Keithley 2010 Repair

[KungFuJosh]

The LT1124 has quite higher power consumption, as it has a rather high supply current. It is anyway a slightly odd choice, as there is no real need for such a fast amplifier at that position. However it is one of the few parts with a rotated pinout in SO8 - so no so easy replacement. AFAIK the LT1013 would fit, but it could be slightly high noise.
One could consider adding some heat sink for that part.
The temperature rating should be for environmental temperature and allow for a little self heating on top.

Yes, I saw your conversation about this. I also saw this in that conversation:

So I replaced U139 with the high-end LT1124AMPS8 because I chickened out with the test
And it has been fine now for a couple days.

So it seems that part should work fine, right? ...regardless of whether or not it fixes anything.

Thanks,
Josh

[picburner]

My FLIR is from a CAT S60 smartphone, so it may not be the best, but it gives me much lower temperatures.

[KungFuJosh]

I used a self-calibrating Fluke TiS. Tested again with less warmup time was still around 96C. Also checked ambient room temp at 22C, spot on with the thermostat next to the test point.

[KungFuJosh]

Here it is again after some more warmup time. I think I'll replace it. Whether it's the source of the issue or not, something that can handle the temp would be better.

[KungFuJosh]

On a separate note, a Keithley rep gave me this info for the TF-245 which might be useful to somebody:

Absolute: Accuracy = +/- 0.1%, TCR = 25 +/- ppm/C, max
Ratio: Accuracy = +/- 0.1%, TCR = 5 +/- ppm/C, max

[KungFuJosh]

I dunno if this is a stupid question or not, but...

Any reason I wouldn't want to switch the reference to an ADR1399?

This will need to go out for calibration either way.

[Kleinstein]

A newly aded ADR1399 would need some time to stabilize / burn in. Whis would be a few 100 hours of operation, before one should do the final adjustment / calibration.
Otherwise there should be nothing wrong with changing from the LM399 to an ADR1399  (still need to add the RC and increase the ref. current to 3 mA).

I don't know how much of an improvement one would see. The meter uses the extra zener ref. for some form of digital filtering of the LM399 ref..  This should work (otherwise , why bother), but it also causes problems and I don't know how much of the noise is from LM399 noise and what is from the filtering (7 V readings).

[KungFuJosh]

A newly aded ADR1399 would need some time to stabilize / burn in. Whis would be a few 100 hours of operation, before one should do the final adjustment / calibration.

Agreed- that's expected with any reference change.

Otherwise there should be nothing wrong with changing from the LM399 to an ADR1399  (still need to add the RC and increase the ref. current to 3 mA).

Am I correct that by the RC, you're referring to 5.1Ω to 1μF from Out to ground from the datasheet? If so, replacing C167 should be simple enough with a cute little RC tower.

Assuming that's correct, is there any reason I wouldn't want to change the 100nF to 1μF for this application? Is 1μF really needed, or could I go a little lower to 0.068μF or less to get a C0G in that size?

The current to 3mA I've currently got no clue about. Does anybody know what the current current level is on the 2010? How can I increase the currently available current to 3mA+?

I don't know how much of an improvement one would see. The meter uses the extra zener ref. for some form of digital filtering of the LM399 ref..  This should work (otherwise , why bother), but it also causes problems and I don't know how much of the noise is from LM399 noise and what is from the filtering (7 V readings).

I dunno how much of a difference it will make, but it seems like it would be fun, assuming I don't make it worse.

Thanks,
Josh

[GigaJoe]

i have 2010 a bit sick one. my plans of building complete schematic and heavy modding , since 94 a lot of better components exist ..  ( i see my lovely NE5534 as null detector , OPA177 2K-2K AD744 - as integrator , and 2.2nF as integrated cap,  and if 2001 integrated cap - MKP 650V poly,  here looks like ceramic dip ..  a lot of stuff to  do modding    )

for reference, consider at least 5000 hours for some drift minimization, better a 1 year

[Kleinstein]

The damping capacitor at the ADR1399 should not be that critical with the capacitor type. So one should get away with a X7R capacitor or a tantalum cap.
Going much lower than 1 µF may not be that good - more like a bit larger.
AFAIK from the plans floating around the reference current is supplied via a resistor from the ~12.4 V reference level from the ADC. For a higher current one can get a smaller resistor (quality should not be super critical) there.

