Somewhere earlier I think I explained this a little better, maybe it was a discussion with flg (possibly), can't remember but it must be close to that one.
I was very subtle in my explanation back then, I didn't want to just baby feed the cloners.
Last edited:
Hi Generg
After looking at the posts again I realised Oreo may have got it back to front, and I didn't correct him.
He understood the idea that to compensate for less degeneration you use jfets of different idss values but did it in the completely opposite fashion.
Sorry for not noticing the error or for not correcting it, I was very protective of releasing too many ideas back then publicly.
I can't recall if I just didn't notice Oreo's error or I chose not to correct it.
Either way, positive or negative phase, I am sure his amp still sounds very nice.
In the end it will depend on the parts you have as to what the circuit should be.
I have exicon parts which I suspect are the same as other laterals with regards to their Vgs threshold hold values but I don't know that for a fact.
So the circuit could be different for each person if the laterals are significantly different.
After looking at the posts again I realised Oreo may have got it back to front, and I didn't correct him.
He understood the idea that to compensate for less degeneration you use jfets of different idss values but did it in the completely opposite fashion.
Sorry for not noticing the error or for not correcting it, I was very protective of releasing too many ideas back then publicly.
I can't recall if I just didn't notice Oreo's error or I chose not to correct it.
Either way, positive or negative phase, I am sure his amp still sounds very nice.
In the end it will depend on the parts you have as to what the circuit should be.
I have exicon parts which I suspect are the same as other laterals with regards to their Vgs threshold hold values but I don't know that for a fact.
So the circuit could be different for each person if the laterals are significantly different.
Last edited:
I have a sim with a 13ma J74 and an 11ma K170.With 13.1 ohms on the K170 and 7.5 ohms on the J74 I get a 2nd harmonic phase of +57 degrees.I have experimented and found that in Ltspice a + 2nd harmonic as viewed in the spice error log equates to a negative phase 2nd harmonic.My jfet models are good ones so I'm not sure why you make it sound like you cannot use a greater Idss J74.I don't have a distortion analyzer but I can't see how spice could be completely wrong with such a simple circuit.Both jfet power dissipations are around 150mw.
Ideally you want more gain coming off the negative rail than the positive rail so it's operating more like a J2 circuit than a push pull.
If you had more gain coming off the positive rail it would be more like an inverted J2 circuit.
Yes normally a positive phase lag at a full 90 degrees usually gives a negative phase harmonic.
But there is also a point where it is not positive or negative, you have the 2nd harmonic with peaks and dips when the fundamental is "neither up nor down" - The grand old Duke of York.
The phase lag in ltspice can be very misleading, I always visually check the residual with a notch filter.
I wouldn't trust ltspice unless you have checked that Vgs threshold and transconducatance match your lateral mosfets.
As well as the jfets obviously.
This circuit is both sensitive to Vgs threshold and transconducatance.
Since both the Vgs threshold and transconductance affect the gain on the input stage. On my lateral mosfets having Vgs threshold of 0.15 V and 0.55 V, N to P, this leads to a situation where the drain resistance on the N channel Jfet is 50% or more higher than on the P channel. Just lowering the source resistance at the P channel Jfet doesn't help so much since the current then increases and you now have to reduce the drain resistance over the P channel Jfet to adjust both dc offset and bias of the output stage. The way to counteract this effect is to use a P Channel Jfet with a lower Idss. Hoping at least from an explanation point of view this is making sense.
Having said all of that, I think the effect of having perfect negative or positive, or somewhere in between (which has never been discussed), is probably not worth losing sleep over.
I believe the best qualities about this amp is the zero degeneration on the output, the positive current feedback loop, low feedback, and the almost zero temperature coefficient needing no thermal compensation.
And finally I am not saying your amp is wrong just that I probably should have explained myself better but felt I couldn't since I had originally built this thing before the product had even been released or even before it was properly reviewed.
I never tried to produce a carbon copy of the F7 I just used the circuit shown in the manual and added a positive current feedback loop as well as the information Nelson freely gave us and I designed it the way it made sense to me or how I envisaged it should be designed.
If you had more gain coming off the positive rail it would be more like an inverted J2 circuit.
