Keithley 2015THD - Re-Capping

[yngndrw]

Sorry for yet another Keithley 2*** re-capping thread, but I have a few questions which I haven't seen answered in other threads. (Or I've missed the answers !)

So I have purchased a working Keithley 2015THD on Ebay and it was last calibrated Oct 2010 with a cal count of 6, which suggests that it's about 12 years old. Its version is: B15 A02, whatever that means !

I should stress that I really don't want to have to get it re-calibrated so whatever I do mustn't affect that. The Keithley manual stresses that it should be re-calibrated even after just removing the boards, although I'm sure mine is far enough out of calibration for that not to be an issue. Either way, I think I should be using gloves and making sure that I clean off any flux residue after I'm done.

So my questions:
1) What benefit is there to the later firmware versions ?
2) What's the latest firmware that's available to me without affecting my calibration data ?
3) Does anyone have a full part-number list (Ideally from Farnell or some other supplier who deals with the UK) for a full set of 2015 capacitors ? I'm looking for replacements which are an exact fit, I don't want to have them mounted on wires. (I've seen some for the 2000, but none of these were low ESD which seems incorrect and I believe there are more in the 2015 - Also I've heard that the list in the service manual is also incorrect ?)
4) What's the correct flux-cleaning method ? I never usually clean flux after soldering but I think I should be in this case ! I believe I should just be using Isopropyl Alcohol and some lint-free swabs - Is this correct and do I need to use anything after this ?
https://www.amazon.co.uk/500mL-Pharmaceutical-Isopropyl-Alcohol-Isopropanol/dp/B00WLAS9W8/ref=sr_1_1?ie=UTF8&qid=1472257138&sr=8-1&keywords=Isopropyl+alcohol+pure
https://www.amazon.co.uk/KTC-Computer-Technology-Cleaning-printers/dp/B01KS97MG2/ref=sr_1_8?ie=UTF8&qid=1472253486&sr=8-8&keywords=lint-free+swab
5) Any other advice ?

From what I can tell from the manual (http://exodus.poly.edu/~kurt/manuals/manuals/Keithley/KEI%202015%20Service.pdf), the following are required: (Sadly it doesn't give an information about pitch or ESR - Also doesn't mention anything about C260 and C261 which are populated on some, will have to open it up and check but wanted to order the replacements before removing any boards I can confirm that C260/261 is not populated in my unit - Part numbers and sizes below are from the datasheets)
Main board

• C104 - 100uF 63V 20% [Confirmed value in my unit - 85C / Nichicon VR(M) - P/N: Nichicon UVR1J101MPD / Size: 10x12.5 (Could replace with 105C version but this has a lower ripple current of 215mArms vs 300mArms: UVZ1J101MPD, no idea what effect this would have)]
• C131/148 - 2200uF 35V 20% [Confirmed value in my unit - 105C / Nichicon VZ(M) - P/N: Nichicon UVZ1V222MHD / Size: 16x25]
• C146 - 3300uF 16V 20% [Value in my unit is 2200uF just like in another thread - 105C / Nichicon VZ(M) - P/N: Nichicon UVZ1C222MHD / Size: 12.5x20 (The 3300uF version is 5mm taller)]
• C156 - 6800uF 16V -20+100% [Confirmed value in my unit - 85C / Nichicon VX(M) - Does not seem to exist in the datasheet]

THD board

• C326 - 100uF 16V 20% [This is an SMD part, will not replace]
• C346/347 - 47uF 100V 20% [Confirmed value in my unit - 105C / Nichicon PR(M) / Size: 10x16]
• C562/565 - 1000uF 50V 20% [Confirmed value in my unit - 105C / Nichicon VZ(M) / Size: 12.5x25]
• C568 - 2200uF 25V -20+100% [Confirmed value in my unit - 85C / Nichicon VR(M) / Size: 16x25]

For C156 it looks like a VR/VZ replacement of value 10000uF 16V or 6800uF 25V would be the same physical size and hopefully the same lead spacing. I'm thinking the higher voltage rating is better so that the value is the same. I doubt the ripple current will be an issue due to increasing the rating, therefore the 105C VZ is probably best. (P/N: Nichicon UVZ1E682MHD / Size: 18x35.5) Sadly Farnell don't stock the 105C VZ version so I might have to get the VR version which they do stock: Nichicon UVR1E682MHD I've decided to keep it as close to the original fitted part as possible, which means 85C where they had 85C capacitors and hence this for C156: Nichicon UVR1E682MHD Where I have to make an adjustment due to a part which no-longer exists, I'll go for an increased voltage range in order to get the right physical size.

Nichicon Datasheets:

• Nichicon VR: http://www.nichicon.co.jp/english/products/pdfs/e-vr.pdf
• Nichicon VZ: http://www.nichicon.co.jp/english/products/pdfs/e-vz.pdf
• Nichicon VX: http://www.techno-star.biz/capacitors/uvx.pdf (Nichicon don't seem to have the datasheet on their site)

I don't have an precision references so the most I can do in terms of calibration and verification is the nulling.

