Priceless tube tester voltage weirdness

[Phil_S]

I've built the Priceless tube tester. I used a rather small derelict chassis that already had a pair of octal sockets. I drilled it for a pair of noval sockets, too. I fixed the doubler circuit so it works as it should. With a pair of cheapo Russian 6p3c 6L6 wannabes, B+ drops from almost 290 to around 255-260 at the B+ takeoff point at the doubler. This drop didn't surprise me. Remember the PT primary winding is 1.9Ω, so I think it has plenty of capacity for a pair of 6L6's.

The design of this thing is pretty simple. It basically wires 4 sockets in parallel and uses a switch to select which cathode is being monitored at the test point. The B+ supply is simple, with no additional filtering.

The socket layout is octal-noval-octal-noval, let's call these V4, V3, V2, and V1. Don't ask why I labeled backwards. It is just how I see the orientation of the layout and how I labeled the cathode selector switch.

I've been playing with my new toy today. I've got 10 or so of these 6p3c tubes. They are so cheap, I don't feel a sense or risk, so they are good for testing the tester. Well, after running through 10 of them (the other two I've got are installed in an amp), with B+ at about 255 and bias voltage at -20, I noticed about half are testing at ~70mA and the other half at about ~110mA. Testing done always with two tubes in he sockets. Suspicion takes hold and I start putting the 70mA tubes in the other socket. Results don't always follow the tube. At this point, I flip it over so I can get voltage readings at the pins rather than test points.

Here's where I observe weirdness. It was weird enough that I did the measurements with 3 different meters, two of which are old Flukes, one a 25B and the other a 8600A bench meter. Both of these, as well as the Amprobe are good reliable meters as far as I'm concerned.

I'm not sure if relevant, the bias chain flows in this order: the take off point at the divider, the test point tip jack, V4, V3, V2, V1.

First thing, I check the daisy chained bias supply. With -20 at the voltage divider take off point, no tubes, it's fine down the line. I put two 6p3c in V4 and V2 and then I see weirdness on the bias voltage. V4 does not have the same bias voltage as V2. V2 meters the same -20 I expect. V2 meters -26. I completely stripped the bias wiring and did it over. I also simplified the layout as it was fanned from a single point originally. No change as a result.

The other weirdness is B+ at V4 is about 265, which is ~ 10V above B+ at the takeoff point by the doubler. B+ runs in the opposite direction of the bias feed: V1, V2, V3, V4, tip jack. But it's a parallel daisy chain, so what would it matter? In any case, how can plate voltage rise above B+ supply? To be clear, I measured at the plate and at the B+ take off point. I got different readings moments apart and going back and forth to verify.

I recognize that plate voltage will rise as bias voltage goes more negative, but I don't see how it can increase about the B+ feed. IOW, if I see 265 at the plate, I expect 265 at the doubler. No, not seeing that.

When I run this rig with only one tube in V2, there is no voltage weirdness, In fact, there is no meaningful amount of voltage weirdness on V2 when there is also a tube in V4, except perhaps to the extent that two tubes pull more current from the PT.

Nothing is getting hot, there's no smoke, 1A slo blow fuse is holding just fine.

Any ideas?

[Phil_S]

I know you guys want pictures. No excuses...not the tidiest or best planned. <shrug> It's from the bone pile!

[martin manning]

My first thought would be oscillation somewhere. Note the added components in the "modified matcher" as shown below...

[Phil_S]

Martin...oscillation, the thought crossed my mind, but without a speaker, I was lost. There is a certain instability in the voltage readings.

On the added parts, how critical are parts values? I'm not sure what's in the parts bin, but I expect I'll find 100&#937; resistors, not 82&#937; resistors. I assume these are more like "stoppers" than plate load, is that right? Same for the caps, I might not have .02 and .002 but I've got something. I know I'm overloaded with .47's, are they too big?

I'm not using cathode fuses as you can see. Do you think that's unwise? As noted, this was a parts bin build, with some amount, particularly the more expensive stuff recycled.

LOL, I see your note about test for oscillation with AM radio. I doubt I've got one of those, even though I'm surely old enough to have stashed a cheap Japanese transistor from the 60's.

Thanks very much for your post.

Phil

[martin manning]

I found that sheet somewhere on the web, I don't recall who the author is. Sure 100R would be fine. May be try that first and see if the weirdness clears up. Tacking in a film cap of some kind will be ok for an experiment.

I'd put fuses in to protect against a shorted tube.

[premiumplus]

I'm suspecting oscillation too. Put an oscilloscope on the individual sockets and look for high frequency oscillation. A high impedance ac voltmeter would be nice here...do you have a VTVM?
I see the original schematic doesn't have bypass caps on the plate of V1. Why did you add it to only one socket?
Are the sockets in good shape? You mentioned that it's made from old parts, and there may be flaky contacts in a socket. I've also had trouble sometimes soldering to old tube sockets, they need cleaning to avoid cold solder joints.
I think the first thing I'd do is find out why V4 has 10 volts higher plate voltage than the B+ takeoff. Kirchoff made a law, not a theory.

Lots of questions...It's a cool circuit and would be a useful tool.

[Phil_S]

I think we are reasonably well settled on oscillation as the likely problem. It was my first thought as both bias voltage and plate voltage are wobbly. Working space on the bench is rather tight at the moment...let's not get into that...I'd just as soon avoid moving the scope to where I can use it.

This is one of many things I've built with recycled parts. I recognize the sockets could be wonky, but I am reasonably satisfied they are making good contact. I had one cathode that needed to be retensioned and that one was very obvious. I thought about replacing the sockets when I encountered the problem, but decided for the time being that it would not likely solve the problem. Pins were well cleaned before reuse.

