-
Hello volt-nuts.
It's probably sickness, but one day I was thinking that having multiple KEI 2001's and 2002 ain't enough, and wanted to have some addition
to not so convinient Agilent 34970A at work, so I randomly bidded on abused Keithley 2000 from Israel, with a hope that somebody else would buy it.
As usually, Merphy gotcha me, so I won it with my single bid for $80USD. Meter arrived packaged nicely in big box, powers on and exactly as described.
VFD is intact, desprite missing front face plastic. I will use my donor 2001 front panel later, so not a problem here.
Functions are pretty much broken, shows overflow on ACV,ACI,4W, ramps readings like crazy on DCV,DCI,ohms.
Cranked it open, aha....
Pesky capacitors leaked their electrolyte causing major disaster on poor PCB.
According to date codes, this meter is already 20 year old
It fails pretty much almost every self-test.
So no wonder, one of power supply rails probably gone bad.
Will see how this worklog will go.
Did I got another endless repair project, or this baby will be up and running in no time?
Bonus included, for Keithley 2000 our chinese friends from bbs.38hot.net forums reverse-engineered schematics for analog part.
I got it mirrored on my site here
And as usual, all my data and photos are always available on related project page here.
In worst case I got very well aged LM399 voltage reference, ha-ha
--------------------------------------------
Current state
Gang switch desoldered, can measure voltages,resistances if connected to R115 input resistor and AGND.
Relay, power seem to be ok.
6800uF 16V need to be replaced
17V zener VR109 require replacement (SOT23-3 device), currently use pair of 11+6.2V zeners
AD706 need replace?
Errors in 40x series tests.
Chassis need some metalwork to fix bending
Almost in calibration, even after all that butchery
-------------
Article & docs
xDevs.com Repair log for Keithley Model 2000
, ?? radiokot.ru - for russian folks
Documentation section about Keithley 2000
jhzyou 126.com from bbs.38hot.net schematics
-
so I randomly bidded on abused Keithley 2000 from Israel, with a hope that somebody else would buy it.
I'm looking forward to reading and following your repair efforts.
With that particular seller, many here will look and shake their heads in disbelief at how such equipment can be abused, but not bid due to normally ridiculously high asking prices. -
You still not replaced caps in your 20-year old gear? That's what happens...
Meanwhile dumped firmware ROMs here -
Hello,
if you can repair the instrument you can find the latest firmware (A20) for the 2000 model here before they
repent of having made it public for free....
http://forum.keithley.com/phpBB3/download/file.php?id=1922&sid=b1e6c26ea52bbf76cd35f47f31259b3b
-
Thanks for firmware, but mine 2000 have non-erasable OTP EPROM chips, so I can't reflash it.
Not sure if I have UV 256KB EEPROMs in PLCC32 anywhere
Added few more photos for everybody enjoyment.
Teared apart meter..
Top side of mainboard
Bottom side of mainboard
Back of front panel board
Self-test status as received:
Last calibration date: 03/13/1995
Calibration counts: 1
Failed: 100.2, 201.1, 201.2, 300.1, 301.1, 301.2, 302.1, 302.2, 303.1, 303.2, 304.1, 400.2, 401.2, 402.2, 403.2, 500.1, 500.2, 600.1, 600.2, 601.2
Replaced caps, 7815, 7915, some JFETs, U109 (14094B register), U111 (DG211DY) from my donor 2001 yesterday night, few errors go away
Few self-test errors gone, now failed: 100.2, 301.1, 302.1, 302.2, 303.1, 304.1, 400.2, 401.2, 402.2, 403.2, 500.1, 500.2, 600.1, 600.2, 601.2 -
After looking at all the other items from that seller all I can say is .... WTF? I like the obviously run over fluke 189 for $90 shipped to the US though.
-
Necrothread revive!
No, this is not 7001.
It's broken 2000 hooked up to front panel from 7001.
Obviously, technically it's possible to use 7001 or 2001 front panel in 2000, since it's same interconnect.
Both instruments series (2000,2700,2750,2790, name it, all single-line Keithley stuff) and dual-line using UART to send data from main DMM 68000 to front panel MC68HC705C8.
