AudioBanter.com - View Single Post

wrote:

Patrick Turner wrote in message ...

Patrick Turner wrote below: "Most pentode/tetrode guitar amps are mostly
class B amps, with a small % of their power in class A."

I have never seen such a statement in any EE book?

And you probably won't ever.
But one has to describe the distinctions between class A and class AB somehow.

You are accidentally introducing some confusion, Patrick, by
re-defining operating class away from engineering criteria and toward
audiophile aesthetic criteria. In engineering terms, there is no such
thing as you describe, though many of us here do know what you mean by
it.

Its quite difficult to describe pp amplifier operation to people who
don't do the research to find out from text books, and never bother
to examin of the wave forms in the amp.

You also refer to "bias current", another well-intended confusion in
terms. Surely you mean the idle plate & screen currents of your class
A amps, which have their max dissipation at idle ("where they are
biased")? Tubes operating in A or AB1 draw no bias (grid) current at
all, until/unless they start to sh*t the bed.

By bias currents, I do mean the idle currents in the electrodes
The class A amps have the highest bias currents, lowest Ea,
while AB tend to have higher Ea and lower Ia.
Class B amps theoretically have zero bias current,
but in fact there are no real class B amps. All class B amps
have some minimal bias currents to minimise crossover distortions,
and so are technically class AB amps.
But such amps which are biased with the minimum Ia, just enough
to keep Xover Dn low are regarded by most as substantially class B amps.

The only valid definition of your A, is that the tubes are biased so
that both exhibit signal flow for 360 degress with no cutoff during
any part of their operation. Any & every amp meeting this criterion,
tube or solid state, is A.

The cut off character as I said in other posts is not always sharp,
especially with PP triodes, but a triode class AB amp with minimal
bias current, and where substantial cut off occurs
could never be said to be a real class A amp. Its an AB amp.
The criteria is the distortion in the tube current flow, and once the
2H exceeds 5%, or what it would be if each of the PP tubes
was loaded by 1/2 RLa-a as an SE triode, then you have moved from class A to AB.

The first 20% of the total AB watts is produced by class A action.
The final 80% by one of the pair, then the other, working
substantially alone in the power production.
The load seen by one tube during the class A part of the working is
1/2 RLa-a, and during the cycle when the other tube cuts off
the RL seen by this tube becomes 1/4 RLa-a
The class AB amp tubes see a changing load value, and the load line for
one of the tubes is not a straight line across the plate curves,
but a kinked or bent one.

Anything else beyond this, while perhaps
of sonically descriptive interest to audiophiles, is irrelevant to its
class, and sows confusion in the guitar amp world.

I suggest the "guitar world" ppl do some serious study about how amps work.
I have no arguments with the opinions of audiophiles or musicians about the
sonic aspects of the sound they hear.

I am not confused about how amplifiers work, and
its because I have taken the trouble to find out by
years of study, and experience building and repairing them.

Perhaps the eventual grid current problems you are seeing/refer to are
caused by eventual cathode stripping?

Some say its leakage across the mica, which becomes ever so slightly
changed / contaminated, and when I see 0.5v across a 100k bias R, I
know I have 0.005 mA of current flow.
If the grid is at -35, and screen /anode at +400,
then there is 435 v across the mica, so the resistance is
87 Meg ohms.

We agree about the appeal & tastefulness of clean playing (mentioned
elsewhere in your comments). Many (not all) players arrive at such a
point after a lot of experience & maturing of technique. But this
does little to assist those trying to kick butt with a 400-PS, which
some styles require.

Kick butt music has its place, but I am speaking to the whole group,
and discussions would be rather boring if we stayed rigidly
to the one subject, and never strayed.
Some will say straying OT is boring. I can't please all the ppl all the time.

I think it is becoming fairly obvious that people are cooking 6550's
by exceeding design plate and/or screen dissipation ratings under full
power, combined with a fair amount of mechnical punishment.

Some folks are, for sure.
I have seen "courageous" commercial designs of amps which
boast huge powers from tiny boxes, and I am one who thinks that trend
is all BS.

A pair of
6550's is in fact good for 100w into a correct load & fed properly,
and sometimes a bit more.

Well, good for PA, or a muso.
Not for hi-fi. The thd is too high. To get 100W or more from any 6550
means you cannot have a worthwhile amount of class A power.
The amount of class A power before the amp goes into AB mode
is trifling.

Certainly 135w is beyond the pale. But the
6550 is essentially a 6146B, and all these tubes are notorious for
being subject to damage from excessive screen voltages, just as you
correctly imply elsewhere. We proved all of this by experience in the
'50's & '60's, and it hasn't changed any, no matter who makes the
tube, as you also imply.

Well, some tubes have had the screen construction changed
so it can take more current, and its distance from the cathode increased slightly.
The 6L6 which first appeared had a low 300v screen rating, but the 6L6GC
will run with higher ratings.
I found the EH 6550 runs quite well with screen voltage at 300 to 350v
and the plate voltage could be anything up to 800v.

