Author Topic: Repaired: Keithley 2000  (Read 20200 times)

0 Members and 1 Guest are viewing this topic.

Offline nidlaXTopic starter

  • Frequent Contributor
  • **
  • Posts: 663
  • Country: us
Repaired: Keithley 2000
« on: November 12, 2015, 07:41:02 am »
In this post, I'm going to transcribe my repair log for my Keithley 2000. I think the various repair topics posted by different people across the net have helped me tremendously, so I figured I'd contribute to the pool of knowledge. If you're sick of reading Keithley 2000 repair posts, feel free to skip right to the pictures or flame me. :blah:

References:
Repair Manual - https://doc.xdevs.com/doc/Keithley/2000/repair_2000.pdf
38hot Reversed Schematics - https://doc.xdevs.com/doc/Keithley/2000/K2000.pdf
TiN's Repair Log - https://xdevs.com/fix/kei2000/
Keithley 2000 Gain Repair - https://www.eevblog.com/forum/repair/keithley-2000-gain-repair/
Keithley 2000+2015 THD - https://www.eevblog.com/forum/testgear/keithley-20002015-thd/
My TF-245 Replacement Topic - https://www.eevblog.com/forum/repair/keithley-tf-245-replacement-options-(k2000-models)/

To start, I purchased the unit from surplus with no information about its history of use. The unit had powered on successfully and displayed measurements, but it was also failing a variety of self tests. It was pretty dirty on the outside, was missing the rear bumper and scanner card cover, and had some evidence of rust and corrosion on the GPIB and RS232 connectors. The Front/Rear input selection switch felt noticeably sticky (this would prove to be a "sticking point" later on). Keypad buttons were all functional, but you had to press pretty hard to get them to work.

Upon disassembly, the interior looked relatively clean with no signs of blow outs. PCB components were mostly 94/95 date codes, and the unit uses the 2000-802-A01 Altera EPM. A visual inspection of the PSU capacitors turned up nothing alarming - I could see no leakage below the capacitors and there were no traces of electrolyte on the PCB. There was a lot of surface rust on the transformer and also a high pitched buzzing noise after the meter finished the power-on phase, but I figured it wasn't necessarily abnormal. One interesting sign: the wiring for the two rear sense HI/LO terminals were polarity swapped as compared with the Keithley service manual, suggesting that the mainboard had been previously removed.

With the PSU area looking cosmetically OK, I decided to power the unit on again to perform testing.
Initial self-test fails:
-100.1: A/D
-100.2: A/D
-101.2: TestCal
-200.2: A/D Mux Lo
-304.1: Input/100
-600.1: Ohm/Amp
-600.2: Ohm/Amp

As you can see, it looked like one sick puppy! I checked DCI and read OVERFLOW on any range below 3A, and the 3A range was just noise with no response to changes in current from my DC power supply.

Next, I checked the power supply voltages:
+ 5V @ U144, pin 2: +4.932 V
+37V @ U101, pin 7: +36.76 V
+15V @ U125, pin 3: +15.67 V vs AGND
-15V @ U119, pin 3: -15.69 V vs AGND
+ 5V @ U124, pin 3: +5.064 V vs AGND


Nominally, nothing seemed out of tolerance. However, I did not check for ripple.

Next, I decided to troubleshoot the A/D gain circuitry because errors 200.2 and 304.1 both pointed to an issue with the X100 gain circuit. At this point, switching the self-test mode to running single tests with pause on fail became very helpful. The AD711JR op-amp at U166 looked OK and was receiving +-15V, and the output voltage under self-test was stable, but wrong. Further examination of the op-amp circuit schematics as well as referencing threads on EEVBlog led me to suspect a problem with the TF-245 resistor network at R271. Measured in circuit, path 9-6 on the TF-245 was ~1.2 kOhms as opposed to the ~1.0 kOhms specified. A test fix was conducted in which ~6.2 kOhms was bodged in parallel with path 9-6, resulting in a self-test pass of 200.2 and 304.1.

Conclusion: Resistor network R271 is out-of-spec and must be replaced.

Next, I decided to troubleshoot the current input circuitry, starting from the Amps jack and moving towards the A/D section. Due to some mis-reading of the circuit schematics and possibly a lack of good judgement, I suspected relay K103 was bad and decided to de-solder it. As it turned out, K103 was just fine when tested under external power, so I soldered it back. :palm: At this point, I recalled that the Front/Rear input selection switch S101 felt sticky, and I measured high resistance between 21-19 which is the input path from the Amps jack to the current shunts when the switch is set to Front. I flushed S101 with liberal amounts of DeOxit and actuated it about 100 times, causing some black gunk to come out. Thereafter, self-tests 600.1 and 600.2 passed, and the various amps ranges returned normal looking readings.

