I remeber that Wrenchone talked about it (gate leakage) with the RO85 in his thread. I know much less than you, so the dynamic effects of gate leakage are indeed beyond me, but we know it is something to consider, as Nelson put low 10k gate resistor. I wonder if the differences in fets is due to yfs at the lower Vgs levels seen during ac operation. At a certain Vgs, you get a bias current of X. In my samples, these are very close. Signal rides on this point, up and down the transfer curve. The curves are also close above this point. Perhaps the issue is what is happening below this point. At lower Vgs levels, I tend to see more variations than at higher Vgs. I would assume that this is one of the underlying factors that cause higher distortions at lower bias levels. You get into the bending part of the curve and away from the linear portion.
Last edited:
Hi tea-bag, I looked for a pair that gives k2 lower or at the same level as k3 without Rs in the output. And then I implemented the R in the lower side to get higher k2 with positive phase and the same on the other side....
As I said one channel works... I am looking for the second channel pair....
Not to forget many people did this already Iquham, flocchini and so on....and reported on the sound..I am only following their footsteps...
Tea-bag had at the first try very good distortion values, I think he cares for Vgs and transconductance! Like buzzforb!
I am still waiting for the probes to use with my Mac to look at the output spectra.
My adjustments have been done by ear and of course what I am listening to is what sounds best to me in my room, with my speakers etc. I am using the F6 to drive the PHL 1120 in my system. This speaker covers the range from 300 -5000 Hz. To me the F6 is an incredible match. It replaces a F3 which I previously used. I have also tried an Aleph 30, a ZenLite and my F4 modified with interstage transformer there.
I use no front end degeneration, no degeneration on Q2 and .025 ohms on Q1.
This is the same on both channels. I did try to put different settings on Q1R and Q1L but it was difficult for me to hear differences. I think it is best to say nothing jumped out
When I learn to measure the spectra I plan to vary Q1 from 0 ohms up to .05 ohms to try and determine what mix I like. I then will try to get that mix on each channel with its own unique degeneration R for Q1. What I am hearing is that it may not be possible because JFet differences in parameters. If that is the case I have to say - guys, I am going with sounds best to my ears - I have to spend some time enjoying the music even though I want to tinker.
Best to all and most of all thanks for what you all continue to teach me
My adjustments have been done by ear and of course what I am listening to is what sounds best to me in my room, with my speakers etc. I am using the F6 to drive the PHL 1120 in my system. This speaker covers the range from 300 -5000 Hz. To me the F6 is an incredible match. It replaces a F3 which I previously used. I have also tried an Aleph 30, a ZenLite and my F4 modified with interstage transformer there.
I use no front end degeneration, no degeneration on Q2 and .025 ohms on Q1.
This is the same on both channels. I did try to put different settings on Q1R and Q1L but it was difficult for me to hear differences. I think it is best to say nothing jumped out
When I learn to measure the spectra I plan to vary Q1 from 0 ohms up to .05 ohms to try and determine what mix I like. I then will try to get that mix on each channel with its own unique degeneration R for Q1. What I am hearing is that it may not be possible because JFet differences in parameters. If that is the case I have to say - guys, I am going with sounds best to my ears - I have to spend some time enjoying the music even though I want to tinker.
Best to all and most of all thanks for what you all continue to teach me
Second harmonic phase
On each channel I swapped devices until both channels had positive phase on the second harmonic.
I’m not sure how you would do this without a spectrum analyzer. I used visual analyzer as it has a nice phase window to show phase.
I did play with a pot across the positive coil and could make it flip from negative to positive second harmonic phase, and in the end I just swapping the devices and this gave the same results. I have not tried this with the source resistors and the zen pots.
On each channel I swapped devices until both channels had positive phase on the second harmonic.
I’m not sure how you would do this without a spectrum analyzer. I used visual analyzer as it has a nice phase window to show phase.
I did play with a pot across the positive coil and could make it flip from negative to positive second harmonic phase, and in the end I just swapping the devices and this gave the same results. I have not tried this with the source resistors and the zen pots.
Tea-Bag. The following explanation of F6 operating in two additioal modes maybe such a diagnostic circuit.
