I have just finished switching from my M2 to my version of the F7 (thanks to Mr.Pass and other members for the hints).To my ears and speakers the F7 is fantastically balanced top to bottom.Bass is tighter than any of the other amps I have built(AJ,F6,M2).Highs are as good as the F6 in my opinion.Mids are a perfect balance,female vocals are stunning. For those not diy inclined the F7 is probably the best money spent period.With Mr.Pass's version I can imagine you get the best possible sound out of this topology.I myself cannot afford one but if I could it would be a lifetime amp.
Last edited:
Arriving very late to this show, but my sims indicate 0.33 to yield a very modest DF. Much higher values seem appropriate, at least in sim. Have you tested the DF of your build?
What is the bias and jfet IDSS you have your sim set at?That makes a difference.Also for a given bias different jfet IDSS makes a difference.For eg. my amp with 1.25A bias 0.33 ohms and 13/11ma jfets gives a sim df of 133.Amp with 1.25A bias 0.33 ohms and 8/8ma jfets gives df of 61.No I haven't actually measured my actual value,my ears tell me it is pretty high.
Built a prototype with parts from my drawers. Did not have a good sounding 0.33, so used a 0.25. The measured DF for both channels varies from 10 to 20 @8ohm. Why? Because resistors need to be fine tuned for a specific Df. It is interesting how sensitive the Df is to Rsen. Of course it also depends on the fets Gm.
Here is the simulated output impedance. Vertical axis is in miliohms, Rsen is varied from 0.25 to 0.35ohms.
Here is the simulated output impedance. Vertical axis is in miliohms, Rsen is varied from 0.25 to 0.35ohms.
Attachments
Something looks odd in your simulated response. For example, why does the DF for Rsen = 0.29 appear between 0.35 and 0.34?
It mostly affects the area of bass impedance resonance. The frequency response of an amp with a real speaker as load shows a distinct difference between low and average DF.
From memory my speakers have a bump of around +3.5db @30Hz when driven from a push pull 2A3 amp, not huge compared to room modes, but definitely leading to flabbier bass.
There are other, subtler effects as well, but generally most speakers are unhappy with a Df < 20.
Plumpy, valvey bass? No, thanks.
Last edited:
Not sure if this is odd, but yes, the impedance reaches a minimum at Rsen = 3.2ohms. That was my point, the best Df can only be reached by tweaking the value of Rsen.
Tested in the prototype. As the output impedance is very sensitive towards the value of Rsen and can easily turn negative i aimed for Df within 50-70. One channel Rsen had to be 0.434ohms and the other 0.456. Will have to do some longer term testing to see if this is stable.
It seems very unlikely that a commercial product is so finicky, which makes me think it probably is not.
Anyway, had some great fun and learned a thing or two. All thanks to the helpful contributors to this thread and of course the essential pictures in the review.
It seems very unlikely that a commercial product is so finicky, which makes me think it probably is not.
Anyway, had some great fun and learned a thing or two. All thanks to the helpful contributors to this thread and of course the essential pictures in the review.
Some observations.
Attempt No1 was running the outputs @1.3A, Df=60, distortion set for minimum. The sound was, no exaggeration, a complete disaster: flat soundstage, no dynamics.
Attempt No2 committed a cardinal sin, two changes at a time, but those were desperate times
Bias reduced to 1.1A as per F7 specs, distortion raised, so that F2/f3 - 10db @1W and phase checked as negative. As a result of the bias change Df came down to 20.
An almost complete transformation. Voices became textured and palpable. Females, specifically, lost an unpleasant edge, sound became much more open, free an dynamic. The lower ranges perhaps suffered a bit.
Attempt 3. Raised Df to >100. For lack of patience used the same Rsen in both channels, so one is about 100 and the other closer to 200.
Listened only on speakers which don't care much about Df, have not yet tried it on the Wilsons. Not entirely convinced with the high Df, but more listening and comparisons are in order. And anyway: a very enjoyable sound.
Final tweak which brought in some further improvement: a bypassing switch on the power on thermistor. Yes, it requires to observe a power on sequence, but the small inconvenience has a major sonic upside.
Attempt No1 was running the outputs @1.3A, Df=60, distortion set for minimum. The sound was, no exaggeration, a complete disaster: flat soundstage, no dynamics.
Attempt No2 committed a cardinal sin, two changes at a time, but those were desperate times
Bias reduced to 1.1A as per F7 specs, distortion raised, so that F2/f3 - 10db @1W and phase checked as negative. As a result of the bias change Df came down to 20.An almost complete transformation. Voices became textured and palpable. Females, specifically, lost an unpleasant edge, sound became much more open, free an dynamic. The lower ranges perhaps suffered a bit.
Attempt 3. Raised Df to >100. For lack of patience used the same Rsen in both channels, so one is about 100 and the other closer to 200.
Listened only on speakers which don't care much about Df, have not yet tried it on the Wilsons. Not entirely convinced with the high Df, but more listening and comparisons are in order. And anyway: a very enjoyable sound.
Final tweak which brought in some further improvement: a bypassing switch on the power on thermistor. Yes, it requires to observe a power on sequence, but the small inconvenience has a major sonic upside.