Restoration glory of Keithley 2001 DMM

[TiN]

Time has come. 
After seeing this new big-LCD DMMs like 34461A which Dave recently reviewed, and mm, new Keithley SMU with touch-screen..
Let's show some respect to good-old instruments, with bright nice VFD with 180°C viewing angles and simple interfaces which don't require
reading user manual to begin work.

Small intro - I work mostly with digital stuff, various microcontrollers projects and last few years PC components design. As with most of folks who started with electronics, first tool was some pos cheapo DMM. Then it become full-time stuff and I got decent Fluke 87V, which still working as first day, and covering 99% of needs for DMM usages. But I always wanted to get some more accuracy for analog circuit needs, and needed some real system/bench meter with nice display, good accuracy and interfacing to capture and process measurement data. But since there was no actual project-critical need, one could never justify spending hundreds on great meter. Till now.

But I do like messing with hardware, digging PCB design solutions and repairing stuff as hobby, so once I saw broken Keithley 2001 7.5 DMM on sale, I decided - now or never and pull a trigger. After couple weeks I got it in my hands, beast had some burnt components on PCB, but was able to power on. I loved it's screen, function set right a way. After this my DMM soul was sold to Keithey, as that was years ago.

After few weeks of digging thru service manual, messing with analog PCB I got it "working", it was able to take sort-a meaningful measurements but still failing some self-tests. Few years pass... I got somehow addicted to Keithley and bought total 7 (yes, seven) broken 2001's and even managed to get SMU 2400 with broken VFD (had to salvage one of 2001 screens to fix it). And because some of 2001's were missing small ADC board, i have now 3 complete meters and some parts. All meters had front-end failures due to leaky capacitors which corroded PCB traces.



Now my plan is to repair at least these 3 meters to fully functional state, to have self-test pass and to be able of running calibrating routine.
It's rather difficult for me, because i have minor experience in analog and precision stuff, but I'm eager to learn



Schematics reversing is in progress currently:
Schematics in PDF and DSN
I use Cadence Orcad for schematics, but export to Adobe to PDF as well.

---

Toolkit I have for aid this project:

HP 33120 15MHz ARB
TLA715 logic analyzer (I had plan to reverse how ADC board is interfaced so bought a TLA for it . Maybe can do DIY ADC board later with ALTERA FPGA's i have lying around)
CSA7404 scope (have TDS5034 with various probes at work lab too)
Keithley 2400 SMU (calibration state unknown, but assumed working in spec) (UPDATE: Completed calibration at January 2014)
Keithley 2002 (uncalibrated)
EDC MV106 (calibration unknown, but matches voltage cell within 140uV )
Eppley 809684 Standard Voltage Cell 1.01909VDC @ 70F (history unknown, bought off ebay. It's resting already for couple month on table)

So second non-easy task will be to get resistance and voltage sources for calibrating K2001's. Per service manual, it needs 2VDC, 20VDC,20.75k,1.75Mohm for DC and Ohms calibration. I'm don't have access anything precise like 3458a, and because I currently work in Taiwan full-time, there are some language problems to get verification

I hope it's okay if I link my related keithley-repair redmine project here.
Also I have mercurial repository to keep all related files here.
Anyone is free to download everything, it's open project for fun.

Let's make it alive 

EDIT 8.APR.2015



Model 2001 Multimeter, Calibration Manual, 2001-M-905-01 Rev. H / Mar 2011
Model 2001 Multimeter, Repair Manual Addendum Schematics part 2, 2001-902-07 Rev. C / March 2011
Model 2001 Multimeter, Operator's Manual, 2001-M-900-01 Rev. K / March 2011
Model 2001 Multimeter, Repair Manual Addendum Schematics part 1, 2001-902-06 Rev. C / March 2011
Model 2001 Multimeter, Repair Manual Addendum, 2001-902-05 Rev. C / March 2011

[TiN]

UPDATED 30.12.2013 08.04.2015

Hardware (Digital board)

Main CPU: Motorola MC68302FC16C
RAM: 2x128K

Revision K is same as one used in Keithley 2002.

Hardware (Analog board)

Voltage reference: LM399 binned by Keithley (according to repair manual)

Hardware (ADC board)

3 modules, same PCB PN Keithley 2001-162-04G

Top:



Bottom:



Integrating ADC implemented in U808 ALTERA/Cypress CPLD (MAX EPM5128JC or CY7C342B) and interfaced via SPI thru optocoupling to digital board.
Onboard power: Q812, 78L05AC in TO92 package
Clock: CMOS oscillator, 7.680 MHz
Output amplifier: BB OPA602AP, High-speed precision Difet OpAmp, 6.5MHz, +/-15V, 35V/us, +-1mV Vofs, +-15uV/°C Drift
Onboard VREF Opamp? : 2 x LT1097CN8
Onboard VREF? : IN4579 Zener 6.4V
Timer: NE5534

Current schematics in PDF (not finished, work in progress).

