Restoration glory of Keithley 2001 DMM

[benvei]

@benvei,

Is there any reason not to diagnose and repair the AC selector circuit? Did you destroy any of the traces on the pcb?  Odds are that repairing the defective parts would be easier than rewiring the primary side of the meter.

Yes, i did, but i think i should be able to fix it. I just wanted to know, if its even possible, to hook up the transformer directly to the 230V mains.

Thanks!

[ManateeMafia]

It might be possible to hook it up straight to the power input connector. I just don't know if you will bypass anything important on the analog board. I believe the Freq Out signal may come from that section but I am not sure.

I am paranoid about modifying anything on the primary side of a transformer. I will do direct replacement of parts, but I would not bypass anything. I have read enough posts about input filters burning up and I am concerned my homeowners insurance will blame me if it happens and causes damage to my home.

Hopefully you can get it straightened out. I can get you some pictures of anything that may be missing or unreadable.

[ManateeMafia]

I am one step closer to fixing 2K1 #2. I stole the A/D converter board from #2 to get #1 working.

The defective board from #1 had the following errors that were confirmed on both working meters...

200.1 No A/D communication and/or noisy A/D.
201.2 Cannot measure 7V at A/D IN

The 200.1 is a showstopper since some of the other individual tests run this first and will throw it as an error. The meter seemed to read fine and I was certain it would be difficult to track down.
I started probing the on-board voltages and everything seemed fine. All transistors "seemed" fine compared to another board using the diode test function.

I happened to notice that Vcc (+15V) had a low impedance to the GndA(S Com) signal path. After checking C828,C809, and C812 I started looking at resistance readings across component legs and eventually found Q806 had low resistance between the source and gate pins. I dropped in a replacement and now the board passes its self tests.

Unfortunately, I seem to have a new error 409.6 on #1. I am certain I had it before but it wasn't there earlier today. It throws the same error with two different converter boards.

[TiN]

I have 409.6 come and go on one of analog boards too. It's unrelated to ADC board, just for note.

[ManateeMafia]

I figured the same thing. I tested what I could so far with my handheld meter, but the circuit around C554 needs a higher impedance meter due to R557. I will check again tomorrow with a 8846A and hopefully find the part(s).

C554-C556 are available from Digikey, and it wouldn't be too much trouble to measure them. I confirmed that U526A (pins 2-3) measured within tolerance (33ohms) which matched my other meter.

Have you measured the input impedance on your meters that had the failed capacitors? I am wondering if the electrolyte has created any issues and I will have to measure it when I get these put back together.

[benvei]

Hopefully you can get it straightened out. I can get you some pictures of anything that may be missing or unreadable.

Hi!

Thanks for your advice. I wanted to connect the transformer directly, but yeah ... i thought about your post. And i went on searching for the problem. The Transistor that was smoking was Q103, which was switching the 110V relay. Since we have 230V I don't need it, so i just desoldered it. And? Oho - Its working fine!

I've attatched a picture of my test. Its just insane how accurate this thing is. It was built an calibrated in 1993!

I'm quite happy with it. Thanks to all of you, and thanks for reengineering and drawing the circuits, that was quite useful. And thanks to keithley for their great Repair manual!

[ManateeMafia]

It is nice to see it was a simple fix. I am sure it will be tempting to get a second.

I think Keithley did a great job with these meters. I would like to have TiN's setup but I think the eBay sellers are asking too much for non-functional units and I am happy with what I have. At least for today...

TiN and the other xDev's members have done a great job. They have certainly been a big help with their notes and schematics.

[TiN]

Probably most of them saw this thread and bumped prices up once to twice  .

Anyway, be careful and get meter proper repair/treatment it worth. Accuracy of these machines is mostly on voltage/resistance ranges, not frequency  .

Also if you have access to photocamera, photos of your 2001's internals would be good to add into collection.
What is firmware version in your meter?

[benvei]

Also if you have access to photocamera, photos of your 2001's internals would be good to add into collection.
What is firmware version in your meter?

Yes, i've got a photocamera, i'm going to take some pictures for you in the next few days. I just cleand up my workbench, and yeah ... if students start cleaning their desks without their parents shouting at them, it has to be a *real* mess. And ... it was.

Is the value displayed at the boot the hardware rev or the software rev? If its the software, its firmware "B06".

[TiN]

It's software, thanks. I already have B06
Hope to see photos of your digital and analog boards.

