OTL Amp

[km6xz]

Juan, no fair introducing facts and engineering.....those old fashion uncool notions

"Tube frequency"? Can I get a translation on that...?

It appears the OP is building exactly the opposite sort of circuit that he thinks it is, the least "tube" sounding of tube topographies, and getting rid of one of the most tube sounding elements...the transformer. A bad transformer is usually more in keeping with characteristics that are assumed to be due to tubes.
The suggestion of a low level tube distortion generating stage, followed by a solid state current booster stage is probably closer to what would match his stated goal of "tube sound" as heard without a transformer but driving a clean current booster operated well under clipping.
A cathode follower would be clean and neutral, sort of hi-fi sounding, just the opposite of an instrument amp that is needed to supply both power and gobs of distortion, even in "clean".

[Steve Conner]

Here's my take on the "OTL" philosophy, which some of you may have seen before. It uses a complete Champ-sized tube amp driving a transistor output stage. (Yes, I found a couple of $5 power transistors a more attractive option than 300 12AT7s. )

The tube amp's output transformer is now an interstage transformer, but it still gives the "transformer saturation sound".

http://scopeboy.com/scopeblog/wp-con...HybridRev2.pdf
Ninja Corvette Hybrid guitar amp demo - YouTube (not really a fair comparison, as the camera can't handle the sound level of the high power setting, and produces some weird compression that some people blamed on the amp)

Finally, the guy who makes "Devil Dog" OTL guitar amps used to hang out here. They use 6AS7/6080 power tubes, and speakers connected for 32 ohms. I agree with Stan that he is missing the point.
SEMPER FI AMPS

[km6xz]

Transistors are great for current boosting, if you have the transfer function you want before feeding the neutral output booster. I am using the term "booster" not only because it is what it does but also since it gives the impression that it is not very influential on the tone that is created at lower levels.

Some years ago I built some Stasis amplifier which worked well. They were high power solid state amps designed to be hi-fi, in that they gave a faithful representation of the drive signal. But they were driven by a comparitor which managed the combination of feedback from the speaker, the feedback around the power amp and the output of a small optimized 6AQ5 class A or a part of sections of 12AU7's in a AB pushpull, transformer coupled mini amp or a copy of an old Champ. That isolated amp was driving an appropriate load and sounded, at low levels, how the large full amp ought to sound like. It was tricky to get the speed of the various loops fast enough with minimum phase shift but those who heard it were blown away by how a 10 second model swap changed the amp to sound like a 300 watt Champ, or early VOX or whatever distinctive amp desired.
I build one for a bass player that was 500 watts modeled on the transfer function of a Bassman, using 6AU6s. The breakup of the little original style 5" speaker was evident in the resultant sound of the main amp. That was very different from just adding a booster to a small amp, because the main speakers were reflected in the FB loop, as was the looser coupling of the reference amp stage and its small speaker loaded into a small damped box. I did not have the time or desire to produce them but there was a lot of interest from guitarists and bass players who wanted to reproduce the feel and response of some classic small practice amps.
It would be easy to whip up a small reference amp scaled any way needed, using a variety of low power tubes driven as, proportionally, hard as large tube amps were and getting the same sort of sound. It is pretty easy to come up with 75-300volts at 15ma, and transformers are cheap, light and small in that scale. That was before the massive but light power available from Class D and switching supplies. Now, such an amp could actually be built cheaper than the conventional heavy iron way, besides being 1/10 the size and weight.
I am the distributor for ZT amplifiers in Russian speaking countries and have a LunchBox on my bench for my own utility amp functions. The weigh of that little loud amp is most from a large magnet and a moderately large transformer. It sounds really cool stock but as an experiment, I took out the transformer and installed....temporarily...a switching supply and 100watt Class D amp. The 4.5kg dropped in half and it sounds just as good. I did not have to build anything, just off the shelf modules and an hour. Now, to build a Electro-Dynamic speaker using the high current capabilities of tiny switchers for the field coil. Imagine a 100-200 watt 1-2 12s combo with 10 lbs. and almost all of that being the case? It is all back to stock now because I have other uses for the supply and amp module but it was an interesting hour experiment. I really do not have the time for serious construction work, nor the machine tools to do a good job. There are priorities in life....partying, dancing, 28y.o. beautiful tall, slim surgeon's(this last weekend's lucky find), good food and better drink...etc.;>)

[J M Fahey]

Cool experiment.
So they have separate pre and power PCBs?
Do they use a chipamp?
Can you post a couple gut pictures?
Very interested in packaging and available space distribution in such a small amp.
If not possible because of contractual constraints, please send the microfilm to Praha disguised inside a smoked sturgeon can, where our agent .047 will smuggle it out of the Country sewn into her underwear.

Pssst !! you ... yes, you ... pretend we are not talking .....
That 28 y.o. "surgeon" is really the dangerous KGB agent Tatiana Románova, she's setting you a trap.
"They" will film you both through a two way mirror and blackmail you.
Tell her that "the secret" was sent to Buenos Aires.
She'll follow it and when she arrives I'll deal with her for good.
Don't worry, we are backing you up.