Update for analog section of ADC: PDF Schematics of ADC analog parts.

Hardware (front panel board)

It's very similar to some other Keithley products as well.
Keithley 2002, 2400-series SMU, Keithley 7001, 7002 all use same VFD.

Hardware typical failures

First step should be replace leaky capacitors on front-end drive circuitry. All 7 meters which I had were made in 1992-1994 accoring to date codes on parts and had nasty corrosion from electrolyte leakage on capacitors C114,C115,C116,C117 (470uF 63V and 1000uF 35V) near mains transformer and LDO's.

Firmware (ROM)

Using standalone programmer I read EPROMs with firmwares from my units:
Firmware A05,A06,A08,B07,B08,B10,B15.

B07 had two EPROMs with even and odd bytes. So had to write small tool to merge two binaries into single ROM. Compiled by GCC in freebsd 7.2.

Funny to see, that there are names in open text (Firmware A06) of project engineers who probably worked on Keithley 2001 design.

Code: [Select]

Dave Skerkoski,Mike Marks,Robert Fitch,Mike Dahlhausen ,Andy Creque ,Roger Chaplin - The Software Team: Please report occurrance
 

And much more people in Firmware A08 from offset 0x741AA (probably whole team)

Code: [Select]

George Topola     Mike Szalkowski   Bob Spinks        Larry Spilar      Greg Sobolewsky   Dave Skerkoski    Dave Sherrets     Karl Sasgen      Paul Sally        Ken Reindel       Bill Pelster      Gerald Naujoks    Scott Oberlin     Steve Mihalick    Marla Mock        Michael Minneman Mike Marks       Peter Lui         Dave Lewis        Warren Kumley     Dave Kruszewski   John Kraizel      Larry Klubert     Meindert Kleefstra  Doug Kirsop       Fred Hume         Jim Horvath       Todd Hayes        Theresa Harlan    Jay Hardway       Kevin Hansen      Kevin Graham      Robert Fitch      John Ferman       Mike Dahlhausen   Mark D'Ambrosia   Jeff Cunningham   Andy Creque       Grant Coyle       Dale Cigoy        Roger Chaplin     Mark Brinich      Charles Bona      Robert Bennet     Gary Bish         John Banaska      Kevin Ballou      Mike Atwell   
The Project Team
Thumbs up to all these people for great meter! 

Options (MEM1,MEM2,STD)

Hacking options on 2001

Secret goodies!

Secret menus in Keithley gear

[TiN]

Test setup

Components

PC software / GPIB Intefacing (LabView)

Scanner card

TBD later, sorry

[cyr]

I have upgraded my 2001 (firmware B06) from STD to MEM2 using a DS1245Y, works fine. Did you move the jumper to the correct position?

I think your firmware B07 dump is incomplete BTW, each EPROM should be 512kB of data and the merged image 1024kB, instead you have two 512kB files with half the bytes blank in each and a merged 512kB file.

[free_electron]

I have adc modules for those things...

[macboy]

I have upgraded my 2001 (firmware B06) from STD to MEM2 using a DS1245Y, works fine. Did you move the jumper to the correct position?

I tried doing this too, and it worked ... almost. The meter identifies the RAM and displays "MEM2" on power up, and lets me take over 30000 readings into memory, but they are lost at power-off. Also, it does not allow me to save settings other than slot "#0" (MEM2 should add #1 to #9). Can you save 10 sets of settings? If you save settings to memory are they retained after power-off? I think my firmware is A06.

[cyr]

Actually, you are correct. I never bothered testing anything beyond logging a lot of samples, and my unit behaves exactly the same. Interesting...

[TiN]

My MEM2 with DS1245Y with A08/A01 firmware allows to store 6017 readings, and loses stored data when powered off as well (does not react on Recall button).
No issues with saving settings to #9 slot tho. Restores saved settings after power toogle from slot #9 too.

I'm hooking LA now to capture ADC module comms, to see what rates and data is used.

[macboy]

My MEM2 with DS1245Y with A08/A01 firmware allows to store 6017 readings, and loses stored data when powered off as well (does not react on Recall button).
No issues with saving settings to #9 slot tho. Restores saved settings after power toogle from slot #9 too.

I'm hooking LA now to capture ADC module comms, to see what rates and data is used.

