Keithley 2002 DMM repair

[plesa]

More than year was waiting for repair - Keithley 2002. As a sideeffects of working on LTZ1000 project I needs 8.5 digit DMM again.
The meter is working properly on DCV ranges, so it can be used on working on LTZ project, but it is not possible to calibrate such a meter at cal lab.
Ohms and current measurement is not working.

Following self test failed:
304.2   2MW Open-circuit Ohms
306.1   9.6uA Ohms Source
306.2   1.92uA Ohms Source
308.2   Overload Ohms
500.1   96uA Ohms/200uA DCA Range
500.2   960uA Ohms/2mA DCA Range
500.3   7.2mA Ohms/2mA DCA Range
500.4   20mA DCA Range
500.6   200mA DCA Range
500.7   2A DCA Range
501.1   Amps Protection
501.2   Amps Bootstrap
502.1   AC Amps Switch

I checked the current in Ohms measurement which is not correct (only ranges above 20M are working), same with the current measurement (20mA and 200mA seems to be OK).
For this meter schematic is not available, but lot of is common with 2001 meter where schematic is available.

Suspected parts are the DG411 switches or amps range selection.
Does anybody have experience with 2002 repair?

[crispy_tofu]

Here's the service manual if you don't have it (credits to TiN)

[TiN]

Aha. Mr. plesa ready to roll, at last.

Credits to Keithley for official service manual, I just mirror their data 

I'm ready to sacrifice my 2002 in holy fire to aid this repair project as much as possible.

Err, not really    but can take measurements for reference, if needed to. Maybe as end result I'll finally bodge up calibration jig to DIY-cal my 2002, as it's ~30ppm low on DCV.

plesa

Was was already done to diagnose these 30x.x errors? We talked about this year or two ago, but I'm not remember everything performed back then.
First would be to check actual signals around ohm sources (can be spot not far from precision resistors near LTZ1000, near MAX326 switches)

I also leave here link to my review with teardown photos in case someone need photos. Thread here as well.

For 501.x, check diode bridge, CR226, not long ago someone had similar issue in 2001, and fixed by replacing borged diode bridge.

[dacman]

Have you checked U220?

[plesa]

I checked the rectifier CR218 and  diode CR226. Both are OK.
So I continued with Amps measurement circuit. Because it can affect the ohm failed tests.
The Vishay foil resistors are all OK.
2 mA range measurement completely out
20mA range measurement is OK
200mA measurement is out , same for 2A range.

Based on this I suppose that it is related to U226(DG408DY) and U222 (LT1007ACN8).
Components above are ordered.Stay tuned:)

[plesa]

Have you checked U220?

U220 is OK

[plesa]

Exchanged MAX 326 switches ( U232 and U223 ) and errors still the same 

[TiN]

Hm, so I'd look into current source circuits then.

[plesa]

Thanks TiN, but first of all I needs to fix current measurement, after that I will focus on Ohms measurement.
Self test is using current measurement and it is not fully functional (2mA and 20mA ranges working only).
Today replaced U234 (DG411) without change in DCI measurement.
Today I desoldered the U221 and U224 (shift registers) and put it into the IC tester and the U224 pin 7/Q7 seems to be permanently high.
New shift registers ordered including LM339 comparators which quite often are bad as well.

[plesa]

Success - after replacement of U224 amps mesurement works OK.

Following self test failed:
304.2   2MOhm Open-circuit Ohms
306.1   9.6uA Ohms Source
306.2   1.92uA Ohms Source
308.2   Overload Ohms
500.1   96uA Ohms/200uA DCA Range
500.2   960uA Ohms/2mA DCA Range
500.3   7.2mA Ohms/2mA DCA Range
500.4   20mA DCA Range
500.6   200mA DCA Range

Replacing U224 solved theese test which failed:
500.7   2A DCA Range
501.1   Amps Protection
501.2   Amps Bootstrap
502.1   AC Amps Switch

I measured current for ohms measurement and except 20Ohm range it is about 10x lower.So only 20Ohms range works +/- OK.
So now it is time to fix current source.

[dacman]

I suppose you know that several of the mA units should be µA.
How are U230, U231, & U233?

[plesa]

I suppose you know that several of the mA units should be µA.
How are U230, U231, & U233?

Thanks, of course, but for someone else I corrected it.
U230 - LTC1043 - it is quite unlikely
U231 AD706
U233/U238 AD705
are ordered, but did not arrived yet. In this I agree with TiN  to replace all analog switches, comparators and other available components.
Good news is that the Vishay hermetic foil resistors are all spot on.

[TiN]

In one of my 2001 which had 304.x issues (resistance current source for ohms), issue was in dodgy TO92 JFET (TG128 Kei marking).
Might be similar thing as well.

[plesa]

In one of my 2001 which had 304.x issues (resistance current source for ohms), issue was in dodgy TO92 JFET (TG128 Kei marking).
Might be similar thing as well.

