bunch of keithley 2015 on ebay for $350

[usagi]

seller has sold 107 of them so far. must be sitting on a warehouse.

http://www.ebay.com/itm/KEITHLEY-2015-6-1-2-Digit-THD-Audio-MultiMeter-fully-tested-w-warranty-/371582307445

[usagi]

oh yeah, if you buy one - replace the electrolytics in the power supply immediately. they are known to frequently go bad.

[cncjerry]

I know a lot of people bought them.  I checked them out thinking they might have value for Audio work but though I don't have one, that was about the limit of my interest.  If they were dialed-in, calibrated exactly, maybe they would interest me more.

I thought there was a report a while back on them.

[agdr]

Those guys (Parts Connexion) have been selling those for a year or two now.  Their website says  "..with a basic calibration performed..":

http://www.partsconnexion.com/78916a.html

http://www.partsconnexion.com/product26900.html

http://www.partsconnexion.com/78916b.html

I emailed him in the past asking about the details and he said they have an 8 digit calibrator.  I had already bought one elsewhere by then or I would have likely bought theirs upon learning that.   I didn't ask what the "refurbishing" included, and most specifically if it meant those electrolytics that leak over time have been replaced!

I'm going to be sending mine into Tek for calibration soon.  I got a quote last year.  I think it was something like $250 and it did include calibrating the audio THD functions.  Plus shipping which is likely to be another chunk of change in both directions.   So there is something to figure into the pricing, if he really is still calibrating them with a traceable (didn't ask if it was) calibrator.  In my view the whole point of a meter like this is accuracy.  If it isn't (recently!) calibrated then it is just a big metal doorstop. 

[alm]

I emailed him in the past asking about the details and he said they have an 8 digit calibrator.

I would be skeptical. Pretty sure the multifunction calibrators pretty much top out at 6.5 digit multimeters. Higher precision multimeters tend to need a more complicated calibration involving multiple instruments, not just a big box that says 'calibrator' on the front. And I am pretty sure none of the multifunction calibrators would calibrate the audio part.

I had already bought one elsewhere by then or I would have likely bought theirs upon learning that.   I didn't ask what the "refurbishing" included, and most specifically if it meant those electrolytics that leak over time have been replaced!

No experience with this particular seller, but in general on eBay you are lucky if 'refurbished' means that they wiped some dust off the front panel. The response from one seller when I asked what refurbishment entailed was that the unit looked nice without major scratches, so he marked it as refurbished. I wouldn't count on the caps being replaced, especially not with high quality ones.

In my view the whole point of a meter like this is accuracy.  If it isn't (recently!) calibrated then it is just a big metal doorstop. 

Well, there is also precision. When matching two resistors for a differential application, for example, you may want the resistors to be within 0.05% of each other, but if they are 995 Ohm or 1.005 kOhm is not going to affect your circuit in any meaningful way. In addition there is the thing that many of these instruments are quite stable, and tend to get more stable as they age. It would not surprise me if many of the instruments would still be very close if not within the 1 year specification in most ranges. Of course, if you need traceable calibration, then you should send it to a competent party. No idea about the stability of the THD stuff.

For audio applications I believe the distortion of the internal oscillator is pretty terrible. I believe people have tried using a better (lower distortion) oscillator with the 2015, but I do not know the results.

[RGK]

Description says they verify functions and zero with a special Keithley plug.  Not sure that really qualifies as a basic calibration, but better than nothing and probably good enough for many.  I've been considering one.

BK

[agdr]

No experience with this particular seller, but in general on eBay you are lucky if 'refurbished' means that they wiped some dust off the front panel.

Lol!  Very true! 

[tpowell1830]

Here's one for $250.

http://www.ebay.com/itm/Keithley-2015-6-12-Digit-THD-and-Audio-Analysis-Multimeter/361848524745?_trksid=p2047675.c100005.m1851&_trkparms=aid%3D222007%26algo%3DSIC.MBE%26ao%3D2%26asc%3D20131003132420%26meid%3D8c96265693ab4d5e83a623de525637b4%26pid%3D100005%26rk%3D3%26rkt%3D6%26sd%3D371582307445

[1audio]

I think these are left over from some cell phone mfr line. I got one. It has the remains of a China cal tag on it. I'm not sure how accurate it really is. The audio part really doesn't need any special calibration since its all software. In loopback it gets to .004%, from a low distortion source I get between .001% and .000% (a limitation of the software?).