Meanwhile, here's draft article about secret menus in Keithley gear:
Keithley secret menu and settings

Tried B-version 2001 with secret menu, it have little more stuff , and MESSAGES/GAME item, but it does nothing when I enter it 

[benvei]

Hope to see photos of your digital and analog boards.

http://cl.benvei.at/public.php?service=files&t=5975efac14949058f8775e8744de03c0

Didn't upload them to the forum because of the large image size. Sorry for the bad light quality, should i do them again when i have some sunlight? Its 8pm here in Austria.
I've also got a Scanner Module, if you are interessted in it. I also replaced all the caps.

I know, cap 103 is a mess, didn't have one that fitted, going to change it as soon as i get one that fits. I could make photos of the other side of the PCB when i'm changing the cap (since it works now, i didn't really want to take it apart, sorry).

//edit:

Tried B-version 2001 with secret menu, it have little more stuff , and MESSAGES/GAME item, but it does nothing when I enter it 

Same here.

[mmarks3141]

Nice article about the secret menus - brought back a few good memories.
The references and menus about a phase card are for a peripheral that we had considered making, but ultimately never was completed or sold. Not sure whether there wasn't enough market to justify it, we couldn't get the accuracy we wanted, or the designers got pulled off to other projects.
--
Mike

[TiN]

What is about TC-Ohms option?

Also I am still puzzled about what is proper Dallas NVRAM for 2002, since it does not like DS1251 (say's no clock installed) even tho 1251 do have PhantomClock in it.

[mmarks3141]

TC Ohms was a Temperature-compensated ohms card.  Not sure if it ever got sold or was just another feasibility/development project.

I don't recall the Dallas part number for the 2002 NVRAM, but it does have to have a real-time clock.  The 2001 does not have a real-time clock and the two parts are different, but that's all I remember.

[macboy]

What is about TC-Ohms option?

Also I am still puzzled about what is proper Dallas NVRAM for 2002, since it does not like DS1251 (say's no clock installed) even tho 1251 do have PhantomClock in it.

DS1251 is 512kB, but MEM2 option is 128kB, so it's unlikely to be right. DS1251 has 2 extra address pins over the 128k version. DS1646 is a Dallas 128kB with RTC. It's the right size at least, so it's more likely the correct part than the DS1251. Access to the clock portion of those two chips is completely different.

[Galaxyrise]

I'm wondering if I should drill out the affected board areas?  There's not too many traces in the C117 area, but there's a lot of stuff going on around CR513.  Or is there a better way to restore the integrity of the board?

I couldn't convince myself that the isolation on that ground trace was still good, so I cut it and replaced it with an ugly bodge wire.  It seems that what actually finally stopped the catastrophe was the transformer secondary dying, so I bought a replacement transformer from Tek (only slightly more expensive than I expected, but not cheap)  And now the unit boots!

Should the fan start when the unit is powered up?  The next thing I did was let it sit there while watching it thermally, and the fan never came on.  Without anything to compare to, I don't know if this is because the case was open or is another fault.  It would make perfect sense if the fan died, causing the caps to overheat and die.

Do you remember what the hot spots were from your thermal analysis?  I've got several parts operating around 60C.  The two parts clipped to the chassis were also quite warm, with the power resistor at 65C.  Ambient was around 18C.

Before plugging in the new transformer, I compared winding resistances with the old transfer just to confirm that the old one was dead.  Here's the new resistances in case it might help someone else troubleshooting their unit: (Note I swapped the Orange/White pin from the connector order.)

Code: [Select]

BLK   -+                  PINK    -+
       |- 3                        |- 2.7
BLK/W -+                  ORNG/W  -+
       |- 14.3                     |- 2.6
WHT   -+                  PINK    -+
       |- 16.5
BRN/W -+                  PURP    -+
       |- 3.9                      |- 0.75
BRN   -+                  PURP    -+



[TiN]

You talk to yourself?
You could just shoot me PM, I had extra xfrmr from donor 2001.

Fan powers up instantly, it's running all the time. I got feeling that some of my 2001's died because of fan too, because it was locked on few units, while very noisy and near failure on others.
So just swap it out to standard 40x40mm 12V 0.3A fan and you will be ok.



Unit does run quite warmish, that's normal to have linear regs and mains part around 60's. R100 and Q528 basically sit on rectified mains, as you can see from my schematics (posted before)

[Galaxyrise]

Fan powers up instantly, it's running all the time. I got feeling that some of my 2001's died because of fan too, because it was locked on few units, while very noisy and near failure on others.
So just swap it out to standard 40x40mm 12V 0.3A fan and you will be ok.
...
Unit does run quite warmish, that's normal to have linear regs and mains part around 60's. R100 and Q528 basically sit on rectified mains, as you can see from my schematics (posted before)

Thanks! 

