Tube Matching Overhyped?

[HotBluePlates]

The "rules" for output tubes are pretty well known: use matched output tubes or you risk hum, distortion, low output, less punch.  But how tightly do those tubes need to be matched?

It's a judgment call for sure, but can quickly get carried away and obsess over the degree of tube-matching, or waste money buying tube sets when the ones they have might be perfectly serviceable.

On another forum this topic came up, and knowing I had some 6L6s with widely different idle current draw I wanted to see what would happen.  Theory is one thing, but practical experience lends context & perspective.  So here's what happened when I used two 6L6s in a Super Reverb with nearly 61mA difference in idle current.

[SILVERGUN]

 The sky didn't fall & the ground didn't open up to swallow my heretical tube set.

... but can quickly get carried away and obsess over the degree of tube-matching, or waste money buying tube sets when the ones they have might be perfectly serviceable.

I can't wait til someone asks about a 5mA difference and I can reference this post.

How dare you try to take away the magic bullet from billions of amp tweakers?....with actual evidence?

I'm offended.

[Latole]

Gerard Weber's book say matched between 6 to 10 ma is ok . Amp sound is less sterile than if "perfectly" matched.
I often do bias close to that.
To hear some hum, distortion or low output you need a almost dead tube.

[BrianS]

How would running the mismatched tubes over an extended period affect the reliability of the output transformer?  How long would a mismatched pair of modern tubes last in any given amp?
"It's a judgment call for sure, but can quickly get carried away and obsess over the degree of tube-matching, or waste money buying tube sets when the ones they have might be perfectly serviceable."
I agree with you 100%, but when buying a new set of tubes for a customer's amplifier, I'm going with a matched set from a reliable source.  For me that is Apex Matched, Russian made tubes, almost always Tung-Sol or EHX branded ones as I've had zero failures with them as of yet. 

[Latole]

How would running the mismatched tubes over an extended period affect the reliability of the output transformer? 

If output tubes are not bias very hot ( more ma than OT specs ) you can't harm OT.

[Raybob]

I usually build with bias pot for each power tube.  The only real difference I notice with mismatching is with idle hum.  In my testing, more than a 5mA mismatch, idle hum increases.  Under 5mA difference, 0 db increase of idle hum.  Tests were with 6V6 and with 6L6 amps.

[Latole]

I usually build with bias pot for each power tube.  The only real difference I notice with mismatching is with idle hum.  In my testing, more than a 5mA mismatch, idle hum increases.  Under 5mA difference, 0 db increase of idle hum.  Tests were with 6V6 and with 6L6 amps.

Good to know. Class A or /and class AB amp ?

[j_bruce]

Mythbusters eat your heart out.

[vampwizzard]

I usually build with bias pot for each power tube.  The only real difference I notice with mismatching is with idle hum.  In my testing, more than a 5mA mismatch, idle hum increases.  Under 5mA difference, 0 db increase of idle hum.  Tests were with 6V6 and with 6L6 amps.

Two ways to attack it.. silverface fender amps with their hum balancing pots or separate bias pots.

Allegedly this wasn't a problem in the 50's and 60's due to tighter tolerances on tube manufacture? Then became a problem when mil-spec was abandoned? Im a little too young to have first hand knowledge of that era.

[shooter]

this wasn't a problem in the 50's and 60's

you didn't notice since it was the 2 prong, roll your own "hot side" single wire heater days

[PRR]

> The sky didn't fall & the ground didn't open up....

+1

[ALBATROS1234]

i dont believe leo or jim bothered to match tubes did they. i was wondering that when i built my supro 24 and i have 6,used 6973s i plugged them into each of my testers and split the difference , put the ones with the closest readings together which is a very approximate way of matching. it sounds golden. my favorite amp has a pair of original sylvania 6L6gb which are not matched and that thing is suuhhwheeet

[Raybob]

I usually build with bias pot for each power tube.  The only real difference I notice with mismatching is with idle hum.  In my testing, more than a 5mA mismatch, idle hum increases.  Under 5mA difference, 0 db increase of idle hum.  Tests were with 6V6 and with 6L6 amps.

Good to know. Class A or /and class AB amp ?

ABFender circuits

[vampwizzard]

this wasn't a problem in the 50's and 60's

you didn't notice since it was the 2 prong, roll your own "hot side" single wire heater days

  hey at least drinking til it sounds better is still universal.

[HotBluePlates]

I can't wait til someone asks about a 5mA difference and I can reference this post.

Gerard Weber's book say matched between 6 to 10 ma is ok . Amp sound is less sterile than if "perfectly" matched.