As a bench meter its not bad. Its slower than my older 5 1/2 digit Fluke (8860a) and with all the functions it can take a moment to select what you want.

Does anyone have a 2015-p firmware? Would it work in a straight 2015? That opens up the harmonic analysis capability. Then we need someone to write a simple app to pull the info out of the meter. Given the number of them around a crowd funded project for good software seems quite doable.

[Tony_G]

It's really easy to get the harmonic data. Here is a simple app I wrote to get it when I was testing my 3325B:

Program.cs

Also these can be completely calibrated and adjusted electronically using a Fluke 5700 series calibrator and a Stanford Research DS-360, so it isn't out of the question to think they plugged them in and ran the Cal program either from the manual (Appendix C) or from proper Cal software.

TonyG

[usagi]

I'm going to be sending mine into Tek for calibration soon.  I got a quote last year.  I think it was something like $250 and it did include calibrating the audio THD functions.  Plus shipping which is likely to be another chunk of change in both directions.   So there is something to figure into the pricing, if he really is still calibrating them with a traceable (didn't ask if it was) calibrator.  In my view the whole point of a meter like this is accuracy.  If it isn't (recently!) calibrated then it is just a big metal doorstop. 

these guys do calibration of keithley 2000 for $75, maybe they could do 2015 thd?
http://4gte.com/calibration-and-repair/

[Tony_G]

I took a quick look at the Keithley 2000 calibration manual and the equipment is also spec'd to be a Fluke 5700 series calibrator (as an aside the 5720 will do 8.5 digits and the 5700A 7.5 digits if the Fluke marketing bumf is to be believed).

The question is, do they have the DS-360 which is needed for the THD part of the meter. I would assume so, even our cal lab here at work has one.

TonyG

[alm]

Fluke 5720A (II) uncertainty 90 days best DCV: 3 ppm
Fluke 8508 1 year: 3 ppm
Keithley 2002 1 year: 10 ppm
HP/Agilent/Keysight 3458A 1 year: 8 ppm, less with some options or Fluke branding

Good luck with that test uncertainty ratio. Even the 5720A 24 hour uncertainty (2.5 ppm) is borderline for the Keithley 2002 1 year spec. Never mind the 90 days spec. Hence my statement about needing multiple instruments, like voltage standards or standard resistors, for calibrating a 8.5 digit DMM.

[Tony_G]

I think you and Fluke might be in agreement, though obliquely.

They specifically talk about 7.5 & 8.5 digit DMMs in their descriptions:

5720A

Customers can get all the performance they need to calibrate their most demanding workload of multimeters up to 8 1/2 digits quickly, easily and reliably. This improvement in performance results from factory testing to even tighter tolerances, and from a variety of hardware and firmware improvements.

5700A

Considered the calibration standard worldwide, the 5700A delivers high value as well as accuracy covering 5 1/2 to 7 1/2 digit dmms.

They claim they achieve TURs from 4:1 to 10:1 through characterization - The have a detailed write-up of the process in this document:

Can a multifunction calibrator really verify a long scale digital multimeter that is more accurate than the calibrator? This paper describes methods to characterize a multifunction calibrator to better than manufacturer’s 24 hour specification at all points that are used to verify 8.5 digit multimeters. Using automated processes, data collection and statistical methods, it was found that a calibrator can be characterized at all required values to uncertainties capable of verifying long scale DMMs. Sources of error that had to be overcome include thermal EMF, loading errors from the devices used to measure the calibrator output and loading errors from DMMs when measuring the characterized output. Data from repeated measurements of the calibrator shows 10 V will drift less than 0.3 ?V/V for 30 days. For resistance, 10 k
can be shown to stay within 0.25 ?
/
for the same period. As the uncertainty of the calibrator was reduced, standards and techniques used to verify the multifunction calibrator had to be re-evaluated to decrease their uncertainty also.

