OTL amplifiers
Doug Flynn
It's a tad quiet around here (the Government must finally have got around to
putting anti-psychotic medication in the water...LOL).
Anyway, output-transformerless (OTL) amps. I'm thinking of building one to
power my Quad ESLs. I like the look of circoltron designs using triodes*
(eg. 6AS7) a la Atma-Sphere.
Any thoughts, experiences or suggestions you want to share with us? I've
never heard an OTL so feel free to chime in with your impressions if you've
ever lived with one.
Cheers. Doug
* Cue Patrick Turner to start ranting about the superiority of pentodes...
Patrick Turner
I don't need to rave and rant about the superiority of beam tetrodes
such as the 6550 which Doug so ardently hates for a litany of quite
irrational reasons.
I have nothing against pentodes or triodes either.
There are some fundemental misconceptions about OTL amps. The first is
that where ppl are in favour of them, they like to justify leaving out
the output transformer which they claim is a veil or even a blanket
between the tubes and their ears.
They don't understand that even the most suitable triodes for OTL such
as the 6AS7/6080 or the 6C33C have an output resistance, ie, plate
resistance Ra, many times the speaker impedance.
They do not understand that to obtain an OTL amp with a good damping
factor one must use an enormous amount of global negative feedback. To
avoid the OTL tubes becoming severely overheated while connected to
loads for which they were never designed, the OTL tubes are biased
with very low idle current and thus can only be used in class AB very
close to pure class B operation where open loop distortion is
horrendous, and a lot worse than a pair of mosfets.
So in OTL you have a basic badly designed class B amp which is flogged
to death by a drive amp circuit working too hard and the whole shit-
pile is heavily corrected by NFB application, and you still end up
with an inferior amp that costs a fortune to run, and has woeful
efficiency.
If you have ONE triode connected EL34 pentode, it is usually set up
with Ea = 400V and Ia = 70mA and the ideal load for maximum triode
class A power is about 3.5k ohms and the power is 7 watts of pure
class A and it sounds exceptionally good. A transformer can match the
3,500 ohms to 8 ohms. The OPT ratio needed has a 437:1 ratio.
Now should anyone omit the transformer and still desire to use EL34
tubes and have them work in class A, and directly drive an 8 ohm
speaker they would need 437 EL34 all connected in parallel. They would
be able to get 3,059 watts in pure class A.
Anyone can use 8 ohms with ONE EL34 but you only get 0.02 watts in
class A.
If PP is used and class B biasing, the number of tubes needed without
a transformer is less but when you do the necessary calculations and
you do the load line analysis you'll find you need way too many EL34.
Even when you use 6AS7 or 6C33C the number of tubes is still way too
many and distortion and Rout way too high so huge amounts of NFB must
be used and you still end up with a far worse performing amp than
something with a transformer. There was a guy in Sydney who swore by
OTL circlotron with 6C33C. At least the 6C33C didn't burn out as quick
as when he used 6AS7. But finally the 6C33C gave up anf turned red
hot. So he asked me to design an OPT and when he used the pair of
6C33C in parallel SET, he finally got MORE power than the OTL, and it
was all class A power at low distortion and it sounded beautiful. He
finally had to face the fact that tubes need an OPT between them and
the speaker load just like a car needs a gearbox to allow it to drive
up a hill with 6 fat smiling police seargents.
I can see the gleam in Doug's eyes now that I have said OTL amps are
an abomination; He'll now try to build one just because I said it was
so crook.
The Quad-II amp was a reasonable amp for ESL57. My pages explain how
Quad-II amp performance can be lifted, see http://www.turneraudio.com.au
A Williamson amp with pair of 300B or pair of trioded 6550/KT88/KT90
would be fine for ESL57.
Patrick Turner.
Doug Flynn
>I can see the gleam in Doug's eyes now that I have said OTL amps are
>an abomination; He'll now try to build one just because I said it was
>so crook.
Actually, I am just about to build a Rozenblit amp which uses several 6C19
triodes in parallel. I'm not doing this to spite you; I'm genuinely
convinced it's a workable design. Have you actually *heard* an OTL amp
(apart from in your dreams)? Just because something is theoretically crap on
paper does not mean it will sound bad. SET amps using low powered directly
heated tubes look dodgy on paper, but they are orders of magnitudes better
than crap like 6550s.
Cheers. Doug
Trevor Lees
0400307489
"Doug Flynn" <doog...@bigpond.com> wrote in message
news:C9vIm.52290$ze1....@news-server.bigpond.net.au...