I had studied those parts of the schematics and the repair manual (section 2.10.2) before asking. I think I understand what's going on there, and I did expect Q528 was usually in the linear region (or it wouldn't be much of a pre-reg) but had no sense for the normal amount of current .  If mine were running hot, then I had another failure to go find. 

The schematics posted here are a part of why I undertook this repair and they've been very useful already.  I really appreciate all the information in this thread! 

[TiN]

How about your unit's selftest? Any error's? I have two units open on table now, waiting for more debug and schematics work for U5xx chips.

Also what is your analog board Rev? From thermal pics looks like it's older one with avalanche diodes on top near caps.

[Galaxyrise]

How about your unit's selftest? Any error's?

I measured voltages, put the shield back on, hung a 50mm fan on the PSU section, let the unit stabilize for an hour, and ran the built-in tests.  No failures! 

Adding airflow brought the top-side power components down into the 30-40C range, which was reassuring.  However, the bottom-side 5V supply (CR622 and U629) still runs 60-ish C.

I attached it to my 5xLM399 reference for awhile, and it has 15uV more noise in its measurements than my 2000s do (all using the default "slow" speed; 2Hz?)  I'm hoping that goes away when it's properly reassembled, but I'm not optimistic.

Also what is your analog board Rev? From thermal pics looks like it's older one with avalanche diodes on top near caps.

The board is marked:
2001-102-04L

On boot, it displays:
Rev. B06  A01

All the date codes on board components I've found so far are 92 or 93.  The old transformer's serial number ends in 9412, if that's not coincidence. So I think it's safe to say this is an older unit

[Zucca]

Guys,

I am getting soon a 2001! I am so excited!!!!!

Question on the side, some EE brothers here:

https://www.eevblog.com/forum/testgear/tektronix-2465b-oscilloscope-teardown/msg583136/#msg583136

managed to upgrade a DS1225Y NVRAM with FM18W08 FRAM:



Any idea if we can do the same here with the MEM2 chip Dallas DS1245Y? Sorry if it was a stupid question, I am not an expert here.

Can´t wait to post something here... thank you all for the hard work!

[TiN]

Galaxyrise

Great job.
Can you send B06 as well for my collection, since I don't have it?
Try NPLC 10, auto-zero = synchronous (will auto zero before every new measurement), line-sync - enabled. You can enable 8.5 digit for fun too.

U629 mounts to chassis via screw for thermal transfer. Poor LDO powers whole digital board and front panel board including DC-AC converter.

Zucca,

Welcome, I can send you a 1245, have few extra. I thought you already had 2001 before?

Which unit you have got? All you people might bring some life again in this poor thread I wish could have some quality time to work on schematics too.

[Zucca]

I can send you a 1245, have few extra.

Thanks for your offer TiN, I will consider that.

I thought you already had 2001 before?

Unfortunately no, I got only 6.5 DMMs on my bench never a 7.5 one.

Which unit you have got?

The Unit will be shipped to me, but there some pictures:



I don´t know if it has the B or A Firmware.... ah, it has the  2001-342c 10 ch Scanner card 
Do you guys have an extra front handle for me?

PS: the owner was a wine producer in Canada, he was using the 2001 to control the wine temperatures  . I got it pretty cheap, but let´s see when it arrives in my hands.

[ManateeMafia]

I was watching that 2001. That was a good price with the included scanner card. It will probably be a rev B firmware. The serial number is between my B10 and B13 versions, but closer to B10.


TiN,

Will upgrading the FW blow out the calibration constants and/or serial number? I should have the adapter for my programmer in the next couple of days and I would like to back up the eeprom before I upgrade.

I also have to fix the serial number on my 2510 that was lost when I installed your uploaded version. BTW, I finally got time to upload ver A03 for the 2510 on xDevs.

I have been emailing Keithley for prices on new replacement parts to spruce up the external appearance. I will post an update when I can get my hands on some of them with updated part numbers from the service manual. A few parts have slightly different part numbers.

[Zucca]

Very Interesting:

Replacing SRAM With FM1608B,FM16W08, FM1808B, or FM18W08
http://www.cypress.com/?docID=48118

Does anyone know how the chip-enable transitions to LOW is managed in a Keithley 2001?

The serial number is between my B10 and B13 versions, but closer to B10.

Thanks for the Info!