Thanks for hint.
According to manual it should be 2N4392 in TO-92 package and in real it is Siliconix J2611.
Unfortunatelly I cannot find datasheet. Vishay/Siliconix 2N4392 seems to be still available.
Did you replaced them with Vishay 2N4392 or by J2611?

[TiN]

I replaced with J2611 remarked as TG128, from dead donor 2001 board. Yah, I cheated a bit 
Make sure to replace JFETs in pairs, if they were matched. Also how about front/rear input switch? Old switches are not rare to have gunk inside/bad resistance causing leaky signals and faults on most sensitive ranges. Does meter show constant Overflow on max 1Gohm range, or showing random charge resistances? Should be overflow on all resistance ranges with open inputs..

[plesa]

I replaced with J2611 remarked as TG128, from dead donor 2001 board. Yah, I cheated a bit 
Make sure to replace JFETs in pairs, if they were matched. Also how about front/rear input switch? Old switches are not rare to have gunk inside/bad resistance causing leaky signals and faults on most sensitive ranges. Does meter show constant Overflow on max 1Gohm range, or showing random charge resistances? Should be overflow on all resistance ranges with open inputs..

Q207 seems to be ok. The switch is OK ( on all ranges is overload).
So needs to check pairs of JFETs.

[plesa]

Small progress in repair. Checked all JFET and all seems to be OK (Q226 and Q227). But according to repair manual and Keithley 2001 schematic U244 were suspected.
And bingo after replacement U244 ( DG444) all errors are resolved.

DCV,DCI,2Wire Ohms seems to be fully fixed and according to check are spot on.
I hope this will helps to someone in future ( on eBay are K2002 with errors quite rare, which is interesting compare to 3458A.

[TiN]

Great, happy to hear. I think most prefer more common 3458A instead of tricky Kei2002.
Now jump in to DMM noise testing thread, and let's see how good it is.

Also here's I tried something.

[plesa]

Great, happy to hear. I think most prefer more common 3458A instead of tricky Kei2002.
Now jump in to DMM noise testing thread, and let's see how good it is.

Also here's I tried something.

Thanks for link, I did not notice that you are working on it.
Components ADA4627 and AD797 are ordered. So within 2 weeks we can have first comparison.

[plesa]

During the check of all basic function I noticed wrong ACV with selected AC coupling. When selected AC+DC it seems to be more close but still not spot on.
ACI is OK.Frequency is also not working with AC coupling.
Repair needs to be continued
Strange is that all self test passed.
Suspected is AD637J RMS to DC converter, in case this will be confirmed I'm going to replace it with AD637B with better specification.
AD637J Total error ±0.5mV ± 0.1% of reading
AD637B Total error ±0.25mV ± 0.05 % of reading

[Kleinstein]

With just ground buffering, I am supprised to see a AD711 and a JFET OP of any kind. The natural candidate would a OP27 or similar. Should be better than AD711 and ADA4627, and less expensive and power hungry than the AD797. Also bias is much lower than AD797 just in case it matters.

[plesa]

Progress on ACV readout solving:
After replacement all switches (DG444, DG411 and DG408) and shift registers 14094 problem still persist.
Self test also started to indicate following errors:

400   Calibration/Frequency Compensation DAC
400.1   -4.36V DAC Output
400.2   -2.15V DAC Output
400.3   0V DAC Output
400.4   +2.33V DAC Output
400.5   +4.48V DAC Output

401 Series  Regulator
401.1   Regulator

402 Series  Switching
402.1 Frequency Switch

405 Series   Absolute Value (X10 Gain):
405.1    Gain Comparison (Large +DAC Output)
405.2    Gain Comparison (Non-inverting)
405.3    Gain Comparison (Small +DAC Output)
405.4    Gain Comparison (Non-inverting)
405.5    Gain Comparison (Small -DAC Output)
405.6    Gain Comparison (Inverting)
405.7    Gain Comparison (Large -DAC Output)
405.8    Gain Comparison (Inverting)

406 Series   Sample and Hold Circuit:
406.6   Sample and Hold Voltage Comparisons

Seems that I screw something or the issue can be related to U433 aka AD7528.

[dacman]

Did you use an ESD mat and wrist strap tied to ground?

[plesa]

Good new is that AC+DC coupling after replacement of all switches can simultaneously measure frequency also on voltage terminal.
AC coupling no, only on amps terminal.
During investigation of ACV circuits I found that it is copy exact of K2001, so schematic of Keithley 2001 is quite handy.
First of all I needs to understand how it is working.
AC coupling gives value 7-8V on all ranges
AC+DC coupling is more close and able to measure frequency.

Yes, I used ESD mat.

[plesa]

Small progress on solving AC measurement errors. The Q428 transistor  (Vishay VN0605T) is dead and has short circuit (50 Ohm).This explained the permanent 7V.