Is yours a "genuine" MEM2? Can you dump the NVSRAM contents with a device programmer and post them? I am guessing that Keithley initializes the chips with some special data, like some "empty" data structures to hold the extra 9 front panel settings.

I share your affliction; I have 4 2001's on my bench right now   One of mine has B15 firmware, which is almost the newest (there is a B16). Are you interested in the B15 firmware? I need an excuse to finally buy a MiniPro programmer.

[TiN]

It was genuine MEM1. And I need same excuse to buy standalone programmer, as I just borrowed one for a day from work to read my ROMs, lol.

On units i have digital boards have different revisions, some with two ROMs (even and odd), other seem simplified a little. How old are your meters? Analog board from two meters i have is different too, i assume Keithley had fix front end bias curcuitry which was failing on initial revision.

I got schematics for ADC module almost done, working on capturing transactions on digital bus. Will post it here when get some cleanup.

[TiN]

Almost finish ADC module schematics - PNG file
Added front panel schematics - PDF, ,need use black background, as i had black in CAD

Looks like front panel MCU which controls VFD and scans keypad is interfacing by simple UART.
This makes Keithley as neat chassis for own devices. For example Keithley 7001 switch units go quite cheap, and feature same front panel.

I already have idea to take one of those and assemble small ITX computer with Core i7,16GB DDR3,SSD in it. Add GPIB card and will be neat controller/storage unit for measurement devices in rack, with great VFD and keyboard on front.

[echen1024]

How much do you typically pay for a Keithley?

[TiN]

That depends on patience and condition. I got most of them without ADC module as parts, for two hundred or so per unit. Still more than some decent handhelds might go for.

[TiN]

Project is going slow, but going

Got ADC modules for rest of 2001's. Now can bring all meters except one (lots of damages/missing parts on analog board, was used for parts anyway).
They also have now pin definition on silk



Thanks, Vincent!

Interesting enough, old ADC modules I had were Revision G, with date codes on IC's 1992, while new ones are Revision H and Revision J, with datecodes 1992 and 1994. There are few layout differences and extra OPA177GS (U813) near to integrating amp U816 (OPA602AP). I will update schematics soon, to see what was changed.

I wish there could be some forum member which own fresh Keithley 2001 and don't mind to unsrew three screws and take a photo of internal boards of current 2001 hardware

[macboy]

I have four 2001's. Three work fine (one of those has issues with AC current, but fine otherwise). The forth one had been working perfectly, no self test fails or strange behavior, but suddenly stopped working recently.

When turned on, it makes a few clicks after a few seconds, then stops. No display or other signs of life. I suspect the power supply of course. Nothing looks strange inside (no leaky/bulging caps). Do you have any suggestions as to where to look? Your reverse-engineered schematics will prove very helpful I think.

When all four were working, they all had read within +/- 3 ppm of the average at both 2 VDC and 20 VDC ranges (didn't cross-check them in other ranges/functions) even though none of them is recently calibrated, AFAIK.

[free_electron]

@macboy remove the powerr supply caps. you can bet your ass they have gone bad. these caps don;t bulge .. they leak at the bottom ,eating the traces !

[macboy]

@macboy remove the powerr supply caps. you can bet your ass they have gone bad. these caps don;t bulge .. they leak at the bottom ,eating the traces !

Thanks, I will do that as soon as I can, certainly before I attempt to power it up again. No need to add electricity to an electrolyte spill, the traces would be gone it no time (if not already!). In your experience do the bad caps measure as low on a simple cap meter? I don't have a meter with ESR yet.

[TiN]

What PCB revision of analog board? Its letter on top pcb label near rear input connectors.

 Early units often have issues with floating ground curcuitry. I would replace four big caps, four diodes and two zeners on analog board just near mains transformer output connector. As general precaution, I would replace electrolytes anyway, because even if they might look OK still can be faulty and start leaking from bottom, corroding PCB.

Also there are voltage check test pads near ADC board connector, with +/-15v, digital 5V and +/-8V voltages.

Worth checking your display board too, just hook working one to digital board, and check if anything changes

[macboy]

What PCB revision of analog board? Its letter on top pcb label near rear input connectors.

 Early units often have issues with floating ground curcuitry. I would replace four big caps, four diodes and two zeners on analog board just near mains transformer output connector. As general precaution, I would replace electrolytes anyway, because even if they might look OK still can be faulty and start leaking from bottom, corroding PCB.

Also there are voltage check test pads near ADC board connector, with +/-15v, digital 5V and +/-8V voltages.