Yes normally a positive phase lag at a full 90 degrees usually gives a negative phase harmonic.
But there is also a point where it is not positive or negative, you have the 2nd harmonic with peaks and dips when the fundamental is "neither up nor down" - The grand old Duke of York.
The phase lag in ltspice can be very misleading, I always visually check the residual with a notch filter.
I wouldn't trust ltspice unless you have checked that Vgs threshold and transconducatance match your lateral mosfets.
As well as the jfets obviously.
This circuit is both sensitive to Vgs threshold and transconducatance.
Since both the Vgs threshold and transconductance affect the gain on the input stage. On my lateral mosfets having Vgs threshold of 0.15 V and 0.55 V, N to P, this leads to a situation where the drain resistance on the N channel Jfet is 50% or more higher than on the P channel. Just lowering the source resistance at the P channel Jfet doesn't help so much since the current then increases and you now have to reduce the drain resistance over the P channel Jfet to adjust both dc offset and bias of the output stage. The way to counteract this effect is to use a P Channel Jfet with a lower Idss. Hoping at least from an explanation point of view this is making sense.
Having said all of that, I think the effect of having perfect negative or positive, or somewhere in between (which has never been discussed), is probably not worth losing sleep over.
I believe the best qualities about this amp is the zero degeneration on the output, the positive current feedback loop, low feedback, and the almost zero temperature coefficient needing no thermal compensation.
And finally I am not saying your amp is wrong just that I probably should have explained myself better but felt I couldn't since I had originally built this thing before the product had even been released or even before it was properly reviewed.
I never tried to produce a carbon copy of the F7 I just used the circuit shown in the manual and added a positive current feedback loop as well as the information Nelson freely gave us and I designed it the way it made sense to me or how I envisaged it should be designed.
Last edited:
I build the my F7 some time ago and I found the difference between a h2 neg and pos phase very big in sound terms.
But I confess I had no problem hearing the pod phase either!
:--))
Measuremt with Diana will take some time, in the case resides the J2 at the moment.
Thank you for your long descriptions and care!
But I confess I had no problem hearing the pod phase either!
:--))
Measuremt with Diana will take some time, in the case resides the J2 at the moment.
Thank you for your long descriptions and care!
I used the jfets I had on hand,I didn't purposely buy them ,I made do with what I had,i'm retired and on a budget.With an fft probe,software and some P3 adjustment my amp would be 90 degrees,i'm just happy the way it sounds now.
Last edited:
and me... I liked even the pos phase more!
So Mr. G,what difference did you hear? If I may ask.
F7 clone for LX Minis
I decided to embark on an LX Mini speaker build and was looking for a class A amp for the high-end drivers. I will be using Papa's analog crossover - board and parts from the Store.
I was about to start on an F6 clone build but got intrigued by the F7 Review thread. I started (like most of you guys) with a bunch of sim runs. But what finally got me off my butt and gathering parts was when I got LTspice to spit out this lovely FFT. The THD at 30 watts is around 0.5%, but it is almost entirely even harmonics (fine with me). The 3rd is 120db down!
I decided to embark on an LX Mini speaker build and was looking for a class A amp for the high-end drivers. I will be using Papa's analog crossover - board and parts from the Store.
I was about to start on an F6 clone build but got intrigued by the F7 Review thread. I started (like most of you guys) with a bunch of sim runs. But what finally got me off my butt and gathering parts was when I got LTspice to spit out this lovely FFT. The THD at 30 watts is around 0.5%, but it is almost entirely even harmonics (fine with me). The 3rd is 120db down!
Attachments
Oops - I goofed. Obviously the third is NOT that far down! But still not bad.
I should add that I am using the LTspice built-in models for LSK170B/LSJ74B for the fets and the Ian Hegglun/keantoken lateral models found here for 10P20/10N20 Exicon parts. And here is the LTspice log for the previously mentioned FFT.
I should add that I am using the LTspice built-in models for LSK170B/LSJ74B for the fets and the Ian Hegglun/keantoken lateral models found here for 10P20/10N20 Exicon parts. And here is the LTspice log for the previously mentioned FFT.
Last edited:
