Brett said:
Agreed...
I'd go one further, and say that matching is rather over-rated. Matching was required to get the idle current close in an amp w/o ability to adjust individually. Build in that capability and it isn't all that necessary. Even less valuable with SE( but that isn't something I do anyway...heh-heh-heh ).
cheers,
Douglas
hilbert_mostert said:
Correct.
Or even individual tube bias (for fixed bias amps).
True about the matching points. An example is beautifully matched quads of 6L6 used in guitar amps.... at their recommended 455V or so B+ and 36mA each, they will be within a mA or so of each other.
BUT... if you want to bias them colder (or hotter), the difference between them *really* can become pronounced.
Cheers!
Most "auto bias" circuits use a common cathode resistor for a pair of tubes. These should get a matched pair. For stereo, two matched pairs. If four share a single cathode resistor (or a single bis supply) a matched quad is best.
With a batch of 100 new tubes, they're probably all pretty well matched. With 100 NOS tubes, (unless it's an original factory batch...) they're probably 10 different manufacturers made over a 20 year period. So picking out matched ones will be time consuming, and will leave a lot of singles... Actually the NOS ones still left may be primarily the singles left over from years ago... the tails of the bell curve.
With a batch of 100 new tubes, they're probably all pretty well matched. With 100 NOS tubes, (unless it's an original factory batch...) they're probably 10 different manufacturers made over a 20 year period. So picking out matched ones will be time consuming, and will leave a lot of singles... Actually the NOS ones still left may be primarily the singles left over from years ago... the tails of the bell curve.
Good... so if I, for any reason whatsoever, would like to have a fixed bias using potentiometers in the schematic, then I would have to have four, one for every power tube in the amplifier.
( This is not something I would do, too much hassle when the tubes need to be replaced ;-) )
I thought the matching was done because the tubes had different tonal characters and was not aware that that was because of the idle current.
( This is not something I would do, too much hassle when the tubes need to be replaced ;-) )
I thought the matching was done because the tubes had different tonal characters and was not aware that that was because of the idle current.
hilbert_mostert said:
They can charge you more for those matched sets. Also, there are more EL** and KT** based designs out there than 807-based designs. The other question is what are they matching? Is it plate currents for designs that use cathode bias? Or is it dynamic characteristics such as g(m)? How long will they stay matched once put into service? All these nasty little problems seem to be overlooked.
You are better off with a design that allows for some bias adjustments in order to equalize plate currents with PP pairs. You are never going to balance exactly dynamic characteristics such as g(m) and r(p). Not gonna happen. You can get reasonably close, however, if you stick to one brand, such as not putting an RCA in one hole and a Sylvania in the other, for example.
If I'm building a PP fixed bias amp (Poinz' EL34 Music machine), is there any downside (other than doubling the number of adj pots req'd) to having 4 individual bias adjustments instead of 2 pairs?
If having individual adjustments allows more flexibility with regard to tube matching, why do most designs only adjust in pairs? To save the $$ for extra pots and resistors?
I actually made this mod on my breadboard the other nite and it seemed to work fine, although I have since undone the mod while chasing another issue.
If having individual adjustments allows more flexibility with regard to tube matching, why do most designs only adjust in pairs? To save the $$ for extra pots and resistors?
I actually made this mod on my breadboard the other nite and it seemed to work fine, although I have since undone the mod while chasing another issue.
Sorry, but....when tube works, input amplitude must be = [Ubias].
For normal work PP , -Ubias 1 tube must be the same -Ubias 2-d
tube, it mean that Ia must be the sames. Ia depend Ubias.!! It's and for SE! If the tubes are no matched Ubias or Ia
are not the sames and tubes works different . /sorry for my bad engl./.
For normal work PP , -Ubias 1 tube must be the same -Ubias 2-d
tube, it mean that Ia must be the sames. Ia depend Ubias.!! It's and for SE! If the tubes are no matched Ubias or Ia
are not the sames and tubes works different . /sorry for my bad engl./.
