Sorry - I meant to say IN CIRCUIT (blah.- it was late here).
The JFETS are matched -BL range, 7-8 ma, from punkydawgs. I did not measure them out of circuit and don't think I saved his package which had the actual measured values, but they were within .03 ma Idss of each other and 4 mV, according to his eBay page, FWIW.
I recall reading somewhere (maybe the F7 review thread) that matching of the JFETS was not particularly useful in this circuit. Perhaps deliberately mismatching in one direction or another would help with phase, but I have not figured that out yet.
I may get brave and unsolder them to measure.
Last edited:
Thanks - I missed that. So deliberate mismatching is the ticket!
What I don't understand though is the FFT of the actual live circuit is telling me that the phase of the second harmonic is -94 degrees (with matched JFETS).
See attached. This spectrum analysis is using REW software on an EMU 0404 USB and a Macbook Pro, both running on batteries. I am using the internal REW generator for the 1KHz sine wave, sampled at 96KHz and level adjusted to produce 1 Watt output from the amp into 8 Ohms. The FFT was done using the pot settings I mentioned earlier.
I can document my test setup better at some point, but I am still refining it. At any rate, the performance is very close to the published specs, so I am not sure how much further to take this.
What I don't understand though is the FFT of the actual live circuit is telling me that the phase of the second harmonic is -94 degrees (with matched JFETS).
See attached. This spectrum analysis is using REW software on an EMU 0404 USB and a Macbook Pro, both running on batteries. I am using the internal REW generator for the 1KHz sine wave, sampled at 96KHz and level adjusted to produce 1 Watt output from the amp into 8 Ohms. The FFT was done using the pot settings I mentioned earlier.
I can document my test setup better at some point, but I am still refining it. At any rate, the performance is very close to the published specs, so I am not sure how much further to take this.
Attachments
In LTSPICE -90 degrees actually gives a positive phase harmonic, and +90 gives a negative phase harmonic.
The angle is just an indication of phase lag, whether it is positive 90 or negative 90 does not guarantee whether it is a negative or positive phase harmonic.
You really need a notch filter or you can calculate it out with some math and using the data sheets etc but even that is a bit of a rough estimate since parts can vary a bit from the data sheet.
The angle is just an indication of phase lag, whether it is positive 90 or negative 90 does not guarantee whether it is a negative or positive phase harmonic.
You really need a notch filter or you can calculate it out with some math and using the data sheets etc but even that is a bit of a rough estimate since parts can vary a bit from the data sheet.
Last edited:
Just out of curiosity.
Set source trimpot back to midpoint.
Reset dc offset and bias, and do another measurement and tell me what angle you get.
More beastly than SOT227
See link
On A Hippie Trail, Head Full Of Zombie
https://www.mouser.com/datasheet/2/240/VBO52-08NO7-1549562.pdf
Last edited:
I dedicate this song to myself.
Choir! Epic! Nights with Colin Hay - Men At Work "Overkill" - YouTube
The lyrics are highly appropriate for my diyaudio condition.
Hahahahaha
Choir! Epic! Nights with Colin Hay - Men At Work "Overkill" - YouTube
The lyrics are highly appropriate for my diyaudio condition.
Hahahahaha
Don't think so.
Both bridges show the same problem.
Nichicon Electrolytic, and TDK Multilayer Ceramic.
I saw the discussion in the Analog Engineer Zone. Where's Jim Williams (RIP) when you need him?
Yeah
Hopefully they can provide some additional tips.
So here is the FFT with the FET source pot centered (within 1% or so) and the bias pots readjusted for balanced DC zero out at 1.2 A. (BTW my pot is a 25- turn Bourns 20 Ohm nominal but actually measures about 21.3 Ohms total).
The first thing to note is that overall THD is lower: 0.035% vs. 0.053% at 1 Watt.
The second thing is that H2 and H3 are now equal: 0.024% distortion, roughly 72db down from the fundamental. Before the H2 was about 7db higher than H3 - I recall I had adjusted the pot to get that result (on the theory that having H2 about 10db higher met Nelson's criteria).
The third thing is that H2 is still negative, now showing -98 degrees vs. -94 degrees before.
REW has some built in notch filter capability, but I have not played around with that - I may try to get some accurate readings on the harmonic phases.
The first thing to note is that overall THD is lower: 0.035% vs. 0.053% at 1 Watt.
The second thing is that H2 and H3 are now equal: 0.024% distortion, roughly 72db down from the fundamental. Before the H2 was about 7db higher than H3 - I recall I had adjusted the pot to get that result (on the theory that having H2 about 10db higher met Nelson's criteria).
The third thing is that H2 is still negative, now showing -98 degrees vs. -94 degrees before.
REW has some built in notch filter capability, but I have not played around with that - I may try to get some accurate readings on the harmonic phases.
Attachments
Only slightly changed - N-channel pot old wiper to source was 165, now 159.
P-channel pot was 270, now 265.
These are pretty sensitive - I don't know how you could do this with single turns.
Looks like I'll be ordering some unmatched JFETs from punkydawgs (cheaper anyhow).
Oh - and a big DUH moment. I am using differential inputs to the EMU so I can accurately measure across the non-grounded load. If I reverse the differential leads, it flips the phases of all the harmonics. Unfortunately, REW does not tell me the phase of the fundamental(!!) so I did not even notice until I played around with different leads (to see if I could reduce noise) and accidentally reversed the leads. So all my concern about phase was a red herring. Sorry about that. I will look at notch filtering to see if I can find out more.
P-channel pot was 270, now 265.
These are pretty sensitive - I don't know how you could do this with single turns.
Looks like I'll be ordering some unmatched JFETs from punkydawgs (cheaper anyhow).
Oh - and a big DUH moment. I am using differential inputs to the EMU so I can accurately measure across the non-grounded load. If I reverse the differential leads, it flips the phases of all the harmonics. Unfortunately, REW does not tell me the phase of the fundamental(!!) so I did not even notice until I played around with different leads (to see if I could reduce noise) and accidentally reversed the leads. So all my concern about phase was a red herring. Sorry about that. I will look at notch filtering to see if I can find out more.
Last edited: