Woow, thought that KC fans are mostly dead & buried! Good going Zene!
My 70's KC vinyl collection traveled 10000km with me from BGD to the oldest continent, 20+ yrs ago
Still my favourites. And Yes. And Jethro Tull...Lhquam: I copied the CCS schematic which uses TL783C [attached] from the Texas Instrument Data Book . Both CCSs need to be powered from the positive voltage rail; especially the lower one [actually positive rail to negative rail] as both have a 1 uF at the input port. Simple but effective.
buzzforb: you have an impressive build. Two output devices per channel are readily visible. Very neat. Which Pass amplifier is it, so I connect with its schematic?2
I am going to go listen now
50+mA ccs current is pretty high for a bootstrap capacitor, and besides, the idea behind current through the coil was to avoid capacitors altogether.@buzz
in BA2 OS , same as in F4 , bias gen is bootstrapped , so it's mid can swing above and below rails
@lhquam - same trick - bootstrap ; somewhere ; maybe across that additional resistor or ....
but - as I'm not for CCS through coil solution , I'll not give you complete schm
If I were to use bootstraps, I would be thinking in terms of a capacitor in the path driving the gate.
Power dissipation is the problem here. About 50mA current is required thru the transformer winding with an average of over 20 volts across the ccs when using 24V rails. The means 1 watt average power dissipation. We need a high power regulator or a ccs circuit which uses a power transistor for the main current.
and then hasta la vista baby ......... to simplicity and elegance
having 2 caps is small price for elegance , especially considering that their influence is exactly opposite in character , in every moment
2 caps and few resistors , without need for bigger xformers , which eventually will be needed , considering fact of pushing twice 40++mA through tiny nickel cored bstrd
having 2 caps is small price for elegance , especially considering that their influence is exactly opposite in character , in every moment
2 caps and few resistors , without need for bigger xformers , which eventually will be needed , considering fact of pushing twice 40++mA through tiny nickel cored bstrd
CCS noise problem?
One aspect of constant current sources that has not been discussed is noise. The noise current of the ccs multiplied by the DC resistance of the winding will be an additional noise voltage on the FET gate. I am not sure how this might compare with other noise sources, but it is likely to be much higher than the noise injected from the circuits shown in methods A, A1, and A2 of this post post http://www.diyaudio.com/forums/pass-labs/216616-f6-amplifier-78.html#post3126634.
One aspect of constant current sources that has not been discussed is noise. The noise current of the ccs multiplied by the DC resistance of the winding will be an additional noise voltage on the FET gate. I am not sure how this might compare with other noise sources, but it is likely to be much higher than the noise injected from the circuits shown in methods A, A1, and A2 of this post post http://www.diyaudio.com/forums/pass-labs/216616-f6-amplifier-78.html#post3126634.
At this point I am not advocating any particular circuit, but evaluating "out-of-the-box" alternative approaches. I thought that the ccs/dc-winding-resistance idea was worth looking at, nothing more.and then hasta la vista baby ......... to simplicity and elegance
having 2 caps is small price for elegance , especially considering that their influence is exactly opposite in character , in every moment
2 caps and few resistors , without need for bigger xformers , which eventually will be needed , considering fact of pushing twice 40++mA through tiny nickel cored bstrd
You are right about that. Will it be the reflected impedence of the front end driver? If so, the impedance might be fairly low.
TL783 is a high power positive voltage regulator [TO-220AB]. It has a published free air dissipation of 2 W at 25 Celsius, and 20 W on an adequate heat sink. The maximum recommended current is 0.7A at 125 V. Plenty!
Clearly lhquam you are chasing an experiment which you thought has value. As Mr. Pass will say: put it together and see what happens!
ZM: many methods to tan a skin to make leather. Each leather has its unique balance of valuable properties; an outcome of experiments.
My apology, the underlined is a mistake as written. [at] is the problem. The current [0.7A] times 125 V will violate the 20 W upper dissipation limit of TL783. Instead, operate the regulator at 0.7A such that the In-Out voltage differential is 28 VDC. Or operate the regualtor at an In-Out voltage differential of 120 V such that the output current is ~0.16A. Each of the above 2 situations do not violate the maximum dissipation constraint of 20W.
The TL783 appears to have a problem for this application. The datasheet shows a graph of output current vs. in-to-out voltage. The current rapidly falls off below 25v. A Spice simulation for the part shows that below 12V in-to-out the part will not maintain 50mA current.