So if you add the characterization process into it then yes, you need the calibrator and other stuff but, at least according to Fluke, you can operate multi-function calibrators in such a way that they can calibrate 8.5 digit DMMs.

You have now officially exceeded my knowledge of this scenario so I don't have anything more to add. I just found it interesting that you could use calibrators up to this level.


TonyG

[pigrew]

The question is, do they have the DS-360 which is needed for the THD part of the meter. I would assume so, even our cal lab here at work has one.

You sent me on a wild goose chase to figure out how the DS360 creates a low distortion signal. It turns out that it's doing DDS with an outdated audio DAC followed up with fancy filters.

Based on its manual (which provides a component list and theory of operation), it uses the (obsolete) 20-bit BiCMOS AD1862 DAC fed into "a seventh order Cauer anti-imaging filter". Huh...  The DAC can run at up to 17 MHz. Then, it has a adjustable relay-controlled (lowpass?) filter to block the harmonics, and a PA using a standard linear 3-stage PA. I'm sure that there was a bunch of tuning involved in its design, but it seems like something that I'd just put together in the lab one day.

[G0HZU]

I still don't understand the fascination (obsession?) with calibration for stuff like this. I bought one of these meters several years ago and I bought it for the design integrity of the overall meter and the AF DSP not the absolute calibration. Adding an expensive annual calibration will add very little 'real' day to day value to the meter unless you are one of the volt nut types who like to connect loads of these together on a workbench and mentally jerk off if they all agree with the calibrated one over xx minutes/hours/days. (and then do the same comparison a few days later... and few days after that... and repeat for years and years)

I doubt I will ever have my 2015 calibrated. I expect it to die before 'calibration' becomes an issue.

Maybe some of you think I'm being rude or trolling here... but in all of my career in electronics I've never met any engineers who get excited when a DMM comes back from calibration. It generally seems to be volt nuts and some (misled) hobbyists who seem to think having a megadigit DMM with a calibrator AND a calibration is an important part of running a successful electronics bench at home.

The other thing I'll add here is that the classic Keithley 2015(THD) isn't a nice meter to use for everyday electronics work. The display is very small and the fonts are poor and quite a few of these meters will suffer from a dim/dying display because they will have been run for many hours in an ATE system. The meter is big (as in very deep for a shelf)  and heavy and it runs quite warm. The user interface is poor if you want to use any advanced features and this meter is really meant to be used in an ATE system via remote control. I usually use mine via GPIB but this DMM doesn't normally live on my workbench because the human factors are so poor. It comes into its own when the advanced features are exploited via remote control from a PC and also the 4 wire resistance measurement feature is very useful from time to time

If someone has >100 of these for sale I would expect that most of them will be ex ATE meters that could have been running all day every working day for quite a few years. So watch out for tired and dim displays

[usagi]

i couldn't resist, so i bought one. 

here's a dave video teardown of his.

[slurry]

I've got one of those a year ago from partsconnexion, it is a really good benchtop multimeter and a fairly good distortionmeter although it's nothing like a AP or R&S UPL.

As G0HZU mention, they have probably been run in a ATE-system as most of the displays are a bit dim,
also, it's not the best display around either and it is a bit advanced meter for most of the work i do,
i also have a Keysight 34461A which i seldom use as it takes forever for it to boot up!

Usually for most measurements i go to the venerable Fluke 45, standard setting: dual display, DC with AC-component 
I really wish there was a new Fluke 45mkII with larger digits and LXI around the corner, i would buy one without hesitation.

[G0HZU]

Usually for most measurements i go to the venerable Fluke 45

I do the same if I need a bench meter I much prefer the user interface of the Fluke 45. Obviously the Fluke 45 has much lower performance specs but it is an excellent meter for most tasks. But I still use my old 3.5 digit H/H meter the most of all.

[usagi]

if the display is bad i'll get a replacement DD-52.

