help with repairs Keithley 2400

[tagchen]

Hi,

in same year this forum helps me to repair some of defect devices and in the next i will write about some finish repair projects.
But now i need help by an 2400.

all devices are calibrated annually and any time some are goes wrong. calibrated with selfbuild tool.
its devices bevor year 1999, also more than 20years running.
first i changed the caps (würth) and the fan (1570-1064-ND digikey).

the problem is ->  Calibration by selfmade box testet with load. with load condition everytime the current end so near -37mA in negative voltage quadrant. positiv quadrant is good.
i solder same of the mosfets and transistor and caps with oxidized solderpoint. test voltage ,see in block picture.
calibration protocol attached
 protocoll.pdf (251.18 kB - downloaded 237 times.)

but next??? what controls the current?? defect mosfet? im not sure what i can do in this error.

device off -> power on, measure an F-middlepoint is -165v to finish the boot, after that 0v.


Thanks,

[Zucca]

https://xdevs.com/fix/kei2400/

in case you were not aware of that, it could help.

So only the current in the negative voltage quadrant is bad right?
I skimmed the manuals and did not found any deep self test procedure, rather than the one at powering up.

To me without schematics is a big adventure to fix this device.

[tagchen]

So only the current in the negative voltage quadrant is bad right?

mostly, current end to round 37mA

and

[Zucca]

The red dots are all the failed test in your protocol.

I would suggest to investigate if always when the current must be below -37mA it fails.

As next step I would test these points:

200V, -100mA
20V, -100mA
-2V, -50mA
-20V, -50mA
-200V, -50mA

and see what happens.

Interesting would be also the following points:
-2V, 50mA
-20V, 50mA
-200V, 50mA
-2V, 100mA
-20V, 100mA
-200V, 100mA
-2V, 10mA
-20V, 10mA
-200V, 10mA
-2V, 1mA
-20V, 1mA
-200V, 1mA

[goaty]

Too bad noone has a schematic so far, my 2420 isn´t working although I´ve replaced many parts and still no luck ;-(

[tagchen]

its not the first, some of 2400,2410,2420 are running, but change mostly of ic.
but here i think the driver is correct working, setup worked. unload worked mostly.

Think so mosfet or so .

[Jens01]

I've repaired a K2400 and documented some hints here;

TP221 '+si' referenced to TP222 '-si': feedback current. Range dependent, should linear rise with measured output current.

In your case you can find out if the power-stage is broken by checking if the feedback current (voltage between TP221/TP222) linearly rises with the output current. if not: try different ranges (sink/source low current in high-current range) to see if only part of the feedback circuitry is broken.

[tagchen]

thanks,
i measure the first things ans see the voltage are end by same point.
can i say the feedback is ok and the power-stage is defect?




run as source failed and current sink are limited her

[tagchen]

cpu


blue resistor r377-r378 and transistors (but more the two transistore behind the resistors are hot)


u210


u211


u221


u263


u500


u501


prog-cam

[openloop]

"Thou shalt check voltages."

2400's output stage is powered separately from the main board.
Power is coming from T500 transformer - the big one, in the corner.
There are two diode bridges nearby:  CR500-CR503: +- 225V, CR504-CR507: +- 30V
(actual voltages can be higher)

Check that the voltages are there (against "common" - TP501)

EDIT: Just noticed that piece of paper implying that you already did. If so, then please disregard this message.

[openloop]

OK, it's getting interesting. 

The whole output circuit is controlled by U500 op-amp (AD847JN). It's an 8-DIP package near all those power transistors.
Check that it's "in regulation" when your output is wrong: voltage on pins 2 and 3 should be the same and be equal to SMU's output voltage.
Also check it's power pins - this op-amp is bootstrapped around SMU's output voltage.

[Le_Bassiste]

yep, that should be a resistor. this particular block diagram is erroneous in all KEI24xx docs, as far as i have seen. 

[udok]

- check the Supply voltages first
- check that VFB and IFB (voltage and current measurement amplifier) corresponds with Vout and Iout as measured with a DMM.
- IF VFB is ok with negative voltage outputs, and the output buffer amplifier is ok, then the error could be in the range switching.
- check that the DAC output is ok
- check connectors

Note: The complete analog circuitry is floating and voltages are referenced to the floating ground.
The floating ground is on the output of the power amplifier.

VD voltages are referenced to digital GND.