I also have a HP 6032A which is also likely in need of re-capping, but that's another thread.

Thanks.


Update

After opening it up, it looks like there are a few parts which aren't listed in the manual at all. (C279, C280, C350) There are a couple where the series doesn't exist anymore (C346, C347) and therefore I've substituted a similar part that's slightly shorter but the same diameter and ratings. Additionally C326 is an SMD capacitor so I won't replace that at all. (I don't fancy using hot air around a precision piece of kit - That and I'm not 100% sure about the spec of the existing part.) The table below represents what I'll be ordering and replacing - Any comments ?


Parts List

Component Reference	Manual Specification	Original Part	New Part Number	Quantity
Main Board
C104	100uF 63V 20%	Nichicon VR 100uF 63V 20% 85C 10x12.5	Nichicon UVR1J101MPD (100uF 63V 85C)	1
C131,C148	2200uF 35V 20%	Nichicon VZ 2200uF 35V 20% 105C 16x25	Nichicon UVZ1V222MHD (2200uF 35V 105C)	2
C146	3300uF 16V 20%	Nichicon VZ 2200uF 16V 20% 105C 12.5x20	Nichicon UVZ1C222MHD (2200uF 16V 105C)	1
C156	6800uF 16V -20+100%	Nichicon VX 6800uF 16V 20% 85C 18x35.5	Nichicon UVR1E682MHD (6800uF 25V 85C)	1
C279,C280	Not Listed	Nichicon VR 47uF 50V 20% 85C 6.3x11	Nichicon UVR1H470MED (47uF 50V 85C)	2
THD Board
C326	100uF 16V 20%	SMD 100uF 16V 5x5 (Markings: "ON 100 16F")	Will Not Replace	1
C346,C347	47uF 100V 20%	Nichicon PR 47uF 100V 20% 105C 10x16	Nichicon UPS2A470MPD (47uF 100V 105C)	2
C350	Not Listed	Nichicon VZ 4700uF 16V 20% 105C 16x25	Nichicon UVZ1C472MHD (4700uF 16V 105C)	1
C562,C565	1000uF 50V 20%	Nichicon VZ 1000uF 50V 20% 105C 12.5x25	Nichicon UVZ1H102MHD (1000uF 50V 105C)	2
C568	2200uF 25V -20+100%	Nichicon VR 2200uF 35V 20% 85C 16x25	Nichicon UVR1V222MHD (2200uF 35V 85C)	1

[singapol]

I don't have an precision references so the most I can do in terms of calibration and verification is the nulling.

I don't have a Keithley so I don't know what this nulling you refer to but calibrating a 6.5 digit DMM is not trivial. The precision references for 6.5 digit DMM needs more thought than those available to calibrate 3.5 or 4.5 digit DMMs. At best I can confidently trust 2 or 3 places after the decimal without proper calibration. It's best to know something about metrology ( the science of measurement) which is one of the forums here.

Have a read of this- Although it's old but the principles are still very much in use.

Calibration- Philosophy in practice- Fluke:

http://www.introni.it/pdf/Fluke%20-%20Calibration%20Philosophy%20in%20practice.pdf

[yngndrw]

Perhaps nulling was the wrong word to use, by that I just meant the open / short circuit calibration step which I thought you could do without doing the rest of the steps but looking through the manual it seems that you have to do DC voltage, current and resistance calibration with it which I'm certainly not equipped to do. I think I'm just going to have to be careful to not affect the current calibration status and if I mess it up I'll have to send it off.

Thanks for the link, looks like an interesting read but it might take me a little while to get through it !

[yngndrw]

A DigiKey BOM if anyone's interested: https://www.digikey.co.uk/classic/RegisteredUser/BOMBillOfMaterials.aspx?path=1&exist=1&id=4208854 (Hopefully you can all access that, I think you'll need to login though)

[azer]

Thanks, this was rather convenient.

While you have the device apart, could you check the firmware revision numbers printed on the eeproms?
There is a lot of information on different firmwares and upgrading them for the top board, but I have not seen information on if there are newer versions of the ones on the bottom board.

Top:
U157 803B15 - odd
U156 804B15 - even

Bottom:
U330 805A09
U315 806A01
U346 808A02
U326 807A01

[yngndrw]

It's all back together now but I'll be ordering the parts tonight and will check for you once I receive them all. (In a week or so I expect)

[yngndrw]

The parts have arrived today which is pretty quick considering one capacitor was back-ordered and it was shipped to the UK, top marks to DigiKey on that one. I'll be doing it this weekend but before I start, does anyone know how to unclip the front/back rod from the switch on the main board ? I wasn't sure how it comes off when I last opened it up and don't want to break it.

