Sprague Tel-Ohmike TO-6A Capacitor Analyzer

[innkeeper]

I hit an odd problem while checking out the leakage test on this analyzer...and could use some input from people familiar with the circuit used.

I got this Sprague Tel-Ohmike TO-6A Capacitor Analyzer in "unknown" condition, it appears to be rev 2 so a later model.  I did some basic switch and pot cleanup and checked the power supply caps, all seemed good enough to bring up on the variac, and it came up fine. it is out of cal, and likely needs to bridge caps repalced and an adjustment at a minimum.

What I see on the leakage though is on the .06 ma scale, while testing a known good cap, the needle will hover above and below zero. I can seem to simulate this by slightly adjusting the line voltage up and down. so I am assuming this is related to voltage regulation/stability.

Is this normal with this piece of gear. id be highly surprised if it is normal.  and then the second question is if it's not normal, what might I look at as the likely cause.
Schematics and manual here - http://www.mcmlv.org/Archive/TestEquipment/Sprague_TO6A.pdf

[coromonadalix]

Have you done the section 2.2 and section 4.2,  the operating procedures ??

[innkeeper]

section 2.2 has to do with testing capacitance and power factor. but none of those settings have any control of the leakage test as those controls, as best I can see int he schematic, are out of the circuit during the leakage test. i have tested those functions and they do need some work as described above and am working through that.

section 4.2 is the leakage test, and I did follow the procedure and is what I am commenting on as being an odd behavior. The observed behavior continues indefinitely and with any electrolytic I test.

what would you like to know or point out from 2.2 and 4.2?
Is there something that might explain the erratic leakage measurement reading on the .06ma scale?

[Cyberdragon]

Define hovering. Do you mean twitching/oscillating or slow drifting?

4.3:

It will be noted that the voltage reading will tend to increase after a short time as the leakage current begins to decrease to a stable value...The measurement of leakage current should be made only after a stable value is reached.

Also check C15

[innkeeper]

I tacked in a new cap for repalced C15 just to see if that might be the cause. (lifted one lead of the cap then tacked in a new cap temporarily)...and it made no difference.

Video of the behavior: https://imgur.com/a/a6PufU5

[innkeeper]

fill video of the issue:
i didn't realize imgur only showed the first few seconds.

in this video, I had a fairly new high-quality Nippon chemicon capacitor,  low ESR, high life, 105C which has near 0 leakage.  the voltage was set to 30 volts for the test and had been on the test rig for over 5 min at the time of the video. Note that this behavior occurs with any high-value cap I test, so it is not the device under test that is causing this.

I've got some replacement filter caps on the way, but doubtful this is the issue. What I have observed using a digital meter is that the variations in the leakage current meter track with minute changes in voltage to the cap.  So far I only see this behavior on the .06ma scale on the leakage test. 

[Cyberdragon]

That's 60 MICROamps though, that's probably just noise, like imperfect switch connections. Theres nothing in the leakage circuit but resistors and switches. You'd have to trace the source of the spikes.

[innkeeper]

tube maybe?

[Chris56000]

Hi!

With the 0-60uA F.S.D. range you're using, any random changes in the leakage resistance of the component you're testing will inevitably show up on that size of meter - it is large enough to show a current-change of 100 nano-amperes if you watch it carefully enough, and you've also got random variations of emission of valves, H.T./heater or mains voltages, etc., to take into account, as the H.T. power supply to the instrument's circuits is not voltage-stabilised!

Obviously such variations ought to be less apparent on the instrument's higher leakage-current ranges, if you saw this amount of variation of the meter pointer on the 0-6mA or the 0-60mA leakage ranges, there's obviously a fault, either in the component being tested or the instrument itself!

Chris Williams

[innkeeper]

I tested the idea that line voltage variances are causing it... I happened to have a vanguard Constant Voltage Transformer I've never actually had a use for until now.... it is good to be a packrat .

Using the CVT, i got variances of about +- .5 uA as compared with  +/- 4uA or more plugged directly into mains.
CVT isn't perfect but it defiantly compensates pretty good for line voltage variances... I did put it into a variac just to see how well the CVT worked and changed the input voltage, and it has a small but opposite variance, lower the voltage the output would go up just slightly and vice versa.  it would certainly be enough change to account for the +/- .5 uA but it id regulate the output ac pretty darn well. within about 1v or so on a 20v input change, didn't measure the input, just the output and adjusted the variac, so take those numbers with a grain of salt.

ok so, unless someone out there that has one or has used one says this is normal for the Tel-Ohmike, id have to believe this is just an indication of something not working as designed.   
I am not familiar with the regulation circuit there using. so not sure what to check or if this is just inherent in the design.  I can only do capacitance checks and cant do leakage tests on the caps. I'll go through the caps in this circuit chain and check their capacitance, and check resistance values.  I guess that's a start I just wish I understood how the regulation circuit works.

i've got a pair of filter caps on the way, should be here Monday.. that s kind of shotgunning it, but I figured it can't do any harm. if anyone can point me to some theory on how the 6bl7gt is doing regulation, that be nice or can explain that part of the circuit.