TP3 on the power supply PCB is the Digital GND
TP4 is the floating ground for the analog circuitry - TP3 floats with the output voltage - DO NOT connect with digital GND!
TP2 is +12VD Digital
TP1 is +5VD Digital

TP6/TP7 is +15VF and -15VF (floating)
TP8/TP9 is -30VF and +30VF (floating)

If you enter 1 Volt and 100 mA  at the front keys,
the voltage and current DACs should change accordingly:

TP201 is the negated Voltage DAC control voltage
TP203 is the negated Current DAC control voltage

There is a special "Approximate Feedback" part.
TP204 is the negated control voltage for this part.

There is a "Polarity" voltage on TP202

The 2400 regulates the difference between the control voltage (current) and the measured voltage (current).
There is a lot of switching for the different quadrants depending if the unit is in voltage or current mode.
Test voltage and current mode independently with No-Load or a small resistor at the output.

The result of this magic switching is on TP213, which is the error drive going to the output buffer stage.
The main buffer drive is on TP214

The switching signals are on TP205-TP212

The output voltage is measured with a differential amplifier.

The input of the differential amplifier is SENSE_H and SENSE_LO
or a local version which comes from internal nodes.

TP227 is the buffered SENSE_HI input
TP228 is the buffered SENSE_LO input after the voltage range divider
TP230 ist the VFB (voltage feedback signal), the difference of TP227 and TP228
TP231 is the CFB (current feedback signal).
TP233 is the ADC input signal, after the multiplexer which switches between VFB and CFB.

There is a special  supply +5VFL on TP218 and -5VFL on TP219 for the SENSE_LO voltage buffer (not floating)

If you measure +-1 Volt between the input pins with a DMM and measure a current of +-100mA through a load resistor,
the VFB and CFB voltage should correspond to the DMM voltage and current measurement, scaled by the voltage and current range.

ISENSE_HI ( is at TP221
TP220 is ISENSE_LO (=output of power amp after current sense resistors), which is buffered at TP222
The difference is the current feedback signal IFB

There is a current guard voltage on TP224

And some parts have a driven guard supply which floats in TP223

[Zucca]

Also since you have a good working 2400, please open it up that too and compare voltages against the good one.

[Kleinstein]

Hi,

in same year this forum helps me to repair some of defect devices and in the next i will write about some finish repair projects.
But now i need help by an 2400.

re block.jpg
the zener to the left of the Q518 doesnt seem to be connected correctly ... the cap on top of it looks weird 

The schematics shown seem to be not 100% accurate. The positive side does not make very much sense, e.g. shortening one of the NPN is shorted emitter to collector. The zener and cap in series is another point. I would consider the negative side more practical and the positive side is likely symmetric.

[tagchen]

Also since you have a good working 2400, please open it up that too and compare voltages against the good one.

yes, i have much more of 2400. and i will compare the point.

thanks for the answers, i will check today and post.

[macboy]

- check the Supply voltages first
- check that VFB and IFB (voltage and current measurement amplifier) corresponds with Vout and Iout as measured with a DMM.
- IF VFB is ok with negative voltage outputs, and the output buffer amplifier is ok, then the error could be in the range switching.
- check that the DAC output is ok
- check connectors

Note: The complete analog circuitry is floating and voltages are referenced to the floating ground.
The floating ground is on the output of the power amplifier.

VD voltages are referenced to digital GND.

TP3 on the power supply PCB is the Digital GND
TP4 is the floating ground for the analog circuitry - TP3 floats with the output voltage - DO NOT connect with digital GND!
TP2 is +12VD Digital
TP1 is +5VD Digital

... <snip>

Fantastic information! Your post has much more useful information that the entire service manual for this thing. Thank you for this.

I am inspired to take my faulty 2400 off the shelf and have another go at fixing it.

[JohnPi]

I didn't want to start another thread, but wanted to document what it took to fix my Keithley 2420.

I wouldn't turn on. Measurements of the Vicor supply (large DCDC with an Aluminum heatsink) showed it was bad, but a replacement didn't help. Turns out it wasn't bad. Some of the power FETs on the power supply board near it were bad (burned) -- so I replaced those. It then did power on, but gave an OVP error when the output was enabled. I measured some of the test points (as in the service manual) -- the +/- 85 V were OK, but the +/- 18 and +5 were bad. I removed some of the transistors from the output stage  -- that didn't help. I was about to give up when I noticed some surface mount 1A fuses on the power supply board.
- F3, F4, F5
There are 3 of them and all measured open. I put wires across each (they are nearly hidden under a transformer) -- I didn't have any actual fuses. The the machine came back up and seems to work fine 

[JohnPi]

Just a follow-on -- it's still working fine and calibrated well (14 years since last cal !).

The fan is quite loud; I might replace it (80 x 80 x 25 mm^3; 24 V 2-wire).