Keithley 199 -- FIXED!!!!

[c4757p]

It's easy to test out of circuit: install a replacement, see if it works. Hey, the original part is technically out of circuit, right?

[staze]

It's easy to test out of circuit: install a replacement, see if it works. Hey, the original part is technically out of circuit, right?

lol! I guess that's just as good... =)

[staze]

alm, or PedroDaGr8,

So, something still isn't quite right. I did replace the inverter, and that seems to be working now. But...

On the 30V and 300V range, with no voltage applied to the inputs, or the inputs not shorted, the voltage is pretty stable at 00.0000 or whatever. But on the 3V or 300mV range, if there's no voltage applied, or the inputs aren't shorted, the reading creeps up (down) toward -1.5V or more. On the 300mV range, this relatively quickly leads to an overflow, but on the 3V range, it just gradually creeps. If I hook anything up to the inputs, like my FG, even with the output off, the impedance is low enough for the creep to not happen or be much slower.

Sounds like some type of capacitance on the inputs?

Do either of your working meters do anything like this? I know some drift is common from it picking up transient voltages... but to pick up over a volt...

I will say that it reads accurately when I give it a voltage, and it zeros properly when I short the leads together...

Thoughts? And thanks!

[ve7xen]

This is normal. The input impedance for (good) bench meters in the low ranges is in the Gohm (for all intents and purposes it is completely open) so builds up charge from the environment in short order. If you short the inputs with a megohm or ten as a handheld meter has, you should see the charge buildup is basically gone.

This is a desirable feature, and a good reason to own a bench meter, as you can measure high impedance circuits without having an impact on them.

Good job on the repair

[alm]

As ve7xen stated, this is the expected behavior. Many bench meters will actually range up (because the voltage exceeds the maximum of the range) until they reach a range with a low 10 Mohm input impedance. This will discharge the parasitic capacitance, and then they will downrange again. This is repeated indefinitely, and can be quite annoying and eventually wear out relays.

[staze]

This is normal. The input impedance for (good) bench meters in the low ranges is in the Gohm (for all intents and purposes it is completely open) so builds up charge from the environment in short order. If you short the inputs with a megohm or ten as a handheld meter has, you should see the charge buildup is basically gone.

This is a desirable feature, and a good reason to own a bench meter, as you can measure high impedance circuits without having an impact on them.

Good job on the repair

As ve7xen stated, this is the expected behavior. Many bench meters will actually range up (because the voltage exceeds the maximum of the range) until they reach a range with a low 10 Mohm input impedance. This will discharge the parasitic capacitance, and then they will downrange again. This is repeated indefinitely, and can be quite annoying and eventually wear out relays.

Huh. Okay... for some reason I thought the input impedance of the 199 was only like 1M, so I was a bit confused. So if I'm going to leave the unit on for long periods with no input (not planning on this, but even so), I should probably set it to a specific range and not leave it on auto.

Got it.

[alm]

Yep, I usually set it to 30 V (where is has an input impedance of 10 Mohm instead of the Gohms it has on the lower ranges) when not in use. If only to silence the annoying sound the relay produces. The same happens on the HP 3478A, for example, although with a lower frequency.

[c4757p]

It would be nice if these meters could sense this (either sensing the behavior after a couple cycles, or even a mechanical/optical sensor on the jacks) and automatically switch to a low-impedance range...

[staze]

Yep, I usually set it to 30 V (where is has an input impedance of 10 Mohm instead of the Gohms it has on the lower ranges) when not in use. If only to silence the annoying sound the relay produces. The same happens on the HP 3478A, for example, although with a lower frequency.

Perfect. I'll set that and save it so it comes up in that config. For interest sake, is it possible to measure the input impedance of the various ranges? Just set it to DCV in a given range and hook another meter up to the inputs measuring ohms?

Thanks!

[c4757p]

You're not going to measure the high-impedance ranges without something like a picoammeter. Of course, if you just want to know what is "high impedance" and what is the usual 10M, then yeah, just stick another DMM on it, it'll read infinity for the high impedance ranges.

[staze]

It would be nice if these meters could sense this (either sensing the behavior after a couple cycles, or even a mechanical/optical sensor on the jacks) and automatically switch to a low-impedance range...

Yeah... does seem like a weird thing to not address. The sensor on the jacks would solve it for no leads, but if you just left the leads plugged in and dangling...

Programmatically, it wouldn't be that hard to address this, like you say, keep a counter of the cycling, and how long it takes to do the cycle, and if the cycle takes in the order of tens of seconds, and if the charge disappears once you hit the 30V or 300V range, then you're likely in a loop... so just stop and disable "auto". =/

[staze]

You're not going to measure the high-impedance ranges without something like a picoammeter. Of course, if you just want to know what is "high impedance" and what is the usual 10M, then yeah, just stick another DMM on it, it'll read infinity for the high impedance ranges.

Oh, that works. =)

I would have thought the manual would cover the impedance of the ranges, but I couldn't find it... ah well.

Thanks all!

[c4757p]

I'd actually just be happy with not autoranging into the hi-Z ranges at all. Most of the meters with a high impedance range have more than enough resolution to be useful for almost everything up in the 20V/30V/whatever range - I always set my 3468A to the 30V range to shut it up, and that range has 100uV resolution! I can manually switch if I want gigohm impedance.

Of course, the real bastards are the ones with nonstandard impedances, like my B&K Precision 2712, which actually has 2.3M! It took me a while to figure out what the hell was wrong with my circuit (I was probing a high impedance spot and calculating out the effects of the meter from an assumed 10M....) As long as HP and Keithley don't start doing that shit, I'll let them do whatever they want with the hi-Z behavior......

[alm]

I think some of the recent Agilient handheld DMMs had an odd input impedance (something like 2 or 5 MOhm) in some modes. Good fun when using accessories like high voltage probes, indeed.

[ve7xen]

I would have thought the manual would cover the impedance of the ranges, but I couldn't find it... ah well.

In the specs section under DC volts, you'll see 'input resistance' noted as > 1Gohm for the 300mV and 3V range. Not sure if they mention it, or the consequences, anywhere else.

[c4757p]

I think some of the recent Agilient handheld DMMs had an odd input impedance (something like 2 or 5 MOhm) in some modes.

Do they? Hmm... I've been considering getting one, and I didn't notice that! (Not that it'll make me not want one, just surprised I didn't catch it while reading datasheets.)

[alm]

The Agilent U125x series is the one I had in mind:

[c4757p]

Is it? That looks normal to me for a dual display meter - double the load for double the display. It's definitely a nice-looking meter, though I can't afford it right now...

I do love the dual-display setup. My Fluke 45 has that and it's quite handy. I'd much rather dual AC/DC than a single AC+DC, for instance, or AC+freq. V/I is also useful, though it annoyingly flutters the relays to do that... I'll have to probe my Fluke and see if the impedance halves with two displays active.

[staze]

Finally found a spec sheet that shows the input impedance for the ranges. I just figured 3V would be 10M like the others, but it's >1G, same as 300mV range.

Spec sheet is attached, since it took a bit of work to find (found it on one of the calibration/repair sites). Just incase anyone wants it. =)

[ve7xen]

KO4BB has archived the full user manual, which includes this spec as well, just assuming you didn't find it. Manuals for your equipment are always good to have handy, sometimes you need to RTFM just to get a sensible reading from the instrument!

Keithley kindly even includes schematics in this one (their newer equipment they are very stingy about technical information). The manual is actually available on the Keithley site directly, but it's behind a registration wall, so this is better

http://www.ko4bb.com/Manuals/09)_Misc_Test_Equipment/Keithley/Keithley_199man.pdf

Edit: fixed URL

[staze]

KO4BB has archived the full user manual, which includes this spec as well, just assuming you didn't find it. Manuals for your equipment are always good to have handy, sometimes you need to RTFM just to get a sensible reading from the instrument!

Keithley kindly even includes schematics in this one (their newer equipment they are very stingy about technical information). The manual is actually available on the Keithley site directly, but it's behind a registration wall, so this is better

http://www.ko4bb.com/Manuals/09)_Misc_Test_Equipment/Keithley/Keithley_199man.pdf

Edit: fixed URL

Oh, I have the manual, that's how I fixed the thing (with the schematic)... but for the life of me, I can't see the specs in that manual.

Or... I'm completely blind... I just found it. Fair enough... well, thing I attached should show up in google if someone's lookin'. =)

[staze]

PedroDaGr8,

Any luck fixing yours?