Hi,
I played around to simulate the circuits posted here and came up with this circuit.
This looks good on sims but not sure if the "un-labled" resistor will pose any issues, although it improves the THD from .010% to 0.004% @1watt 1Khz sine
Attached also the output sine and square wave.
I was attracted to this F7 due to its simplicity and based on reviews i read and watched and seems really worth to build one.
I do not have the K170 and J74 on the software I use but in general those should work, atleast on sims?
output waveform below is 1Vp sine and square. @1Khz
I played around to simulate the circuits posted here and came up with this circuit.
This looks good on sims but not sure if the "un-labled" resistor will pose any issues, although it improves the THD from .010% to 0.004% @1watt 1Khz sine
Attached also the output sine and square wave.
I was attracted to this F7 due to its simplicity and based on reviews i read and watched and seems really worth to build one.
I do not have the K170 and J74 on the software I use but in general those should work, atleast on sims?
output waveform below is 1Vp sine and square. @1Khz
Interesting design . . . but I would need to play around with it in spice to understand how the current through the load is translating to extra voltage i/p
I would be interested to know:
What the 10khz square wave looks like
Damping factor without Positive Feedback
Damping factor with Positive Feedback.
My understanding is that these are some the key parameters that define the F7 functionality
Alternatively the open loop gain & closed loop gain with & without positive feedback would tell a similar story
mike
I would be interested to know:
What the 10khz square wave looks like
Damping factor without Positive Feedback
Damping factor with Positive Feedback.
My understanding is that these are some the key parameters that define the F7 functionality
Alternatively the open loop gain & closed loop gain with & without positive feedback would tell a similar story
mike
rawbas,
If you run a sin wave in spice at 50hz then measure the peak output first with the 8R load connected then with it disconnected
Then repeat these two readings but with R12 set to 0.001 ohms
Make the peak output somewhere between 5-10V
From these readings we can work out the output impedance with and without the positive feedback.
If you run a sin wave in spice at 50hz then measure the peak output first with the 8R load connected then with it disconnected
Then repeat these two readings but with R12 set to 0.001 ohms
Make the peak output somewhere between 5-10V
From these readings we can work out the output impedance with and without the positive feedback.
There is an easy way to plot the output impedance and the DF over a frequency range direct in LTspice.
1. Define resistance Rload as parameter {Xload}
2. Set spice directive: .step param Xload list 8 4
3. Set AC Anaysis for example: .ac oct 101 10 100k
4. Enter following expression to plot DF: 8ohm/(8ohm*(V(out+,out-)@1-V(out+,out-)@2)/(2*V(out+,out-)@2-V(out+,out-)@1))
where V(out+,out-) means measured voltage on load resistance Rload
1. Define resistance Rload as parameter {Xload}
2. Set spice directive: .step param Xload list 8 4
3. Set AC Anaysis for example: .ac oct 101 10 100k
4. Enter following expression to plot DF: 8ohm/(8ohm*(V(out+,out-)@1-V(out+,out-)@2)/(2*V(out+,out-)@2-V(out+,out-)@1))
where V(out+,out-) means measured voltage on load resistance Rload
Last edited:
Here is the results from my sims.
This is 0.57Vp sine @50hz
R12 0.33
8R - connected - 5.978Vp
8R - disconnected - 9.932Vp
R12 .001
8R - connected - 9.993Vp
8R - disconnected - 10.082Vp
Last edited:
My calculations say:
With no Pos FB you have a damping factor into of about 1.5
With Pos FB you have a damping factor of just over 100
Nelson suggests a DF of about 5 without pos FB. With a DF of 1.5 the amp should sound exceedingly relaxed but with so little Feedback the Harmonic Distortion will almost certainly be higher than a standard F7
The damping factor of just over 100 with the Pos FB means that this part of the circuit is working very well.
I think we would have to ask Nelson how it might sound . . . .
One question I have is this:
When calculating the DF with the load connected do you have to add the resistance of the sense resistor to that of the load resistor to get a true result ?
(Long-time mostly-lurker.)
I just built an ACP+ pre/headphone amp, adding the small P098ZB power filter in front of the filter on the board. I'm using it as a preamp, not as a headphone amp. I posted this entry in the ACP+ thread, but since I also recently bit the bullet and bought a used F7 amp, I figured I would post it here as well. FWIW it's about the sound, not about trying to reverse-engineer the F7 circuit.
The combination of the ACP+ and F7 (and SVS Pro 2000 subwoofer) sounds very good. The speakers are 87dB sensitive.
No hum or hiss coming out of the speakers with the volume control all the way up. I was wondering if the amp/subwoofer combination would work well simply using a Y-adapter out of the preamp, and it does.
