I have compared it to the amp with class A output stage without GNFB. All equal, the second was better in CCIF IM components. Same iddle current, supply voltage and output voltage and load. They are both few dB under limitation, no signs of limitation, just output stage great signal non-linearity effect. I would have expected that the FB design had to win.
I do not know the JC-1 plot at some 20Vpeak. JC-1 has 9 output pairs, PM-A1 has one output pair. The bias current of PM-A1 is 1.6A approx, JC-1 has 1.35A, which is similar. The output power is not similar, though. It is 11:1 for JC-1. But CCIF IM 2dB under limitation is better for PM-A1 than for JC-1 2dB under limitation.
But you probably meant to compare JC-1 and JC-3
But you probably meant to compare JC-1 and JC-3
PMA for a measurement expert, I am disappointed. Why would you take a little amplifier and drive it so hard that it will obviously distort? Then you compare it to a completely different amplifier.
IF you would take note, the JC-1 power amp has the SAME basic topology as the JC-3, and the same idle current. IF you measured it at the same voltage that you measured the JC-3, you would find little, if any significant distortion. WHY? They are the same topology, the same idle current, and the same slew rate.
Attempting to make the JC-3 measure badly, is bad form, in my opinion.
IF you would take note, the JC-1 power amp has the SAME basic topology as the JC-3, and the same idle current. IF you measured it at the same voltage that you measured the JC-3, you would find little, if any significant distortion. WHY? They are the same topology, the same idle current, and the same slew rate.
Attempting to make the JC-3 measure badly, is bad form, in my opinion.
john curl said:
I did not try to attempt to make it measure badly, I was asking how it measured in similar conditions. This was only simulation and we know it does not necessarily has to be same as reality. I do not think there is something like "very unfair plot". This is just a simulation plot several dB below limitation. Maybe it fits, maybe not, according to model accuracy. It is you to judge.
john curl said:
The answer is easy. Much higher supply voltage and nine output pairs make output transistors to work in a smaller and more linear portion of h21e = f(Ic) curve.
BUT - in case of one output pair, it would be possible to design much more linear output stage than the typical 2EF one. Then the IM distortion (esp. higher order) would be much lower for higher amplitudes.
I don't know what you are simulating, but it seems to be a bad model, as it does not reflect what I would expect in the real world. But then, I don't use SPICE simulation with the abandon that some do around here, even though I have been involved with circuit analysis for more than 40 years. This is probably one of the reasons.
john curl said:
The simulated circuit has been shown few pages before.
Could you specify what does not reflect what you would expect in the real world? For the lower voltage, the CCIF IM is between -110dB and -120dB for 2nd order dist. and better than -120dB for 3rd order distortion. These numbers of course go higher for the high amplitude plot, but still better than -80dB for both 2nd and 3rd order distortion. And this is few dB, like 2, under limitation. What is wrong with these results?
PMA, let me rephrase this: I don't know what you are measuring for, or what your results mean. They just don't make sense to me. Maybe, it is too much information presented out of context. I am not against your efforts in general, but when you get specific about the JC-3, with a simplified schematic, artificial models, and alternative parts, I just don't know if you are doing anything useful.
I can see it now, but was unable to bring it up earlier.
Not that it will have any effect on the PCB, but the DC balance voltage divider was 10k/1k pot/10k.
If you're planning on running much current through the front end differential, the pot will need to be much less...more on the order of 20 Ohms than 500. Otherwise it will be twitchy as hell, with all the adjustment coming in the last 2 degrees of rotation. (I get something on the order of 6mA per side with 2SK389/2SJ109s biased with a fixed 10 Ohm resistor, though they're the lower Idss units of what I have on hand. Higher Idss units will take a little more resistance.)
*****
John,
How hard did you run the front end on the JC-3? Looks to me as though it will make a difference in how the VAS comes up. Otherwise the resistor values may need fine tuning.
*****
The bias voltage divider was 1k/5k pot/1k. Again, no effect on layout.
Got to earn my keep. I'll try to look at it again later.
Grey
Not that it will have any effect on the PCB, but the DC balance voltage divider was 10k/1k pot/10k.
If you're planning on running much current through the front end differential, the pot will need to be much less...more on the order of 20 Ohms than 500. Otherwise it will be twitchy as hell, with all the adjustment coming in the last 2 degrees of rotation. (I get something on the order of 6mA per side with 2SK389/2SJ109s biased with a fixed 10 Ohm resistor, though they're the lower Idss units of what I have on hand. Higher Idss units will take a little more resistance.)
*****
John,
How hard did you run the front end on the JC-3? Looks to me as though it will make a difference in how the VAS comes up. Otherwise the resistor values may need fine tuning.
*****
The bias voltage divider was 1k/5k pot/1k. Again, no effect on layout.
Got to earn my keep. I'll try to look at it again later.
Grey
I have been lurking around since the beginning of this thread. Having just finished a D'Agostino KSA50 amp with great succes, makes me want to try a Curl-amp also Usually I am a "more-than-dedicated" Pass-fan, but the KSA was great. I beleive that an amp with "John Curl approval" would easily be better I am not sure at all 'bout this, but since the KSA is pretty old I think it could easily be the case (Watching everything from the outside!!)
In any case, I would like to join in on the building of this amp! If MikeW intends to pop out a boardlayout, I am in
Usually I am a "more-than-dedicated" Pass-fan, but the KSA was great. I beleive that an amp with "John Curl approval" would easily be better I am not sure at all 'bout this, but since the KSA is pretty old I think it could easily be the case (Watching everything from the outside!!)In any case, I would like to join in on the building of this amp! If MikeW intends to pop out a boardlayout, I am in