Tube tester - Sencore MU150 - socket contamination leakage - cleaning it?

[jamesmafyew]

I know tube testers aren't indispensable for this kind of work, but I like this one a lot and use it frequently enough that I want to get it working right...

I've got a Sencore MU150 tube tester, which I completely overhauled about a year ago (replaced almost every component on the board, all 1% metal film resistors, full calibration). It was flakey before that, but since then has been rock solid, until this summer. I'm in Philadelphia and my shop has no A/C, so the humidity gets pretty high. All the tester's functions (shorts, emission, gm) are stable and repeatable EXCEPT the grid leakage function, which shows the all-too-common symptoms of reading leakage with no tube in any socket. When the humidity is low, it's fine. When the humidity goes above about 65%, I start having problems.

When I overhauled the tester, I made a point NOT to clean the face or the sockets with anything that would leave a residue -- only 99.8% anhydrous IPA and a toothbrush. Like I said, it's been fine for a while now. I've never cleaned it with anything else, but who knows what was used on it before I got it. I have isolated the problem to the sockets -- with the switches isolated, there are no leakage issues. My Fluke 8050A measures about 50-100 megohms (10-20 nS conductance) between most socket pins and between any pin and chassis, with the sockets disconnected from the rest of the circuit. Same meter, when measuring switch contacts, sees over 1000 megohms (<1 nS), regardless of humidity. So the socket panel is definitely the location of the problem.

Anybody have experience with this sort of stuff? I have a thing for old test gear and am particularly attached to this tester, but the leakage issue is a tough nut to crack.

What I was hoping to learn is:
Is anhydrous IPA the right stuff to use on the socket panel and socket pins? Should I be using something else like the Techspray Blue Shower stuff to clean it? Is there something super-strong I can try?
Is this normal and my expectations are just too high?
Is there a safe way to clean the sockets out, dry them out when the humidity is low, and then seal them with some sort of acrylic or lacquer, to prevent the problem from recurring? Or would that just be a waste of time?

The tester in question:
http://25.media.tumblr.com/tumblr_m1...257o1_1280.jpg

[soundguruman]

I think I would rather test the tube in the circuit.
To me, tube testers are rather "dull."
Tube testers don't detect even half the potential problems.

[soundguruman]

I know tube testers aren't indispensable for this kind of work, but I like this one a lot and use it frequently enough that I want to get it working right...

I've got a Sencore MU150 tube tester, which I completely overhauled about a year ago (replaced almost every component on the board, all 1% metal film resistors, full calibration). It was flakey before that, but since then has been rock solid, until this summer. I'm in Philadelphia and my shop has no A/C, so the humidity gets pretty high. All the tester's functions (shorts, emission, gm) are stable and repeatable EXCEPT the grid leakage function, which shows the all-too-common symptoms of reading leakage with no tube in any socket. When the humidity is low, it's fine. When the humidity goes above about 65%, I start having problems.

When I overhauled the tester, I made a point NOT to clean the face or the sockets with anything that would leave a residue -- only 99.8% anhydrous IPA and a toothbrush. Like I said, it's been fine for a while now. I've never cleaned it with anything else, but who knows what was used on it before I got it. I have isolated the problem to the sockets -- with the switches isolated, there are no leakage issues. My Fluke 8050A measures about 50-100 megohms (10-20 nS conductance) between most socket pins and between any pin and chassis, with the sockets disconnected from the rest of the circuit. Same meter, when measuring switch contacts, sees over 1000 megohms (<1 nS), regardless of humidity. So the socket panel is definitely the location of the problem.

Anybody have experience with this sort of stuff? I have a thing for old test gear and am particularly attached to this tester, but the leakage issue is a tough nut to crack.

What I was hoping to learn is:
Is anhydrous IPA the right stuff to use on the socket panel and socket pins? Should I be using something else like the Techspray Blue Shower stuff to clean it? Is there something super-strong I can try?
Is this normal and my expectations are just too high?
Is there a safe way to clean the sockets out, dry them out when the humidity is low, and then seal them with some sort of acrylic or lacquer, to prevent the problem from recurring? Or would that just be a waste of time?

The tester in question:
http://25.media.tumblr.com/tumblr_m1...257o1_1280.jpg

Well, you see people bring tubes that test good all the time, even the NOS.
But I reject at least half of them because of noise, microphonics, hum, intermittent, loose internal construction, etc...
just because it tests good, does not mean it will work right in a guitar amp.
Tubes in a guitar amp are very different from, say, a hi fi amp, or a TV and radio.
The tubes are subjected to physical pounding and vibration from the speaker, bass drum, etc...
and they are over-driven, far past the manufacturer ratings and voltages.
that's where they give up, even though they test perfectly good, they just don't cut it.

There is no tester that checks for "any" of that. That's why we have tube amp techs.

[km6xz]

Yeah, I have not used a tube tester, or even had one in the shop for 30 years but do have a curve tracer...that is useful but so is the circuit the tube is intended for. But that is not the question at play here, one of leakage.
anhydrous alcohol is only anhydrous until you open the bottle. Storage and use of it makes the waster content climb quickly. There are drier solvents but they also have the problem of increasing corrosion and oxidation. If you are sure the leakage path is in the sockets alone(which ones?), replacement is always a reasonable solution for leaky sockets but exact replacements might be a problem to find since only a few companies still make them. After working with condenser mics and the extremely high Z's involved, any path for leakage is a problem for noise and it can be a headache. Since you are not dealing with 60db of post circuit gain, noise is not your problem, a reading that seems too high for residual is. Does a truely gas leaky tube displaying considerably higher on the meter? What I am asking, is the residual reading high enough to mast actual defective or high leakage tubes? If not, you might be better off ignoring it;>). Another way would be to use a opamp in a DC servo to cancel the leakage that does not come from the tube itself. That is another issue however........


