Question about an EICO 667/tube tester

[dude]

I was given this tester by a good friend. After reading the manual and adjusting the calibrations it seems to work well, testing for leakage, shorts and quality of the tube.

You'll see a novar socket for testing eight pin penatodes. correction, Novar is for 9 pin as FYL mentions below but next to the Novar socket is an eight pin socket with a 6V6 in it.

You just follow the included chart for the tube and set the levers. I was wondering if the same 8 pin socket is also used for rectifier tubes?

Of course this tester doesn't test any matching just if the tube is good.

To match good tubes can this method be used:

Use know working amp with adjustable fixed bias with 1 OHM resistors off the cathodes to ground on the power tubes. And match the tubes voltage drop across the resistor.  Matching tubes that drop the same voltage (millivolts) within 5 points. Example, two 6V6s with a drop of 22 and 23 millivolts would be considered closely match.

A reliable method?

One other question, testing the voltage drop of tubes (in the amp) that tested good have had readings way off, is this just that the tubes have drifted over time? I have had old RCA 6V6s which have read one tube, 5 millivolts and the other 18 millivolts but testing the lower reading tube on the tester it's very Good actually both tubes are good. Drifting over time? I'd have to raise the biasing resistor 10K past the schematic to get a 20mv reading on that 5 mv tube hoping to find another tube that is matched, all normal?

thanks as always,

Al

[FYL]

You'll see a novar socket for testing eight pin penatodes.

Novar = 9-pin

I was wondering if the same novar socket is also used for rectifier tubes?

You may of course test Novar rectos such as the 5BC3 or 6DW4.

A reliable method?

Yup, but plate voltages are too low for real-world matching.

[dude]

A reliable method?

Yup, but plate voltages are too low for real-world matching.

Using a Princeton amp:

I'm keeping one tube in the socket, and keep pulling the other with the bias set the same and reading the drop on each tube I try, marking them. Then I group them together according to the closest voltage drop. I take the pairs and put them both back in the amp and then bias them to about 22mv for 6V6s in a Princeton. (all the pairs do measure close @ 22mv biased)

When I started, the original 6V6s in the Princeton had reading of 5mv and 22mv, way out but the tester said good for both. The amp sounded OK clean but bad up the volume. I figured the tubes had been in the amp for years and drifted. Just have a bunch of used USA 6V6s that all say Good on the tube tester and trying to match them up with this method.

I don't know what you mean by "plate voltages too low" as I'm checking the mv drop in the amp with 380v on the plates...?

Thanks
al

 

[FYL]

You didn't mention the Princeton amp  . The post was about an Eico 667...

[dude]

Sorry I wasn't clear.

So, this method is pretty accurate?

Thanks,
al

[FYL]

So, this method is pretty accurate?

Yup. And IMO one of the best for real world current matching.

[PRR]

> wondering if the same 8 pin socket is also used for rectifier tubes?

The EICO and similar _ARE_ "rectifier testers". They connect any tube as a rectifier and see how much current it will flow. The difference 6L6 or 5U4 is just pin connections, heater voltage, and target current.

These little testers probably can't hurt big tubes 6V6 and up. They may be able to over-current some smaller tubes, especially if you get the settings wrong and hold it for more than a couple seconds.

> leakage, shorts and quality of the tube.

BAD leakage, yes. Short-test can be quite valuable (but can't test for a short that happens only when HOT-HOT). They do not meaningfully test "quality", whatever that is.

Tubes may read "weak" yet be totally functional in-amplifier because the test condition is very different and more extreme than many amplifier stages.

They often find "dead" tubes. They can't prove a tube is "good".

> two 6V6s with a drop of 22 and 23 millivolts would be considered closely match.

Excellently, and perhaps pointlessly, "matched".

Unless you need show-off THD numbers, a 2:1 difference is often acceptable in fix-bias amps (as long as the hot tube isn't over-dissipating).

6V6 is a peak 120mA tube. What you would like to know is that it will pass 120mA with a G1 voltage very near zero.

In the "classic" 6V6 amps, the tube idles at 40mA. If this happens near "normal" G1 bias, if this happens for both tubes in a pair, you can be pretty sure they will work together and give rated output cleanly.

In hot-rod guitar amps, we jack the voltage up to 400V or more. To stay within the 12W rating we need to idle down at 30mA or even 20mA even though peak current increases to 150mA or 180mA.

Since the tubes are optimized for 40mA-120mA, assembly tolerance slop tends to happen at "very low" current.

We can't be real sure, from a 20mA idle point, how it will handle 150+mA.

> one tube, 5 millivolts and the other 18 millivolts

The 5mA is pretty darn low. That bothers me more than the "un-balance".

I'd bump it up and see if it could be (depending on B+) say 25mA and 10mA.

[dude]

 

> one tube, 5 millivolts and the other 18 millivolts

 The 5mA is pretty darn low. That bothers me more than the "un-balance".
I'd bump it up and see if it could be (depending on B+) say 25mA and 10mA.   




