I have few different sets of output tubes and I perceived differences in sound so I measured the distortion under the same operating conditions. The amplifier is a PP UL with A431 style OPT, 470V B+, 15dB gNFB and the tubes biased for 20W each at idle. The graphs are distortion sweeps at 5W into a non-inductive 8ohm load. The %THD legends are measured at 1kHz.
Last of all I did the same test with a Firstwatt J2, 5W into a 8ohm non-inductive load
Last of all I did the same test with a Firstwatt J2, 5W into a 8ohm non-inductive load
I prefer the tube amp with the 6CA7s but included the J2 as a sanity check on distortion. I think the dip in 3rd HD in the midrange is what I perceive but that's why I decided to measure things and see if I could understand it better.
You can also measure differences in distortion and power between brands of the same tube. For instance, in my Williamson amplifiers, Gold Lion KT66s put out 22 watts of power at 1% THD, but TAD Chinese versions only reach 19 watts.
Should the biasing not be optimized for each tube type?
I would think that comparing different tubes all with the same bias is not 'fair'.
Jan
I would think that comparing different tubes all with the same bias is not 'fair'.
Jan
I don't think I'm trying to achieve the best THD measurement, but more trying understand how different HD responses are perceived and why certain combinations may be preferred. The J2 has a fairly flat HD response and it doesn't sound bad, and my Purifi amp also doesn't sound bad and measures extremely low THD.
Yes, try again with the bias at 55mA. Also your second harmonics are high, it is possible that the bias is not set perfectly right and that doesn't cancel out 2nd harmonics
In my recent tests at a lower B+ I obtained 0.02THD total at 5 Watts, only 3rd harmonics.
Imo, the lower the better, it is more important in the 60 - 500 hz region especially IMD, the THD figures are simply an indication of the performance of the circuit.
When I changed the UL to pentode it made a huge difference in clarity and coherence in the sound. I never though it would be the sound which I always sought, a full, analytical and truth sound where I can listen to as many things at the same time.
My solid state is a bit like that too. I found class T sound way too soft and lethargic. The only tube which I like is the kt120.
In my recent tests at a lower B+ I obtained 0.02THD total at 5 Watts, only 3rd harmonics.
Imo, the lower the better, it is more important in the 60 - 500 hz region especially IMD, the THD figures are simply an indication of the performance of the circuit.
When I changed the UL to pentode it made a huge difference in clarity and coherence in the sound. I never though it would be the sound which I always sought, a full, analytical and truth sound where I can listen to as many things at the same time.
My solid state is a bit like that too. I found class T sound way too soft and lethargic. The only tube which I like is the kt120.
Last edited:
You should do these tests at also 50hz and 10khz to get a better idea. I find the 1khz useless, it is only a damn specification.
I find the THD at high power at 50hz + send a 2khz will reveal a lot more how the amp will sound. The First watt is a real 'soft' amp.
I find the THD at high power at 50hz + send a 2khz will reveal a lot more how the amp will sound. The First watt is a real 'soft' amp.
The grid bias voltages are individually controlled, good question.
The tests are a sweep, 20Hz-48kHz signal - maybe I don't understand your suggestion? I do like the idea of an IMD test, what do you suggest for frequencies and balance?
Here are the measurements at 55mA, the B+ dropped to 455V
The fact that dist is frequency dependent tells me that the circuit itself ( transformer etc) is the real
non-linearity here. The tubes themself has NO frequency preferences.
non-linearity here. The tubes themself has NO frequency preferences.
It would be much easier to use the "Load Line Calculator" to find the right B+, idle current, and Load impedance for low distortion output.
Johnny
Johnny
Not sure what conclusion you can draw - you need to indicate how you drive them and the bias conditions. I would say the differences in second harmonic will be more to do with tube matching and you would need to measure this.
Ah yes, it’s sweeps , I was confused.
I like 40+500
50+2k
100+20k
Then closed signals like 60+80, 2000+2050 and 9k+9500
I like 40+500
50+2k
100+20k
Then closed signals like 60+80, 2000+2050 and 9k+9500
I use also harmonics generator on the internet, but it just gives an idea, it’s not very reliable
I agree with Peter Tub, and the other who says the feedback works wonders, why?
The thd is quite high and looks like transformer limitation by the shape of the curves, some ringing or reactance, It could be the transformer limitation which the feedback is unable to correct with phase shifts not 180 in all spectrum
Try disconnecting all feedback and measure again, it would be appreciated!!
The thd is quite high and looks like transformer limitation by the shape of the curves, some ringing or reactance, It could be the transformer limitation which the feedback is unable to correct with phase shifts not 180 in all spectrum
Try disconnecting all feedback and measure again, it would be appreciated!!