instead of three power =/-12v and enough Vb, the input can be made cap coupled with divider/polarizing for the input buffer or p-mosfet; and ditto the gate polarise of the n-mosfet. Simplifies things. Man, you can even use an LND150 for that 😉
Thanks! Yes, making the gate of the PMoS furthere neg can also work to get out of the region of high input capacitance. Ordered a bunch of new low Crss MOSFETs to play with. We’ll see how it goes.
Meanwhile, the CCS-loaded version with a 1k output resistor sounds really promising.
Meanwhile, the CCS-loaded version with a 1k output resistor sounds really promising.
Very interesting ra7! I did a Spice, I could not find the models for the FQA19N20 and the FQA12P20 so I had to take the IRFs.
I had to lower the voltage to get the 40mA and so the voltage at the upper drain is only 40V. I do not know why.
Here the Spice data, of course your real values are more important.
I had to lower the voltage to get the 40mA and so the voltage at the upper drain is only 40V. I do not know why.
Here the Spice data, of course your real values are more important.
With the smaller FQP and their available Spice models the values get closer to what ra7 measured. Current is 43mA.
Thanks Generg! Interesting sim. I didn’t do it in Spice and built it instead. I can sweep from 25V all the way to 110V using the pot. So, maybe the IRF models aren’t representing the FQP FETs very well but in practice it should work just as well.
I did open Spice yesterday after a long time, mainly to play with the Schade feedback resistors and see what it does to the curves. But I don’t think I am getting meaningful results. Will have to just plug it into the Frankentracer and get curves the old-fashioned way.
edit: just saw your next post. Congrats!
I did open Spice yesterday after a long time, mainly to play with the Schade feedback resistors and see what it does to the curves. But I don’t think I am getting meaningful results. Will have to just plug it into the Frankentracer and get curves the old-fashioned way.
edit: just saw your next post. Congrats!
I only wanted to do a first step I can do at the moment....
:--))
Bandwidth in Spice is very high, you have no oscillation problems? I have the FQA at hand, so a build is possible, but what is the Salas PSU you used?
:--))
Bandwidth in Spice is very high, you have no oscillation problems? I have the FQA at hand, so a build is possible, but what is the Salas PSU you used?
Bruder - fine tuning of upper mosfet gate voltage level is very important - point is that you want to meet sum of lower mosfet Ugs and upper mosfet Ugs, where they need to be
I was getting about 50 kHz but that is with the output cap. As far as I can tell, there is no oscillation but I don’t know exactly how to test for it. The soundcard doesn’t have the bandwidth tomeasure that high. Bias stays steady after initial warm up in both channels.
I used the SSHV2. It is in the LuminAria chassis in rats nest form. I want to built it with a zener follower like the BOSOZ so that I can test at higher currents. Right now limited to 40 mA because of the shunt reg.
Oh what am I saying… the output cap is a high pass filter so it doesn’t affect the bandwidth up top.
Strictly speaking, the choice of output coupling caps can affect the upper frequencies. Large electrolytic capacitors will be inductive at high frequencies. That’s why you sometimes see them bypassed by smaller value film caps. Phase offset can be noticeable as well.
Interesting concept of a signal feed with a p-fet, and the schade feedback on the upper device.
- This should also work with a SIT without a feedback being needed . . . at a higher current . . into a choke, the Ben Mah type. For instance I have a 2 amp P-Fet that has a threshold of 7 volts. It will also retain the correct phase.
- Good frequency behaviour, the gate being tied to earth, the 10nF gate charge vanishes and is out of the equation.
- No bias needed, it will auto-everything.
- [including flashing
the FET?]
Last edited:
it is tempting, but not so simple
SIT at top - that probably means power amp, means higher Iq, means fat mosfet down (T0247 being fat enough)
then , even if working as SFollower, lower mosfet is still better if driven with low impedance, means some buffer in front
if there is buffer in front, why not driving SIT directly with buffer ...........
in the end - matter of choice, technical and from likeness perspective
SIT at top - that probably means power amp, means higher Iq, means fat mosfet down (T0247 being fat enough)
then , even if working as SFollower, lower mosfet is still better if driven with low impedance, means some buffer in front
if there is buffer in front, why not driving SIT directly with buffer ...........
in the end - matter of choice, technical and from likeness perspective
Agree, if discuss may take this thread.
but the tempting part is - the very good ‘drive’ of gate, nowhere else to be gotten.
I must try it with a to-3 SIT. Do ; not think 🤢
but the tempting part is - the very good ‘drive’ of gate, nowhere else to be gotten.
I must try it with a to-3 SIT. Do ; not think 🤢
The main motivation for me was to get triode curves from non-triode, cheap parts. Schade is the answer for that but how do you overcome the input impedance? Drive it from the source and free the gate. That’s where this comes from.
I actually tried driving the KP926 using the source, i.e., this topology, but it didn’t do any better than directly driving the gate. If you have an SIT, buffer is a fine choice.
I actually tried driving the KP926 using the source, i.e., this topology, but it didn’t do any better than directly driving the gate. If you have an SIT, buffer is a fine choice.
Luminaria is gone, I sent it to friend, who died and I did not get it back.
It had a really lovely sound, but for me the bass was too shy.
So I did not build one more.
It had a really lovely sound, but for me the bass was too shy.
So I did not build one more.
yes, and it is very impressive, and repeatable. Compare hat with SIT that you buy, they may have operating points all over the globe, have different curves hence different sound - and thus different way of projecting that cherished image into the room.
[Currently i am listening again to my 300B that is getting better by the day. Totally different from my SITamp that is recieving som surgery - a new more powerfull driver - DN2540/BD140 - and power supply - SMPS. It is fun but not yet that barrel of monkeys]