Old Fluke Multimeters

[Dave Wise]

I didn't see the third 3D button template, what is it's Thing Number?

[blue_lateral]

The third one is 4681912.

[Dave Wise]

Just following up on the 8060A C19 RMS cap.  My main board is Rev H and assembly Rev H.

My C19 is installed matching silkscreen and schematic, that is, plus facing the display and connected to U2 pin 6.  Measuring voltage using a hi-Z meter, I see -2mV in DC mode, +5mV in AC mode with a non-overloading 1kHz sine, and +0.5V with 2x overload.  So polarity matters but leakage won't be an issue.

I checked my C19 candidates at room temperature.  I only tested one sample of each type.  Charge one hour at 10V, measure leakage with a 10 Meg meter, discharge 10 seconds, record peak voltage to estimate DA.
Nichicon UKL1E470KEDANA is 9nA and 0.19V (1.9% DA), Chemi-Con ELE-350ELL470MF11D is 5nA and 0.27V (2.7%), Murata ceramic RDEC71E476MWK1H03B is 25nA and 0.70V (7%) - to my surprise, it's the worst as well as the most expensive.  Of these three parts, I would use the long life Chemi-Con aluminum electrolytic.

And my beeper boards are on their way!

[blue_lateral]

I've had one of the polymer contenders for 8060A C36,  Kemet A759, connected to 10.5 Volts for about a day and a half now measuring leakage current at room temperature to see how bad the polymer capacitor leakage really is under the conditions it would probably see in an 8060a laying on my bench. It is connected across the 9 volt battery. If I remember correctly, drtaylor mentioned earlier in this thread that he expected polymer cap leakage to be negligible in the places where electrolytics were originally used in the 8060a. That tracks with these capacitors as C36. Leakage at 10.5 Volts at room temperature is below my ability to accurately measure. It appears to be less than 20 nanoamps, maybe a lot less. I'll probably check a couple more of these as time goes on.

Spec is 120 microamps at 20C, not to exceed 120 microamps at 20C after damp heat or surge testing (see datasheet). Kemet A759 22uf/35VDC , Mouser 80-A759BQ226M1VAAE80 , datasheet here: https://www.mouser.com/datasheet/2/447/KEM_A4072_A759-3316694.pdf

These are almost too tall but do clear the processor board.

[blue_lateral]

I checked my C19 candidates at room temperature.  I only tested one sample of each type.  Charge one hour at 10V, measure leakage with a 10 Meg meter, discharge 10 seconds, record peak voltage to estimate DA.
Nichicon UKL1E470KEDANA is 9nA and 0.19V (1.9% DA), Chemi-Con ELE-350ELL470MF11D is 5nA and 0.27V (2.7%), Murata ceramic RDEC71E476MWK1H03B is 25nA and 0.70V (7%) - to my surprise, it's the worst as well as the most expensive. 

Could you go into a little more detail about your testing method? Thanks.

[Dave Wise]

I don't see what I left out, except possibly the well-known trick of using a voltmeter with known resistance to measure small currents.  I put the meter in series with power supply and DUT and let the meter's input resistance be the ammeter shunt.  With a 10 Meg meter, 10mV is 1nA.  Since I had three DUTs in play, I used three meters, a Fluke 8000A, a Fluke 8100B, and a Monsanto 2000.  After measuring the leakage, I shorted the meters.  Then I turned off the power supply, which has a few kilohms of bleed at the terminals, and after 10 seconds I pulled off the meter shorts and watched the readings go negative, level off, and fall.  This isn't exactly like the IEC standard since the meters load the recovering caps, but since all I wanted was a comparison it doesn't matter.  It was quite a surprise to see the ceramic cap perform worse than the electrolytics!

I see that the typical leakage of the polymer caps, like that of the regular caps, is far below the spec sheet limit.

[Dave Wise]

D.R. Taylor, if you're listening, there's text missing from your Hidden Features paper.  At the end of page 1, you write, "The accuracy loss is due to a".  At the beginning of page 2 you write, "out in the normal 10 megohm volts mode".  I'd love to know what went in between.

And it seems to me that one can only get hi-Z by having the MAC disconnect the OREF- pin, because otherwise there's a sneak path to ground through R2, RT1, Z1.2-4, Z5.7-6, OREF-, OREF+, and Z5.8-10.

[blue_lateral]

I don't see what I left out, except possibly the well-known trick of using a voltmeter with known resistance to measure small currents.  I put the meter in series with power supply and DUT and let the meter's input resistance be the ammeter shunt.  With a 10 Meg meter, 10mV is 1nA.  Since I had three DUTs in play, I used three meters, a Fluke 8000A, a Fluke 8100B, and a Monsanto 2000.  After measuring the leakage, I shorted the meters.  Then I turned off the power supply, which has a few kilohms of bleed at the terminals, and after 10 seconds I pulled off the meter shorts and watched the readings go negative, level off, and fall.

That did occur to me. Years ago I often used that trick to measure leakage with a HVPS and a VTVM in tube based equipment. After posting I tried it with the power supply and DMM (Tektronix TX-3). I wondered at the time if you were timing self-discharge somehow, but now I see. It never occurred to me to also check for dielectric absorption until I read your post. Now I probably will.

It seems to have stabilized at around 3.4 mv, give or take a few counts as it drifts. It's been there most of the day. It does get a little worse if you heat it up with your fingers, but we are down where it shouldn't matter for C36.

I just got an email from Mouser that the Kyocera caps ( RPF0509220M035K ) have finally shipped. I'll post a comparison when I get them.

It was quite a surprise to see the ceramic cap perform worse than the electrolytics!

I bought some of those same series Muratas in 22uf as a possible C36, but they have a huge capacitance rolloff with voltage according to the datasheet, about half of it at 10 volts as I recall. I've not tested it. It occurred to me that maybe 47uf would be closer to a 22uf electrolytic when connected to a 9V battery, but I sort of gave up on that idea. I too am surprised they were the worst performers of the 3 you tested. Can you see anything about them (other than cost) that is bad enough to kick them out as a possibility for C19?

[Dave Wise]

If my instrument is representative, C19 sees only a few millivolts in normal operation and did not have to be a special low-leakage part.  Although its leakage and DA is worse than basic aluminum electrolytics, a Type II ceramic should do fine, and outlive your great-grandkids besides.  I won't bother; a 10000-hour electrolytic would outlive me and that's plenty.