A host of tube vendors will match tubes and sell'm just on quiescent currents with a mere mA diff. A real time waste. No-one really bothers on the more important AC gain (transconductance) when real signal current flows. My experience this can be 50% out, so the DC with signal makes the conceptual quiesent value look silly.
richy
richy
Real matching is hard, especially when you buy Russian tubes, but if you buy RCA or Telefunken from the same batch they are already well matched. I use 3-point measurements to match tubes.
I tried asymmetrical combination, when only one pentode was passing the signal while another one delivered the same quiescent current balancing output transformer's bias. However, it was class A amp, but it sounded better than a complete P-P class A one though output power was lower. Later I replaced that "idle" tube with a SS current source, the result was the same. But after I counter-modulated that SS CCS I got the same SE sound, but more of output power (Alligator project).
In PP amps matching helps to get the last fraction of power from tubes that is not audible anyway, so better save your money for better tubes.
I tried asymmetrical combination, when only one pentode was passing the signal while another one delivered the same quiescent current balancing output transformer's bias. However, it was class A amp, but it sounded better than a complete P-P class A one though output power was lower. Later I replaced that "idle" tube with a SS current source, the result was the same. But after I counter-modulated that SS CCS I got the same SE sound, but more of output power (Alligator project).
In PP amps matching helps to get the last fraction of power from tubes that is not audible anyway, so better save your money for better tubes.
Matching can be vital if you replace the output valves in some guitar amps.
not for the sound, but - anode (plate) power dissipation.
I found this out the hard way when I put four unmatched JJ EL84 in a Vox AC30 recently. This amp takes EL84 x 4 (each rated 12W anode disspation) and runs them at 12W each! The unmatched JJs measured 20% or more variation across the 4, so that one was dissipating ~15W... this would make for a very short lifetime.
Other, more recent amps run EL84s even hotter (call it bad design).
As a general observation, modern valves seem to draw more current for the same cathode resistor (or bias voltage) than the original (1960s..1980s) designs expected. This makes the matching problem worse (less margin on Pa). I have often had to increase the cathode resistor in EL84 amps.
not for the sound, but - anode (plate) power dissipation.
I found this out the hard way when I put four unmatched JJ EL84 in a Vox AC30 recently. This amp takes EL84 x 4 (each rated 12W anode disspation) and runs them at 12W each! The unmatched JJs measured 20% or more variation across the 4, so that one was dissipating ~15W... this would make for a very short lifetime.
Other, more recent amps run EL84s even hotter (call it bad design).
As a general observation, modern valves seem to draw more current for the same cathode resistor (or bias voltage) than the original (1960s..1980s) designs expected. This makes the matching problem worse (less margin on Pa). I have often had to increase the cathode resistor in EL84 amps.
richwalters said:
Please don't say no one does matching of Gm - I do.
Responding to the other posts...
1. It's not an economic issue. Matched tubes cost more because of the time invested in the matching process. I could do much better productivity-wise if I didn't have to match tubes. Yes, the matched tubes are SLIGHTLY higher priced (maybe a buck or two for a quad), but it doesn't cover the costs of doing it. So Miles, your accusation doesn't hold water, at least as far as I'm concerned.
2. Why no matched 807s? As Tom Bavis said, there aren't enough out there to match them. There is no sinister unethical reason I have KT-88s matched and not 807s!
3. Actually, the current production tubes are as tight as the old stock in terms of "out of the box" matching. But both will have variances of as much as 100% in cathode current and 30% in Gm.
4. Stu Hegeman (the Citation series designer) said there was a statistical probability of matched tubes aging more evenly than unmatched tubes. He was pretty darn smart. I give his statements credence.
Honestly, it really bothers me when people accuse me (by making broad sweeping generalizations) of borderline thievery:
richwalters said:
The people YOU deal with may overlook them, but I don't. And I don't appreciate that sort of comment... there are people out here who do a very good job, the sort of comment you made is inaccurate and just serves to damage the reputations of good vendors.
I can't speak for others, but don't go painting me with that broad accusatory brush. I don't deserve it and you have no basis to say it does.
Your da*n right I'm mad.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.