Conclusion: Input selection switch S101 was corroded and caused high impedance in the current input circuit path. Being a cheapo, I decided to clean the switch rather than replace it.

Now, errors 100.1, 100.2, and 101.2 remained. Since the meter was able to change between its various measurement functions and ranges and take somewhat sensible readings, I posited that a failure in the A/D mux or the A/D converter itself was unlikely. The 7V reference was also present on the A/D input under the required self-test conditions. At this point, I decided to de-solder R271 to check for more failures that could be causing various problems in the A/D amplifier section.

Based on the information in the service manual and on posts across the net, it seems that the TF-245 resistor network for R271 is a precision part with an anecdotally high rate of failure among Keithley 2000 owners. This is potentially supported by the fact that design changes in later Keithley 2000 revisions switched to using a combination of different precision resistor networks in this part of the circuit, although Keithley's official word on the matter is that they are unaware of TF-245 being a high failure rate part.

In any case, I began my search for a replacement TF-245 by contacting IC suppliers in China. Eventually, a few places quoted me prices ranging from $25 - $90 USD per piece excluding shipping costs. Next, I contacted the Tektronix sales team in Oregon via their main 800 number and was promptly quoted a price of $17 per piece excluding shipping. It is worth noting that both the Tektronix support / sales team in Oregon and the Keithley support team in Ohio provided excellent service with no hassle, knowing that I am an individual end user and that the unit itself is second hand. I was even put in contact with Dale Cigoy, Keithley's "Lead Application Engineer" (who's also in the credits for the Keithley 2001/2 8)), who answered my random questions about the 193. Anyway, I ended up ordering a TF-245 and a 2000-802-A02 ASIC just in case the Altera EPM was also a dud (happily, that was not the case). TIP: Use your own UPS account number when ordering to pay the actual shipping cost which may be less than their flat rate.

Here is how my failed TF-245 measures in comparison with the replacement part (using a UT71B):

Path #    New (0819)    Old (9511)
1         1.000 kOhm    1.214 kOhm
2          9.00 kOhm     9.10 kOhm
3         10.00 kOhm    10.25 kOhm
4          19.6 kOhm     19.8 kOhm
5         15.00 kOhm    15.20 kOhm
6          29.8 kOhm     30.3 kOhm
7          29.4 kOhm     29.7 kOhm
8          29.8 kOhm     30.3 kOhm
9          27.0 kOhm     27.3 kOhm
10        1.000 kOhm    1.206 kOhm
11         9.00 kOhm     9.19 kOhm
12         89.8 kOhm     90.2 kOhm

(New #4 path at 19.6 k? Hmm... :-//)

After replacement, the multimeter passed all self-tests! I cleaned and reassembled everything, and now I have a nice 6.5 digit DMM which will hopefully become my main workhouse on the bench. For the next step, I plan to upgrade the firmware to the latest revision and calibrate the unit since last cal was in 1995. In the future, I am considering replacing the three big Nichicon filter caps just in case. :-BROKE

Pictures (click for high res):

Replacement parts arrived from Tektronix! The ASIC was actually unneeded, but I figured that buying the unobtanium before it runs out was a good idea, and I also needed to meet Tek's minimum order requirements. ^-^


Original TF-245 made in HK on 9511, new TF-245 made in Indonesia on 0819.




New TF-245 soldered (shittily) at R271.


All cleaned up.


I decided that $15 is too much for a scanner card cover, so I designed and manufactured my own superior replacement instead.


Nice display!


Hope my write up was useful and/or entertaining!
« Last Edit: November 12, 2015, 01:20:45 pm by nidlaX »
 
The following users thanked this post: DenCraw, jlo

Offline tec5c

  • Frequent Contributor
  • **
  • Posts: 423
  • Country: au
Re: Repaired: Keithley 2000
« Reply #1 on: November 12, 2015, 07:48:16 am »
Great to see such a well documented repair. Excellent work.  :-+
 

Offline crispy_tofu

  • Super Contributor
  • ***
  • Posts: 1124
  • Country: au
Re: Repaired: Keithley 2000
« Reply #2 on: November 12, 2015, 07:55:27 am »
Nice and thorough, you make it look so easy!  :clap:
 

Offline eas

  • Frequent Contributor
  • **
  • Posts: 601
  • Country: us
    • Tech Obsessed
Re: Repaired: Keithley 2000
« Reply #3 on: November 12, 2015, 08:18:56 am »
Thanks for writing this up.