As an aside, to SGM....did my sister's check arrive? She sent it a couple weeks ago I beleive and she sent the transformer to me but I do not expect it for another week or so.......Thank you very much and the very cute girl singer thinks you also...!

[soundguruman]

How cute?

[km6xz]

Julie is a very cute smart girl who is fleuent in a number of languages.. She sings in a jazz vocal group of 3 other girls but her career is as a researcher in organic chemistry. She as a masters but will start her Phd program soon.

[jamesmafyew]

One of the best hijacks I've ever seen.

There's overlap between the socket leakage (somewhere between 500M and 50M, depending on ambient humidity), and the leakage of, say, an EL34 that will redplate in a Marshall. The tester puts the "good/?" boundary at 200M, and the "?/bad" boundary at 100M. I've found that to be a reasonably helpful guide. Most of the time, the socket leakage with no tube in the tester is at varying places in the "good" area, so no big deal. But at humidity around 70-80%, it will fairly rapidly go into the low "bad" area.

Again, my interest in this problem is not primarily for the unit's usefulness, but because I like old test gear and would like to have it working 100%... plus I don't only work on guitar amps, but also rehab old test gear, hi-fi, whatever...applications where tubes are being used within their datasheet parameters. It's possible that I'm just being too demanding of a 50-year-old piece of mobile service gear.

[Steve Conner]

Is it the sockets leaking, or the big bundle of wires behind them that connects them to the switches?

Either way, hard to see what you could do short of replacing the offending parts or getting AC for the shop.

[jamesmafyew]

Can't tell if it's the wires or sockets. I did disconnect the wires at the terminal strip between the switches and the socket plate. Switches didn't leak at all, but the sockets/wiring were unchanged. In places where sockets have pins that are not connected to anything, I can't measure leakage between those empty pins and the adjacent connected pins. So any individual socket doesn't have measurable leakage -- I disconnected a couple of them and tried to measure -- but the whole thing does.

I've heard mention of "electrolytic leakage" before -- but don't know anything about it. Could that be what's happening here? What kind of contaminant would cause that?

[Steve Conner]

I'd suspect the wiring, especially if it has that funky vintage cloth-covered rubber insulation and is tied into neat bundles.

I've never heard of electrolytic leakage.

[jamesmafyew]

Definitely not that old-skool wire on the socket panel. There are only a few cloth wires on the transformer leads, but they're nowhere near the socket wiring. This tester was probably made in the late 60s as I understand it -- I don't know enough to know what type of wire would have been used in shop gear in 1968, but it looks like ordinary PVC to me. May have had mold or mildew at one point, though it smells ok now.

Here's a gut shot: http://local.anchorstates.net/images...50-gutshot.jpg

The grid switch that isolates a pin for leakage testing is the second from the right. The leads go from there, to each of the switches to its left (screen, then plate), then to the ten-terminal strip in the center, and on to the sockets. I disconnected the socket wiring at the terminal strip for testing, and the leakage was on the socket side. That's as narrow as I've gotten it.

I suppose I could replace all that wiring next time I have a vacation. :P

Here's a schem: http://local.anchorstates.net/images...mu140schem.jpg

[Steve Conner]

Hmm. I don't know of any issues with the PVC wire, and it isn't bound together over any great distance, so there aren't many opportunities for leakage. I'd have to agree the problem is in the sockets. Do you see any solder flux residue on the socket bodies between pins? That stuff can absorb moisture and get conductive.

Your meter might not measure the leakage of one socket by itself, but in use they are all in parallel.

If you were feeling particularly brave/suicidal, you might like to try removing the socket panel from the tester, putting it in the dishwasher, then drying in a low oven for several hours. Bob Pease recommended that for precision analog PCBs.

[jamesmafyew]

I might actually try the washing at some point.

There WAS some solder flux gunk on the sockets when I got it, but after many baths of IPA, it's gone now. Use to be that the run off from flushing the sockets was yellow, now it's clear.

I'm beginning to think that the best use of the device is as a totem, on the shelf, to establish my trustworthiness with the hi-fi crowd. Although, they'll complain that it's not a Hickok.

[km6xz]

Sorry about the hijack but sometimes electronics needs a little injection of non-nerdiness. There are so many places where leakage can occur in such point to point wiring. You isolated it to the sockets? Did all the cleaning with alcohol make measurable differences? If not, look elsewhere since any changes to the sockets, if the culprits would have shown a difference before and after, quite off, worse. That switch in the lower right in phenolic which is hygroscopic, has that been eliminated as a source? To measure such hi z, the meter needs a lot of gain, have you eliminated the meter amplifier, in other modes it is looking across lower Z circuits and any meter amp drift would be swamped. Are all sockets showing the same leakage or just one or a couple?
Maybe I will have to go "consult" with Julia on this one;>)

[km6xz]

Thanks for the schematic, it appears to have even more points of possible contamination than I thought. Have you looked at the noise from the meter bridging amp to see what sort of leakage it is? When a gas regulator gets old, it can generate a lot more noise, that is the reference for the meter bridge. Since the same mode switch is used on all sockets with many parallel wire connections to various sockets, it could be anywhere. If you disconnect the wire from the socket selection switch that goes to the metering mode switch position 4, does the leakage or noise go away? It is noise or current leakage? If AC hum, look for leakage between the 3rd and 4th position of the metering mode switch on the gang that selecting the input to the meter bridge.. If higher frequency noise, look for a noisy reference that is derived from the regulator tube.