# Power Transconductance Amp with Tubes, How?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.

#### CLS

A little background...

Supposedly most of you have read this:
http://www.passdiy.com/pdf/cs-amps-speakers.pdf

It's fascinating to drive those low Q driver with "additional power" to the region of bass resonance peak. Doing it elegantly with high output impedance. While "normal" amp can only provide a fraction of rated power on these low frequency high impedance peaks with their voltage output...

Why do I need this? I'm going to build an OB bass module, I guess this might be a good alternative to the "regular amp +massive EQ" combination.

So I'm thinking how to build this "power transconductance amp" with tubes.

Or, simply doing it by modifying an existed amp?

By reading the text book (or the likes), a transconductance amp can be formed by "current series" feedback. Simply put, taking the output current to feed the input.

So, here it is.

This is a simplified circuit of Altec 1568A:

You can see the NFB is formed by a separate winding on the OPT, and feed into the cathode circuit of the 1st stage by the 47R/1k5 divider.

Except for the independent feedback winding, it is a typical voltage feedback type, thus a "voltage amplifier".

Now I'm trying to modify it into a "transconductance amp" by:

Taking the output current to feed input, simple enough?

But how can I figure out the "??" in the cathode circuit?

Will this work?
Am I too naive on this?
Any foreseeable disaster?

#### planet10

Paid Member
My understanding is that a no feedback pentode is a natural transconductance amp.

dave

#### CLS

Thanks.

But pentodes have too much intrinsic distortions, wouldn't they?

#### jan.didden

Paid Member
I would look at the voltage you are feeding back in the voltage feedback case. Suppose that for a 1V output on the speaker you feed back 0.1V (I'm just making the numbers up). Then, in the current feedback case, with the same 1V output, I want the same feedback level. So, with the 1V into the nominal 8 ohms, I have an output current of 125mA. This 125mA through the ?? should give that 0.1V so ?? = 0.8 ohms. Sanity check: we are looking for 0.1 times the drive as feedback. With a load on nominal 8 ohms, we came up with a sense resistor of 0.8 ohms so this looks allright.

Then, at resonance, the nominal speaker load increases above 8 ohms -> the current through ?? decreases -> less feedback -> more output level -> that's what you wanted.

Jan Didden

#### planet10

Paid Member
But pentodes have too much intrinsic distortions, wouldn't they?

I believe that might only apply when using them in a constant voltage amp.

dave

#### jan.didden

Paid Member
planet10 said:
I believe that might only apply when using them in a constant voltage amp.

dave

Dave, I don't understand your comment?
The way I explained it, would lead to two amps, one voltage source, the other current source. Both would play equally loud at mid frequencies where the speaker impedance approaches a nominal value. Then, at resonance, where the speaker imp would rise, the current amp would increase the drive. That was the idea, no?

Jan Didden

#### planet10

Paid Member
janneman said:

:^)

I was actually commenting on the post before yours... i wasn't expecting you to be writing a post at the same time i was, so i didn't bothrer to quote. I've fixed that.

dave

PS: how is tomorrow morning in BE?

#### jan.didden

Paid Member
Tomorrow morning is drab, as is fitting for the season..... Just had a global-warming-blamed storm. I was hoping to lose a few trees that are too close to the house but they didn't give in

Jan Didden

#### CLS

agent.5, I found that article, and studying it, thanks a lot. (but I always have some problems to convert those circuits in the articles into reality )

janneman, you should be a teacher, and you can be a good one I understand you clearly and I'll follow that for my own calculations.

plante10, Actually, I once tried pulling out the feedback resistor of this amp and had a listen. (virtually no NFB)

It was very interesting, almost funny, and really scaring. I just hooked it up to my bass box (not an OB back then), and the sound was heeeuuuuge, vague, overly expanded, and almost no control at all. (And the gain is just too much)

Maybe I'll try it again with OB to see if any good

#### planet10

Paid Member
janneman said:
I was hoping to lose a few trees that are too close to the house but they didn't give in

Watch what you wish for... a tree thru the roof of the house is not fun... we had 2 last year that just missed my parents house -- it would have been nice if the one had fallen the other way and crushed the trailer (we just can't talk my dad into getting rid of it)

dave

#### Attachments

• tree-fall-2006.jpg
87.5 KB · Views: 542

#### SY

CLS said:
Thanks.

But pentodes have too much intrinsic distortions, wouldn't they?

At low bass, yes, that's a problem, not because of the tube distortion but rather the output transformer. For use above 100Hz or so, pentodes with a well-chosen operating point do very well in this application- I've tried it with EL34 loaded 4300 ohms plate to plate, 410V rail, 55mA cathode current per tube. At 1W/1kHz, the distortion was predominantly 3rd order and pretty low (IIRC, about 0.05%).

#### CLS

Alas, I NEED that current drive for bass!

And what's the problem in OPT?

#### SY

OPT distortion at low frequencies rises with increasing source impedance from the output tubes.

If you need to drive bass, then the feedback arrangements that you and Jan have discussed would be better. Also, look through some of the old literature for "variable damping factor," quite a hot topic in the late 50s and early 60s.

#### SCD

Hybrid?

I just read the review article in 6moons on Nelson Pass's F4. In the article the fellow discusses his impressions of driving the F4 from the outputs of a 45 SET. From his discussion I get quite excited about a hybrid amp using the F4 and a SET.
Does this fit the need

#### kevinkr

Paid Member
Something no one has mentioned is that impedance maxima in bass drivers occur at driver and box/port resonance(s) - in my limited (mostly to BR and Onkens) this is usually the point where acoustical output is greatest.

Don't confuse the need for lots of output current with the need for a current source output, they aren't the same thing.

I'd measure the driver impedance and acoustical output of your driver on an OB and then determine what combination of current and voltage feedback is required.

#### Tube_Dude

CLS said:

Taking the output current to feed input, simple enough?

But how can I figure out the "??" in the cathode circuit?

Will this work?

Yes. it will work.

With a 1 Ohm resistor (value of the ?? resistor ) you get 1 Amp output with 1 Volt input.

Regards

#### Wavebourn

kevinkr said:
Something no one has mentioned is that impedance maxima in bass drivers occur at driver and box/port resonance(s) - in my limited (mostly to BR and Onkens) this is usually the point where acoustical output is greatest.

I was waiting till somebody comments this point.

Also, what is "transconductance power amp" vs power amp: does a power amp suppose to amplify a power ?

#### SY

Understand what the transconductance amp is doing- at constant current, the output voltage at the speaker's resonance rises. So if you have a driver that is overdamped, the transconductance amp equalizes the bottom end. It's a pretty old concept, but the more rigorous examination of the other advantages (e.g., distortion reduction) by Mills and Hawksford make it attractive- for the right speaker.

#### Wavebourn

Constant current output is no more constant power output than a constant voltage one.

At lest a constant acceleration of a woofer's cone is closer to a constant acoustical power than constant current or constant voltage, does not matter which extreme to prefer.

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.