Alpha Nirvana 39w 8ohm Class A Amp

Andy, why do you need to run that many earth wires? especially those from the heatsinks and CPU coolers. Aren't they already bonded to the case? Those wires seem redundant and unnecessary unless your heatsinks and CPU coolers are isolated from the case.

I tried my AN without and with wiring the heatsinks to ground. I found no audible difference and removed them.
 
Member
Joined 2014
Paid Member
Andy, why do you need to run that many earth wires? especially those from the heatsinks and CPU coolers. Aren't they already bonded to the case? Those wires seem redundant and unnecessary unless your heatsinks and CPU coolers are isolated from the case.

Correct EmeryBB.

Andy,
Please check for continuity between the heatsinks/CPU coolers and chassis earth ground just to make sure. Then you can remove those wires and clean up some clutter. You should only have 2 SLB and 2 AN chassis ground wires plus your Mains earth ground tied to a single bolt. (Trafo shield wires if equipped also)
 
Andy, why do you need to run that many earth wires? especially those from the heatsinks and CPU coolers. Aren't they already bonded to the case? Those wires seem redundant and unnecessary unless your heatsinks and CPU coolers are isolated from the case.

I simply followed the wiring diagram which X had produced.

I tried my AN without and with wiring the heatsinks to ground. I found no audible difference and removed them.

Correct EmeryBB.

Andy,
Please check for continuity between the heatsinks/CPU coolers and chassis earth ground just to make sure. Then you can remove those wires and clean up some clutter. You should only have 2 SLB and 2 AN chassis ground wires plus your Mains earth ground tied to a single bolt. (Trafo shield wires if equipped also)

That's good to know, guys - thanks. :)

And makes a lotta sense.

Andy
 
Member
Joined 2014
Paid Member
In certain situations if the heatsinks the MOSFETs are mounted to are not grounded they can oscillate. When X was prototyping the AN build he was using an open plywood base framework therefore he had to run dedicated wires to the heatsinks to make sure everything was grounded and eliminate any chance of oscillation.
 
In certain situations if the heatsinks the MOSFETs are mounted to are not grounded they can oscillate. When X was prototyping the AN build he was using an open plywood base framework therefore he had to run dedicated wires to the heatsinks to make sure everything was grounded and eliminate any chance of oscillation.

Aah, OK.

After having unscrewed the earth bolt and freed all the ground ties, I measured the resistance between:
* the 2 CPU cooler ground wires - 1 ohm, and
* the 2 SLB heatsink ground wires - 'OL'.

So the CPU coolers certainly are earthed - so I've removed these ground wires. However, the SLB heatsinks would appear not to be earthed - so I'll keep their ground wires.

Andy
 
Founder of XSA-Labs
Joined 2012
Paid Member
Some chassis parts are hard anodized aluminum. Unless there is a specific sanded/ground bare metal surfaces touching each other - any ground connection you get is inadvertent and unreliable. With grounding strategies for hum and noise elimination, every ground connection has to be deliberate. When the ground connection is accidental or happenstance, there is a chance that a ground loop can be set up if it is not truly a low impedance path like a dedicated bonding wire. If you remove it and there is no hum, that’s probably ok. An output MOSFET heatsink that is not grounded can actually cause oscillation, in addition to causing hum or ground loops.
 
Finally, I am listening to the first of my AN 4Rs! It's powering my 3.2 ohm mid panels.

Dead quiet – thank goodness (through the speakers)! So the fact that the 240v mains feeds to each traffo:
• pass along (and near) the -ve side of the L channel SLB and the +ve side of the R channel SLB,
• and crosses some of the earth cables at the chassis bolt

… has not caused a problem! :) So now I can solder each mains active wire onto its NTC lead (I had croc-slipped them, in case I had to put these mains wires into an earthed shield).

And I don’t think I can hear fan noise at the listening position (which is all that matters)! :)

It’s running a bit hotter than I had measured before (when I had no spkr load):
• R & L P-channel MOSFETs (the edge) are 40 & 41 deg – whereas before, I measured 30 & 33 deg
• the back of each CPU cooler bracket is also 40 & 41 deg – whereas before, I measured 33 & 37 deg.

Traffos are ever-so-slightly warm.

Turning up the volume louder than I normal listen at … no sign of the red clip LED! (Even though this amp is ~half the power of the de-rated N80s that I’ve been using.) And turning up the volume … does cause heatsink temps to drop a bit! :D

Listening to an LP - Joe Cocker "Sheffield Steel" (MFSB). Very authoritative and detailed mids. :) But I'll turn it off in a while and, tomorrow, I’ll swap the AN to my (2 ohm) ribbons. That will be the "torture test"! :D

How do I get the clip LED to light up? Do I have to use a sig gen to feed it a mid-range signal (say 1500Hz) and put a dummy load on the amp, so I don't ruin my mid drivers when the clip LED lights?

A very happy Andy! :)
 
Last edited:
Founder of XSA-Labs
Joined 2012
Paid Member
Congratulations on first sound, Andy! :cheers:

My clip light flickers sometimes when I play loud, and it’s not bass kick drums that makes it light (as I guessed), but usually it’s either female vocals or piano fortissimo crescendos. My speakers are very inefficient though at 82.5dB so it can happen. But the clip is very graceful and soft as I cannot hear it - only see the light once in a while. Without the light, I would have never known it was clipping.