I should have some of my own Keithley 2000 refurb/repair notes to share soon.
 

Offline nidlaXTopic starter

  • Frequent Contributor
  • **
  • Posts: 663
  • Country: us
Re: Repaired: Keithley 2000
« Reply #4 on: November 12, 2015, 08:33:53 am »
Thanks for the kind words! Looking forward to reading your notes eas. :)
 

Offline ElektroQuark

  • Supporter
  • ****
  • Posts: 1244
  • Country: es
    • ElektroQuark
Re: Repaired: Keithley 2000
« Reply #5 on: November 12, 2015, 12:02:33 pm »
Price of the 2000-802-A02 ASIC?

Offline TiN

  • Super Contributor
  • ***
  • Posts: 4543
  • Country: ua
    • xDevs.com
Re: Repaired: Keithley 2000
« Reply #6 on: November 12, 2015, 12:47:04 pm »
Sweet. I'll add this thread into references  ;)
I suggest to link https://doc.xdevs.com/doc/Keithley/2000/K2000.pdf for schematics (file is same) instead, it's faster to serve (I know it's nitpicking...)

Puts me into shame, as mine 2000 still laying around in bits and pieces, collecting dust.
« Last Edit: November 12, 2015, 12:49:55 pm by TiN »
YouTube | Metrology IRC Chat room | Let's share T&M documentation? Upload! No upload limits for firmwares, photos, files.
 

Offline WN1X

  • Supporter
  • ****
  • Posts: 79
  • Country: us
Re: Repaired: Keithley 2000
« Reply #7 on: November 12, 2015, 01:20:40 pm »
Excellent repair posting. Well done!  :-+
- Jim
 

Offline nidlaXTopic starter

  • Frequent Contributor
  • **
  • Posts: 663
  • Country: us
Re: Repaired: Keithley 2000
« Reply #8 on: November 12, 2015, 01:24:30 pm »
Price of the 2000-802-A02 ASIC?
$17, min $30 per order, $19 ground shipping for every $50 of parts ordered (use your own shipping account to get cheaper shipping).

Sweet. I'll add this thread into references  ;)
I suggest to link https://doc.xdevs.com/doc/Keithley/2000/K2000.pdf for schematics (file is same) instead, it's faster to serve (I know it's nitpicking...)

Puts me into shame, as mine 2000 still laying around in bits and pieces, collecting dust.
Updated link.  8) I'm sure that pile of 2001s, 2002s, and 2400s makes you feel better about it. ::)
 

Offline 3roomlab

  • Frequent Contributor
  • **
  • Posts: 824
  • Country: 00
Re: Repaired: Keithley 2000
« Reply #9 on: December 03, 2015, 02:59:14 pm »
i guess it is my turn to post some repair notes  :-DD :-/O (unfortunately  :-DD)
i got my K2015 from a korean seller, it appears the seller has sold 66 units !

the repair is still in progress, i have so far identified what i broke as AD822 and possibly LTC1050 (zero volt buffer/generator/guard) + DG408(U163)

there is an additional pair of power lines not usually identified, they are the raw rectified rails +/- 20v. however in my case, they measure +22.2v/-22.3v. these unregulated rails very amusingly powers the VREF multiplier and the zero buffer, through a combination of SOT23 zeners (2x 2.5v to give the LTC1050 zero buffer 5v, and 2x 6.2v coupled with opamps to produce +/- 14v VREF). on 1 occasion i measured the ripple to be somewhere in the region of 5-10mV, but it is highly inaccurate with my UNI-T61E

my +/- 15v rails also are looking somewhat off as they read 15.6v and -15.9v. it is a likely i will modify the 15v and 20v rails to exactly 15v and 20v. but i am still figuring out the 22v to 20v part, the margin is so tiny :P, maybe i will do 19.5v?

the entire troubleshooting was using partially the K2000 schematic from 38hot/TiN, BUT there are many version differences. example the LTC1124 in my version is missing, it becomes U139+U140 2 single opamps. i attach a pic which was before any mods, i hope it is useful

as of now, the DCVO rail is sitting at -15 switched by DG408
by manipulating the manual range, DG408 proceeds to flip around TPP104. i didnt happen to have a DSO plot of TP104 before the breakdown.
and also the signal on pin6 U166
before the AD822 is broken, the open air operating temp is FLIRed at about 43C

(i hope i will post more abt my repair later if all are successful)
« Last Edit: December 03, 2015, 03:14:20 pm by 3roomlab »
 

Offline 3roomlab

  • Frequent Contributor
  • **
  • Posts: 824
  • Country: 00
Re: Repaired: Keithley 2000
« Reply #10 on: December 03, 2015, 08:28:16 pm »
hmmm there is something i dont understand about the function of the DG408 (U163), why is the DCVO idling at -15v rail.

if i were to choose to change the DG408, i found that there are a few variants, does any 1 know which version to get? (vishay / maxim?)
 

