Not yet. I don't get much time to play in the garage so I'm slow at this.Hi
Thanks for answering.
.
Did You measure Primary inductance of Output transformer maybe?
Thanks
Thanks for the kind words. I miss the kenpeter/Michael Koster discussions here. I learned so much from them.
I'm sure you can clean up the distortion with the right feedback scheme on those 100TH. Beautiful tubes!
@trobbins,
Ok, I think I have the compensation fairly optimized. I ended up with a 300pF cap and 34.2k resistor. Yellow trace is the plate of the output tube (where the feedback is sampled). Blue is the output transformer secondary. First picture is with 8 Ohms on the secondary. Second picture is with the secondary open. Stand by for tests with capacitors on the secondary...
Ok, I think I have the compensation fairly optimized. I ended up with a 300pF cap and 34.2k resistor. Yellow trace is the plate of the output tube (where the feedback is sampled). Blue is the output transformer secondary. First picture is with 8 Ohms on the secondary. Second picture is with the secondary open. Stand by for tests with capacitors on the secondary...
Attachments
Next set of plots are with the secondary loaded with capacitors only. 1.5nF, 22nF, 100nF, and 330nF. 1.5nF looks pretty similar to an open secondary (slight change on the rising edge). As capacitance increases, you see increasing magnitude of ringing on the secondary with decreasing frequency. Makes sense. Since the transformer is outside the feedback loop, the feedback is only trying to control the yellow trace.
Attachments
I figured out that my scope could measure phase between the two inputs, so I gave FR/Phase plots a shot. I just swept, recorded interesting points, and plotted from a spreadsheet. I probably need to buy a better signal generator that is flat to a higher frequency so that I can take these types of plots further up.
First plot is the amplifier input to output tube plate (everything in the feedback loop). Second plot is the output transformer primary to secondary.
I measured the output transformer primary inductance at 35.6H at 60Hz/1W.
Edit: I ordered a chassis from Landfall. Time to start working on getting this in a real box.
First plot is the amplifier input to output tube plate (everything in the feedback loop). Second plot is the output transformer primary to secondary.
I measured the output transformer primary inductance at 35.6H at 60Hz/1W.
Edit: I ordered a chassis from Landfall. Time to start working on getting this in a real box.
Attachments
Well done for making a gain-phase plot - that's tough without automation. It was interesting to see the OT secondary plots, since they are decoupled from the feedback loop, so of no consequence for stability. What do you think is causing the -1.5dB droop and plateau? There must also be a main cause of roll off above 100kHz, apart from the 100kHz RC you recently introduced.
I think the -1.5dB droop and plateau is caused by the load that the OT presents, although the flat plateau area seems strange to me. The same feedback circuit with different compensation and the Electra-Print OTs had a much more ideal 10kHz square wave. All I changed was the OTs and I noticed the change in shape to the bump/dip/slow rise of the square wave above. After disconnecting the 8R load from this OT, I have a nearly ideal square wave on the output tube plate.
These are quite a bit bigger than the Electra-Print OTs. I also noted that the datasheet for the HQ-5090s shows all of the testing done on the 4 Ohm tap. I suspect that this OT is optimized for 5k:4 or 10k:8 operation, not the 5k:8 that I am testing. I noticed that my testing doesn't quite match the datasheet performance for FR/phase. Also the single small wire (compared to multiple wires for all the other taps) makes that tap look like an afterthought.
After I get this into a chassis, I will put together another amplifier with the Electra-Print OTs and the 826 output tubes I have. That should be fun.
Actually, the OT with the best HF performance that I played with was probably the Edcor, but it has been a while so I'm not totally sure.
I traded more than 6dB of feedback for more gain to get a bit lower input sensitivity and to make the biasing of the input tube easier, but the differential feedback on the op-amp that @zintolo mentioned might be a way to use the op-amp to get some more feedback in there, since I can't really increase the gain of the input stage any further. I'll have to think about it and see if it could work.
These are quite a bit bigger than the Electra-Print OTs. I also noted that the datasheet for the HQ-5090s shows all of the testing done on the 4 Ohm tap. I suspect that this OT is optimized for 5k:4 or 10k:8 operation, not the 5k:8 that I am testing. I noticed that my testing doesn't quite match the datasheet performance for FR/phase. Also the single small wire (compared to multiple wires for all the other taps) makes that tap look like an afterthought.
After I get this into a chassis, I will put together another amplifier with the Electra-Print OTs and the 826 output tubes I have. That should be fun.
