[Cat Month] Amp Cat ! (17/31) by Tiido

[Cat Month] Amp Cat ! (17/31)


18 July 2015 at 19:46:16 MDT

Neighbors complaining ? The Amp Cat will meow so loud you cannot hear them !

I'm cheating here to catch up with the real calendar. I end up having like 28 days in this month rather than 31 and that explains my falling behind here. It isn't like I'm skipping days, I just haven't had as many days !

This is one of the two hybrid amp modules I'm building that go into my brother's amp. There's a bit of work left to do like connect the output stages to the opamp and connect feedback stuff and some ground wires... This schit has so low noise floor that you cannot hear it, only see on scope, and there's no audible distortions even at multi-hundred watt output ! It'll be really beefy once done !

That damn marker I was using is running dry, and it was so super uncomfy to doodle this on that heatsink lol.

The next pic is gonna make up for this, I promise !

Submission Information

Visual / Photography


  • Link

    Cool. I don't see a ground wire, is the heatsink grounded? I know that wiring would make some EMC engineers shudder. If all the high-current wires were just snug against the heatsink then magnetic radiation/reception might be reduced down to a low frequency because it would be a very thick ground plane (especially with enameled wire). Of course, if the distortion is inaudible, then this is just a mental hamster wheel, albeit one I enjoy.

    • Link

      I can make wires cross to reduce the area and I can always put shield plates everywhere, but even as is the physical distance is more than enough to give good characteristics, plus less material hahaha. If I could have chosen transistors location and orientation on the heatsink I would have done things much differently.
      The heatsink is grounded but it isn't used as a ground. I have several options to use as far as ground goes. I can make a star-point on the heatsink, and connect my wires there, or I can make a star-point elsewhere and have multiple ground wires going there. My other amp design has the latter going on but I realized I'm not taking any advantage of the 3 galvanically separated outputs of my transformer that way, so I will experiment with local starpoint in this one to see if it is better or not, and if it turns out better I'll make my other amp match this.


      I enjoy those mental hamster wheels too :P

      • Link

        I'm not sure why connect the trafo shield to load star ground, since you're just putting the noise right back in. I think it would be better to connect it to chassis, ground the source RCA shields to chassis, and use that as star ground. I think this would short out all paths for common mode noise through the circuitry. It still flows through the input cable shields though, and few attempt to solve that problem.

        • Link

          I get a nice ground loop when I have input GND to the chassis, bad buzzing in the sound from that. I have nothing connecting to chassis besides speaker protection logic as it needs no stable 0V reference. Star point remains a solid 0V and the power rail fluctuations relation to it won't be a problem in any case, PSRR and CMRR are both very high in the used parts.
          The transformer actually doesn't even have shield, and if it did it would connect to the chassis anyway, just on the schematic in ended up where it was. The literature I have used as reference show the shield going to star point also :


          My own experiments seem to show it is also way to go.

          • Link

            If you get buzzing if you don't isolate from socket ground, then it could be your source is a problem and is trying to push current through the shield, or there is a ground noise source somewhere else. The best thing to do is to separate all grounds, measure AC voltage between them to find ground noise sources, then connect each of them through an ammeter to see if there are large ground currents. Once you know the actual source of the noise you know exactly what to do. I think with a good ground scheme shields to chassis should be silent. It is easy to have accidental chassis and ground connections.

            • Link

              The big problem is the input filtering caps of switching power supplies of the modern devices. I wish laptops and others AC bricks were made to be openable, I would love to yank out two caps on the AC input that sit between 240V and system GND... that's where all the noise normally comes from. (Also the joy of ungrounded outlets, yay old soviet wiring).