Let me dig through my notes and the thread to get a current best set of values and go from there. It's been a long time since I looked it over or even touched this thread 😞
All right, it is almost a year ago since the last response, in the end Ling was so kind to sent me the Gerber files to produce a PCB of his design.
But in the mean time I have drawn my own and ,yes, using his file which he was so kind to sent me, I have made my own.
But after populating my pcb with the relative components, I discovered that I do not know at which parameters I
have to judge the design: for instance, at which power (250mW into 8 Ohms load) can I judge the THD at for instance 10kHz?
I am now using the Block RDK15-2x6V as OPT, which gives a 11.8k Ohm load primarily at 8 Ohm secundair.
On the whole it behaves quite good.
My standard is to measure THD & PO when the amp is driven by 1kHz just before clipping.
Yup, 1khz sine is about standard for the testing. If you have the means to look at the waveform I'd say rated output power is right before the tips of the peaks begin to flatten out. THD is gonna be more tube and bias dependant but 250-400mW output is about where the Russian tubes measure best. Good NOS American tubes can net a bit better THD. You can get better low level numbers and less overall output by running the output stage hotter but I see no need as most of the time output is unlikely to be over 250mW in my use cases. Higher output impedance can net lower THD with lower output.
3rd harmonic will be higher than 2nd due to the output stage design unless you do something non ideal to the first voltage gain stage triode biasing. Maybe an interesting thing to play with?
3rd harmonic will be higher than 2nd due to the output stage design unless you do something non ideal to the first voltage gain stage triode biasing. Maybe an interesting thing to play with?
I measure maximum power @ 30Hz... I find it to be a more useful test since I might need full power @ 30Hz but rarely @ 1 kHz.
11.8K loadline is much too steep for a 6SN7 running Class A. 25K much better.
Refer to some of my comments earlier in this thread. 🙂
Plenty of real bench test info beginning at post 233, page 12 of Lingwendil's thread.👍
Covering pros & cons of Garter Biasing. determining proper load, driver alternatives,
using other tubes. Real test results, not just simulations.🙂
Covering pros & cons of Garter Biasing. determining proper load, driver alternatives,
using other tubes. Real test results, not just simulations.🙂
But did somebody run simulations of that PP amp using LTspice?
I cannot match the measuring results of the real amp and the simulations.
I cannot match the measuring results of the real amp and the simulations.
A perfect simulation program is only as good as: The imperfect models of the parts you use.
Example: the Output Transformer.
Until you come up with perfect models of all the parts, the simulation versus the real measurement results will never match.
'Nuf Said
Example: the Output Transformer.
Until you come up with perfect models of all the parts, the simulation versus the real measurement results will never match.
'Nuf Said
"Example: the Output Transformer."
Exactly. I am using the Hammond 125A as pp transformer. And after some measurements (prim-sec induction values) I cannot get these results matched with practice when
simulating the PP-circuit.
Exactly. I am using the Hammond 125A as pp transformer. And after some measurements (prim-sec induction values) I cannot get these results matched with practice when
simulating the PP-circuit.
How do you have the transformer configured? If I'm not mistaken when configured for high impedance the 125A can roll of the higher frequencies considerably. Secondhand info, but as I don't have one in hand I can't comment further.
With the 8 Ohm speaker impedance I use the 3-5 secondary connection which gives a total winding impedance of 17k6.
JoeAlders,
What measurement(s) is/are not matching the simulation result(s).
Be specific:
Low frequency response?
High frequency response?
Power output?
Harmonic distortion?
Intermodulation distortion?
Damping factor?
Square wave response (rise time, fall time, slope, ringing)?
Hum and Noise?
How large are the differences?
1%, 5%, 15%, etc.
What measurement(s) is/are not matching the simulation result(s).
Be specific:
Low frequency response?
High frequency response?
Power output?
Harmonic distortion?
Intermodulation distortion?
Damping factor?
Square wave response (rise time, fall time, slope, ringing)?
Hum and Noise?
How large are the differences?
1%, 5%, 15%, etc.
Its the THD what is mismatching at large signal levels
When running the transient response and simulating large input signal behaviour, simulation results of the THD
is much larger (around 1.5 to 3%) than I am measuring the actual print board.
For instance at Rload=8 Ohm and increasing the input signal up to around 900mW Pout, measuring result is 0.42% THD on the board, measured with my
HP339A distortion meter. LTspice calculates very much more.
This of course when using the 3-5 output terminals of the 125A where (according to Hammond), primary impedance is 17600 Ohm.
But I think that the large signal behaviour is inaccurate when using the LTspice tube models.
When running the transient response and simulating large input signal behaviour, simulation results of the THD
is much larger (around 1.5 to 3%) than I am measuring the actual print board.
For instance at Rload=8 Ohm and increasing the input signal up to around 900mW Pout, measuring result is 0.42% THD on the board, measured with my
HP339A distortion meter. LTspice calculates very much more.
This of course when using the 3-5 output terminals of the 125A where (according to Hammond), primary impedance is 17600 Ohm.
But I think that the large signal behaviour is inaccurate when using the LTspice tube models.
Probably. I'm not a simulation guy so can't comment too much there.
.42% THD ain't too bad at 900mW, especially with the original intent of the amp swimming along at 100mW or so average in daily use.
.42% THD ain't too bad at 900mW, especially with the original intent of the amp swimming along at 100mW or so average in daily use.
At 50 mW output I first got a 0.63% THD due to the 50Hz mains component. Whatever I tried, I did not manage to get that figure down until I
used 6.3Vdc in stead of 6.3Vac filament voltage (using the Arta spectrum analyser software together with the soundcard of my computer). Then it dramatically dropped to .18% THD.
So all THD figures are ranging from .18% (50mW) to .42% (900mW) and using 6.3Vdc filament voltage. Did a simple sound check with one loudspeaker.
Now I have to populate a second board to do a final listening test.
used 6.3Vdc in stead of 6.3Vac filament voltage (using the Arta spectrum analyser software together with the soundcard of my computer). Then it dramatically dropped to .18% THD.
So all THD figures are ranging from .18% (50mW) to .42% (900mW) and using 6.3Vdc filament voltage. Did a simple sound check with one loudspeaker.
Now I have to populate a second board to do a final listening test.
Last edited:
Some progress because of holiday, just need to power up
