EEVblog #478 - Keithley 2015 THD Multimeter Teardown

[EEVblog]

Teardown Tuesday
Inside the Keithley 2015 THD 6.5 digit bench multimeter.
Not just a multimeter, it does low noise signal generation and automated audio THD system measurement.

[ChadSeibert]

Excellent teardown as always, Dave! Do you know whether there is a niche application where a meter with all this functionality would be useful?

[EEVblog]

Excellent teardown as always, Dave! Do you know whether there is a niche application where a meter with all this functionality would be useful?

Production testing of any device that has audio/voice capability.
It's not designed for high end hi-fi type testing, but phones, radios, or any other such voice channel quality sound is required it would be ideal.

The datasheet for example lists applications like
• Wireless communication device audio quality testing
• Component linearity testing
• Lighting and ballast THD limit conformance testing
• Telephone and automotive speaker testing

[chickenHeadKnob]

Great, just great, another quality multimeter/o'sillyscope that is going to get the Dave endorsement price inflation.

Yes I did get moist while you were performing the tear-down.

The one question I had was does this meter have one of those accursed battery backed RAM for CAL. data. You go fast sometimes and I thought maybe I missed it. So I checked on the internets. Assuming model 2000 and 2015 share the same main board, the answer is no.
The 2000 model repair manual lists U136 and calls it NVram in the block diagram and eprom in the parts list, but it is just a jelly bean 24lc16b 16K bit EEprom.

[carloscuev]

Great, just great, another quality multimeter/o'sillyscope that is going to get the Dave endorsement price inflation.

Just as the FE-5680A Rubidium Atomic Frequency Standard Oscillator on eBay, you could get them for $50 before dave's first videos about them, now ou're lucky if you find them at $80

[nukie]

I know you mention cap, but most LM399 and LTZ1000 references have wobbly cap, the pins however are securely mounted on the sockets.

Good choice with the Keithley I prefer the Keithley 2000 over HP34401A. The Keithley 2015 is ALOT MORE MEAT than the 2000 model.

[ChadSeibert]

Great, just great, another quality multimeter/o'sillyscope that is going to get the Dave endorsement price inflation.

Just as the FE-5680A Rubidium Atomic Frequency Standard Oscillator on eBay, you could get them for $50 before dave's first videos about them, now ou're lucky if you find them at $80

Eh, I got my $45 shipped three weeks after his review/teardown . Maybe this is just a bad time?

[EEVblog]

Hmm, I thought there was software to go with this meter that could do general purpose THD testing bode plots etc, but I can't seem to find anything, just some drivers.
Anyone?

[FrankBuss]

Interesting teardown. Looking forward to your test of the THD measurement. Once I've used an Audio Precision device for audio development (not my own, at a client site, too expensive for me) and measuring distortions of -100 dB and less was no problem at all.

[EEVblog]

Interesting teardown. Looking forward to your test of the THD measurement. Once I've used an Audio Precision device for audio development (not my own, at a client site, too expensive for me) and measuring distortions of -100 dB and less was no problem at all.

I used to have an Audio Precision in my lab until SWMBO found out it was worth 5 figures and I had to sell it 

[synapsis]

Long time viewer, first time poster. Thanks for all the education you're giving for free, this programmer appreciates it.

Here's a photo of the top board of my Keithley 2000 for comparison. Where the siggen board is on the 2015, mine has an option slot which I believe is for a multichannel scanning/switching array of thermocouples.

[cyr]

There seems to be a lot of these around now on ebay, most probably from some automated test rigs. Mine originally comes from Rohde & Schwarz according to the cal sticker (two T&M companies for the price of one).

I don't think I will ever use the THD function, but it's a very nice general purpose bench meter...

[EEVblog]

Here's a photo of the top board of my Keithley 2000 for comparison. Where the siggen board is on the 2015, mine has an option slot which I believe is for a multichannel scanning/switching array of thermocouples.

Thanks.
As I suspected, you can't have the scanner on the 2015. The 2015 is essentially a 2000 with the THD option instead of the scanner option.