- Keep Pass NFB
- Disable the AC drive to the upper FET; but maintain its Vgs bias. Now this FET operates as a constant rather than a modulated current load to the bottom FET. Do your tests and determine whether you like its music performance.
- Repeat step 2 for the bottom FET instead
whatever - short and sweet THD spectra measuring Cook Book would be beneficial to plenty of Dodos , lurking though this tread
could not find THD Spectra Measuring for Dummies
for myself
Last edited:
Tea-Bag: Here is potential diagnostic test with the F6 circuit:
- Disable Pass NFB
- Reverse the phase of one of the secondary windings. In an ideal world, the power output of this amp will be zero. But; measure it anyway and listen to its performance [signal and music] with headphones.
- Enable Pass NFB, test and listen again as in bullet #2
Use headphones where possible to disable room effects and standing waves so as to better control the variables under test.
I suspect that transconductance and internal source resistance are the most important parameters of the SS-R100s for use in the F6. Out of 6 R100's in my possession, I have been able to only find one well matched pair.
My experiments with Fairchild FQA28N15 MOSFETs in the F6 indicate that their gate capacitance is too high -- bad hi frequency roll-off.
My experiments with Fairchild FQA28N15 MOSFETs in the F6 indicate that their gate capacitance is too high -- bad hi frequency roll-off.
You have as much experience with it as anyone, so I cant argue with your thinking.
buzzforb and lhquam. Maybe matching of FETs can also be done dynamically in addition to the standard static one. Suppose that the hardware preceeding the FETs are matched; especially the output voltage, its frequency response, phase etc.. of the secondary windings of the transformer. Reverse the phase of one secondary winding relative to the other. Slowly increase the input signal for the safety of the FETs because both will turn on and turn off simultaneously. The simultaneous turn on is the problematic one in this dynamic test; because they have 46 VDC supply across them. If the FETS are matched, there will be no output voltage. A lack of exact matching will generate an output voltage which may be proportional to the mismatch; small or large.
If the simultaneous turn on of the FETs turns you off then maybe use a variac to lower the power rails and limit the inadvertent over dissipation of the FETs.
O.K......
it took me some hours to find out of my 8 R125 the two pairs that gave me around 0,012% THD and a distortion spectra where k2 and k3 are at the same level.
I wanted this position to go later more easily for negative and positive phase.
If it is possible to find the pairs only by measuring the transconductance of the parts outside of the circuit then I am an idiot.....not to build the circuit Nelson suggested 1993...!
So buzzforb and tea-bag can measure the transconductance, so it would be very interesting to see, if they get a predictable distortion picture.
Tea-bag please insert in your F6 two pairs of SS with same transconductance and let us see the distortion picture. k2 should be reduced, shouldn't it?
Buzzforb can you see distortion?
and of course Ihquam did you see a correspondence between equal transconduction and a reduced k2 compared to k3?
when my 10 Ohm pots are coming I will continue....
because changing the source resistors changes as well the bias, as the offset values...... ! Too much work! Am I correct to suppose that using the pots the bias and the offset is not changed?
Gute Nacht!
it took me some hours to find out of my 8 R125 the two pairs that gave me around 0,012% THD and a distortion spectra where k2 and k3 are at the same level.
I wanted this position to go later more easily for negative and positive phase.
If it is possible to find the pairs only by measuring the transconductance of the parts outside of the circuit then I am an idiot.....not to build the circuit Nelson suggested 1993...!
So buzzforb and tea-bag can measure the transconductance, so it would be very interesting to see, if they get a predictable distortion picture.
Tea-bag please insert in your F6 two pairs of SS with same transconductance and let us see the distortion picture. k2 should be reduced, shouldn't it?
Buzzforb can you see distortion?
and of course Ihquam did you see a correspondence between equal transconduction and a reduced k2 compared to k3?
when my 10 Ohm pots are coming I will continue....
because changing the source resistors changes as well the bias, as the offset values...... ! Too much work! Am I correct to suppose that using the pots the bias and the offset is not changed?
Gute Nacht!
if you meant on ZAP , then that's correct ; they're working strictly in AC domain