Author Topic: Keithley 6517A Electrometer: dead voltage source; Replacment for 2SK1412 MOSFETs  (Read 1248 times)

0 Members and 1 Guest are viewing this topic.

Online DB4UCHTopic starter

  • Newbie
  • Posts: 7
  • Country: de
Hello together,

I recently took on repairing a Keithley 6517A Electrometer with a broken voltage source. When I got this unit, some components in the voltage source section were already missing, including one 2SK1412. I then checked the remaining large MOSFETs and two of the three 2SK1412 had failed (low resistance between Gate and Drain). So, I will need to replace those MOSFETs (2SK1412). I will also replace most of the semiconductors around those FETs, because I can't really be sure what exactly killed what. (the supply voltages for ±1200V and ±, 120V seem to be in range according to the service manual, so most likely not those)

Is there any "trusted" source for those MOSFETs or for the possible replacement (NDFPD1N150CG, which is EOL as well and seems even harder to source)?

Are there any other (available) FETs that might work in this application?

Thanks!

Greetings,

Simon
 

Offline Jetecnet

  • Contributor
  • Posts: 25
  • Country: us
Last time i bought these 2SK1412 I got them from Radwell, price has gone a bit crazy though.
https://www.radwell.com/Buy/SANYO/SANYO/2SK1412
JC
 
The following users thanked this post: DB4UCH

Offline voltsandjolts

  • Supporter
  • ****
  • Posts: 2547
  • Country: gb
 
The following users thanked this post: DB4UCH

Online squadchannel

  • Frequent Contributor
  • **
  • Posts: 423
  • Country: jp
  • deepl translate user
Not trusted but cheap!

https://www.ebay.co.uk/itm/226356193374

TO220, but the shape of the molding is very unique.
for fake parts, most of them are just re-marked with a laser. is not possible to change the shape of the package.
legs are oxidized and solder plated.
I think it is genuine.
 
The following users thanked this post: DB4UCH

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 15149
  • Country: de
There are other high voltage MOSFETs available. The problem seems to be finding small ones with a low gate capacitance - no problem with 1 or 2 A types (e.g. STW3N150 or IXTH1N200). It depends on the circuit how well it would handle the higher capacitance.

If one can accept a slightly lower voltage rating, the IXTP02N120P could be an option. Still a little higher capacitance, but not that much.
« Last Edit: October 29, 2024, 01:51:30 pm by Kleinstein »
 
The following users thanked this post: DB4UCH

Online coromonadalix

  • Super Contributor
  • ***
  • Posts: 7001
  • Country: ca
the originals   rds is somewhat high for sure,

there are better ones out there

not sure if this help

https://alltransistors.com/mosfet/transistor.php?transistor=22309
https://alltransistors.com/mosfet/crsearch.php?&struct=MOSFET&polarity=N&pd=20&uds=1500&ugsoff=1.5&id=0.1&rds=200&caps=TO220FI&cf=on

upon a search this one came out     2SK3745LS ??   10x  lower  rds   ...  seems available
« Last Edit: October 29, 2024, 04:54:46 pm by coromonadalix »
 
The following users thanked this post: DB4UCH

Online DB4UCHTopic starter

  • Newbie
  • Posts: 7
  • Country: de
Thanks for all your answers!

I have decided to try ordering the MOSFETs from the eBay link that voltsandjolts had posted and to compare their characteristics with the ones specified in the datasheet. (To some degree, I don't really have a proper high voltage source, yet)
Currently I'm also in the process of reverse engineering the voltage source to find out if one can get away with MOSFETS that have higher capacitances.
Using MOSFETs with a much lower Vds isn't really an option (I think) as the voltage source outputs up to ±1kV and operates partially from an unregulated ±1.2kV supply. (according to the service manual ±1.1kV to ±1.3kV is the regular output range of this supply)

I will keep you updated on the repair and the reverse engineering.

Greetings,

Simon
 

Online DB4UCHTopic starter

  • Newbie
  • Posts: 7
  • Country: de
So, I think that I have mostly reverse engineered the output stage of the voltage source. It isn't too beautiful and might still contain mistakes (including wrong symbols), so use it at your own risk.

I have ordered all the parts to (hopefully) fix it. I will keep you updated.

Greetings,

Simon
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 15149
  • Country: de
The symbol for the TLP590 is wrong. This is an photovoltaic opto-coupler, so a controlable current source that can diliver voltage.
There are some parts missing / wrong around the gate of Q311 and CR303. As shown there is no path to discharge the gate. There is also an issue with the gate of Q317 - no control there, just a zener for the protection.