I hear ya loud and clear, idling at 22mA then playing watching the meter I can see the mA's going up to your specs. I did the math and I'm at about 85% dissipation for a 6V6.

Your right about perfectly matched tubes, as sometimes they don't have the harmonics as tubes about 5 to 10 mA off. I tend to want a difference, just like a balanced PI is often not the best tone as my ears tell me. It's all about harmonics for me.

But the method I used to match tubes seems to work well "to my ears", just that I don't match them dead on.

Anyway, I kept the 5ma tube and found an old Sylvania 7408 that had the same voltage drop, then I dropped the biasing resistor in the amp to get about 22mA at idle, both tubes were idling at 20mA after playing  while. I checked out all the voltage and they were dead on to the schematic. The amp's distortion on 10 wasn't anywhere as muddy as the old set-up. I'm sure my friend will be happy.

Your right about those testers not being that great, I calibrated the tester but those resistors and pots are old and most likely off plus the testing voltages are low. So I take the results like a grain of salt.

I put a known brand new set of 6V6s on the tester, no shorts, merit test went to right side of good, 140% (whatever that is) so I figured testing other 6V6s might be reliable...? At least I could tell if they were weak or strong, shorts and leaks.

I can say one thing, I can hear tubes that are 15mAs or more off from each other, of course an opinion.

Thanks, for the knowledge

al       

[PRR]

> idling at 22mA then playing watching the meter I can see the mA's going up...

That's what happens. At FULL power, a 400V 8K amp will pull 120mA sine, 170mA square-wave (or gross overdrive distortion). Whether it idles at 5mA or 25mA doesn't really affect the FULL power condition.

And you didn't build it just to watch it idle.

So why care? Because as you come up from silence, or for guitar as you fade to silence, as you come down below ~~1 Watt the idle bias controls the gain and distortion.

> sometimes they don't have the harmonics as tubes about 5 to 10 mA off.

Yes, you'll get a rise in even-order harmonics as the power fades.

> I can hear tubes that are 15mAs or more off from each other

It's relative. When we ran 6V6 at 250V near 50mA, I doubt you would hear a 40mA/55mA or 38% mis-balance. When you have 5mA/20mA or 400% unbalance, assuming you play/listen in the soft zone, that's significant.

> those testers not being that great, I calibrated the tester but those resistors and pots are old

It doesn't even test the tube as an amplifier.

Look. Say I build a "Clarinet tester". All it does is blow a fan down the pipe and read the flow. If there's a sock jammed in the bore, the meter shows "dead"; if there's a half-inch of scum in the bore it reads "low". And of course a scummed-up clarinet has a problem. But if it reads "90%" flow, or "140%" flow, is that good or bad? Some players don't use full flow, they just don't need full flow. Flow is not "tone": clarinet makers tweak the bore for a tight or loose voice. My "tester" doesn't even say if a clarinet actually PLAYS, or if it plays like a frog.

> I take the results like a grain of salt.

Bucket of salt.

The shorts-test is valuable. If you can see leakage in a dark room, that's bad. Seeing the heater light-up is re-assuring.

Then I tap the TEST button. If nothing happens, hmmm. Check settings, try a similar tube. If the needle does move, and may wind-up near the green zone, that's good enough. The test condition (ALL non-cathode parts tied together) and the test voltage and target current are totally bogus for audio amplifier work.

> At least I could tell if they were weak or strong

I don't think this type tester even tells you that. Remember the Grid is the input and the Plate is the output... but this tester ties both TOgether! A tube with a dead plate and a grid that really sucked (or a weak plate and sucky grid) could read "OK" even though it would be useless (or poor) as an amplifier.

And tube sellers hate these things. The tester may expect 12AX7 to gush 3mA at 60V. Some tubes may "fail" that, but easily pass 1mA at 200V, the way we use them in an amplifier. Moreover, these bogus target currents were picked for cheap simplicity, and "may" exceed tube ratings. Prolonged leaning on the TEST button may damage perfectly good tubes. Some vendors (including our host) decline to replace "bad" tubes tested in these low-cost testers.

The only real test is in the actual apparatus.

A wide-open Champ with a few extra sockets will test most gitar-amp tubes in real-life conditions. To calibrate, run a handful of probably-good tubes through it. Go by the sound and the electrode voltages. Say five out of six 12AX7s read 170V-195V at the plate, one shows 110V and sounds different. That's your "good" zone for 12AX7. 12AT7 will be less-loud and may show 100V-130V at the plate; anything else is dubious. The 6V6 hole will take 6L6 and 6550, give cathode voltage a little different from 6V6 but pretty consistent for each type. Jump in 1 to pin 8 and you can test EL34. Add a 9-pin socket to test EL84 and 6AQ5. EL84 in a 6V6 bias will read a lower cathode voltage, but 95% of not-sick EL84s will show the same voltage within 30%.