They do seem to be quite good, but if using NFB around the transformer you will need to put a zobel network on the feedback and also on the anode of the first stage, to cancel the overshoot and ringing you'll otherwise get.
The LF distortion you see is due to saturation of the transformer, not crossover distortion. Both of these artifacts have a different characteristic signature and that is what points to saturation. If you look at the Hammond datasheet, at 30Hz it's -1dB down i.e. just 11.9W. The negative feedback plus the design headroom saves the day makes the amp work harder to compensate for the loss and so keep the power at 15W. However the THD and IMD will rise dramatically. What has been have measured is not so much the Hammond transformer but the entire system. Assuming the frequency response of the amplifier is fairly flat, a better test to evaluate the transformer would be to disconnect the feedback.
Youre not wrong :) Come on Adam, we want to see how the TRANSFORMERS fare, not your amp - you just gotta pull one little wire (or unsolder from the board) Hammond make good looking competent Iron, but some say they're outclassed by things from Edcor & Transcendar (at a similar price) so tests are always interesting!
Howdy. Nice presentation. I believe You are correct. The low freqency buckling is likely to be cross-over distortion. You might consider running the end tubes slightly warmer. The square wave overshoots can possibly be remedied by connecting a capacitor in parallel with the feedback resistor. Try, say, 220 pF. Settle for the smallest value that kills the overshoot, or almost kills it, not to sacrifice too much slew rate. Otherwise the design looks good. There is no ringing after the overshoot. Regards.
Yep - you are completely correct :) There's a 100pF cap in parallel with the NFB resistor in the test unit in the video, but in my current iteration of this design, that is 220pF - so you're right on the money. The unit I'm running these tests on is the first prototype of this amp. There's also a zobel network on the anode of the first stage, which makes the design much more resilient against capacitive loads. I did a blog post on this at the time I was constructing another amplifier, similar design but higher power, you can see that here atrad-audio.co.nz/index.php/2017/07/17/fine-tuning-the-nfb/
@@A_RosnerNZ Howdy again. Personally I don't like global NFB for two reasons. The time shifting of the OPT. Also. I feel that upholding the galvanic isolation between the primary and secondary provides shock protection should the amp develop a fault.. I use cross feeding of the NFB from the end tube anodes to the drivers. Works nicely but the setup is proned to oscillations which must be investigated and managed. Actually. I only recently learned that I should have used grid stoppers for the driver tubes too. I need to fix that. Perhaps I can interest You to check my blog eugpoh44.blogspot.com where I have a description of my amp. Regards.
Excellent low and response even at the low power level what type of drive circuit are you using you have three tubes per section are they all triodes or watt ?
I have no idea. I don't know how to measure inductance. I had a method which I thought worked, but it gave me such improbable and inaccurate results it must have been wrong. So that's a hard IDK on that question for now.
@@A_RosnerNZ it’s mostly an academic exercise in your case. Looks like you were getting good response down to 30Hz which is the lower band of Hammond’s published spec. I wind a lot of line and mic level transformers using M6, Metglas, Permalloy, and MPP. Never made an output transformer before, but I’d love to try it.
Looking at a triangle waveform: Adam: im not super impressed with the square wave.. :) Just made a 6n8s amp, now about to start a gu29, kinda scary to check the reproduction of the waveforms on my amps:)
I’ve always wanted to build an amplifier with the Hammond 1650W. Might have to do it now.
They do seem to be quite good, but if using NFB around the transformer you will need to put a zobel network on the feedback and also on the anode of the first stage, to cancel the overshoot and ringing you'll otherwise get.
The LF distortion you see is due to saturation of the transformer, not crossover distortion. Both of these artifacts have a different characteristic signature and that is what points to saturation. If you look at the Hammond datasheet, at 30Hz it's -1dB down i.e. just 11.9W. The negative feedback plus the design headroom saves the day makes the amp work harder to compensate for the loss and so keep the power at 15W. However the THD and IMD will rise dramatically. What has been have measured is not so much the Hammond transformer but the entire system.
Assuming the frequency response of the amplifier is fairly flat, a better test to evaluate the transformer would be to disconnect the feedback.
Youre not wrong :) Come on Adam, we want to see how the TRANSFORMERS fare, not your amp - you just gotta pull one little wire (or unsolder from the board)
Hammond make good looking competent Iron, but some say they're outclassed by things from Edcor & Transcendar (at a similar price) so tests are always interesting!
Brings back memories. I used to be the organ guy at our shop and we were Hammond factory authorized.
The organs were made by a different Hammond company
@@ortzinator Yes, this Hammond is from Canada 🙂
Very interesting, thanks.
Nice demonstration, thank you!
Howdy. Nice presentation.
I believe You are correct. The low freqency buckling is likely to be cross-over distortion. You might consider running the end tubes slightly warmer.
The square wave overshoots can possibly be remedied by connecting a capacitor in parallel with the feedback resistor. Try, say, 220 pF. Settle for the smallest value that kills the overshoot, or almost kills it, not to sacrifice too much slew rate. Otherwise the design looks good. There is no ringing after the overshoot.
Regards.
Yep - you are completely correct :) There's a 100pF cap in parallel with the NFB resistor in the test unit in the video, but in my current iteration of this design, that is 220pF - so you're right on the money. The unit I'm running these tests on is the first prototype of this amp.
There's also a zobel network on the anode of the first stage, which makes the design much more resilient against capacitive loads.
I did a blog post on this at the time I was constructing another amplifier, similar design but higher power, you can see that here atrad-audio.co.nz/index.php/2017/07/17/fine-tuning-the-nfb/
@@A_RosnerNZ Howdy again.
Personally I don't like global NFB for two reasons. The time shifting of the OPT. Also. I feel that upholding the galvanic isolation between the primary and secondary provides shock protection should the amp develop a fault..
I use cross feeding of the NFB from the end tube anodes to the drivers. Works nicely but the setup is proned to oscillations which must be investigated and managed. Actually. I only recently learned that I should have used grid stoppers for the driver tubes too. I need to fix that.
Perhaps I can interest You to check my blog eugpoh44.blogspot.com where I have a description of my amp.
Regards.
Excellent low and response even at the low power level what type of drive circuit are you using you have three tubes per section are they all triodes or watt ?
Ultralinear push-pull, 2 X EL84 per ch
@@A_RosnerNZ not outputs but drive a Schematic is needed
@@A_RosnerNZ out put was obvious, I ask about the drive the section before the coupling caps
@@harryconover289 it's on my site atrad-audio.co.nz/wp-content/uploads/2019/07/Matariki_color.png
Looked at the page nothing odd but watt you label d as d5 , also how does it do near its clip point ?
What’s the inductance of the primary? Electromotive force formula tells me it should be about 42H if LF response is flat down to 30Hz
I have no idea. I don't know how to measure inductance. I had a method which I thought worked, but it gave me such improbable and inaccurate results it must have been wrong. So that's a hard IDK on that question for now.
@@A_RosnerNZ it’s mostly an academic exercise in your case. Looks like you were getting good response down to 30Hz which is the lower band of Hammond’s published spec. I wind a lot of line and mic level transformers using M6, Metglas, Permalloy, and MPP. Never made an output transformer before, but I’d love to try it.
Looking at a triangle waveform:
Adam: im not super impressed with the square wave..
:)
Just made a 6n8s amp, now about to start a gu29, kinda scary to check the reproduction of the waveforms on my amps:)
A gap in a push-pull amp? Not serious...
TEST signals....................