L'Amp: A simple SIT Amp

[Rodeodave]

Okay, so I reworked the bias supply and changed the grounding scheme a tiny bit. Input ground, bias ground and the VFET's source ground-connection now meet at the VFET and go to the respective star ground from there.

With the inputs individually shorted to ground I get zero hum through headphones (HD800, with 10R in parallel) connected to the output. As soon as I leave the inputs open or connect a source, there's audible hum. I'm kinda running out of ideas.

The output ground is connected to the star ground. Should I route it to the VFET (source) and ground it to the star ground from there? But if that connection is what's causing the hum, why does shorting the units silence the amp?

 

[Zen Mod]

catch is in post #3071

care to elaborate with some sketch , wtf. is that?

 

[plasnu]

Ground loop. Have you read this before? This article helped me to understand about grounding a lot.

Audio Component Grounding and Interconnection

 

[Rodeodave]

Oh boy, here we go. This is a sketch I drew up before starting with the build. It's surprisingly accurate, albeit ugly.

Edit: I must be getting close. Like I said, with the inputs shorted they're sub-mV noise. It must be the grounding.

 

[Zen Mod]

see enclosed ; bottom is same as top

do you see some logic in there ?

 

[Rodeodave]

If I'm reading the pcb correctly, POS and NEG is the input side, and B+ and B- are the outputs. The ground plane is slotted to route the ground currents in a start-ground fashion.

My amp build quite visibly has a star ground too (which is not working 100% it seems). What am I missing?

 

[Ben Mah]

Looking at your diagram in post #3084 and looking at ZM's board, your star is at the input side instead of the output side.

Your star is also not centred between the left and right sides.

 

[Zen Mod]

If I'm reading the pcb correctly, POS and NEG is the input side, and B+ and B- are the outputs. The ground plane is slotted to route the ground currents in a start-ground fashion.

My amp build quite visibly has a star ground too (which is not working 100% it seems). What am I missing?

no Bubba

Pos. GND and Neg are side to load

other side is side leading from diode bridges

 

[Rodeodave]

Okay, I thought that the first cap as seen from the rectifier is the lowest impedance point of the PSU and is thus used as the start-ground location. Guess I'll move the star-ground point to the ones furthest away and see what happens. Should be easy enough with the screw terminals and all.

 

[Zen Mod]

your GND take out ( leading to load) point must be equidistant to terminals of these two central caps , and also need to be mounted not in area of greatest pulsating currents

one Al plate would do the job - with 3 holes ; two holes for caps terminals , one hole for GND take out point

these 3 holes pretty much forming a triangle

see enclosed

 

[Rodeodave]

Okay, so here's the revised grounding scheme.

Edit: If this works, I think I'll be getting this on a T-shirt.

 

[Zen Mod]

that could work

if some buzz persists , try fat wire bridge between two channel main GNDs , meaning , fat wire tying two sources directly

 

[Dennis Hui]

That would make a great T-shirt!

 

[Rodeodave]

Aight, I rerouted the amp's ground connections and fired it up again. Sub-mV noise with a source connected, just barely audible through the HD800. On most classical records the background noise is above that...
I'm not sure if I can get it any more quiet.

Time to vectorize that t-shirt print.

 

[Zen Mod]

 

[Rodeodave]

Oh, and I forgot to add...

Mighty Zen Mod saved the day!

Having the star ground at the PSU output side really was news to me.

 

[Zen Mod]

it's irrelevant who helped , what's important is that Amp is Happy One

 

[maxlorenz]

Mighty ZM also wrote:

and also need to be mounted not in area of greatest pulsating currents

Think currents.
Add this feature to your PS caps with a copper or Al plate, as ZM so beautifully showed.

Good luck.
M.

 

[Zen Mod]

in this case he's safe , simply because each (of these two) cap is on single rail , each is having own separate ground feed and there is no current flow between them

in short - Lucky Bztrd

 

[Botte]

Is the amp quieter without a source connected, shorted input?