Grounding scheme and loop breaker, need advice..

[endia]

Hi,
for a while, i'm building a SymAsym amplifier.
http://www.lf-pro.net/mbittner/Sym5_Webpage/Images/Symasym5_2_eagle.GIF
before starting to chassis work, i'm trying to finalize chassis layout and grounding scheme..

so my question is;
the 10R between signal and power ground is confusing me,
should i connect both IGND and VGND on the amp board to the star point? or this layout is quiet fine?

also i made a loop breaker board and set the star point onto the board.. board will be secured directly to the chassis screw..
what is your opinion and suggestions?
thanks in advance..

edit: diodes on the lb circuit are 1N5408, are they enough for the purpose?

 

[TANDBERGEREN]

The starpoint on the board directly to chassis?

I wouldn't do that.

That starpoint should be connected directly to the "branches" on the earth-card on the left there.

I would have soldered the connection though. (Screwterminals on that board?)

 

[endia]

sorry, my explanation is not clear,
no, star point is not connecting to chassis, loop breaker safety earth point is on the board and will be secured to the chassis..

 

[jcx]

monoblocks are so mcuh easier to gnd "right"

each channel's output and output gnd return should be twisted pair terminating on each channel board; no reason to make big loops with the loudspeaker load return current going to the PS before getting back to the boards

if your pic really shows the amp boards then my condolences – shows the classic poor layout with +,- power entering at opposite edges of the board

this makes it hard to do a proper low loop area PS, pwr gnd return – so you will have the half-wave rectified Class B currents possibly creating nonlinear magnetic coupling distortion

twisting PS gnd return with each of the +,- PS leads, you would have to hack the amp board to connect each PS gnd to near the respective on-board PS cap “gnd” side
the on board PS C are big enough that this may not matter as much as just getting very low Z amp board to PS gnd connection


since you don't show a volume pot there may be signal gnd options - is this driven from a low output Z preamp?

 

[endia]

Hi jcx, thanks..
moved speaker gnds to the amp boards as your suggestions..

twisting PS gnd return with each of the +,- PS leads, you would have to hack the amp board to connect each PS gnd to near the respective on-board PS cap “gnd” side the on board PS C are big enough that this may not matter as much as just getting very low Z amp board to PS gnd connection

but i'm not sure, if i understood correctly what you mean..
moved +, gnd, - connections to the underside of the caps.. does this solve the problem?

by the way pcb is from the original design from the Michael Bittner,
http://www.lf-pro.net/mbittner/Sym5_Webpage/Images/symasym5_board_preview.GIF
and as far as i know, so many people built this amp with no problem..

since you don't show a volume pot there may be signal gnd options - is this driven from a low output Z preamp?

i'm planning to use just a 22K pot which is what MikeB suggested.. but i'm planning to put it in a dac enclosure..
thanks..

 

[jcx]

I was suggesting the extra pwr gnd wires to the amp board on-board 1000u caps - like the "hot fix" on-board amp pwr gnd jumper this could reduce some THD in careful measurements - may not make a big "sound" difference

the holes in the middle of the PS board gnd are fine on that end

any chance you would be willing to make a “Ground Canceling” buffer/driver in the DAC after the volume pot ( shown in Doug Self's books)

you should really have some sort of buffer after the volume pot - the high Z of the pot wiper will make the signal susceptible to noise pickup and will give varying fequency response with longer cables



a system may be acceptable with compromises vs “best” practice – sometimes you just have to try it and see what works in your system – decide if the “next level” is too inconvenient or expensive

 

[jaycee]

Signal grounds should go to their individual channels ground where they are disconnected by the 10 ohm resistor.

The speaker grounds should go to the GND of your power supply board. So should the VGND's from the amp boards.

Then, run one wire from the power supply board's GND terminal, to chassis ground, via your loop breaker. You may not even need the loop breaker. Try it without, and if you get hum when you connect a source to your amp, add it.

Make sure the grounds on your signal input sockets are isolated from the chassis. Make sure the ground on the IEC connector connects *directly* to your chassis - use a toothed washer to ensure grip.

 

[endia]

i think, understood finally
but i should split to long jumpers because they are already on the top layer..

i'll search for a ground canceling buffer but my knowledge may limits me to choose a suitable one.. would you suggest me a specific circuit?

thanks for your helps..

 

[endia]

above picture should be this..

 

[endia]

jaycee, thank you,
i think last scheme is corresponding with jcx's and your suggestions..
also thanks for the toothed washer tip, never thinked that before, i'll be consider it..

 

[jcx]

Self's web page: Balanced Line Technology shows a ground canceling circuit and lots more balanced line circuits, you do need a buffer between the volume pot and the divider resistors in most of those circuits
the ground canceling circuit should work OK with RCA, cold=shield - bypassing the gnd R at both ends with some C reduces the chance tht RF could be a problem - or could used shielded twisted pair XLR with "single ended" signal for higher RF immunity

could be done with your own op amps of choice, matched R or as an integrated example the INA111 is a possible instrumentation amp probably OK for audio

 

[endia]

well.. i have lots of reading parts..

i think i can add ground cancelling and buffer circuits later when the amp is running without problem..

jcx, thank you very much for helping..
regards.