Offline sync

  • Frequent Contributor
  • **
  • Posts: 799
  • Country: de
Re: Repaired: Keithley 2000
« Reply #11 on: December 03, 2015, 09:28:18 pm »
my +/- 15v rails also are looking somewhat off as they read 15.6v and -15.9v. it is a likely i will modify the 15v and 20v rails to exactly 15v and 20v.
The +/-15V rails are fine. The 15V regulators have a diode (CR115) in series to ground. So the designed voltage is not 15.0V. More like 15.6V. And what is the point to bring the 20V unregulated rails to exactly 20V?
 

Offline 3roomlab

  • Frequent Contributor
  • **
  • Posts: 824
  • Country: 00
Re: Repaired: Keithley 2000
« Reply #12 on: December 03, 2015, 09:37:48 pm »
my +/- 15v rails also are looking somewhat off as they read 15.6v and -15.9v. it is a likely i will modify the 15v and 20v rails to exactly 15v and 20v.
The +/-15V rails are fine. The 15V regulators have a diode (CR115) in series to ground. So the designed voltage is not 15.0V. More like 15.6V. And what is the point to bring the 20V unregulated rails to exactly 20V?

 :-//, a possible improvement, maybe lower noise? going thru some LDO VREG? (in any case, i did manage to poke around and remove some noise, not very significant, but say about 0.25ppm)

on the other hand, reading this http://bardagjy.com/?p=1167, it could be that a certain JFET refused to turn off. MOAR TESTS NEEDED !
« Last Edit: December 03, 2015, 09:48:02 pm by 3roomlab »
 

Offline sync

  • Frequent Contributor
  • **
  • Posts: 799
  • Country: de
Re: Repaired: Keithley 2000
« Reply #13 on: December 03, 2015, 10:26:52 pm »
:-//, a possible improvement, maybe lower noise? going thru some LDO VREG? (in any case, i did manage to poke around and remove some noise, not very significant, but say about 0.25ppm)
Then try it. I doubt it will make a big difference. Otherwise Keithley hadn't used this design.
 

Offline 3roomlab

  • Frequent Contributor
  • **
  • Posts: 824
  • Country: 00
Re: Repaired: Keithley 2000
« Reply #14 on: December 04, 2015, 01:30:40 pm »
anybody know of a GNU license type PDF editor? (like kicad style?)
i would like to update that K2000 schematic and post a revised version with error corrections (and add waveforms / voltages etc)
 

Offline krivx

  • Frequent Contributor
  • **
  • Posts: 765
  • Country: ie
Re: Repaired: Keithley 2000
« Reply #15 on: December 04, 2015, 02:15:51 pm »
anybody know of a GNU license type PDF editor? (like kicad style?)
i would like to update that K2000 schematic and post a revised version with error corrections (and add waveforms / voltages etc)

I believe Inkscape can open and edit PDFs but it is kind of a pain. This is sort of the point of PDF, I think.
 

Offline 3roomlab

  • Frequent Contributor
  • **
  • Posts: 824
  • Country: 00
Re: Repaired: Keithley 2000
« Reply #16 on: December 04, 2015, 03:59:54 pm »
anybody know of a GNU license type PDF editor? (like kicad style?)
i would like to update that K2000 schematic and post a revised version with error corrections (and add waveforms / voltages etc)

I believe Inkscape can open and edit PDFs but it is kind of a pain. This is sort of the point of PDF, I think.

hmmm ok. i tried a few. seems no "kicad" style. also seems like the 38hot pdf is partially picture conversion, cant select symbol and delete  :palm:
VR105 n VR106 is swapped around in my PCB.

**edit , if anybody could verify, 2W ohm mode, the DMM sends out a wave form like this? (100R to 1M range)
« Last Edit: December 04, 2015, 04:05:35 pm by 3roomlab »
 

Offline eas

  • Frequent Contributor
  • **
  • Posts: 601
  • Country: us
    • Tech Obsessed
Re: Repaired: Keithley 2000
« Reply #17 on: December 04, 2015, 11:42:27 pm »
You could try Inkscape.

Whatever you use, it is likely to be brutal though. PDF isn't intended to be edited.
 

