Author Topic: This Keithley 2001 IS NOW saved! (Read 10809 times)

Zucca · « **Reply #25 on:** November 26, 2019, 01:26:15 pm »

Thanks Kleinstein to stay with me in this madness.

I am planning to re-calibrate the unit at the end anyway also to have the certanty it works properly.

Moreover this device with full schematics makes a lot of fun and I want play a little bit with it.. so it's more a learning experience than a proper fix it and forget it project. That's also the reason why I keept this 20 years old DMM and not turned in a DMM6500 or so.

If someone have some idea to try it out just let me know.

Zucca · « **Reply #26 on:** November 26, 2019, 06:34:23 pm »

Order placed.

Zucca · « **Reply #27 on:** November 30, 2019, 12:50:43 am »

All right, Digikey delivered.

One minor step forward tonight.

Replaced U522, U532. Nothing changed.

Well not so fast.

Everytime after de- or soldering a chip, always power on the unit to check if and where something changed.

I traced the problem at U526 (the next after U532 in the signal path).

Replaced U526 and bingo the 400.x Problem went away.

Now I got:

402.1
404.1-5
405.2, 4 ,6 ,8
406.6
407.2-3
412.1

before was

400.1, 2, 4, 5
404.1-5
405.2, 4, 6, 8
406.6
407.1-3
410.1
411.1-2
412.1

I could be more happy that I bough all the DGxxx chips.
The plan is to get the unit error free again and then replace all the chips with new ones.

It takes time but it is fun.

Zucca · « **Reply #28 on:** November 30, 2019, 01:09:04 am »

oh oh..

402.1

Quote

OUT B of the DAC (U531) is set up to output 4.33V at PRECOMP+ (U528 pin 1). The
operation of the frequency switch, U522 (FREQ pulled low), is veriÞed by dividing the
PRECOMP+ voltage by the voltage ratio across R560 and R558. The "on" resistance
(approximately 25Ω) of the analog switch (U522) is added to the resistance of R558
since it is part of the ratio. Again, as in the 400 series tests, this voltage is routed to the
A/D buffer (U322) which is conÞgured for x1 gain. Measure 32mV at A/D IN.

my new fancy improved U522 has a Ron of 3Ω max....

Tempted to replace R588 with a 75Ω to compensate

Zucca · « **Reply #29 on:** December 05, 2019, 02:01:01 am »

404.1 on hold, I start to suspect the U509 (IC, MCHAN LAT DMOS QUAD FET, SD5400CY (SOIC)) is gone...

U508 and U515 were replaced with brand new ones....

Zucca · « **Reply #30 on:** December 05, 2019, 08:33:57 am »

Two SD5400CY are on the way to me. I paid them 9€ incl. shipping

Kjelt · « **Reply #31 on:** December 05, 2019, 09:43:17 am »

Darn looks like your 2001 was a zombie.

Zucca · « **Reply #32 on:** December 05, 2019, 10:35:28 am »

This is a painful fun, the worse it is the more happy you will be at the end.

Kjelt · « **Reply #33 on:** December 05, 2019, 12:46:39 pm »

Check and doublecheck internal pcb trace corrosion when the ics are removed with a bright led lamp below it.

Zucca · « **Reply #34 on:** December 05, 2019, 01:13:12 pm »

Quote from: Kjelt on December 05, 2019, 12:46:39 pm

Check and doublecheck internal pcb trace corrosion when the ics are removed with a bright led lamp below it.

Generally yes, in my case the leacking damage was very little on the surface, so I don't think it will help.
What is mandatory is to check if the same voltage is on the pins connected togheter, and if there is something not OK power down the unit and check the traces resistance with a DMM.

Zucca · « **Reply #35 on:** December 09, 2019, 06:48:58 pm »

Desoldered Q501, between D and S 1,2KOhm with nothing on the gate.

Maybe I did/do not need a new SD5400CY, oh well...

Kleinstein · « **Reply #36 on:** December 09, 2019, 06:56:07 pm »

For testing the MOSFETs, one should use a defined signal at the gate, more like gate connected to source and substrate: should give high impedance.

Testing with an open CMOS gate is like a random number generator.

Zucca · « **Reply #37 on:** December 10, 2019, 12:27:49 am »

R gate to source = R gate to drain = 1,03 MOhm

I don't like it.

Moreover I tried to put a negative square wave at the gate and I did not get nothing on the drain out (yes with a 1K pull up resistor to 10V).

Then I applied -10V=Vgs and the R between drain(+) and source was -2,1MOhm, between source(+) and drain was 30MOhm...

With Vgs=0V (gate shorted to source), Rdrain-source 135Ohm.

To me it's dead.

And of course the SST4416 JFET is obsolete.

https://eu.mouser.com/datasheet/2/427/910019-1211380.pdf

Zucca · « **Reply #38 on:** December 10, 2019, 12:40:36 am »

MMBFJ112?

https://www.onsemi.com/pub/Collateral/MMBFJ113-D.PDF

I go in bed, see ya.... $:=\$

Zucca · « **Reply #39 on:** December 11, 2019, 06:29:39 pm »

New SD5400CY arrived today, replaced.

Ah I resolder the q501 back.

Now I get:

402.1
404.1 ,2
405.6 ,8

before it was:

402.1
404.1-5
405.2, 4 ,6 ,8
406.6
407.2-3
412.1

Zucca · « **Reply #40 on:** December 11, 2019, 06:36:27 pm »

Power down and up again.

Now I got:

402.1
404.4, 5
405.2, 4
407.1, 2, 3
412.1

ar the third time power up:

402.1
404.4, 5
405.2, 4
407.1, 2, 3
412.1

looks stable now.

mmmm the temperature is playing a role here?

Kleinstein · « **Reply #41 on:** December 11, 2019, 07:08:56 pm »

xxx4416 FETs of some kind should still be available, possibly in a SOT23 case.

AFAIK J112/J113 are essentially the same as the usual 4391/4392 JFETs typical used as switches.

Getting JFETs is becoming tricky if a TO92 / TO18 case is needed. The same chip in SOT23 is usually still available.
There are quite a lot of similar FETs from the same process, but different testing sold under different numbers.

A good reference for JFET data / replacements can be Fairchild AN6609.

Zucca · « **Reply #42 on:** December 11, 2019, 08:04:00 pm »

Kleinstein Thanks you so much, one day you need to tell me why you know everything about EE componets.

Zucca · « **Reply #43 on:** January 09, 2020, 10:42:31 pm »

Well power it up today, here what I got:

402.1
404.1, 2
405.6, 8

I tried to let it warm it up, turn it off and on again but it did not moved from there.

Is my bench slowly repairing the devices just by holding them against gravity?

No sense, or I have an intermittent contact somewhere.

Zucca · « **Reply #44 on:** January 10, 2020, 05:17:39 pm »

Today's errors:

402.1
404.4, 5
405.2, 4
407.1, 2, 3
412.1

Kleinstein · « **Reply #45 on:** January 10, 2020, 08:36:24 pm »

Different errors coming up could be a parameter that is just boarder-line. Something like a intermittent error (e.g. cold solder joint) is also possible.
The 40x self tests are testing mainly the AC part.

The service manual describes the test in quite some detail and AFAIK one can run them one at a time to bring the meter in a defined state.
A point to check would be the DAC to generate test signals. Also have an eye on some of the CMOS switches - some of the old DG2xx tend to fail with age.

Besides the build in tests one could also create a test case by hand: measure some AC voltage and follow the signal to the RMS converter.

Zucca · « **Reply #46 on:** January 10, 2020, 11:13:47 pm »

Thanks Kleinstein, always a pleasure to read your comments.

Now I know what is going on.
The problem (thank God!) is the zero crossing amplifier, The input signal NETOUT is correct but the output signal COMP- goes bananas.
For some reasons COMP- either is always >0V or <0 so U509 can not switch between the non inverting or the inverting path in the VGA amplifier.

It is a very simple circuit and I am almost sure I will fix it easily.

Stay tuned, I see the light at the end of this tunnel.

Zucca · « **Reply #47 on:** January 11, 2020, 08:55:37 pm »

Do you see the bastard?

DuPe · « **Reply #48 on:** January 13, 2020, 03:21:04 pm »

U514?

RandallMcRee · « **Reply #49 on:** January 13, 2020, 04:05:39 pm »

I'll bet on CR501.


EEVblog Main Site	EEVblog on Youtube	EEVblog on Twitter	EEVblog on Facebook	EEVblog on Odysee

Author Topic: This Keithley 2001 IS NOW saved! (Read 10809 times)

Share me