I do have a 1962 Deluxe with a pair of 6V6s matched to 0.1mA.  But that's more because I have a bunch of American 6V6s to choose from.  I think the first couple American Tung Sols I put in the amp just happened to land that close.

I had previously demonstrated in a thread over there that you could pull a tube from the socket and run only one side of the push-pull stage and still sound subjectively "okay."  This is with that same '62 Deluxe:
https://soundcloud.com/hotblueplates/2-tubes-vs-1-tube

Allegedly this wasn't a problem in the 50's and 60's due to tighter tolerances on tube manufacture? Then became a problem when mil-spec was abandoned? Im a little too young to have first hand knowledge of that era.

Did ya follow the link to my post?

The point was using one tube I knew was roughly-dead (it measures about half the idle current & 2/3 the transconductance that it should when put on my tester) and another tube I know to be defective (it idles near 2x the current it should have when fed a given bias).  I knew these tubes would be about as mismatched as you can get, and both are old American-made tubes.



Note that I'm not advocating anyone use a pair that's mismatched by 61mA, or one tube running 1000% of the plate current of the other tube.  I just hate to see people waste money needlessly because they think they're adhering to some best practice (but they've got it set in their mind as a Law rather than a Best Practice).

I've got 40 or so American-made 6L6 (type) tubes, and maybe 100 American made 6V6s.  In my measurements I've seen +/-20% variance around the "book value" for any given tube.  I don't think the tubes were necessarily tighter-tolerance back in the day, though the gross defects were likely screened out better.  And I think cathode chemistry was likely better, enabling longer useable life.  Beyond that, if you dig into them long enough you'll find the bad apples.

How would running the mismatched tubes over an extended period affect the reliability of the output transformer?  How long would a mismatched pair of modern tubes last in any given amp? ...

No idea.  I doubt my interest would remain long enough for me to do some kind of life test.

I do think you're right to give your customers matched tubes, because their amp will certainly perform better especially if they need maximum output power for stage performance.  I didn't get to crank my Super Reverb high enough, but I'm sure it would have distorted earlier with the mismatched set.

But the audience over at that other forum are mostly guys playing at home, buying too much amp, and then buying a Fryette Power Station or UA Ox to knock the sound level back down.  They don't need maximum output power, and instead thinking they're missing out on TONE! if they don't spend more money replacing serviceable tubes with a precisely-matched set.

The goal of the thread was to impart perspective. 

[shooter]

I doubt my interest would remain long enough

yup

Musician builders are quite set in their ways,  I've done a couple "blind" sessions with a few trying to dispel "myths", with little to no change.  the last one was fun, a 23W PP self biased.  4 of 4 would NOT believe the math, the scope, nothing, they all thought I was jerking their chain, because "THAT HAD to be a 40W amp" 

[HotBluePlates]

...  The only real difference I notice with mismatching is with idle hum.  In my testing, more than a 5mA mismatch, idle hum increases.  Under 5mA difference, 0 db increase of idle hum.  Tests were with 6V6 and with 6L6 amps.

I believe you observed the results you report.

In my case I had no hum in the Super Reverb when matched within 2.7mA, no hum when mismatched by 37.5mA, and a slightly audible hum when mismatched by 60.9mA.  However, that hum wasn't audible in the audio clip I posted (and I hear more hum always with my '63 AC30).

[JayCobie]

I've personally had a speaker thump/pop happen every time I took my Twin Reverb '65 Reissue out of standby when I had a mismatched set of tubes (about 7-10mA if I remember correctly) that went away completely when I installed and rebiased a closely matched quartet of 6L6GCs.
Playing that mismatched set didn't present any other noticeable problems though.

[HotBluePlates]

I've personally had a speaker thump/pop happen every time I took my Twin Reverb '65 Reissue out of standby when I had a mismatched set of tubes (about 7-10mA if I remember correctly) that went away completely when I installed and rebiased a closely matched quartet of 6L6GCs. ...

Standby switches are a sticky problem.  There's a 100% chance that your Twin's Standby will eventually thump no matter what output tubes you use.

Tone Talk with John Suhr

I personally don't use Standby switches on any amp, including my vintage Fenders.

[JayCobie]

That is totally fair. My issue may well have had other origins than mismatched tubes. Coincidence does not equate causation, after all.

[HotBluePlates]

That is totally fair. My issue may well have had other origins than mismatched tubes. Coincidence does not equate causation, after all.

That's what I thought.  At the same time perhaps it was related: could the different-current through halves of the OT have equated to an a.c. thump at the primary (then passed to the speaker)?  I have no idea.