Offline TiN

  • Super Contributor
  • ***
  • Posts: 4543
  • Country: ua
    • xDevs.com
Re: Repaired: Keithley 2000
« Reply #18 on: December 05, 2015, 01:09:58 am »
I can modify or can make extra doc based on your data if you like..
YouTube | Metrology IRC Chat room | Let's share T&M documentation? Upload! No upload limits for firmwares, photos, files.
 

Offline 3roomlab

  • Frequent Contributor
  • **
  • Posts: 824
  • Country: 00
Re: Repaired: Keithley 2000
« Reply #19 on: December 05, 2015, 07:22:59 am »
I can modify or can make extra doc based on your data if you like..

i started to convert to kicad :P (but i guess only parts i tinker with or repaired?).  i hope it will "immortalize" the build, forever repairable?  :-DD (and moddable?)
(rename .c ---> .sch)
« Last Edit: December 05, 2015, 07:26:13 am by 3roomlab »
 

Offline TiN

  • Super Contributor
  • ***
  • Posts: 4543
  • Country: ua
    • xDevs.com
Re: Repaired: Keithley 2000
« Reply #20 on: December 05, 2015, 08:02:06 am »
Emm, not unless you make it readable w/o kicad. Export to PDF is better, just because anyone can view PDF.
YouTube | Metrology IRC Chat room | Let's share T&M documentation? Upload! No upload limits for firmwares, photos, files.
 

Offline 3roomlab

  • Frequent Contributor
  • **
  • Posts: 824
  • Country: 00
Re: Repaired: Keithley 2000
« Reply #21 on: December 05, 2015, 09:58:05 am »
Emm, not unless you make it readable w/o kicad. Export to PDF is better, just because anyone can view PDF.

test print
**doesnt work as well as i thought in pdf, cant search the words like 38hot did it :(
***tried eagle too. something doesnt convert the texts. i think i will leave it in kicad
****hmmm text search works using visio :P, but visio  :palm:
« Last Edit: December 05, 2015, 10:10:29 am by 3roomlab »
 

Offline TiN

  • Super Contributor
  • ***
  • Posts: 4543
  • Country: ua
    • xDevs.com
Re: Repaired: Keithley 2000
« Reply #22 on: December 05, 2015, 11:09:52 am »
That's fine, I'll redraw it anyway :)
Few more details in description would be great.
Perhaps a PCB shot too?
YouTube | Metrology IRC Chat room | Let's share T&M documentation? Upload! No upload limits for firmwares, photos, files.
 

Offline 3roomlab

  • Frequent Contributor
  • **
  • Posts: 824
  • Country: 00
Re: Repaired: Keithley 2000
« Reply #23 on: December 05, 2015, 11:13:52 am »
check out post#9 pic, that U139/U140, it replaces LT1124. theres somemore i cant quite recall off hand :P
oh yes 5v VREG, U144 is LM2940. i didnt went further in to trace what they did to reset pin, there is no reset pin on 2940 :P

i will post kicads as i go along :D

now that i rewatched chris gammells' kicad utubes, the hierarchical feature is really nice to use ! (re-editting into hierarchy, hopefully i can do all)
in hierarchy style, main sch is Keithley2000-2015_r001.sch, and all the blocks are easy to see :P. by R-click enter into each block/sheet, it show further into each block detail.

**edit : 2 shots of a cluster of important resistors which some are not in the 38hot schema (ie : bourns R128/131)
R128 pin resistance 1-16 = 6.63k, 1-15=10.1k (example of 4 pins), 2-15/1-16 = NC
R131 pin resistance 1-16 = 4.39k, 2-15=4.36k, 1-15=7.54k (example of 4 pins)
blue R126. 7k87-7k87-2R8-7k87-7k87 (measured direct from PCB)
blue R119. 3.1k-0.9k (measured direct from PCB)
and a whole bunch of PTF56/CMF55 with NO REF marking. all the CMF55 are 1M28,
***update KKF.jpg --> the difference in AGND connected for x10/x100 resistors (see next post for the schematic)
« Last Edit: December 07, 2015, 11:04:14 am by 3roomlab »
 

Offline 3roomlab

  • Frequent Contributor
  • **
  • Posts: 824
  • Country: 00
Re: Repaired: Keithley 2000
« Reply #24 on: December 07, 2015, 10:43:12 am »
need component googlefu experts to ID the package resistors (bourns n vishay i think) in above post :P (5001/5002 BLF ?)
**edit for comparison purpose, the "measured" and previous schema. its either a mistake in old schema or a new PCB revision
the smd usage smells like massive cost reduction in "newer" PCB
*** or anybody wish to confirm/verify a diff PCB version? ( i am measuring off a K2015 which should share similar top side PCB)
« Last Edit: December 07, 2015, 11:06:38 am by 3roomlab »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf