 was too high to allow me to calibrate S2.   I also replaced R1041 as it was worn out (almost impossible to get a null).

Sadly the beast isn't giving me a nice linear voltage sequence when using just the A decade. 

Input voltage 100.001,4V fed to 0/1.0 Input terminals  All Decades set to zero except Decade A.

Decade A     Output V
 0.1              9.999,88
 0.2             19.967,8
 0.3             29.936,7
 0.4             39.903,8
 0.5             49.874,7
 0.6             59.855,6
 0.7             69.852,4
 0.8             79.871,3
 0.9             89.919,6

Any suggestions on where to look for trouble will be most welcome ...

All sections of the A decade attenuator now measure about 9.999,2 ohms.

Thanks
Dave

[The Soulman]

Sadly the beast isn't giving me a nice linear voltage sequence when using just the A decade. 

.....

All sections of the A decade attenuator now measure about 9.999,2 ohms.

Not linear but a nice pattern nonetheless, something in the next decade loading it down?

[perdrix]

I'm so embarassed

I forgot about the input impedance of the meter - on the 100V range it is only 10M.  So it was loading the KVD 

I'll go hide in a corner for a while

Dave

[tggzzz]

I'll go hide in a corner for a while

... with the rest of us!

[The Soulman]

 

[Conrad Hoffman]

In one of Jim Williams writings he described buying two (I think) KVDs for peanuts because they were tagged defective and showed meter readings like yours. It was just meter loading.