Actually, the OT with the best HF performance that I played with was probably the Edcor, but it has been a while so I'm not totally sure.
I traded more than 6dB of feedback for more gain to get a bit lower input sensitivity and to make the biasing of the input tube easier, but the differential feedback on the op-amp that @zintolo mentioned might be a way to use the op-amp to get some more feedback in there, since I can't really increase the gain of the input stage any further. I'll have to think about it and see if it could work.
@SpreadSpectrum what I've learned on differential feedback is taken from a French DIY site: http://www.audiyofan.org/forum/viewtopic.php?f=60&t=9543
If you don't speak French, use this link: https://www-audiyofan-org.translate...r_sl=it&_x_tr_tl=en&_x_tr_hl=it&_x_tr_pto=wap
In Push-Pull amps it is really a nice feature, as you can adjust the DF with a pot (within a stability range) without affecting the sensitivity of the amp.
I'd like to try it in SE amps too, but by now the amp is on paper only due to some work trips and some other priorities at home.
If you don't speak French, use this link: https://www-audiyofan-org.translate...r_sl=it&_x_tr_tl=en&_x_tr_hl=it&_x_tr_pto=wap
In Push-Pull amps it is really a nice feature, as you can adjust the DF with a pot (within a stability range) without affecting the sensitivity of the amp.
I'd like to try it in SE amps too, but by now the amp is on paper only due to some work trips and some other priorities at home.
As I suspected this is a very interesting amplifier design.
I really like how the P mosfet is used to apply feedback to the input pentode's cathode. This is a innovation for me.
Slick how it separates the feedback voltage from the pentodes AC current and the cathodes impedance to ground.
I was working on a phono input stage where low noise and gain was a key goal and struggled to find a way to add negative feedback to the input device without adding any impedance in the cathode as that degrades the noise figure and gain. The P mosfet may well be a useful answer.
Going to try it out.
I like the way you have paid attention to measured data to inform your design decisions.
I tire quickly of the "sounds great" but no measurement data designs.
Getting 19 watts of very clean power from your relatively low rail of 525 volts is impressive.
The DF with 0.7 ohms output impedance is a workable low for many speakers and this is without feedback around the transformer.
Is the grid of a 826 piticullary rugged that allows the positive drive (and so grid current) without failure?
Wondering about trying other tube types that could deliver high current with positve grid drive in triode mode.
Thanks for the excellent brain food.
I really like how the P mosfet is used to apply feedback to the input pentode's cathode. This is a innovation for me.
Slick how it separates the feedback voltage from the pentodes AC current and the cathodes impedance to ground.
I was working on a phono input stage where low noise and gain was a key goal and struggled to find a way to add negative feedback to the input device without adding any impedance in the cathode as that degrades the noise figure and gain. The P mosfet may well be a useful answer.
Going to try it out.
I like the way you have paid attention to measured data to inform your design decisions.
I tire quickly of the "sounds great" but no measurement data designs.
Getting 19 watts of very clean power from your relatively low rail of 525 volts is impressive.
The DF with 0.7 ohms output impedance is a workable low for many speakers and this is without feedback around the transformer.
Is the grid of a 826 piticullary rugged that allows the positive drive (and so grid current) without failure?
Wondering about trying other tube types that could deliver high current with positve grid drive in triode mode.
Thanks for the excellent brain food.
The 826 is meant for grid drive. The datasheet gives various operating conditions, all of which feature considerable grid power dissipation.
There are probably better tubes to use, though. Lots of NOS 826 end up being gassy. One big name vender got kind of frustrated with me because I kept sending gassy tubes back. Not my fault, they kept sending gassy tubes (but they had no means to test them ahead of time). These weren't just a little gassy either, they were bottles of air and the filament went up in actual smoke when I applied power as it lit up.
Of all the tubes I tested, I think the new production 211 got lowest distortion and were super easy to acquire and reasonably priced. The various 100W tubes made the most power and looked the coolest, but I spent a lot of time acquiring, testing, and returning to build my stock.
There are probably better tubes to use, though. Lots of NOS 826 end up being gassy. One big name vender got kind of frustrated with me because I kept sending gassy tubes back. Not my fault, they kept sending gassy tubes (but they had no means to test them ahead of time). These weren't just a little gassy either, they were bottles of air and the filament went up in actual smoke when I applied power as it lit up.
Of all the tubes I tested, I think the new production 211 got lowest distortion and were super easy to acquire and reasonably priced. The various 100W tubes made the most power and looked the coolest, but I spent a lot of time acquiring, testing, and returning to build my stock.