The MONGREL (supersym II)

[ostripper]

Originally Posted by AndrewT

post257.
The NFB tapping should be from the same output pad as the Output L//R.

BTW,
I think it is better to mount the L//R off the PCB and put it in the route from PCB to speaker terminals. Keep the Zobel C+R very close to the output devices.

Andrew , why did you not advise Quasi to do that ?? (post 254 , picture 2). As a project I am already at 2 boards per amp ... no external L//R. I can still do the NFB routing , heck I will , just to put this to sleep AND be better than the others.
This leaves me with a choice with the PB250 , which is worse taking NFB from halfway between 2 opposing outputs where I do now or running a real long trace right to the output pad. A way around this actually would be to run a grounded shielded wire right from the NFB input of the voltage board directly to the L//R pad , bypassing all the board contamination.


OS

[ostripper]

See pix 1 below and tell me how the signal would be different from the present NFB point or the pad with this style layout. I won't argue the point .. as I will document AND add 1 more pad to run a shielded NFB from the present jumper to the L//R pad (method 2).

Pix 2 is the PB 60 , absolute NFB takeoff , right at the pad.... shorter route.
OS

[keantoken]

Ostripper, when you run that wire, what node should the shield be connected to?

Also, the 220pF input cap makes me nervous. What if you get a high source impedance?

- keantoken

[ostripper]

Originally Posted by keantoken

Ostripper, when you run that wire, what node should the shield be connected to?

Also, the 220pF input cap makes me nervous. What if you get a high source impedance?

- keantoken

If you run the shielded wire just tack the ground shield to the voltage board ground connection on just 1 end of the wire . I did this on the first "frugalamp" (like the AX) no problems.

I am running the 220p on the "supersyms" now, no problems. On both the "supers" and these new ones , the input is shunted by a 100k resistor between the input and signal ground.
Do you see my point on the NFB , Kean. The middle of the op trace is the "fulcrum" for the opposing voltages coming through the emitter resistors .. each 40mm spacing between each p / n emitter pair is a separate amp in itself . To put the shield right to the L//R trace IS better , I agree ... but only by a VERY small degree. On a typical left - right layout (actual traces meeting at a center point) , going off-center by even a few millimeters can be seen as corruption and a bad choice for a takeoff point. I have no traces , almost like hard wiring (emitter to bus ). PB250 is a "wire with gain" Style , able to let even "beefy" MJL21193/4's "drop the load" into that center bus.

I am going to even use small gauge copper solderwick to augment the traces (V+ / V- / OP bus) Overkill in everyway.
OS

OS

[keantoken]

Okay, it's been several posts, but... CCS DISTORTION!

Yes, solid state CCS's have distortion because of their nonlinear output impedance (very nonlinear). I like the thought about avoiding phase/response anomalies in the output impedance, but output impedance itself is important because it determines how much distortion is injected by the CCS.

I suppose the question is, at what point does CCS distortion exceed capacitor distortion? Does it?

Here are THD plots of the CCS's, at 34V pk-pk, 10mA, commonly found in the VAS.

Also, would it not be okay to connect the shield to the output and leave it unconnected at the other end?

- keantoken

[ostripper]

Originally Posted by keantoken

Okay, it's been several posts, but... CCS DISTORTION!

Yes, solid state CCS's have distortion because of their nonlinear output impedance (very nonlinear). I like the thought about avoiding phase/response anomalies in the output impedance, but output impedance itself is important because it determines how much distortion is injected by the CCS.

I suppose the question is, at what point does CCS distortion exceed capacitor distortion? Does it?

Here are THD plots of the CCS's, at 34V pk-pk, 10mA, commonly found in the VAS.

Also, would it not be okay to connect the shield to the output and leave it unconnected at the other end?

- keantoken

Concerning the baxendall CCS , I got it to sim in isolation no problem. when I used it to source a VAS or LTP , it "fell apart". At under 3ma or with a large modulated load (vas) it was quite "squirrelly "

Maybe if you could "plug it in" to an amp I could see it's benefits.

You should ground the shield , as this would negate any inductances from the OP bus or cap returns ...as you would have the wire under the board "sandwiched" between the FR-4 and the heatsink/chassis.



OS

[keantoken]

Did you use the "normal" version, or the tweaked one? The tweaked ones may give trouble... After all it is only an experiment. Try the one that I just posted in the schematic, exactly as shown.

- keantoken

[ostripper]

Originally Posted by keantoken

Did you use the "normal" version, or the tweaked one? The tweaked ones may give trouble... After all it is only an experiment. Try the one that I just posted in the schematic, exactly as shown.

- keantoken

That is what I said , used as a "strait" CCS , as you post them .. they work - all of them. But , I try to plug them into an amp they "poop the bed". BTW , I tried the tweaked one first then the normal one , both worked in isolation perfectly but I had great difficulty with the more complex application.

here, plug it into the AX (2 CCS's), see if you can get it to work.. (below)
OS

[keantoken]

Originally Posted by keantoken

Try the one that I just posted in the schematic, exactly as shown.

- keantoken

IF that doesn't work, try this.

- keantoken

[wahab]

Originally Posted by keantoken

I suppose the question is, at what point does CCS distortion exceed capacitor distortion? Does it?

- keantoken

The question is rather what is the real importance of a CCS
as a THD source..

THD for a classical CCS implemented with a bjt and a voltage
reference is in the order of a few ppm, while the distorsion
due to the VAS itself is about 1 % and soaring with high
voltages swings...