Why do we connect channel grounds together?

[wes-ninja250]

Hey, All!

I can't hear a difference with my current set-up either way, but is there a reason channel grounds are normally connected together?

My GC (which is nearly done -- it needs heatsinks, rubber feet and a faceplate) is the typical two-PS BrianGT LM3886 kit.

The only difference I can think of when the channel grounds are shorted together is that this would provide a ground loop path for whatever's connected on the inputs. How can this possibly be good?

With the channel grounds *not* connected, I effectively have two completely separate amplifiers in one chassis, which happen to share transformer secondaries.

The "star ground" point I decided for each channel is the ground plane of the PCB. As it stands right now, the amp is quieter than my source.... with the inputs shorted, I can't decide if the hiss I hear when pressing my ear against the tweeter is seashell-ocean noise or if it's coming from the amp.. so as it sits right now, this is the quietest amp I have ever owned, including my much-loved Harmon-Kardon HK440.

Wes

[Mick_F]

You dont have to connect the channel grounds. Having two separate amps for the two channels with two separate grounds in one case is ok.

Mick

[wes-ninja250]

Thanks, Mick.

Any idea why this seems to be a common recommendation?

I'm trying to figure out any advantages to the configuration... the only one I've come up with is that a half-broken input connection might still sound okay..

For now, I think I will just leave them disconnected, I suspect that will make for a quieter amp when hooked up to crappy gear.

Oh, one more thing -- does it make a difference for bridging?

Wes

[Mick_F]

Originally posted by wes-ninja250
Thanks, Mick.

Any idea why this seems to be a common recommendation?

I'm trying to figure out any advantages to the configuration... the only one I've come up with is that a half-broken input connection might still sound okay..

For now, I think I will just leave them disconnected, I suspect that will make for a quieter amp when hooked up to crappy gear.

Oh, one more thing -- does it make a difference for bridging?

Wes

I dont think its a common recommendation. There is just no need to make a connection.

Have a look at
this site , where I described the grounding scheme for my Chipamp in detail.


Mick

[wes-ninja250]

Hey, Mick!

Great site, I wish I had found that a while ago!

Your grounding scheme is nice and a tight, going with a P2P construction really helps in that regard. You P2P guys must be very patient, that is some nice intricate work!

Your grounding scheme does keep the channels at the same ground potential, though; the single power supply enforces this.

Is that hunk of aluminium enough for a heat sink, or did you have to add some more? I have a similar-sized piece of stainless steel I am using, it gets quite warm even though it is connected to the rest of the chassis. I had planned on adding more heat sinking, but I snapped my #6 tap trying to put a blind hole in copper. GRR!

I also find the oscilloscope views interesting, I have a 'scope and an audio signal generator here -- but I've never actually tested an amplifier in more detail than 'does it work? Yup!' before.

Wes

[Mick_F]

The aluminium plate (actually its a book end) is enough. Only at very high listening levels an increase in chip temperature can be noticed.

Mick

[wes-ninja250]

Interesting -- can you estimate the running temperature?

These are just guesses, but my chips seem to run at about 50C, the heatsink at 40C near the chip, at about 25C at the extremities (suggesting stainless steel is not a great heat conductor). Ambient is about 17C.

It doesn't seem to alter much between quiet and loud listening levels, although I haven't listened to it loudly for more than a couple of minutes -- I'll get a chance to do that this weekend, though.

Wes

[Mick_F]

I only touched the Aluminium at the outside opposing the chips. There is may have been 60C at max. Away from the chip the temperature falls very quickly as Aluminium is a good heat conductor.

Mick

[AndrewT]

Hi Mick_f,
that rapid fall in temperature is a high thermal gradient in the aluminium heatsink. A good conductor of heat would hold almost the same temperature across the heatsink.
That temperature gradient is telling you that the sink is too thin.


A rough estimate for your heatsink thickness is a tenth of the radius to which you wish to dissipate your heat. i.e. 2mm thick works out to 40mm diameter. 3mm thick out to 60mm diameter. 10mm thick out to 200mm diameter. after you go beyond a sensible radius then the temperature does indeed fall and the sink becomes less efficient. Large increases in radius give little increase in dissipation. Stainless steel will be a lot worse in this respect. I don't have data to hand, but it would be mighty thick, but a big advantage would be thermal inertia.

[Mick_F]

Hi Andrew,

I dont completely agree with you. First, a good thermal conductor will effectively carry the heat away from the chip and therefore will give rise to a thermal gradient. Second, whether or not a heatsink is held at the same temperature is strongly depending on the size of the heatsink, which for a given material determines its heat capacity, and by the power dissipated by the chip which is used to heat the heatsink. If only a small amount of heat is produced by the chip, the "hot radius" will be small and all you observe is the strong thermal gradient at its borders.

Actually my heatsink is grossly oversized. The chip is directly attached to an aluminium plate of about 2mm thickness and about 600 cm^2 surface area.

Mick