Idea for an ECL82 PP headphone amp.

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi their all,
After my last power amp upgrade I have a few ECL82's which are now redundant. I currently have a parafeed partial feedback ECL82 SE headphone amp, but my last experiments have show the weakness of this valve in this circuit. So I was thinking of a new PP headphone amp.

The idea is to use the two triode section of a pair of ECL82's to drive a microphone line driving transformer in parafeed. This steps down the voltage gain of the triode by 10x which should make it more reasonable to deal with the excessive gain. It will also do the phase splitting for the two triode strapped pentodes. The phase splitting and output stage is old hat and I can cope with that. What I want to know is how best to set up the two triodes to give the best drive to the microphone transformer.
I originally thought of using them in an SRPP, but after reading the TUBECAD article, I realise that this is a totally inappropriate application for them. Of course I could just parallel them - but this risks imbalance blurring the signal - seems to be a poor choice - better to stick with a single triode and let the other go to waste.

The question is does anyone have any good ideas how to get the best out of a pair of triodes as a driver ?

Shoog
 
Hi Shoog

Only thing I can think about is a MU stage with one triode as driver and the other as load. But I doubt if that helds any benefit.

Actually I am replying with a question on your planned use of the microphone line driving transformer. I have seen several microphone tranformers, but they can all handle just a couple volts AC, maybe 2 or so. Will you be using one of those?

Erik
 
Ahh the choices...

You could 'equal load' the triode using a low VA dual primary torroid - one primary in the anode circuit and the other primary in the cathode. Top off with a smallish R to provide Bias (remember it has to be the same top and bottom, and you have a nice PP stage, with balanced out.

Or

Operate the triodes in inverted mode, feeding the pentodes - hammers thier high gain, and provides lots of current drive for CMiller (which is low anyway, because it's a 'tode) - They can be fed from the transformer further upstream, where you're unlikely to encounter saturation issues

Or

Use one triode as a CCS on the second triode to linearise the curves, take a parafeed output to a TX that drives the pentodes

Or

use the 'tode as a CCS for the triode, operated in inverted mode - drive as a PP pair to a transformer

Or

use the 'todes as a CCS for the triodes, operated normally with a cathode follower output

Or, or, or


see, lots of choices


Owen
 
Operate the triodes in inverted mode, feeding the pentodes - hammers thier high gain, and provides lots of current drive for CMiller (which is low anyway, because it's a 'tode) - They can be fed from the transformer further upstream, where you're unlikely to encounter saturation issues
Sounds promising.
However if I put the phase splitting TX before the triodes, I would have a step down situation and would need quite a lot of gain to drive the pentodes. I would guess that I would need at least 20x gain in this configuration. What could I expect from the inverted triode ?

Shoog
 
This is my latest thinking.
Use the microphone transformers at the input to phase split and step down. Send the split signal to the triodes (in normal mode) and have them as a LTP with both inputs driven by the transformer. Use this to directly drive the triode strapped pentodes.
The question is a) will this work b) will a 1K cathode resistor be sufficient for the LTP's cathode load. If not can anyone suggest a 2mA CCS for the job.

Shoog
 
Hi Shoog

From theory both your ideas would work, off course. But, as you have had good experiences driving the 6AS7 in differential modus, I would try to use a CCS for the triodes to (although the resistors would also work, as the phase splitting is done by the transformer).

When you came up with the transformer idea, I remembered this
thread by Stuart. I forgot to post it, but now that your thread popped up again...

Erik
 
Interesting thread. Unfortunately implementing it would involve stringing an order together - which I am reluctant to do.

But, as you have had good experiences driving the 6AS7 in differential modus, I would try to use a CCS for the triodes to (although the resistors would also work, as the phase splitting is done by the transformer).

I was thinking of a single resistor in the tails of the two triodes - so combining transformer phase splitting with a conventional differential stage as well. The transformer should reject input noise/ ground loops, and the LTP should reduce heater hum at the sensitive input stage, as well as correcting for transformer inbalance. The question is would the performance be significantly improved by replacing the single resistor with a CCS ?

Nice theory at least.

I was thinking of going differential on the outputs as well - but I think it might not be necessary.

I was talking to Jkeny about the possibility of using this as driver stage for a bigger output stage. Getting the benefits of transformer phase splitting on the cheap. The only issue would be implementing gNFB.

Shoog
 
Hi Shoog

This is probably the difference between 'working' and 'working and fantastic sounding', but, which is which?