Issue is only different protocol commands, so someone persistive enough with plenty free time could bodge a little MCU in between to translate 2000's data/readings into 2001/7001's format and enjoy bigger, nicer VFD on 2000 instrument. I'm not gonna go that road. My 2000 native VFD panel is very dim, so I was wondering if 7001's display would work as is.
As result is no, I will go forward and desolder NEWHAVEN DISPLAY DD-51C glass from 7001 board and solder on 2001-112 front panel PCB, since I lack of one 2001 FP.
Why there is photo of 33120A there? Ahh, glad you asked, it's a trap for young players. Tell me why is that puppy shown?
-
VFD brightness fix
Today seem to be a good day for some fiddling with my broken 2000. One of issues, disregarding broken analog circuits is really really dim display. On usual daylight it was completely invisible, even without dark plastic plexi.
I tried to take photo of original state:
Filament had 8.75 ohm resistance and had 2.11VAC across it during operation.
VFD brightness drop from two reasons:
* Phosporus become aged if segments are on for long periods. This is visible by dim patterns on screen.
* Filament loose emission due oxidation
Not much can do for fixing segments itself, but we can try deoxidize filament and increase current across it to have more emission.
To do so, I used trustly Keithley 2400 SMU in source current mode. Started with 140mA and raised up to 180mA (6.5VDC compliance voltage)
for couple minutes to heat up filament.
Now it become little better, but still not enough for me..
There are current limiting resistors near each filament pin on PCB, stock resistors R405,R412 12.1 Ω and similar on other side.
We can increase operation current by adding parallel resistors on top of original ones. I used three 22.1 Ω in 1206 size on top of each location.
Result is satisfactory!
Now even with panel display can be read easily.
Of course this is temporary improvement, and after few years brightness will drop again, but it's not much else to loose. After all VFDs are giving you that warm fuzzy feeling, and have awesome viewing angles and contrast.
I think now this 2000 have chance to get my attention to actually fix it
-
I would like to try to deoxidize the filaments of the 6517A display. The whole display is verry dim. Just one question, how do you specify the maximum voltage/current for the fillaments?
-
I just carefully adjusted current from 10mA till it glow nicely orange like in photo. Adjust it slowly, and don't stress or shock display, as hot wires are easy to break.
6517 is dualline DD-51 VFD, so it's about double current than for segment one like in 2000.
-
It worked, a little bit. The display is only a bit better, but it was very bad to begin with. I could tell by looking to the bare VFD which pixels where "bad" (the color was significantly different).
Just had to disconnect the display module and remove T901, this transformer drives the filament. Filament is driven on the outermost 2 pins of the VFD module. I had upped the voltage to 8,5V in 3 minutes, on which the the current reached 300mA. The filaments glowed a little bit, I didn't want to destroy them.
Good news, I did find the secret menu in the 6517A C03 FW: Press SEQ + Voltage source UP and the secret menu appears under menu => general. It also has en display burn-in setting, Ill give it a try.
BTW, just put my Keithley 199 FW on your FTP -
Cool, thanks.
If you have time , maybe you can make few photos of secret menu there for article?
P.s. I think worth creating separate thread about 6517's, as we discuss it already in multiple different threads
-
I just carefully adjusted current from 10mA till it glow nicely orange like in photo. Adjust it slowly, and don't stress or shock display, as hot wires are easy to break.
I strongly discourage using a constant current source to do this; use a (variable) constant voltage source.
6517 is dualline DD-51 VFD, so it's about double current than for segment one like in 2000.
The reason is that as the filaments heat up, their resistance goes up too. If you use constant current, the supply will dutifully push the preset current into the now higher resistance, giving more power and more heat, so higher resistance in the filament, which results in more power and more heat, so more resistance... you can get thermal runaway, and the casualty will be the filaments. If you use constant voltage, then as the filament heats up and the resistance goes up, it tends to reduce the current and therefore the power, providing protection from thermal runaway. -
Cool, thanks.
If you have time , maybe you can make few photos of secret menu there for article?
P.s. I think worth creating separate thread about 6517's, as we discuss it already in multiple different threads
Good idea about the 6517 thread.