There was an older thread on pimping a 2010. This also included some reverse engeniering. AFAIK the ADC part is essentially the same as the 2182 and there is some reverse engenierung done by TiN on this, but I don't know if this includes the ADC or just the input amplifier part.
A ceramic integration capacitor may not be that bad: I tested quite some capacitors ( PP, PS and C0G and the best (lowest DA / leakage)  I found was a C0G ceramic from TDK. The NE5534 as slope amplifier is OK (main weak point is power consumption). One could do a little better with the OPA177 (e.g. OPA207 or OPA205).

However one of the real issues is in the software: the Keithley meters seem to average the zero readings over quite some time. For some reason they did not fix it and AFAIK many meters (2001,2002, 2010, likely the 2182, DMM7510/6500) are effected. The averaging may help a little for the short time scale, but really hurts form some 5 to 500 seconds in the Allan deviation curve - just where many precision measurements are done.

[KungFuJosh]

The damping capacitor at the ADR1399 should not be that critical with the capacitor type. So one should get away with a X7R capacitor or a tantalum cap.
Going much lower than 1 µF may not be that good - more like a bit larger.
AFAIK from the plans floating around the reference current is supplied via a resistor from the ~12.4 V reference level from the ADC. For a higher current one can get a smaller resistor (quality should not be super critical) there.

I attached @picburner's schematics for the reference and ADC. I'm not sure what to look at.

Thanks,
Josh

[GigaJoe]

ignore adc ...

so schematic a bit ( tiny) incorrect .. VR122 - actually TC=0 zener , not a basic one. bottom orange diode. Vishay resistor R459 actually low TC and defined current. ( R359 on schematic )

Ideally current for zener  should be set to zener TC=0, 3K its about 2.2mA  ... and it maybe soooo   off . so when LT1124 replaced you need to hook DMM to zener and check TC=0.   otherwise adjust resistor to TC=0 current
you may skip ... but it would be waves when measurement floating up and down due to drift of reference volage and software recalculation of actual value ...   LM399 not a reference for ADC but reference to calculate voltage.  reference for ADC that +- 12V generated from this low noise zener. as well as suppose stable voltage for LM399.

I didn't change LM399 ,  my hypothetic that reference voltage from LM399 averaged in memory from many samples and therefore noise irrelevant ... But long term stable are ..

btw .. ideally that zener should be mounted in stress free mount ,  additional board space and legs loop ...

btw TF-245 drift not really important , as anyway final voltage are software calculated resultant

[KungFuJosh]

ignore adc ...

so schematic a bit ( tiny) incorrect .. VR122 - actually TC=0 zener , not a basic one. bottom orange diode. Vishay resistor R459 actually low TC and defined current. ( R359 on schematic )

I'm not sure I understand. R359 is 3.01kΩ in the schematic, and on the PCB (bottom side). R459 is a 1.28MΩ CMF series resistor (PCB top side), and measures about 51k in circuit. It does connect to VR122. Next time I have the PCB out of the case, I'll see if I can determine the relationship between those two resistors...if I remember.

Ideally current for zener  should be set to zener TC=0, 3K its about 2.2mA  ... and it maybe soooo   off . so when LT1124 replaced you need to hook DMM to zener and check TC=0.   otherwise adjust resistor to TC=0 current
you may skip ... but it would be waves when measurement floating up and down due to drift of reference volage and software recalculation of actual value ...   LM399 not a reference for ADC but reference to calculate voltage.  reference for ADC that +- 12V generated from this low noise zener. as well as suppose stable voltage for LM399.

I didn't change LM399 ,  my hypothetic that reference voltage from LM399 averaged in memory from many samples and therefore noise irrelevant ... But long term stable are ..

Once this is fixed, I will do some long-term stability tests with the LM399. I assume there will still be popcorn noise, and if so, I will try the ADR1399. I will do periodic long-term stability tests with that as it ages and see how it compares. That will give a better idea of whether or not it's a complete waste of time.

btw .. ideally that zener should be mounted in stress free mount ,  additional board space and legs loop ...

Do you know a good part number for a replacement for that zener? If I replace it eventually, I can mount it on ceramic spacers with longer legs...unless that's a dumb idea lol.

btw TF-245 drift not really important , as anyway final voltage are software calculated resultant

Yeah, it doesn't matter for me, but for some people, that's a failure point that causes signifcant errors.

Thanks,
Josh

[Kleinstein]

The VR122 zener in the 2010 is ok, and no need to replace it. The point is that the schematics is showing the wrong zener type.
R359 provides the current for this 6.2 V low noise zener diode. The TC of this zener diode depends on the current and if the TC is high one gets some extra noise from thermal variations.  Once things are repaired, one could check the drift of the 6.2 V or amplified +-12.4 V and if needed replace R359 with a slightly different value.

