150mA is more than plenty. Especially when in Darlington...
Only the SIT/V-FET/Triode need have a real Mu. Ordinary
BJT inherits this property when so controlled.
The anode of the Triode or V-FET device has a low impedance.
This node acts as INPUT for voltage feedback as well as output.
Can be abused to moderate VCE of a bipolar that passes much
larger currents.
Only the SIT/V-FET/Triode need have a real Mu. Ordinary
BJT inherits this property when so controlled.
The anode of the Triode or V-FET device has a low impedance.
This node acts as INPUT for voltage feedback as well as output.
Can be abused to moderate VCE of a bipolar that passes much
larger currents.
Attachments
Gordy said:
No feedback is an absurdity. Mu is essentially a local feedback.
So too are all unity gain followers, the greatest feedback of all...
What really makes Mu and unity followers special, lack of delay
feedback shifted by external time constants...
The feedback happens right there in the channel in real time.
kenpeter said:
Mr. kenpeter, replying without seeking to understand my preceding post is the absurdity. (I can't be bothered to explain. Never mind).
Back to subject: good luck with your circuit development Vladimir.
Gordy, review the papers of O.H.Schade again if you have not done
so recently. To realize WHY a triode behaves linearly as it does.
Yes, his most famous paper is not about triodes, but about the 6L6
Beam Power Tube. But actual behaviors of the space charge within
the envelope of every tube are greatly clarified, and these physics
directly relevant to Triodes.
There is a feedback. The gradient of electron density in the space
charge is an unseen extra screen where distributed local feedback
naturally occurs within a Triode. (Gradient of carriers in the channel
of a V-FET or SIT, same same....)
I'm not sure if we are in any actual disagreement regarding these
facts? Or merely choosing incompatible words to describe similar
ideas???
so recently. To realize WHY a triode behaves linearly as it does.
Yes, his most famous paper is not about triodes, but about the 6L6
Beam Power Tube. But actual behaviors of the space charge within
the envelope of every tube are greatly clarified, and these physics
directly relevant to Triodes.
There is a feedback. The gradient of electron density in the space
charge is an unseen extra screen where distributed local feedback
naturally occurs within a Triode. (Gradient of carriers in the channel
of a V-FET or SIT, same same....)
I'm not sure if we are in any actual disagreement regarding these
facts? Or merely choosing incompatible words to describe similar
ideas???
kenpeter said:
Yes, I have read that some time ago and found it very interesting, thank you.
In my 'simple country boy' view of the world the internal feedback is so fundamental to the device that I assume that it needs no comment or extraordinary consideration. Hence my comments related to global (external component) nfb loops from output to input.
Have a nice day.
homemodder said:
There is no P version from russian manufacturers. For this better to search for SONY 2SK60 - 2SJ18
Gordy said:
Are there any explainations why the triode-like characteristics are good for human hearing response, except for better interstage interaction ? I am sure that it is not related to THD measured with resistive load.
VladimirK said:
Unfortunately I can not provide objective references, therefore I can only offer this as a personal subjective observation.
I guess that it is a combination of relatively low-order distortions combined with (some aspect of) not needing global feedback.
Gordy said:
Yes, intrinsic linearity and low output resistance of a triode (i.e. an easy next stage connection) allow for no or small NFB. This simplifies complete schematics, eliminates multiple external parasitic capacitances, inductivities, NFB "signal echo' effects, etc.
Wavebourn said:
Yes, I also find that specs of KP926A are very so-so, except for triode-like characteristics and high maximum Vds=400V. The sound is not bad, nevertheless. Anyway, I am sure that SONY 2SK60, if they have similar triode character, would sound even better.
In order the triode-like characteristics be able to cause something positive, they should have possibility to fight against some difficulties at the input of the next stage. What difficulties? I guess we should measure frequency dependence of input impedance of the next stage with "difficult" speaker connected. I guess, backward influence of difficult speaker can be measured at the input of power follower or power amplifier.
With reference to the schematic in post #1, I guess it would be okay to use IRFP240 for Q4 and IRFP9240 for Q2 instead?
Q1 should be jFET type, maybe in cascode connection like Zen9, or it could be some modern SiC jFET transistor.
I have a couple of KP926A for Q1 already 🙂
I was wondering about the cap multiplier and cascode FETs.
I was wondering about the cap multiplier and cascode FETs.
Last edited:
I just ordered a few of the Rp926a devices and look foward to building this but wonder about the ccs. Would a ccs like you use be superior to just using a Dn2540 as a ccs ? Oh I know the Dn2540 is only rated at 15 watts so it would need very good heatsink but aside from that any reason not to use it?
Using MOSFET-based CCS is possible. But, in case of KP926A used as a dynamic load for active (lower position) KP926A, we hope for some mutual compensation of distortions. It would be interesting to compare THD with both solutions.