mismatched 6v6 pairs; cathode bias

[deci belle]

I have a bunch of mostly old USA-made 6v6 tubes I'm swapping in and out of my 5e3. The tubes' temperatures are ranging between 230˚ to 360˚ while playing (ie, one tube might be at 230˚ while the other is at 360˚). They seem to heat up a little bit higher that that when I'm not playing. Amp is idling @ 13.4w/ 370v. I'm using real 5y3 rectifiers.

Running a mid 90's EH6v6 (out of a Fender RI) with an old Sylvania 6v6 last night sounded particularly good, but am I risking harm to my OT by running these mismatched tubes? At first I thought cathode-bias would even things out, but now I'm not so sure… because the plate voltages are a little different between the pairs.

[Chuck H]

For starters, part of the reason it sounds good could be because of the mismatch and not entirely the brand pairing. So you do need to keep the mismatch to keep the "sounds good". Since tone is subjective, how much mismatch sounds good is also subjective.

The shared cathode resistor does a little balancing. Mostly at the extreme swing ends. Not as much at idle. Worth noting is that you won't really notice hum when playing, but you will at idle. So the greater standing imbalance at idle will hum. The greater standing imbalance may also be why you get your greatest temperature differential at idle.

Nothing wrong with using mismatched tubes if the amp is up to the task. Yes, you can damage the amp by doing it. Not because of the mismatch itself. But because a mismatch can put a lot of stress on the tubes. And tubes can fail in interesting ways. Taking out other components. Especially true if the amp is built vintage style with no protection circuitry.

FWIW even tubes that are "matched" by current standards often show a mismatch in actual working condition in guitar amps.

I've never known temperature to be used as a matching consideration. And I don't think it's very accurate. Different construction on different brands will have too much affect of this parameter. Good idea for a rough check though. And to be sure, the hotter tube is likely to be the one doing more work.

Maybe add the often seen diode string protection to the OT primary leads and/or hefty conjuctive filter with a small cap size (just enough to pinch spikes) would probably help greatly. Then... Smoke 'em if you got 'em.

[deci belle]

Smoke 'em! haha!!

Ya, I checked and a Westinghouse/EH was what I had in there. The boring-est pair is a set of matched Stromberg-Carlsons. I was measuring temps just so I had an idea what was happening~ hadn't thought about pairing temps, though (the temps I listed were the most extreme examples I'd measured out of eight 6v6 tubes).

I do have diodes on the rectifier socket PT input pins (4+6); and a .1µF 600v tube cap from the OT primary to GND.

Is what you mean by "often seen diode string protection to the OT primary leads" something else, Chuck? If so, please tell me about that!

I don't know what the hefty conjunctive filter is, but I have a doghouse on the back of the chassis with two 47µF 500v caps in series for most rectifier tubes (I can switch in a pair of 100µF 350v caps when I use a 5ar4 w/ JJ6v6s).

Thank you for the response, Chuck!— please tell me if what safety mods you mentioned are different than what I've done to this amp so far— because it's my first scratch-build.

[Chuck H]

A conjunctive filter is a zobel network (which is a fancy way to say cap in series with a resistor) across the OT primary plate leads. IIRC the standard is a resistor of 1.4X the primary Z. The cap can be chosen in a small value just to tame voltage spikes or larger values to actually affect the amps tone. The resistor and cap specs are important. Resistors don't generally like high voltages so you need to be sure that the resistor is rated at the amps wattage and able to handle the plate voltage. This works fine for the resistor. The cap needs to be rated for about four times the plate voltage. I think this is because of specific AC conditions in the cap for this circuit. I just use a 2kV cap. The cap value will depend on the desired result and the OT primary Z. To just tame voltage spikes in a 5E3 I would use a 10k 15W to 25W resistor rated for higher voltages and a 2kV 1500pf cap.

The diode string protection circuit is just three 1n4007 diodes in series from each plate to ground. All cathodes toward the plate. This circuit is supposed to divert amp killing spikes away from the power tubes and OT.

Other than these things I would say that a burly OT is a good idea. I have a burly OT in my own amp and I dime it basically every time I play for the last four years.

Mismatched tubes aren't that different or new. Back in the day Marshall and Fender just bought tubes and stuck them into the amps. Matching was for expensive hi fi stuff. his could be one reason some vitage amps sounded great and others ordinary. Most of the time two tubes from the same MFG made about the same time won't be grossley different. I've never tried intentionally extreme mismatching, but, whatever gives you the sound you like is fine. The circuit should be protected and the OT should be able to take moderate abuse and you shouldn't have a problem. I read somewhere that harp players have been intentionally imbalancing AB1 outputs for a long time to get the tone they like.

[deci belle]

Oh— that's great, Chuck!! Thank you for going into detail describing the conjunctive filter and the diode string!! If I have any questions I'll be back!

[deci belle]

The safety mods suggested have made me feel a lot better about mixing tubes— not that I'm going to any extremes. I eventually removed the conjunctive filter and substituted a gas discharge surge suppressor because the cap in the conjunctive filter was noticeably syphoning away higher frequencies.

One thing I heard about but wasn't clear about installing was a fuse in series with the cathode's ground. I did it with a fixed-bias amp, but with the cathode-bias of the 5E3, I was too intimidated to pursue it on account of all the other things going on in that circuit!

[pdf64]

I prefer to use individual cathode resistor per tube, rather than shared; each tube then finds its own idle point and most all tend to idle around a similar cathode current, although the cathode voltage might vary a little.
It may be seen that they 'self balance' to a suitable static condition.
Each cathode resistor will need its own bypass cap, and the cap value may need to be larger, because there's no 'push-pull' cancellation occurring.
And the resistors will need to be larger than the shared value, twice or 4x, depending on whether it's a pair or quad.
Sound-wise, I can't hear a difference between the shared and individual cathode bias arrangement, though the slightly differing cathode voltages should impact on the clipping point of the control grid signal; also the larger bypass cap value may also affect the overdriven tone.
Pete

[Mick Bailey]

Mismatched tubes aren't that different or new. Back in the day Marshall and Fender just bought tubes and stuck them into the amps.

Marshall still does this. When you buy a 'set' from them, they're individual tubes. The extremes probably get filtered out in the supply chain.

[Mick Bailey]

+1 on individual cathode resistors.