• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Posted new P-P power amp design

Transformer impedance matching question. I am using the 8K Edcors that Pete called out. Below is my speaker impedance curve. Based on this curve, would anyone recommend a different xformer primary impedence (into 8 ohm)?

1706714189630.png
 
jderimig,

1. Welcome to the un-flat impedance versus frequency curve of loudspeakers.
In your case it dips to about 5 Ohms, and peaks at perhaps 30 Ohms or more.

I believe your circuit uses global negative feedback.
Pentodes and Beam Power tubes in their Native Operating Mode: almost always require negative feedback, such as global, Schade, cathode, etc.

Negative feedback helps with the damping factor and ability to deal with varying impedance loads.
Hopefully your global negative feedback is also able to deal with the output transformers gain phase curves, and with the elliptical load of the loudspeaker at many different frequencies.
Stability is the first order of the day.

I do not design global negative feedback into my single ended, push pull, and balanced amplifiers.

But my output tubes are either Triodes, Triode Wired Beam Power tubes, or Triode Wired Pentodes (one amplifier was Beam Power in Ultra Linear).
Most of my 8 Ohm rated loudspeakers have minimum impedance of about 6 Ohms, So I use the 4 Ohm tap to drive the loudspeakers.
Results:
Better damping and Lower distortion, versus a little less power is the tradeoff I make.

If you change an output transformer from one primary impedance to a different primary impedance, there may be many other changes:
Primary Inductance
Maximum primary current
Leakage inductance from primary to secondary
Primary distributed capacitance
Primary DCR
Frequency response

2. CCS works good for Triodes and Triode Wired Pentodes.
For CCS and Pentodes that are Pentode Wired, the pentodes should be reasonably well matched.
2 - 3.8mA/V is Not reasonably matched.

Yes, global negative feedback will help with CCS and unmatched pentodes, but the feedback have to work much harder to do so.
The greatest risk is that one tube will clip early at mid to high power output, negative feedback can not help that clipping.

An initial listening test may not sound bad, especially at low to mid power output.
We are all human, sometimes we hear the differences (whether they are there or, are not there), and sometimes we do not hear the differences (whether they are there, or are not there).
A properly designed and properly conducted double blind listening test is the only way to be sure.

If we never hear any differences between two things, do we care?
 
Last edited:
1. The amp sounds fine, actually its the best sounding amp I have right now. It sound glorious and handily beats my Music Reference RM9 with the same speakers. REW measurement of the frequency response is reasonably flat.
2. I am not using GNB and I want to keep it that way.
3. I have seen worse impedance curves and I think that one is pretty good relatively.
4. @6A3sUMMER I have since matched my driver pentodes and output tubes using utracer. I have alot of 6CB6's and 6JN6's. I did not hear much of difference pre and post matching of tubes.

I am just wondering if there is a better nominal value to shoot for.
 
jderimig,

Change the nominal impedance of the output transformers?

1. Assuming that the primary impedance was correct for the original amplifier design (and that there were no modifications of the amplifier),
Then . . .
If I had an output transformer with a single output tap impedance, and I was going to spend money on another output transformer,
it would either be:

An Output transformer with the same primary impedance, and two output taps, 8 Ohm and 4 Ohm.
OR,
A Monolith Magnetics transformer with the same primary impedance and a 5 Ohm OR a 6 Ohm output tap.

2. If you changed the negative feedback (added, removed, changed the type of negative feedback) . . .
Then the complete modified amplifier circuit needs to be analyzed, before deciding on whether a new primary impedance is a good thing or not.
 
Last edited:
  • Like
Reactions: 1 user
Scattered throughout this long thread you could find my experiments with this board in configurations from about 30 WPC to over 200 WPC, but the most popular build made 125 WPC. Every version I built worked well without GNFB since there is an internal "Schade" output plate to driver plate feedback setup. The impedance of the OPT can be tweaked to pick a point on the power output vs distortion tradeoff curve. A lower impedance OPT will produce more power and a higher THD from the same tubes up to the power limit of the tubes. I would keep the 6JN6's to less than 50 WPC. I don't remember all the details, but I think 5000 ohms would be as low as I would go with 6JN6's, or any of the 18 watt plate dissipation tubes. 200+ WPC requires BIG output tubes, a 2500 ohm OPT and over 600 volts of plate supply voltage. The 125 WPC builds used 6HJ5's with some board mods, a 3300 ohm OPT, and a second power transformer to put 625 volts on the output tube plates.

If you like the way it sounds, and you have plenty of reserve power for anything you want to play with your speakers, don't mess with it. Lowering the OPT impedance will bring a bit more power but will reduce the damping factor and increase THD. Raising the OPT impedance will cost in power output.
 
  • Like
Reactions: 1 user