[Macbeth]

Perhaps nulling was the wrong word to use, by that I just meant the open / short circuit calibration step which I thought you could do without doing the rest of the steps but looking through the manual it seems that you have to do DC voltage, current and resistance calibration with it which I'm certainly not equipped to do. I think I'm just going to have to be careful to not affect the current calibration status and if I mess it up I'll have to send it off.

Thanks for the link, looks like an interesting read but it might take me a little while to get through it !

If you do the calibration/adjustment using the front panel then yes you do have to do the whole DC procedure, including volts, ohms and amps.

However if you adjust it using SCPI (GPIB or RS232) then you can stop at any point and save the cal.

In practice that means I have done the 4 wire short, open circuit and the + and -10 volts on mine. My cheap AD584L based 10V ref is usually 10.00275 on a good day and holds quite well (I have another Keithley 2000 to monitor it). You can program the actual voltage including any small offset from 10.00000V and the Keithley will compensate.

One thing I would do if messing with the calibration is make a backup of your EEPROM before proceeding. Get yourself a TL866! 

[yngndrw]

Thanks that's really useful to know, I didn't know there were additional options when using SCPI.

[yngndrw]

While you have the device apart, could you check the firmware revision numbers printed on the eeproms?
There is a lot of information on different firmwares and upgrading them for the top board, but I have not seen information on if there are newer versions of the ones on the bottom board.

Top:
U157 803B15 - odd
U156 804B15 - even

Bottom:
U330 805A09
U315 806A01
U346 808A02
U326 807A01

I've just opened it up again tonight, mine are the same as yours - Here's the numbers: (Formatted as they are on the stickers)

Top:
U157 2015-803B15 - Odd
U156 2015804B15 - Even (This one doesn't have a sticker - the number was laser-etched into the IC and I had to use a microscope to read it ! It's an Atmel AT27C020 12JC 0420)

Bottom:
U330 2015 805 A09
U315 2015 806A01
U346 2015 808A02
U326 2015 807A01

Hope that helps.

[yngndrw]

Completed the re-capping last night and it all works - The old capacitors had no signs of leaking which was good. (Dated 2004) As mentioned I left C326 as it was as I don't know what make / series it is, but I've contacted Keithley to see if they can tell me exactly what part it is.

I noticed that after cleaning the flux off with IPA and a foam swab and then drying the IPA with my hot air gun, the IPA left a sort of watermark around the area. I left this as is seems as it was only around the electrolytic capacitors so hopefully it won't affect anything, but does anyone have any advise on board cleaning ? How can I get the board as spotless as when it was made ?

[Kleinstein]

The power supply area, where the big filter caps are is not critical to leakage. One could reduce the residue with another wipe. It just takes several passes to get all off. But I would not really mind.

One can test leakage currents (across the inputs and to ground) - so if everything is good there is not need for further cleaning.

[yngndrw]

Ah okay, I guess it's just the diluted flux that's causing the residue. Thanks, I'll just leave it.

[yngndrw]

Quick update, Keithley got back to me regarding C326 - They believe it to be this part:
http://www.illinoiscapacitor.com/ic_search/lytics_products_detail.aspx?icpartnumber=20189

I've not replaced mine yet and I'm not sure if I will do as I can't seem to find a place to get them. Even DigiKey only have them in massive quantities, maybe I'll have a look around for an equivalent when I get a chance.

[usagi]

recapped my 2015 thd today. sticker on inside of the display board says OCT 25 2000.

mainboard:
c146 - mine was 3300uF 16v.
c279 - does not exist
c280 - does not exist

text on the board reads:
KEITHLEY © 1997
2015-102-02C
MADE IN U.S.A.

[macboy]

Completed the re-capping last night and it all works - The old capacitors had no signs of leaking which was good. (Dated 2004) As mentioned I left C326 as it was as I don't know what make / series it is, but I've contacted Keithley to see if they can tell me exactly what part it is.

I noticed that after cleaning the flux off with IPA and a foam swab and then drying the IPA with my hot air gun, the IPA left a sort of watermark around the area. I left this as is seems as it was only around the electrolytic capacitors so hopefully it won't affect anything, but does anyone have any advise on board cleaning ? How can I get the board as spotless as when it was made ?

You can blot the liquid IPA off the board instead of allowing it to evaporate; paper towel / kitchen roll works well for that. Obviously, what the IPA dissolved (mostly flux) is now in solution, and if you allow it to evaporate, that dissolved stuff will be left behind. You can also flush with plenty of fresh IPA from a squirt bottle to help rinse the dirty IPA off. I also have good luck with rinsing then blasting with compressed air, which tends to remove most of the IPA rather than evaporating it. Don't use "canned air", that stuff will always leave residue unless you use an expensive residue free type. I use a regular oil-less compressor with a coalescing filter (0.01 micron) and desiccant to provide very clean, very dry air.

[Bud]

I am having good results with Spray 9 marine cleaner. It was recommended in some other thread and worked very well for my Kester no clean flux. IPA leaves white residue on the boards, Spray 9 came out spotless. It seems to be available in the UK.