When replacing the LM399 with a ADR1399, one would need to increase the current to this reference, so change R216, that is in the original configuration some 2 mA with some 1.7 K to get 3 mA there.

There is averaging for the LM399 reference (at least in 10 PLC and slower AZ mode). However the LM399 popcorn noise also has slower components, that are not full filtered out. So the LM399 reference noise can still matter. This may not be as important with the 2001, but it could be with the lower noise 2010.

[KungFuJosh]

Thanks for the info!

Yes, VR122 being incorrect was obvious as the listed part is SMT.

Does anybody know what part number the schematic should be updated with? I'm in the process of converting the old Altium files to KiCad so I can update them when needed (in process because I'll convert them all, but only need to use 2 of them right now). I'll share them all when they're done.

Thanks,
Josh

[Kleinstein]

Chances are the low noise zener would be a similar type (1N4579 ?)  as in the Keithley 2001.

[KungFuJosh]

Chances are the low noise zener would be a similar type (1N4579 ?)  as in the Keithley 2001.

Oh yeah! I completely forgot- I found that previously as the part number in the service manual. I probably should have written that down.

You're correct; it's 1N4579. The funnier thing is that it's a typo in the manual, and they wrote IN4579.

Thanks,
Josh

ETA: Now I also remember why I didn't buy any. Unit price: $27.38 each in quantities of 100. 

[KungFuJosh]

Test current on the 1N4579 says 2mA. Wouldn't something with a higher test current be better if we're raising the current for the 1399?

Something like the 1N4584A with a 4mA test current (which isn't easily available either) would be better, wouldn't it?

Thanks,
Josh

[Kleinstein]

The current for the zener is separately set from the Lm399 reference. So one is somewhat free in the actual current. if one really is desparate one could try a 1N825. They are normally 7.5 mA, but the lowe grades can quite a bit off. I got 2 with a zero TC near 3 mA. Still a gamble but not as expensive.

Anyway the current zener seems to work. So no need to change it.

[GigaJoe]

my point not to replace zener but adjust current to the minimal tempco.

[KungFuJosh]

my point not to replace zener but adjust current to the minimal tempco.

I understand. My looking for a replacement part is because it's safest for me to remove VR122 to replace U139. If the zener gets damaged in the process, I'll need to replace it regardless. Hopefully, it will be fine.

Thanks,
Josh

[GigaJoe]

in my understanding , the way how this kind of multimeters calculate final result.  This particular Zener not so critical ,  as one of my dmm has 6.4  another 6.2 and it can be 1N825.  i have 2.3mA and from cold to hot it was around 2 millivolt difference.  that  have compensated by internal calculation , but resultant output waiving up and down in around 10 microvolt range  ...

probably for lowest noise need to find zener with TC=0  at maximum current ,  like 1N4580 , 1N821A (3A ,5A, 7A ,9A)
and from my point they all the same just binned by TC deviation at fixed current value .  as this binning process indifferent , current can be adjusted for TC=0  need to find cheapest  and sort by stability.  It good to have flat short time stability..   

[Kleinstein]

my point not to replace zener but adjust current to the minimal tempco.

I understand. My looking for a replacement part is because it's safest for me to remove VR122 to replace U139. If the zener gets damaged in the process, I'll need to replace it regardless. Hopefully, it will be fine.

Thanks,
Josh

Unless it breaks, I would leave U139 as is. It is running a bit hot, but that is normal. It really wanted maybe add a heat sink (e.g. glue on top).
One can still unsolder U139 (in case it would fail) with VR122 still in place. The glass cases are relatively robust too heat.
The point to first fix are the errors that the self test finds.

[KungFuJosh]

Unless it breaks, I would leave U139 as is. It is running a bit hot, but that is normal. It really wanted maybe add a heat sink (e.g. glue on top).
One can still unsolder U139 (in case it would fail) with VR122 still in place. The glass cases are relatively robust too heat.
The point to first fix are the errors that the self test finds.

TF-245 is good, and U139 voltages are not.

The only remaining error is 101.2.

I'm going to replace U139 with the better version that can handle +125C. If that does not fix the issue, where do you suggest I look?

Thanks,
Josh

[Kleinstein]

U139 LT1124CS8:

1: 0
2: -15.8V
3: +6.2V
4: +6.2V
5: 0 (unstable)
6: +15.146V
7: -5.936V
8: 0

The voltages look OK.

A 7 V reference level would come from the LM399.