I'm busy as all get out with work right now. I'm not sure when I'll be able to get back to the project and I'm not sure if I have what I need for the stoppers and caps. Working space is tight...damn, plate pins are near the vertical wall. Lifting the existing B+ connections is likely to require a full rewire of the B+ run as I'm expecting I'll have to clip out what's there. I'm not expecting a quick easy fix.

I will be sure to post results. Patience, please.

Thanks for the input.

[Phil_S]

I finally fixed up the tube tester. 100&#937; plate stoppers addressed the oscillation nicely. Using the voltage doubler, Va is ~265V, I can dial in bias voltage between about -7 and -45V, and the 1&#937; sense resistors on the cathodes are doing the job nicely.

I've had my fun with it the past couple of days. I ran a number of 6L6/variants through it at -20V bias. Some are clearly better than others. The cathode readings were between 29mA and ~60mA, with most clustered in the 40's. Then I did all the 6V6's at -14V and several read 40mA or stronger. I even got a couple of match pairs.

I have one pair of EL84. These are Mullards. I had high hopes, but the ugly truth is, though they are matched, I had to bias them at -10V to get 30mA on each. I'm thinking pretty well used, but maybe just a bit of life still in them.

I'd like to go to work on the 6AQ5's but am concerned my homegrown socket adapter isn't up to the task.

This is one of the more useful projects. Highly recommended!

[martin manning]

From Mullard's EL84 data sheet at 250V Vg2 and -10V Vg1, Ia is 20mA and Ig2 is 2mA, 22mA total at the cathode. Seems like the pair you have there are in pretty good shape.

[Phil_S]

From Mullard's EL84 data sheet at 250V Vg2 and -10V Vg1, Ia is 20mA and Ig2 is 2mA, 22mA total at the cathode. Seems like the pair you have there are in pretty good shape.

Martin, thank you for commenting. For some unknown reason, I did not read RC-30 correctly. It specifies Va=250, Vg1=-7.3, Ia=48mA. I was thinking -12.5V but that is for a 6AQ5.

I did not realize or expect the Mullard data sheet to be so informative, but your comment got me to d/l it from Frank's electron Data site. The Mullard sheet is over 20 pages, with lots of informative graphs. I see the 22mA on the graph.

The whole business motivated me to test them again at Vg1 = -8. Va = 269.5 (can't control this). The graph says Ia=40, Ig2=~4.5. These two tested at 49mA and 46mA. I think if I had let them run for a while longer, the 49mA tube would have settled in at closer to 46mA. It took a while for it to drop to 30mA in the earlier test. Curiously, this put these tubes right on the curve for Pa=12W.

I know the graph assumes Ig2=250 and I didn't measure that as it requires getting inside the tester, but with Va=~270, it is probably closer to 260, but that should be close enough to retain the graph as relevant.

Although these tubes are dirty looking, I see now that they test ANOS. They were a surprise bonus I received in a derelict amp that I can't even remember. Needless to say, I'm quite pleased and this really makes my day!

I'm going to think about rewiring my 18W, which is presently set up for 6GK6. Or maybe these just need a special amp of their own?

I don't know how I lived without this thing! Everyone can build one. It was a very cheap project done mostly with stuff I had lying around.

[martin manning]

You could dial the screen voltage in with a Variac. The anode voltage isn't as critical, but you'd want the filament voltage to be within specification. It would be handy to include some test jacks for the screen and filament voltages.

[Phil_S]

You could dial the screen voltage in with a Variac.

Yes, but that assumes I have one or that I don't mind turning a cheap project into something more expensive.

The anode voltage isn't as critical, but you'd want the filament voltage to be within specification. It would be handy to include some test jacks for the screen and filament voltages.

A capital idea. I think I've still got a few tip jacks left.

Thanks for all your help, Martin.

[Phil_S]

Too bad I feel the need to bring this back.

After trying several times to build a cylinder shaped 6V6 to 6AQ5 converter, I gave up. This is more difficult than it looks to get it right and stable. In giving up, I realized I could buy a cheap plastic box and cover in the electric supply section of the Home Depot for under $2. It is now drilled and mounted on the tester -- a pair of sockets with umbilical cords to the octal sockets and proper octal plugs.

These were working great for about 10 pairs of tubes...until the last pair. On flipping the power switch, there is no power flowing. A quick check reveals a blown 1A slow blow fuse. There's now way I can imagine any pair of any tubes, including filament supply, that could demand 1A on the primary. I imagine (rounding up liberally) a pair of tip top better than new spec might demand 40VA? At 120VAC on the primary, that isn't more than about .333A.

There doesn't seem to be anything wrong with the amp. None of the PT windings are open and all meter at reasonable ohms. The dim bulb tester with 40W bulb is faintly glowing (no obvious short.) I've about scratched my head as much as I can and still have no answer.

I'm not looking to feed fuses to this thing. Any suggestions would be welcome.

BTW, it seems I've got a pile of 6AQ5/6005 tubes that test at or above the spec for a new tube. Anyone in the market for match pairs or quads, I'm thinking I'll have some at a very reasonable price. I can do this provided I get the tester running again. PM me if interested.

Thanks.

[martin manning]

Have you only blown one fuse? If so I'd risk another. OTOH, did you include grid stoppers in your extension box?

[Phil_S]

Have you only blown one fuse? If so I'd risk another. OTOH, did you include grid stoppers in your extension box?

Thanks, Martin. I used my last 1A fuse and it seems to be holding. Grid stoppers? I thought (note to self: thinking can be dangerous) the 100&#937; stoppers on the octal sockets were enough. Are you saying to install another 100&#937; (or other small value) on each of the 7 pin sockets (in the extension box), too, for a total of 200&#937;?