I haven't took any pictures, but I did upload a list of the secret menu settings. And some new versions of Keithley 2000 FW
-
macboy
Good point and trap for young players, thanks for highlight.
I was using SMU, so it was not so obvious, when I say it's constant current, as it have compliance voltage, and I had 200mA set for current source limit, and adjusting voltage compliance setting to get desired amount of current, so even if resistance drops, I would just hit voltage limit by compliance, or current limit by source.
If one use just CC PSU without any limiting on voltage - that may cause situation like macboy point out. Beware.
Thanks Smith! Updated catalog.
-
Will the xdevs site add Keithley picoammeters too? They made quite some versions, they're pretty good.
-
What you would like to know about picoammeters? In some thread I already posted some internal photos of 6485/6485 and SMU 2636.
-
I'd like post everything (not just KI gear), as there is space for many other T&M.
-
Cut from the private message:
"
Some time ago I was repairing Keithley 2000 and Keithley 2700 DMMs.
Both have defective LM339 (quad comparator) which controls FET switches.
Both units had small holes in LM339.
After replacement, both units worked just fine (checked on all ranges).
But how did the failure happened? All surrounding circuits are working fine...
I have only two ideas: defective series of IC's or overvoltage on -15V, +5 lines powering the comparator...
So I have two questions:
1. Did You have some units with broken LM339? Is it a common problem in 2000 series?
2. Two voltage regulators (LM7815M and LM7915M) are extremely hot. Is it normal ?
3. PCB near LM399 (Zener) is also quite hot. Is it normal ?
"
TiN answered:
"
1. I had few broken 339's in one of 2001. I have only one 2000, and it still waiting it's time to repair.
2. How hot? They run usually around 50-60ish C. It's okayish...
3. Yes, LM399 is ovenized zener, it's internal die thermostat runs about 90-95C to keep zener voltage stable and independent of ambient temp." -
The regulators do run quite hot indeed on both the 2000 and 2700's. There are a few revisions with different packages (SMD and trough hole) and different configurations. Mainly due to the nearby capacitors getting to hot and spill their guts on the PCB and componente on the early versions. I heared of quite a lot of units with defect regulators. I only had one (LM7915 on the 2700).
BTW I posted some K2000 internal pictures of several older units on TiN's FTP. -
Since we have gazillion of 2000 threads poping up every week, I'd think someone may be up for grabs on this one:
http://www.ebay.com/itm/Keithley-2000-multimeter-/161698789900?pt=LH_DefaultDomain_0&hash=item25a5ffba0c
All components seems there, just teared down. VFD likely be okay, getters are not white.
And price is okayish, if seller may accept reasonable offer I pass on this one, so who wants 2000 - that's ur chance
-
Since we have gazillion of 2000 threads poping up every week, I'd think someone may be up for grabs on this one:
http://www.ebay.com/itm/Keithley-2000-multimeter-/161698789900?pt=LH_DefaultDomain_0&hash=item25a5ffba0c
All components seems there, just teared down. VFD likely be okay, getters are not white.
And price is okayish, if seller may accept reasonable offer I pass on this one, so who wants 2000 - that's ur chance
Seller accepted $95 shipped
-
Dont forget to read firmware from it and post your findings here for us
I hope we can have most of data in one thread, instead of 50 threads about same meter
-
Time to fix all this mess..
Parts removed for board repair :
Anyone have 2000 for measurements around? I'm interested in resistances of R153 network, since schematics says it's two 3.6K resistor divider, but mine measures 3.60K pin 2 to center pin, while pin 1 to center pin is ~9K. -
Dont forget to read firmware from it and post your findings here for us
I hope we can have most of data in one thread, instead of 50 threads about same meter
With forum software like this, I don't think having one thread instead of 50 is necessarily any better. People who are participating in current/ongoing conversations have to contend with conversations on other subtopics that are taking place at the same time, and people who come later have to navigate the tangle of interwoven conversations to figure out the current thinking on a particular subtopic.
One thing that can help keep threads consolidated and long lived is if someone updates the first post occasionally to include a sort of FAQ and links to useful posts and outside resources (like your excellent XDevs project).