[Steve Conner]

I like the ZT amps. I got to try the larger Club model in my local guitar shop, and it sounded really tasty with a hollowbody jazz guitar. It drowned out the usual annoying metal shredding kid completely. My Ninja Corvette is a kind of low-tech tribute to the Lunchbox.

Tell her that "the secret" was sent to Buenos Aires.
She'll follow it and when she arrives I'll deal with her for good.

Who said the electronics business was boring?

[ba3jar]

a request from J M Fahey to post this question :

you said 4mA plate current at a saturation voltage of 50V , that for one plate of triode , so you saying using two triodes the total would be 8mA of one 12at7 on 50v ?

i have other calculation hope you could correct me:
from the data of 12at7 , current of 0.019A , voltage 200V , on one triode , 8 tubes you have 16 triodes , 16 into 0.019A= 0.304A
0.304A square into resistant of 336 ohms = 31watt peak = 22 watt rms ?

[J M Fahey]

a request from J M Fahey to post this question :

you said 4mA plate current at a saturation voltage of 50V , that for one plate of triode , so you saying using two triodes the total would be 8mA of one 12at7 on 50v ?

i have other calculation hope you could correct me:
from the data of 12at7 , current of 0.019A , voltage 200V , on one triode , 8 tubes you have 16 triodes , 16 into 0.019A= 0.304A
0.304A square into resistant of 336 ohms = 31watt peak = 22 watt rms ?

To predict what a tube does, we must refer to its datasheet.
I´ll use a classic one:
http://www.drtube.com/datasheets/12at7-rca1954.pdf
for this problem we´ll refer to the "Plate Characteristics" on page 3 , which show all possible combinations of Plate Current vs Plate Voltage vs Grid Voltage. All 3 interact.

Let's design our OTL stage.
Basically 2 triodes in series.
At rest, each one has +B/2 from plate to cathode .
When one of them saturates, the other one supports the full +B, which sets our first limitation: maximum +B=300V .
So at idle each one stands 150V.
So max (peak) voltage swing is 150V ... but in practice is less.
We want the triode to pass useful current when saturating, but triodes are very poor at that: low plate to cathode vootage (saturation)= low plate current.
A compromise value must be found: I chose saturation voltage of 50V, which yields a plate current of less than 4mA, *at* 0V grid voltage.
When the other triode saturates , same applies.
So we have tubes which can swing from idle to no closer than 50V either to ground of +B.
We are finding roadblocks everywhere!!:
limited peak current capacity of 4mA per triode; limited voltage swing of 200V PP and that into a quite high impedance load.
How high?: (delta)V/(delta) I ("delta" means "variation" or "difference between 2 values")
Optimum load impedance for a pair of triodes (contained within a single 12AT7): 100V Pk/0.004A=25000 ohms.
As I said before, the *terrible* limitation here is having a very poor current capability.
The 19mA you mentioned, are available only with 200V plate to cathode Volts, which we simply do not have.
We *might* have them at idle and with 400V +B, but as soon as those triodes start to swing either side, we lose current capability.
That's the karma of triodes and why pentodes were invented.
As a side note, Bipolars, Fets and MOS behave like pentodes, meaning they can pass huge current with quite low voltages across them, which makes them much more efficient.

In a nutshell?
You like 12AT7s? Fine with me.
Make a *voltage* amplification stage with them, let them clip, overdrive, change duty cycle, block, whatever.
Just doing their thing.
And *then* take that signal and cleanly reamplify it by whichever classic means to drive speakers.

[Chuck H]

I'm automatically inclined to agree. You really seem to be with the program. Earlier I had some spam bot it's not agreeing with me.

[J M Fahey]

Normal day start: a nice steaming big coffee cup, in front of the monitor, reading what a spambot "thinks"

[J M Fahey]

^^^^^^^^^^^^^^^^^^^^
These stupid spambots are scratching me the wrong way.

[Chuck H]

Steve is telling us not to respond to the bots. It makes it worse somehow.

[J M Fahey]

OK, thanks.
Will do.

[km6xz]

It can be boring for sure....instead of out partying, I am trying to trace down a periodic shut down on a friends Ampeg SVT-7 Pro...errr!
It has a dozen triggers for shut down in its multi-sensor protection circuit and all seem to be doing what they were intended to do, yet after the solo tube in it warms up it shuts down every 15 seconds. No unusual operation of the preamp, it is not based on the current pulled by the tube or signal conditions, (substituted a heater supply, and later a 300volt plate supply) and it still shuts down if the tube in warmed up. Runs full power nicely into loads when sending the signal into the power amp in jack, confirming that the switching supply and Class D amp are working as well as they test.
I would rather be out playing.....

[J M Fahey]

What sensors are measuring anything related to the tube?
or
disconnect sensors one by one, until the protection does not get triggered, then find out what it's detecting

Just thinking out loud, of course.
Good luck.

or go out play with Svetlana in the snow.
No, not the tube factory

[Chuck H]

A friends amp is ALWAYS the one that turns out to be tricky and time consuming.