The circuit seems to use the LM334 as current limits and the LH1056 opto relais to switch between 2 settings for a high and low maximum current. It would help with readablity, if this part and the MOSFET gate control part are drawn more separate, e.g. the gate control on the left and the current switching on the right.
The gate connections for the SK1412 seem to be missing quite some parts.
 
The following users thanked this post: DB4UCH

Online DB4UCHTopic starter

  • Newbie
  • Posts: 7
  • Country: de
The symbol for the TLP590 is wrong. This is an photovoltaic opto-coupler, so a controlable current source that can diliver voltage.
There are some parts missing / wrong around the gate of Q311 and CR303. As shown there is no path to discharge the gate. There is also an issue with the gate of Q317 - no control there, just a zener for the protection.

The circuit seems to use the LM334 as current limits and the LH1056 opto relais to switch between 2 settings for a high and low maximum current. It would help with readablity, if this part and the MOSFET gate control part are drawn more separate, e.g. the gate control on the left and the current switching on the right.
The gate connections for the SK1412 seem to be missing quite some parts.

Thanks for your recommendations, I changed the schematic a bit and added a few missing components. I wasn't really able to find some of the gate control circuitry, every SK1412 has the cathode of a Zener and a 1k resistor directly connected. Through those resistors some control signal reaches the gates. I think there should still be some circuitry connected to those resistors, but I wasn’t able to find all of it. :-// I also separated the current switching from the gate control like you recommended.

Greetings,

Simon
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 15149
  • Country: de
Except for CR303 the positive side now makes some sense. They may be something the other way around or so.

VR305 and likely also C322 should be more to the left as part of the gate control. C322 is however a bit odd.

The negative side could in theory would like a constant current source and thus the amplifier like a class A type.
For this the link from R558 to VR307 should however be on the other side or the diode.

the still open end of C339 is likely going to one of the gate controlls.
 
The following users thanked this post: DB4UCH

Online DB4UCHTopic starter

  • Newbie
  • Posts: 7
  • Country: de
Except for CR303 the positive side now makes some sense. They may be something the other way around or so.

VR305 and likely also C322 should be more to the left as part of the gate control. C322 is however a bit odd.

The negative side could in theory would like a constant current source and thus the amplifier like a class A type.
For this the link from R558 to VR307 should however be on the other side or the diode.

the still open end of C339 is likely going to one of the gate controlls.

You are right, I drew the link between R558 and VR307 wrong, in reality it connects to the cathode of VR307. I triple checked the circuitry connected to the gate of Q311 and to CR303 and it seems to be the way I drew it in the diagram. C339 actually only connects to the anode of the LED in U310.

Greetings,

Simon

PS.
During reverse engineering I found that one of the solid-state relays (K302) was faulty. (ca. 25 Ohms between Pin 4 and 6). Interestingly, all of those solid-state relays (K302,303,304) are a different model from what is specified in the service manual. (LH1540 vs the specified LH1056AT). Judging by the date codes and soldering, those are most likely original to the instrument. And those can still be ordered from Digikey and the likes! So, I will replace all of them. (Another one is suspect as well ca. 500kOhm vs. the specified at least 500MOhm Roff)
 

Online DB4UCHTopic starter

  • Newbie
  • Posts: 7
  • Country: de
An update from my repair attempt:
After finding another broken part, this time a SMD resistor (open circuit) & installing all the new active components, the source now lives! It accurately outputs stable DC voltages in both ranges, so between ±100V and ±1000V.  ;D
(Neither the source nor the meter is calibrated at this moment (or were in the last 20 years), so I'm not really able to accurately judge, if it meets its specs)
I used the 2SK1412 MOSFETs from the link voltsandjolts had sent and they seem to be genuine. (Behave similar in DC and regarding their capacitances)
Attached is an image of the Electrometer measuring the output from its own voltage source.
I will now run the source for some period of time to "burn it in" and to see if it fails again.
Thanks for all of your help!

Greetings,
Simon

P.S.
What are your opinions on recapping the whole unit. (I think Keithley 2000s from this era suffer from leaking caps) Does anyone have experience with leaky caps in those electrometers or are they generally not that problematic?
 
The following users thanked this post: edavid, voltsandjolts

Offline voltsandjolts

  • Supporter
  • ****
  • Posts: 2547
  • Country: gb
Thanks for the update on your successful repair. Very nice unit, with perfect looking VFD display too.

IMHO recapping a 20 year old, yet valuable unit like this would be sensible, for both electrolytics and tantalums.
 
The following users thanked this post: DB4UCH


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf