Should I return a Keithley 238?

[leighcorrigall]

Hi Friends,

I am new to this forum, so please forgive me if I am not completely familiar with the ways of posting. The repair thread seemed appropriate for this topic.

I purchased a Keithley 238 on eBay to do some research on low current applications, thinking that it was in good working condition and the price was right (~750 USD).

My concern began when I received a Keithley 238 in the mail today. The legs were missing and some of the screws as well. Big deal, right? On closer inspection of the enclosure, I noticed contamination on the inner surface of the case near the multitap transformer.


I am not an electrical engineer by training, but I know what looks suspicious as I have had experience repairing other Keithleys before. There seems to be contamination from either a poor repair or a hot component that melted some plastic. The capacitor adjacent to the LM323K (see picture) has the same contamination as the board and the case above. The capacitor does not look damaged, but there is some contamination facing the heat sink. I cannot see where the source of this contamination is from. Does anyone have an idea? Maybe the transformer wires have been slow roasted over time by the LM323K heat sink.


From the other side of the circuit board, there is a blank silicone board that might act as an insulator to prevent shorting. It seems out of place and kind of lazy... Would this be original? I ask becuase it does not match the other boards. This area is where I would expect to see signs of modification or maybe some damage on the bottom side since the transformer wires are in the way on the other side.


I removed this shield by taking some screws out of the boards and slipping the sheet away from the support rails. I see that there are post-manufacting solding jobs around the location of where I first saw the contamination on the other side.


The manual specifies that there are three variations of this board: 236 [default], 237 [high voltage], 238 [high amperage]. Would these solder jobs be associated with adapting the Keithley 236 to a 238 board in the factory or has someone attempted to repair this instrument? I am aware that there are forum members who have converted there 236 to a 237 before and understand that the 238 is a modification of the 236. Could the contamination be consider normal? The instrument in my posession is marked as a 238, so I don't think someone has unofficially upgraded the device.

I am considering returning this device, but I want to know if I am overthinking things here. The instrument powers on and apart from the noisey fan, it seems to be in good working order. I do not have a 3 lug triaxial test lead for it yet because I did not want to pay 325 USD until I could inspect the instrument. Was 750 USD a good price for a Keithley in this condition? I want to be fair to the seller.

Thank you for reading my post and I appreciate any advice or thoughts that you may provide as a community.

Regards.

[E-Design]

I cant really comment on whether or not you should return it.. thats an issue of whether or not what you purchased was fairly and clearly represented by the seller.

However, I can say that somebody attempted repair.
 I can say for sure that the Keithley factory would not do a solder job as shown in the pictures.

So the contamination came from the previous capacitor? Did you discover where it came from? It obviously wasnt cleaned up.
Maybe all thats needed is a cleanup job.

[evac]

The clear fibreglass board on the bottom is original, I have the same on my 237.

The flux residue on the bottom of the PCB is likely a sign of a repair, looking at what those solder joints connect to would give you an indication of what was replaced and potentially what the problem was.
The soldering work also seems quite poor, and the fact that they didn't clean the board suggests it wasn't an official repair.
Looking at the location of those solder joints, could it be the transformer that was replaced?

I think the digital board is different between the 238 and 236/237, so not sure if a conversion is possible.

In any case, the affected area is the power supply section, I would be more concerned if I would see such poor repair work near the analog sections.

If the unit would be in operating properly and in spec, I would be okay with such condition, assuming also the rest of the boards didn't show signs of a more extensive (and poorly done) repair.

[leighcorrigall]

Hi E-deisgn and evac.

Thank you for your input. I believe that both of you are correct about the repair job. Without taking the board out of the enclosure, I am guessing that some of the components were capacitors. The transformer was not replaced in this case, but the contamination is very close to it.

Apart from the poor repair job, nothing else seems modified. The seller advertised the unit with the following message: "Unit was in working condition prior to being removed from service." A calibration sticker has been placed over the reset button, but I do not know if it is official. Any way of telling?

Let me know if anyone wants additional pictures for reference to benefit the community. I don't see a lot of Keithley 238 representation out there.

[garrettm]

I'd at least show the seller pictures of the poor repair and see if he/she would be willing to give you a partial refund. Clearly this unit isn't worth as much as the seller thought it should be. In my experience, most sellers are willing to offer a partial refund to avoid paying return shipping on a "not as described" case. So it's worth a try.

Though you really need to test the unit before proceeding any further. I would recommend using only triax when testing, as these units are very susceptible to external noise. I used some Trompeter triax to BNC adapters (E1 wiring) with Pomona low triboelectric BNC cables and kept getting weird noise spikes / random jumps in the output. Once my Keithley triax arrived, all that all went away.

[leighcorrigall]

Hi Garrettm,

You are correct. I plan on negotiating with the seller. Considering the enclosure panels look like they have been warn extensively, he may already know that the unit is modified. I am in the process of carefully photographing the interior. 

For now, I have ordered a Keithley 6171 adapter (2 lug to 3 lug converter) for my Keithley 6011 (2 lug triax to allegator test leads) so that I can perform measurements with my Keithley 2000. For now, I do not want the Keithley 238 anywhere near my precious Keithley 6485 ammeter. In the process of purchasing the Keithley 6171, I was debating whether to purchase a 237-ALG-2, but I couldn't justify the extra 200 USD.

Regrads.

[leighcorrigall]

Looking at the location of those solder joints, could it be the transformer that was replaced?

Now that I have had time to look at the pictures, I think it is the transformer taps. Appologies, evac! 

The flux on the bottom of the board correspond entirely with the coloured wiretaps. I suppose I should lift the transformer and clean around the board with methonal, kimwhipes, and an air can. I wonder how the capacitor got dirty then. 

This could be a very easy fix considering a took the entire instrument apart and know that no other component has been replaced.

Cheers.

[evac]

I don't have a lot of experience with transformer failures, but if something would have caused one of the windings to short internally and overheat, it could be possible that some burned enamel/epoxy would spew out.
The path of least resistance for something to spew out would probably be the openings where the wires are, at the top and bottom of the transformer, and this would seem to match where the stains are.

The question would then be why was the transformer replaced.
Given that there are no signs of repair on other components, I guess we can rule out other sections of the unit failing and shorting the transformer.
Looking at your flag and currency of purchase, I guess the unit came from 110V mains environment so unlikely to have been overvoltage due to wrong line select.
Perhaps it was just an isolated failure due to ageing?

Of course, what really matters is if the unit performs as expected, so once your adapter arrives, I guess you will have the answer to whether the unit should be returned.

Even if the unit works as expected, I would follow garrettm's advice to raise the issue to the seller.
I've had the similar experiences where the item was overall working, but had some damage/issue that was not described, and in some cases the seller was willing to give a partial refund.

Regarding the calibration sticker, there are likely plenty of calibration labs that would be able to calibrate such an SMU, not just Keithley, so even a non-official calibration would be a good sign.
Typically, there would be stickers not just on the calibration enable switch, but also on the top/bottom panels such that the unit cannot be opened with without breaking one of the calibration seals.

You should also find a sticker with the calibration date, from the calibration lab.
In some cases, the sticker might also include a reference number and depending on the lab they may be able to provide you a copy of the calibration report.

[leighcorrigall]

Well, this is turning into a legitimate repair log now...


I ended up cleaning the back of the digital board where the replacement transformer taps were resoldered. This was done with 99 % isopropanol and q-tips because I don't have a way of acquiring methonal for the time being thanks to COVID. The way I clean is I start by dissolving the solidified flux and then I use a dry q-tip to clean towards the flux, starting at the perimeter of the liquid where the contamination is the least. The contamination doesn't spead far this way and I can reduce the amount of q-tips used as I have a recycle pile, in order of cleaniness. The process is repeated until I do not see residue in hard reflected light coming from a lamp. It is a huge improvement over what it looked like when I received the hardware. As it is the digital side of business, I doubt the flux will make an impact on performance, but this is a matter of pride!


The 80mm axial fan also needs attention because it makes clunky noises as if the axal is rubbing. Turns out the 126LF by ETRI is a sleeve type bearing, which breaks down much faster than a ball bearing type and can only be used in a horizontal position. They have a lifespan around 6 years of continuous operation. Sometimes you can fix these by opening them up and lubricating the shaft. This is done by rotating a circular plate on the opposite side of the sticker, removing an end plug that applies pressure to the shaft, removing a retaining ring and some washers. Unfortunately, the lubrication did not fix my problem. 


With some searching, I managed to find the original datasheet for the fan. Obtaining a new one is difficult, so I decided to go with a ball bearing type alternative. The OA80AP-11-1TB from Orion seems to be a good candidate, although I do not know if it has a threaded grounding hole. You can either reuse the original power connector or purchase a ‎3-position connector (0009503031‎) from Molex with chrimp interconnectors (0008500108‎). I took these ideas from a blog: https://www.nicolas.tv/blog/keithley-236-repair.
 126LF-2182-000.pdf (97.88 kB - downloaded 74 times.) OA80AP-11-1TB.pdf (204.45 kB - downloaded 49 times.)

The lithium ion battery from Duracell seems to be functioning as it reads 3 volts when using a DMM. The model number for this battery is DL2450, but you can find one under the name of CR2450 as well.


Another good thing about this instrument has the latest firmware package (I think). Is this a G version?


While preparing to clean the other side the the digital board, I found another problem that I still need to look into. There is a damaged component between the 'int dt 7130 SA100P delta delta U9502P' (Keithley designates this part as U12 in the KE 236 service manual) and the 'M45AB LM323K steel' (VR1). The component in question looks to be a capacitor based on schematics I found in the Keithley 236 service manual (C35). How would it break in this way? For some reason this component is not identified as a replacable part in the Keithley 238 service manual, so I am uncertain about the exact specifications of the component.


Can I assume that the capacitor for the Keithley 238 is the same as the C35 found in the Keithley 236 service manual? 
Circuit Design   Description                               Keithley Part Number
C35                  CAP,.OluF,20%,50V,CERAMIC   C-365-.01

Also, does anyone have schematics for the Keithley 238? I don't have them in the service manual version I own.

[leighcorrigall]

Due to the attachment limitations, I was unable to post the schematics. Here are references for the Keithley 236/237 model.

[garrettm]

The digital board is looking great! Good job.

The capacitor looks like a 10nF or 100nF axial decoupling capacitor for the digital ICs. Not a terribly critical component but should be replaced.

It looks like someone either dropped a screwdriver on it when they replaced the transformer or they were using it as a reference point for measuring voltage and were very rough with the alligator clip. Either way, a very careless act on behalf of the previous owner and/or repairman.

This is why I always pull used equipment apart and thoroughly inspect every inch before turning it on. I remember opening up an HP 6186C HV current source only to find the AC mains was shorted via the neon bulb pulled out of its holder. Good thing I didn't turn it on! Looking further I found the current sense resistor had been cut. After fixing all that nonsense, I turned fired it up and it worked just fine. Goes to show you can't trust anyone who doesn't show the equipment turned on and working.

[garrettm]

Just noticed that your fan was an ETRI. I have two 237s and both use NMB. Maybe yours was replaced previously? I can open up one of them and take a picture of the fan if you would like.

[leighcorrigall]

Just noticed that your fan was an ETRI. I have two 237s and both use NMB. Maybe yours was replaced previously? I can open up one of them and take a picture of the fan if you would like.

Please do. I have not ordered new parts yet and I want to keep myself open for other candidates. Are your fans grounded with a grounding screw like mine?

Come to think of it, the fan might have been replaced seeing the crimp connectors are different colors.

Oddly enough, there are 3 positions on the board termination but only two are used. The centre pin is not connected to ground on my instrument.

[evac]

In my 237, the fan is an NMB 3115PS-12W-B30, apparently NMB has updated the part number to 08038PB-A1L-AA-00.

According to the datasheet it will do 2700RPM/26.5CFM/33dB@50Hz or 3200RPM/32CFM/38dB@60Hz.

I will also need to replace mine as the bearings are worn out, but I think I will try to find something a little bit more silent with not too different air flow.

[leighcorrigall]

Thanks for the info. The specifications you stated are similar to what the ETRI is.