[ChadSeibert]

Interesting teardown. Looking forward to your test of the THD measurement. Once I've used an Audio Precision device for audio development (not my own, at a client site, too expensive for me) and measuring distortions of -100 dB and less was no problem at all.

I used to have an Audio Precision in my lab until SWMBO found out it was worth 5 figures and I had to sell it 

You've got to be shitting me . Was it a down payment for this? http://design.kay-diamonds.com/diamonds/diamond/diam125977/

[KJDS]

Thanks for the teardown, I've not looked inside one before.

[bingo600]

. The Keithley 2015 is ALOT MORE MEAT than the 2000 model.

But only if you use THD , Keithley says it's a 2000 engine with THD added
http://forum.keithley.com/phpBB3/viewtopic.php?p=92667&sid=42116ccd061059f45ab4acb066b32cc2#p92667

/Bingo

[FrankBuss]

I used to have an Audio Precision in my lab until SWMBO found out it was worth 5 figures and I had to sell it 

And I guess what hurts most is that probably she was right, because I guess you don't measure professional audio equipment with balanced analog IOs or AES3 regularly, or need to create characteristics curves of audio amplifiers

[ChadSeibert]

I used to have an Audio Precision in my lab until SWMBO found out it was worth 5 figures and I had to sell it 

And I guess what hurts most is that probably she was right, because I guess you don't measure professional audio equipment with balanced analog IOs or AES3 regularly, or need to create characteristics curves of audio amplifiers

True, but Dave doesn't strike me as the kind of person that would be put off owning a piece of equipment just because he doesn't have a use for it. And imagine the teardown of that machine. 

[ChadSeibert]

It's $850 refurbished at valuetronics, but I'm sure it's not cal'ed. Does anyone know if they are a reputable vendor?

https://www.valuetronics.com/detail/Used-keithley-2015-p.cfm?gclid=CJaixePnzLcCFa5DMgod3j4Axg#.Ua8gtJwizSk

Thanks,
Chad

[plesa]

Hmm, I thought there was software to go with this meter that could do general purpose THD testing bode plots etc, but I can't seem to find anything, just some drivers.
Anyone?

Hi Dave,

I just found Labview program http://www.keithley.com/products/dcac/audioanalyzer/?path=2015/Downloads#1
In attachment you can find prebuilded application. Just install Labview runtime engine.

[EEVblog]

You've got to be shitting me .

Nope. Bought it (2700 series http://www.ap.com/products/2700) from my previous company that was shutting down for $500! Hardly used, and just freshly cal'd.
Interesting story actually:
They knew how much it was worth (we paid like $30K for it or something) and wanted to keep it for their overseas operation (they were just shutting down the oz office). But when I turned up on the day for some other stuff I was buying I was told to bring some cash just in case there was some leftover stuff I wanted. The guy in charge of clearing out the gear said the company had been dicking around about the unit, and technically he was ordered to clear everything, so asked if I wanted to make an offer (he didn't care, he was getting the arse anyway!), but I had to pay and take it right there and then as tomorrow it would likely be shipped out. Hell yeah, I had an extra $500 in my pocket and asked if that would do.
Had to go around and get a few signatures, which raised many eyebrows,  but I walked out with it with a big grin
There was likely someone at HQ who was really pissed!

Eventually sold it on ebay for $13K or so, my biggest score ever! 
Hard to justify keeping something like that.
This was before the blog. I had a lot of cool gear like that I scored before the blog, and sold most of it, so they never made it on video.

[bingo600]

It's $850 refurbished at valuetronics, but I'm sure it's not cal'ed. Does anyone know if they are a reputable vendor?

https://www.valuetronics.com/detail/Used-keithley-2015-p.cfm?gclid=CJaixePnzLcCFa5DMgod3j4Axg#.Ua8gtJwizSk

Thanks,
Chad

It's cheaper (I think) @KJDS (forumuser)
http://www.ebay.co.uk/itm/Keithley-2015THD-6-5-digit-bench-DMM-multimeter-with-harmonic-analysis-/350803796454

The one i bought from him in april was Cal'ed on May-2011 , meaning it's almost valid ....

