Goldmund Mods, Improvements, Stability

Nico, I am sorry for suggesting that you were interpreting the graphs wrong. However I am well aware of how BW and transient response correlate, as I watch it often with the simulator. Also the only way the regulator would saturate is if the amp broke, so there is no worry of long recovery times, which only occur after saturation. Am I wrong?

The basis for a 100mA shunt current is that if we want a regulator to exceed simple electrolytic capacitors in specs, the pass device needs to be passing enough current to lower the reg's output impedance. The other alternative is to add more active components, which would almost certainly adversely affect speed.

A friend sent me some interesting pages:

Power Supplies

Power Supplies

- keantoken
 
Jacco, I've played around with things like this on the simulator. Most series regs, when you REALLY test heir behavior, have pretty foul negative pulse response. I tried to fix this by using a class AB output stage on series regs, just like in amplifiers. As it turns out this is difficult, especially since it's slewing behavior you're trying to correct and at these frequencies class AB is not a good idea. A better way I'll bet is to hang a gyrator (like the JLH Riple Eater) at the output, which can be faster and will absorb part of the load.

- keantoken
 
....Hmmmm

I can not seem to source the LSK389A except seperately from one supplier. And of course RFQ it.
(It's mine! ALL MINE!!!:devily:)

Seriously, anyone else have an input to this quandry?

Sourcing for the other components almost complete with MS Excel BOM. 😱:spin::crazy:

The HS is going to be a 'one-off' dooozie. :shhh::note:

Kean, I thank you for your hard work. Been looking around on the net for a very long time for a mixed amp design, and this is IT!!! Also kudos to Salas, OS, Jay, Andrew, Jam, mlloyd, and a few others (many posts ago
) for the inspiration for such a design. And never to forget Naggy for starting the fire!!!

I am no EE, but a tech only. I can build it. but no ee designs here. 😱
I can't hardly wait for this one!!!!!

Ok, ok, ok, :whip: (me no talkie, i must workie, workie)
 
Last edited:
More specifically, series regs tend to have very asymmetrical slewing.

If by push-pull you mean to rig the CCS to modulate as well, it could be done but might be difficult with MOSFETs because of the reactivity, and would likely decrease input isolation.

Kbeist, you can use the 2SK389, or matched 170's. There are members here who also know how to adjust the circuit to your particular choice of Jfet.

- keantoken
 
2SJ162/2SK1058

Are these good to go for the output of this project as far as FETs?

Will 80 volt rails B ok?😕

The board I am doing will give a choice of where you cut it to have from 2 to 6 pairs. 😀

Input please.

Thanks and lets us all PLEASE not let this project fade out...😉

 
He is talking about the difference between schema in post #1 and the "other side"'s schema, where T6 is connected in another way.

A member explained me that

T6 is set up as a simple diode shunted to the zener , it really serves little purpose. It is not a current mirror. A 2 transistor or diode/led shunted CCS would be more reliable and offer better PSRR.

BTW, the "other side" means http://www.diyaudio.com/forums/solid-state/174468-very-best-amplifier-i-have-ever-heard.html.

Regards,
Max.
 
Last edited:
Krisfr, the top transistors are all TO92, this is not right. The 2 KSA1381's are TO126.

Furthermore, the bottom current mirror transistors are KSC1845's - TO126.

We're getting there...

Jacco and/or others, could you post an example of a very good layout for the FETs?

- keantoken
 
Sharp Optonica SX-9100H amplifier is our new reference, since the best amp is mimesis 3 now. (Mimesis is Greek and means imitation/impersonation).🙂
 

Attachments

  • SX-9100.jpg
    SX-9100.jpg
    21.2 KB · Views: 457