I have been using a Denafrips preamp/amp combination that has two pre-outs, and runs in balanced mode, a very nice setup. The sound of the two systems is a bit different, and both are very good. The ACP+/F7 setup sounds a bit "smoother" or "fuller"; the other setup sounds like a very good Class AB setup with a lot of power is supposed to sound. I presume I am hearing the difference between the emphasis on 2nd harmonic character of Nelson's Class A designs, and the emphasis on extremely low measured harmonic distortion of the Denafrips system.
F6 Comparison: I built an F6 last year. It sounds pretty good, too, but I like the F7 better. I think the F6 probably needs more efficient speakers to come into its own with more dynamics. In my setup it sounds a bit ordinary, perhaps similar to what Nelson describes about the F7 before he added the positive feedback resistor. Or maybe I'm hearing the signal transformer take just a bit of the dynamic edge off?
I just built an ACP+ pre/headphone amp, adding the small P098ZB power filter in front of the filter on the board. I'm using it as a preamp, not as a headphone amp. I posted this entry in the ACP+ thread, but since I also recently bit the bullet and bought a used F7 amp, I figured I would post it here as well. FWIW it's about the sound, not about trying to reverse-engineer the F7 circuit.
The combination of the ACP+ and F7 (and SVS Pro 2000 subwoofer) sounds very good. The speakers are 87dB sensitive.
No hum or hiss coming out of the speakers with the volume control all the way up. I was wondering if the amp/subwoofer combination would work well simply using a Y-adapter out of the preamp, and it does.
I have been using a Denafrips preamp/amp combination that has two pre-outs, and runs in balanced mode, a very nice setup. The sound of the two systems is a bit different, and both are very good. The ACP+/F7 setup sounds a bit "smoother" or "fuller"; the other setup sounds like a very good Class AB setup with a lot of power is supposed to sound. I presume I am hearing the difference between the emphasis on 2nd harmonic character of Nelson's Class A designs, and the emphasis on extremely low measured harmonic distortion of the Denafrips system.
F6 Comparison: I built an F6 last year. It sounds pretty good, too, but I like the F7 better. I think the F6 probably needs more efficient speakers to come into its own with more dynamics. In my setup it sounds a bit ordinary, perhaps similar to what Nelson describes about the F7 before he added the positive feedback resistor. Or maybe I'm hearing the signal transformer take just a bit of the dynamic edge off?
Hi Nelson,F7 is near EOL unless there is a hue and cry from distribution.
In any case I see that I have about 2,000 pairs, so clearly there will be
enough left over for some DIY activity.
So sit tight until after the DIY SONY VFETs are put to bed...
Do you have any plans on showing an official F7 circuit diagram soon? Or maybe preparing a kit for the Store?
So I completed a dual device implementation with dual dies each + cascode. Runs 1.8A bias, 34V rails. Love the stability of the amp, it barely moves the offset and bias from cold to hot, maybe 50mA and 3-4mV offset.
My chosen feedback ratios are quite different due to the higher gm of the output stage. The PFB is deliberately lowered and the NFB is deliberately increased. The DF is correspondingly unchanged from the intent of the amp. I have incorporated a P3 as well for HD tuning, and at 1W it ends up at -70dB H2. This is much higher than I'm used to from most designs, but it does work well.
The quirky dual mono chassis is about 10 years old - I was planning a balanced F5 Turbo implementation but the heatsinks are only good to 2A each. Thus the F7-inspired design (It's not technically F7) is a perfect fit. Power supplies are 34V rails, 1kVA per chassis, with 256kuF of buffers.
The best way to describe the sound is velvet, with a sprinkle of spice on the top. It's more expressive than the Plantfeve design, but less detailed than the F5. It retains control very well, up to about 20 watts there is no increase in distortion.
The amp has already found a new home, so I'll be building a smaller version for myself. Hopefully posting pictures soon, but the amp is not complete and I have a deadline to meet.
My chosen feedback ratios are quite different due to the higher gm of the output stage. The PFB is deliberately lowered and the NFB is deliberately increased. The DF is correspondingly unchanged from the intent of the amp. I have incorporated a P3 as well for HD tuning, and at 1W it ends up at -70dB H2. This is much higher than I'm used to from most designs, but it does work well.
The quirky dual mono chassis is about 10 years old - I was planning a balanced F5 Turbo implementation but the heatsinks are only good to 2A each. Thus the F7-inspired design (It's not technically F7) is a perfect fit. Power supplies are 34V rails, 1kVA per chassis, with 256kuF of buffers.
The best way to describe the sound is velvet, with a sprinkle of spice on the top. It's more expressive than the Plantfeve design, but less detailed than the F5. It retains control very well, up to about 20 watts there is no increase in distortion.
The amp has already found a new home, so I'll be building a smaller version for myself. Hopefully posting pictures soon, but the amp is not complete and I have a deadline to meet.