[innkeeper]

the 6bl7gt was bad, replaced it though i can't say it gave any significant improvement on the issue.
reading some post on the unit it seems the high voltage is not regulated and therefore very susceptible to line voltage variances.

so meter movements like that on higher capacitor values and high voltage caps are to be expected with line voltage changes as the voltage to the cap varies.
any thoughts on how to add some regulation to this?

[ignilux]

I realize the last reply was a couple of months ago, but there is one of these available locally in an unknown condition. Just wondering if you've had any further success with your unit?

[xwarp]

I'm pulling this out of my butt....

Could it be that at that very low level of current and the scale, that you are seeing negative current flow because of the the leakage at that level?

[innkeeper]

I've not been on in a bit so didn't see the followup messages....

So its a regulation issue on the voltage applied.  i used a Ferroresonant transformer to provide a more regulated line voltage to it and the amount of wavering of the reading went way way down.

This is most prevalent on high value capacitances at high voltages, low values do not have the same issue.  This all fits as low values will accept the charge changes faster, and high values will show positive or negative charge changes as the line voltage changes ... basically when the rc charge time constant is high enough that it shows on the meter.

So the answer is live with it or make some kind of regulated supply for it.  Least I know what is causing it now, and really it is pretty easy to average out the movement. ... it only happens when you're looking at microamps of leakage.

[bob91343]

Any slight variation in power line voltage will cause what you are seeing.  When a capacitor is charged from an unstable power supply, its current will surge in and out.  So perhaps the problem is poor power line stability.

Just for a test, perhaps you can replace the internal power source with a stable regulated supply and see if the problem disappears.

[innkeeper]

exactly....

I took a different approach for test purposes though, I was able to make the line voltage much more stable using the Ferroresonant transformer. ... it cleaned up the "issue" significantly enough that I am confident about line fluctuations causing it and due to the lack of regulation in the DC variable supply, the dc voltage follows suit. ... still, at like 400V a very small line change gets magnified. so even though Ferroresonant transformer does a decent job at regulating the line voltage, I still see some effects even with it, just not anywhere near as bad..

if I get some time i'll do a with and without the Ferroresonant transformer.

[bitseeker]

I've got a regular TO-6 in my repair queue. I'll keep this in mind when it gets its turn on the bench. Looking forward to your follow up tests.

[innkeeper]

An update:
On another forum, other owners of the TO-6 and TO-6A report the exact same behavior.
FYI, the only difference between the to-6 and to-6a is diodes to protect the meter, that's a good and easy update to make if you have a TO-6

I got my hands on a dual conversion line conditioner. For those that do not know what that is, it is an online UPS with the batteries being optional.. it converts the AC to DC then back to AC again, so the AC  is completely generated from the unit itself. This one is fairly advanced, microprocessor controlled, uses DSP technology to create and monitor the sine wave. so gives a very clean and stable output ( clary sp560 for anyone who wants to read more on it )

Results:
So after placing the TO-6A on it, and testing the same cap where the readings were bouncing around, it is stable with some small variations that move the needle about one needle with...that are likely in the picoamp range.

So the bottom line here is that the unit is in desperate need of regulation for higher value cap leakage testing. I'm not sure how easily that could be done without ending up totally redesigning it. Maybe someone with a keener eye then I can take a peek at the schematic and come up with some ideas for an easy retrofit.

[bitseeker]

Thanks for the update. I thought there would be more of a difference between the non-A and A versions.

[innkeeper]

To be more specific they used one protection diode on the rev1 and 2 for bidirectional protection on rev2

[bitseeker]

I shall upgrade mine.

[innkeeper]

i misspoke about the to-6 vs the to-6A Sprague TO-6 vs. TO-6A

[bitseeker]

I grabbed both manuals. A quick parts check reveals a few resistor changes, some additional resistors in the A model, and a different switch set. The capacitor leakage tables are also more detailed in the TO-6A manual.

[innkeeper]

from what i gathered, one of the differences is the voltage applied to the cap under test when testing it's capacitance value.

supposedly they lowered it for the TO-6A

as it is for the highest range 45-2000 uf i still get 30v ac on the cap under test.
the rest of the ranges are 3v and under.

thing is if you don't know that, you could kill a low voltage cap inadvertently.
id be curious what voltage the TO-6 applies.

[bob91343]

It appears that most of the service oriented test gear does put a somewhat high potential on the DUT.  I found that to be true on a Solar and an EICO for instance.  So I don't use those instruments in normal work although the leakage test is nice, taking it to rather high voltages.

Having said that, I have found that my GR 1658 bridge puts a very small potential on the tested part, and should I want to measure leakage at high voltage, the GR 1644A can run up to 1 kV and measure leakage in terms of insulation resistance.  Let's see, if I measure resistance of 1 Megohm at 1000 Volts that represents 1 mA of leakage.  Now, instead of looking at an eye tube and saying okay its display says low leakage, now I am forced to decide just how much leakage I want to tolerate.  Frankly I prefer the latter situation.

So while it's nice to make measurements on the cheap, there are pitfalls.  And the 1658 measues ESR as well.

The EICO is put away; maybe I'll sell or trade it.  I suspect the CDE and Heath units were similar in this respect.There was an Aerovox unit as well, if I remember correctly.