Maybe give him a PM , there "might be" some forum discount.
https://www.eevblog.com/forum/blog/eevblog-478-keithley-2015-thd-multimeter-teardown/msg242087/#msg242087

/Bingo

[ChadSeibert]

It's $850 refurbished at valuetronics, but I'm sure it's not cal'ed. Does anyone know if they are a reputable vendor?

https://www.valuetronics.com/detail/Used-keithley-2015-p.cfm?gclid=CJaixePnzLcCFa5DMgod3j4Axg#.Ua8gtJwizSk

Thanks,
Chad

It's cheaper (I think) @KJDS (forumuser)
http://www.ebay.co.uk/itm/Keithley-2015THD-6-5-digit-bench-DMM-multimeter-with-harmonic-analysis-/350803796454

The one i bought from him in april was Cal'ed on May-2011 , meaning it's almost valid ....

Maybe give him a PM , there "might be" some forum discount.
https://www.eevblog.com/forum/blog/eevblog-478-keithley-2015-thd-multimeter-teardown/msg242087/#msg242087

/Bingo

I'll give it some thought, but shipping to the US is pretty expensive. Still cheaper, though.

This was before the blog. I had a lot of cool gear like that I scored before the blog, and sold most of it, so they never made it on video.

Interesting story, but it's probably best that you don't share everything you've sold. Some of us might not be able to take it .

[KJDS]

I'm happy to give forum members a discount on any of my products.

Shipping to the US isn't horrible, and I can remove the VAT which compensates substantially to the shipping cost. It's possible that you'll get stung with import taxes but I have no control of that.

[free_electron]

The opto isolation on the siggen is not so much as to split analog and digital but the fact that the digital is chassis referenced and the analog is floating ( notice the usage of fully isolated bnc connectors that have no metallic connection to chassis)

[Oracle]

really like this multimeter... Added to my future purchases list...

[lewis]

Nope. Bought it (2700 series http://www.ap.com/products/2700) from my previous company that was shutting down for $500! Hardly used, and just freshly cal'd.

Lucky effing Bastard!!!!!!!! AAARGH! Would have given my back teeth, both arms and all four legs for that!

[flodins]

Excellent, but what is this weirdo for? Some pick up antenna? Really strange. http://wstaw.org/m/2013/06/05/8955197658_87404b74c4_o.jpg

[bitwelder]

Dave, talking about Keithley, when do you plan to show/teardown the MegaohmPetaohm bridge (Keithley 515A) you got at a recent auction?
I'm really curious also to see some application for it.

[Dave]

"...you might be able to get one fairly cheap..."
Nope, not anymore.

I have seen a wobbly LM399 creamic case before. I'm quite sure they are mounting it like that on purpose. Does anyone know the reason why?

[Sine]

Hey guys, my first post here.

Carefull with those electrolythics next to the 15V regulators, I have bought three 2000's and on two of them these caps had gone leaky, destroying parts of the powersupply and analog circuitry.

One of them is beyond repair, an other I managed to fix. The accuracy of these meters is just ridiculous.


Inside shots :
http://www.gigawatts.nl/co/projects/Keithley_2000/overall_keith.JPG
http://www.gigawatts.nl/co/projects/Keithley_2000/leaky_keith.JPG

[free_electron]

Dag Sine :p

Yep. i've seen multiple 2000's with bad electrolytics. typically the MELF diodes ( ML4007)
die as well.

[plesa]

"...you might be able to get one fairly cheap..."
Nope, not anymore.

I have seen a wobbly LM399 creamic case before. I'm quite sure they are mounting it like that on purpose. Does anyone know the reason why?

It is not ceramic it it made from plastic. Keithley are using the same refference on K2015,K2000, and following K6485 and K6487.
I suppose that main reason for this capsule is the termal stability. In Keithley 2002 and Keithley 2002 where is used the LTZ1000A  is keithley custom made cap.
http://www.xdevs.com/attachments/download/208/2_3_ace27d41477113b.jpg

This can be useful to someone who score on 2015 or 2000 on ebay:
http://bardagjy.com/?p=1167

If anyone have access to schematic for 2002/2001 please let me know.  One 2002 still waiting under my bench for repair
The DC reading is OK and accurate, issue is in the ohm measurement and amps.

[Dave]

It is not ceramic it it made from plastic. Keithley are using the same refference on K2015,K2000, and following K6485 and K6487.
I suppose that main reason for this capsule is the termal stability. In Keithley 2002 and Keithley 2002 where is used the LTZ1000A  is keithley custom made cap.
http://www.xdevs.com/attachments/download/208/2_3_ace27d41477113b.jpg

You are right about that. I have checked the datasheet for the material of the outer capsule. It's made from a resin called Valox 420 (low thermal conductivity and extremely low electrical conductivity).
I understand why the references are inside a case, but what I'm trying to figure out is why are they mounting it with a large offset from the board? Why don't they just mount it flush with the board, so the case doesn't wobble around? Just by clarifying my original question, I answered myself. The leads of the reference certainly have a different thermal expansion coefficient than the resin case. If they mounted it flush with the board, the changes in temperature would cause mechanical stress on the reference, which kills the precision. Feel free to correct me, if I figured it wrong.

[fmaimon]

How worried should one be with this electromigration problem? I will probably buy another 6 1/2 meter in the next semester and it's hard to buy anything in here, including spares to fix any problems that may arrive.

https://www.eevblog.com/forum/projects/keithley-2000-gain-repair/
https://www.eevblog.com/forum/projects/keithley-2000-gain-repair/msg229818/#msg229818

[Galaxyrise]

How worried should one be with this electromigration problem? I will probably buy another 6 1/2 meter in the next semester and it's hard to buy anything in here, including spares to fix any problems that may arrive.

To fix mine, I ordered a Caddock 1k-9M resistor network.  I haven't done the repair yet, so I haven't updated that thread, but for the resistors that happened to die for me, that was a convenient fix.

But even better, it looks like the 2015 uses different resistors there!  Compare that part of the board from Dave's teardown. Edit: Of course, it could also be that newer rev 2000's work that way too; mine is from 99!

[TMM]

Looks like a nice unit but i'm not sure it's worth paying extra for a THD capability when you only get a -65dB floor. For the applications they list, the FFT capability on most cheapo scopes would probably do the job. I guess the attraction is ease of use and having the function gen built in. It would be interesting to see whether the performance is limited by the function gen or the ADC.

If you don't mind a PC solution, some cheap USB interfaces like the E-MU 0204 are capable of very impressive numbers (<-105dB THD, <-125dB noise floor) with some light modifications, which is about as good as you'll ever need for audio applications.




 

[don.r]

If you don't mind a PC solution, some cheap USB interfaces like the E-MU 0204 are capable of very impressive numbers (<-105dB THD, <-125dB noise floor) with some light modifications, which is about as good as you'll ever need for audio applications.

I'm interested in this approach. Could you provide a link for these mods? Is it something done on head-fi.org by any chance?

[TMM]

If you don't mind a PC solution, some cheap USB interfaces like the E-MU 0204 are capable of very impressive numbers (<-105dB THD, <-125dB noise floor) with some light modifications, which is about as good as you'll ever need for audio applications.

I'm interested in this approach. Could you provide a link for these mods? Is it something done on head-fi.org by any chance?

http://www.diyaudio.com/forums/equipment-tools/231426-emu0204-mods-fft-measurements.html

[don.r]

If you don't mind a PC solution, some cheap USB interfaces like the E-MU 0204 are capable of very impressive numbers (<-105dB THD, <-125dB noise floor) with some light modifications, which is about as good as you'll ever need for audio applications.

I'm interested in this approach. Could you provide a link for these mods? Is it something done on head-fi.org by any chance?

http://www.diyaudio.com/forums/equipment-tools/231426-emu0204-mods-fft-measurements.html

Excellent! Bookmarked.

[Watermelon]

What about W303? High voltage jumper link? Current shunt? Temperature measurement? Anyone?

[free_electron]

W303 : loop to connect ground crocodile clip of scope probe

TP310 standard SMD testpoints as made by keystone and others.

[don.r]

W303 : loop to connect ground crocodile clip of scope probe

TP310 standard SMD testpoints as made by keystone and others.

Always the sign of a quality instrument. You can see test point posts all the time in older HP equipment.