Interesting that you propose a more consistent fan. Good point! 

[Twoflower]

By the way here some pictures from the inside:
https://www.eevblog.com/forum/testgear/keitley-238-restoration/

In the first post, first picture it looks like the same (grounded) fan, while in the last post there's a different fan. EDIT: The other fan type is s grounded as well but at a different place.

I was surprised about the used ground-lugs. The Crimping looks cheap but they seem to be original.

[garrettm]

The ground lead looks identical to my units, red and blue crimped ring terminals as well as orientation with my other SMU. The other one, which I took photos of, used a different orientation for the fan with a longer ground lead.

As evac posted earlier, the NMB model is 3115PS-12W-B30. These are in both of my units.

Now that I look closer, your fan actually seems to be factory original, the mounting orientation and ground lead matches my other SMU and the two pin connector looks original. So maybe Keithley used two different models for the fan.

At any rate, I agree with evac. If you can find a quieter fan with the same CFM for your line frequency you should probably be okay. My fans are reasonably quiet, but if either of you find a better replacement let me know.

[leighcorrigall]

Everything in my Keithley 238 looks original with the exception of the transformer. Whoever replaced it was lazy, but at least they replaced the transformer with whatever was described in the service manual (TR-276). To avoid another future fire hazard, I pulled the wire taps away from the neighbouring LM323K heat sink with some cable ties.


A feature of my Keithley is that it has fake leather coated shields, which I find halarious. I guess that the older versions of this instrument use a white coating only.


I did some browsing and came up with the attached spreadsheet as a list of candidates for an axial fan replacement. There isn't much difference between each model unfortunately, but if you are willing to spend a lot more money (44.40 USD) you can get the 8500N by ebm-papst Inc. It has a flow rate of 1.02 m^3/min [36 CFM] at 34 dB with sintec bearings, which should last longer than sleeve bearings. I will likely use this one as I hate noise and for what I am using it for, it will be running often. Too bad this instrument doesn't have better heat management.
 axial fan candidates.xlsx (11.35 kB - downloaded 56 times.)

[Qw3rtzuiop]

My 236 had a Papst 8830N fan installed. But the fan isn't silent either but there was some corrosion at the fan. Maybe the bearings are already damaged.


Edit: Removed the image. I mixed up some files on my HDD. The image wasn't from me.

[leighcorrigall]

Anyone else know what is inscribed on this capacitor label? Without the Keithley 238 service manual, I don't have a way to varify the specifications are of this part. My best guesss, based on the 236/237 service manual is that it is a ceramic 0.01 uF, 20 %, 50 V located at C35 connected to U12 (7130SA100P). I want to say that the part can be replaced by the SA105E103MAR‎ byAVX Corporation.

I think it reads:
"ASE
1.03

MLD
403"

Any help would be appreciated

[leighcorrigall]

My 236 had a Papst 8830N fan installed.

The specifications for your fan are much lower than original components (0.9 m^3/min [32 CFM]). You might want to upgrade to an even noisier version! 

Before considering an upgrade, you coud try and grease up the shaft or bearings. Sometimes it can give back a little life to it.

[Qw3rtzuiop]

I think it reads:
"ASE
1.03

103 is 10 nF (same as in the 236/237). I dont know why it should be a different value on the 238. Its just a decoupling cap. The exact value isnt mission critical.

[leighcorrigall]

103 is 10 nF (same as in the 236/237). I dont know why it should be a different value on the 238. Its just a decoupling cap. The exact value isnt mission critical.

Thank you Qw3rtzuiop for your input. My guess is that there is nothing different about the 236/237/238 digital board design, so I assumed that the capacitors are the same. I don't know much about capacitors or the functionality of the capacitor for this application. My educational background is in chemical/materials/nuclear engineering, so when it comes to electronics I only know the basics. 

[Howardlong]

I had my 238 on the bench preparing to do a video, so I popped it open. It's a 10nF.

I don;t have the splattering over the transformer, I wonder if it's lamination lacquer, applied after installation? Perhaps the transformer had loose laminations.

Edit: Fan in mine is NMB 3115PS-12W-B30.

[Howardlong]

Fan & transformer pics inside my 238