[Gregg]

Here's one for $250.

http://www.ebay.com/itm/Keithley-2015-6-12-Digit-THD-and-Audio-Analysis-Multimeter/361848524745?_trksid=p2047675.c100005.m1851&_trkparms=aid%3D222007%26algo%3DSIC.MBE%26ao%3D2%26asc%3D20131003132420%26meid%3D8c96265693ab4d5e83a623de525637b4%26pid%3D100005%26rk%3D3%26rkt%3D6%26sd%3D371582307445

I wouldn't exactly trust a seller that is part of Micro Precision Calibration [the banner in the background in the picture] selling an expensive piece of test equipment "as-is checked for power up only".  A quick google search turned up a branch office in Grass Valley, CA.   

[alm]

[...]
So if you add the characterization process into it then yes, you need the calibrator and other stuff but, at least according to Fluke, you can operate multi-function calibrators in such a way that they can calibrate 8.5 digit DMMs.

Thanks for the link! That is actually better than I expected, but still not exactly the straightforward operation as described in the manual, and possibly only for some 5720s (since they are relying on non-guaranteed behavior). For example, they report significant errors caused by rearranging an equipment rack. Still, with an advanced characterization procedure it seems possible to use the calibrators for fairly long times (up to a year) with a decent TUR. I stand corrected.

I still don't understand the fascination (obsession?) with calibration for stuff like this.

No argument here. Even if the ranges were out of tolerance by 10x the 1 year spec (something that usually only happens after a repair), it would still be good enough for pretty much all electronics measurements (excluding metrology).

The other thing I'll add here is that the classic Keithley 2015(THD) isn't a nice meter to use for everyday electronics work. The display is very small and the fonts are poor and quite a few of these meters will suffer from a dim/dying display because they will have been run for many hours in an ATE system. The meter is big (as in very deep for a shelf)  and heavy and it runs quite warm. The user interface is poor if you want to use any advanced features and this meter is really meant to be used in an ATE system via remote control.

I do not fully agree. Yes the fonts are smaller than some other bench meters (the HP/Agilent/Keysight 34401A and Keithley 199 are both much bigger, for example). And as any VFD, they will get dim if left on for extended periods (people too lazy to turn the display off in ATE applications). But if the VFD was not abused, it is readable in pretty much any lighting and has a much better viewing angle than LCD. So I would still put it above pretty much all handhelds and LCD-powered bench meters (e.g. HP 3478A).

I do not mind the controls. Most frequently used functions are available with a single button press, unlike for example the HP/Agilent/Keysight 34401A which puts current and 4W Ohms as shifted functions. I also like the layout of the ranging buttons better than the Hagisight. Only min/max is quite clumsy compared to how many handhelds implement it (store a fixed number of points, than recall the readings to get the minimum and maximum over that time period).

I wouldn't exactly trust a seller that is part of Micro Precision Calibration [the banner in the background in the picture] selling an expensive piece of test equipment "as-is checked for power up only".  A quick google search turned up a branch office in Grass Valley, CA.   

I agree that untested is a reason to be suspicious, and could easily describe something that was tested and found defective. But it is possible that the seller did not want to invest the time because the extra money for a 'tested, but uncalibrated' unit may not be worth it. I have definitely bought equipment as 'untested' from seller that should have known better that worked just fine.

[G0HZU]

But if the VFD was not abused, it is readable in pretty much any lighting and has a much better viewing angle than LCD.

I think there's more to it than just the brightness although I agree that it's a better display than a typical (gloomy hard to read) LCD display as found on some HP meters.
I also find the meter display to be very distracting when run in default/auto modes. It can be configured to overcome some of these issues but I find that the constantly ramping digits and fast update rate are annoying if it is set to have several digits and is on a low DCV range for example. The fonts are poor and this definitely isn't meant to be used as an everyday bench meter in my opinion. By contrast, my Fluke 45 is so much better as an everyday bench meter. Better display, better fonts, no distracting/ramping digits on the display, better user interface, lower power, doesn't run warm, smaller, lighter etc.

[NivagSwerdna]

FWIW the display is replaceable. Keithley still sell spare VFDs for this unit.

"DD-52 Display Unit"?? 

[nazcalines]

I recently received one of these from the ebay seller partsconnexion. It's in pretty nice shape, no yellowing of plastic, display is plenty bright, though I can tell it's been used. Compliments my Fluke 87 nicely.

Does anyone have a BOM for the electrolytics? Not sure if I'm going to replace the display, but definitely will recap it.