Knowing you, I am sure that at the end you will try all those different possibilities (and report us about them).

And now, a bit off topic, about the MOT. I was seeing another thread that talked about the use of a PP tranny in a SE amplifier with a CCS taking cara of 'nulling' current in the OPT. Seeing the MOT thing, I though something similar could be applied to it, passing a current through the 220V (primary) to cancel the magnetising effect of DC through the secondary. I do not know how to calculate, but it could work, improving the inductance...I think

Erik
 
I saw this discussed somewhere else. In the end I concluded that sometimes a square peg is a square peg and shouldn't be forced into a round hole. I think parafeed is a more elegent solution to to a similiar problem.
The MOT's make reasonable chokes but they are big and heavy. Its the secondary that does the choke duty. The problem I could see would be setting the current through the choke - how would you do it without a CCS?

Shoog
 
The way I have envisaged it each channel would use 2xECL82. The pentodes would be for the outputs and the triodes would be for the LTP. A simple neg supply and SS CCS would be adequate.
The problem with fixed bias on the outputs is that I will be driving a toroidal output transformer so current balance is critical.

Shoog
 
Hi Shoog

The MOT's make reasonable chokes but they are big and heavy. Its the secondary that does the choke duty. The problem I could see would be setting the current through the choke - how would you do it without a CCS?

From what I understood of your experience with the MOT's, you used it in a parafeed amplifier. You used the secondary as choke, and let the primary disconnected. The problem I see here is the saturation of the core, as there is DC current flowing through the secondary. The consequence off saturation is that the inductance falls severely.

Reading this

I haven't done it myself, but Jeffrey Jackson (Experience Music) has used the filament current of the output tube to cancel the 'idle' flux in an SE transformer. It takes relatively few turns of wire because the filament current is so high compared to the plate current.

from this thread I thought about applying the same trick to the MOT, but with the current being supllied to the primary, as this is disconnected anyway. I would surely use a CCS, as this seems the easiest way. I was curious if you have maybe tested it, or your opinion about feasibility!

Erik
 
Sounds interesting and not something I had really though about. I run my MOT/chokes at 50mA 350V without any issues of note.

I think your idea is certainly worth a thorough investigation and I would be more than interested in the results. Seems that a simple LM317 CCS might be adequate.

Shoog
 
Hi Shoog

Glad you understood my idea, but this is how far my knowledge goes... I do not know about the maths to calculate how much current has to go through the primary winding to cancel the current through the secondary. And if it improves inductance?? It is just not hard to try when one has the amplifier ready and playing, and even less hard for a hard DIY'er as you :)

Just an idea: I would say that the current through the primary should be the the ratio of the secondary/primary multiplied by the current through the secondary. If the MOT is 220V to 2200V, that implies in a ratio of 10, and 500mA going through the primary. (But, that is just guessing.) I was also thinking about using a lm317..., say that the R of the primary is 10 ohms, 500mA would imply 5V drop over the primary. The lm317 needs a drop of about 3V over it, so one would need 8V. Rectifying and smoothing of the filament voltage could give this value (one could also use a voltage doubler, if higher values of voltage are needed).

As for me testing it... I have been very busy with school last months as a reaction on a chance that appeared with which I may accelerate my graduation (which will be in Chile, maybe I can meet Bas Horneman over there). Besides, I am still working on my 'borg collective' (thanks SY) prototyping rig...sad that it takes two years to develop a prototyping rig - my girlfriend says I could have built about 4 amplifiers in that time, and, well, she is right! But I always liked lego a lot, and this is my new lego!

All with all, I have good memory for 'things to do' so, someday I will try it! (After a CP, a baby huey, a IV DAC stage, a crossover...)

Erik
 
I think the the current conversion would probably be just in the primary to secondary ratio.

My problem is that I can't see myself getting my *** into gear to build anything much apart from a headphone amp. My two current amps are working well and I know the next step up requires serious money which I just don't have. You have to say stop at some point (before the wife walks out).

Shoog
 
Erik said
my girlfriend says I could have built about 4 amplifiers in that time

Shoog replied
before the wife walks out

Some months ago I had a nice meeting in Amsterdam with SY and Morgan Jones and proudly told that my girlfriend allowed me to have large speakers in the living room. But my happiness faded when they warned me that there is a large difference between a girlfriend and a wife. I will try to postpone marriage, but at one point there is no way back...

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