The load *must* be increased as the Ea is raised, unless one runs AB2.

For low values of screen voltage and AB1, the load must be higher value
so the negative voltage swing can be right down to
the knee in the Ra curve at Eg1 = 0V.
So if you had a pair of 6550 with Ea at 800v, and Eg2 at 330v,
then the RLa-a would be 10,666 ohms,
and the peak voltage swing would be 700v,
so the a-a rms voltage is 1,000v, and po = 93.7 watts.
If the screen voltage was lifted to 440v,
then knee in the plate curve is raised somewhat,
and a load of 8k could be used, and a-a v swing is 920v,
and po = 105 watts.
In this latter case, the max AB1 plate current is 350mA peak.
But for either of these AB1 conditions, the load line for one tube
is positioned outside the plate dissipation curve of 40 watts, and
the 8k load is well outside the Pd limit of 40 watts.
Pd = 40 watts is a conservative design figure.

EL34 were used with Ea = 900v, and Eg2 = 450v, and with a 12k RLa-a.

This gives a swing of 1,080 vrms a-a, for po = 97 watts.

If 6 k is used instead of 12k as a load, POOF!,
the tubes turn red very easily!

The max current of the 6550 is about 500mA,
and to get that one can use AB2.
One could have Ea = 600v, and Eg2 at 450v, and RL = 4ka-a,
and the AB2 v swing would be about 678 vrms,
and po = 115 watts.

If Ea = 800v, and RL = 5.3 ka-a, idle current = 30 mA,
then v swing is 961 vrms a-a and po = 174 watts.
peak plate swing is 679v and max peak current is 0.512 amps
From page 589, RDH4, and equation 15, at the bottom of the page,
we get the plate input power for 2 tubes
= 800 x ( 0.637 x 0.512 + 0.5 x 0.03 ) = 273 watts.
If output power is 174 watts, the plate efficiency is 63.7%,
a realistic figure.
The plate dissipation at 174 watts of output = 273 w - 174w = 99 w,
so about 50 watts is the dissipation per tube.

At 116 watts of output, the peak plate current is 0.418 amps,
and plate input power is 225 watts, so
plate diss is 109 watts, so 54 watts per tube.

This is all based on an idle Pd of 800v x 30mA = 24 watts of idle dissipation.

Now if somebody connects
3ka-a RL to the same amp, and we had 0.7 max peak amps in the tube,
the input power is 368 watts, and output power was 239 watts,
there is 64 watts of heat in each tube, and the anodes would fry.
At 134 watts into 3k, the peak tube current is 0.466 amps,
and Pin = 316 watts, so Pd = 182w, so each tube has to cope with 91 watts of Pd.

Should someone bias the tubes up at idle so there is say 35 watts of Pd
instead of the 24 watts, it makes matters worse.

Quite clearly, it can be seen that a quad of tubes would be required for the
174 watts of power to afford some reliability, because someone, some time
will connect a load that's too low, and cause serious damage to the tubes.
I would place the reliable design limit at 75 watts per pair of 6550 or KT88.

I have to wonder why someone needing a 400w reliable bass amp in 2004
is still fooling with tubes at all, but that's another can of worms.

The musicians don't like the sound from transistors, and that's why tubes are used.
But at the really big gigs, the sound comes from a tube amp, then is amplified
by the main solid state system, and the Stones have used around 140,000 watts.....

A guy I know has a system he hires out for dance parties,
with about 100 drivers in two channels forming towers about 8 metres high,
and he uses 2 x 1,200 watt SS amps with smps, quite light and small.
I visited him at a local party he had in a rock climbing venue,
and I couldn't hear myself think inside, and couldn't hear what anyone said to me.
Even with ear plugs, I could only take 5 minutes.
He'd love to have tube amps, but he couldn't afford to have me run up a couple.
The cost per watt plummets when you start building 1,000 watt tube amps,
and someone dedicated could make a 1,000 watt amp a lot cheaper
than something brand new from Crown or Crest.

But such amps are mainly all class B operation.

For class AB with at least 50% of the power in class A, the costs
soon rise.......

Patrick Turner.

Maybe they should build one with 813's. :-) 813's have black plates,
toohahahah. How's 600+w from 2 bottles sound? I suppose with 2.5kv
on the plates, they wouldn't dub with it so much...

813 are not bad for some real grunt.

But there are plenty of quite small triodes with about 400 watt
plate dissipation ability providing one
forces enough air from fans through their copper
anode heatsinks These also need a highish voltage,
but some quite extraordinary power is available.
These triodes are normally used for transmitters, and are in current production,
but they can be used for AF. The ceramic-copper power tetrodes would give heaps.
At large gigs, the sound of the fans is less than the sound of the fans in the audience,
and less than the "music".