Worth checking your display board too, just hook working one to digital board, and check if anything changes

I had only a few minutes to look at it yesterday. All voltages are low, on analog board and digital board. No signs of leakage under capacitors. This meter was definitely built well before 2000, so it did not get a bad batch of capacitors from the early 2000's. Since all voltages are low, I think it is likely a failure on the pre-regulator and/or the voltage selector circuit on the primary side of the transformer. Too bad (for me) that you didn't reverse engineer the schematic for that section 
I'll check the analog board revision later tonight. Can you elaborate on the issues with "floating ground circuitry"?

[TiN]

All my units are 1992-94 year. Bad caps is not really batch issue, it's more just age with elevated temperature usage. I would replace old caps as general measure , when we talk about servicing highend test equipment which is 20 year old

As for schematics, i actually had input side reversed and posted before in this thread. There is electrolytic cap too. One of my units had faulty cap and dead 555 due to corrosion. After cleaning and replacements - come back to life at once

[macboy]

The analog board looks like revision N (labelled 2001-102-04N). Digital looks like revision K, and F/W is B10 on two 32-pin DIPs. Date codes on ICs and Caps indicate manufacture in late 1996 or early 1997. I have probed around and think that something is dragging down the power supply. Dried up caps wouldn't do that... leaking caps? Maybe, if something was shorted. No evidence of leaks. Maybe a bad tantalum, but none of the rails looks shorted out, they are all just low (3 V on 5 V rail, +/- 8V on 15 V rails). A bad shorted rectifier could cause the problem maybe. Which diodes did you recommend replacing? CR109-CR112? Or one of the bridge rectifiers? I don't see any zeners near the secondary connector of the transformer.

[TiN]

Pretty sure problem is on primary side autorange detection curcuitry. Probably you have 110Vac on input but due to failure of selector mains xfrmr is commutated for 230Vac, so you get not enough voltage on secondary side.

Exactly what I was seeing on my unit. Check 555 (or replace it, its cost not worth bothering with test ), relays, caps and mosfet. Schematics was posted earlier as well.

[turbo!]

Now my plan is to repair at least these 3 meters to fully functional state, to have self-test pass and to be able of running calibrating routine.
It's rather difficult for me, because i have minor experience in analog and precision stuff, but I'm eager to learn

If you work on any part of analog section or replace anything on it, it will need to go through a full calibration. As far as moving analog board between instruments, it depends on where calibration constants are stored. Replacement parts need to be the same or better tolerance. (not just 1% resistor. you need to look at ppm/°C, ppm/year which determines drift over time) Some instruments require an environmental chamber that can hold a stable temperature for quite a few hours, so you can test it at say 10C as well as 30C so it can learn the thermal characteristics to use for temperature compensation.

[TiN]

Thanks, and yes, I'm aware of caveats with such sensetive designs, also do have access to temperature/humidity chamber (-40 to +80°C) at work.
I have those units with various problems, and it's about repair to get initial proper function first, then next step will be component testing/evaluation,
and only after all that work to be done - find local cal lab and perform full calibration. It's a long way to that point from now still.
Just because it's for-fun project, I decided go without rush and study it's design and construction methods.

I do have broken unit as a parts donor, so just swapping same specced parts when need to.

[TiN]

Still recovering schematics.

Fixed one of analog boards (datecode 1993), which had burnt PCB around negative terminal of C116 and wrong diodes soldered by previous owner. 

As expected, aged caps leaked and set unit on fire. Replaced C114,C115,C116,C117,C106, avalanche diodes from another analog board donor (CR109,CR110,CR111,CR112), pair of zeners CR513,CR514.
Had to put two jumpwires too, as traces were burnt at connection from CR112.A to C110.K and at C117 from top layer to bottom.



Luckly enough, this seem to be enough to get board to pass self-diagnostics test without errors, which surprised me. All rest analog boards have this or that errors, lol. 



Measurement on various function seem to be nice too, except 1000VDC range.
It reads -1.1222...-1.1227 VDC with nothing connected on terminals.
If i supply 1V from EDC MV106 to it will read 0.1224V, so seem to have offset on 1000VDC range. Lower DC ranges are fine, so not sure if it's cal issue, or some internal component still faulty but not diagnosed by self-test.

Now have problem, where to calibrate unit in Taiwan, without knowledge of chineese language, haha.



Also was curious if it's ok that meter on lower (200mV,2V,200V) VDC ranges have bouncing voltage readout, and 200mV usually shows overflow when nothing connected on terminals. Is it because of high input resistance of meter, and picking up charge? Do fully calibrated and functional 2001's or other 7.5/8.5 digit DMMs do this as well?