I posted in the ss section but since I narrowed the problem down to the ps, thought I'd post here. The issue is the ps in my 3rd old pro amp is cutting in and out to the output stages intermittently. When I tap the supply caps I can replicate the behavior. I reflowed all the connections and the caps but it still behaves the same.
Intermitent faults can be misleading. Often it's a transistor or reg or even a resistor that runs very hot and the solder cracks on the lead outs. It's impossible to diagnose from the info available and if it is a dry joint somewhere the best solution can often be to resolder the whole PCB completely.
Any wiring harnesses/plug/socket interfaces ? that could be a problem.
Any wiring harnesses/plug/socket interfaces ? that could be a problem.
Loads.
It's only the output boards that are affected. The rest of the amp, input boards, meters, protection circuits, fans, work fine. This leads me to believe that it's the connections to the output boards or the main caps as the supply to the rest of the circuits has their own caps. The connectors on the output boards themselves look fine. I managed to get my old multimeter working to the point where I can measure voltages (yes, time for a new one) and the problem appears to be coming from the supply.
The only mechanical support for the wires to the output boards on the ps is solder. There is no glue or strain relief of any kind. This is the same for all three rails, +ve, -ve and ground. The joints look ok but perhaps they have stress cracks?
I'll also add that it is intermittent from cold, usually taking a good boost of volume to get it going.
Last edited:
I'm not familiar with the amp. Is it one or two channel. Sometimes speaker protection relays can be troublesome (if it has them)... just with you saying the volume needs to be turned up to get it going.
If it's a 2 channel amp and both go off then probably PSU related , or, again, relay drive if it's common and shared.
If it's a 2 channel amp and both go off then probably PSU related , or, again, relay drive if it's common and shared.
It's two channel, bridgeable. It's not the relay because that has a particularly mechanical sound when connecting and that doesn't happen. Also, the protection indicator LED does not light. Both channels fail but the left more so than the right. At some points the output sounds crackly, very much like a loose connection.
Thanks for your help, btw.
Thanks for your help, btw.
That smells a lot like a microscopic crack in the pcb tracks, my first bet would be just at the border of a solder pad. I suggest to remove the solder resist close to the pads in question and to extend the soldere to the first few millimeters of the pcb track. This way you would bridge possible cracks.
Excellent idea. I'll look into extending the solder around these joints. Might provide a little more mechanical support too.
I noticed a mistake in my op. It's a 3 year old amp, not 3rd old amp. Damn iPad!
Last edited:
If the contacts are oxidised it can give the fault you describe, and turning the volume up breaks down the insulating layer and it works.
Don't rule it out just because it gives a positive click 🙂 Give the relay a tap.
I'll give it a go when I have reassembled the amp today. I would be surprised if the contacts have oxidized after only three years but I suppose if there is a manufacturing fault in the relay, it might be possible.
Cracks in the solder can best be viewed by using eye loupe, the watchmaker's lens.
If you inspect the areas around hot components(mostly near heatsinks), you will notice a circle formed on the solder, which is a crack. Look for any such circles and solder them and other nearby ones.
gajanan phadte
If you inspect the areas around hot components(mostly near heatsinks), you will notice a circle formed on the solder, which is a crack. Look for any such circles and solder them and other nearby ones.
gajanan phadte
It seems to be behaving itself now. I'll have to bring it back in from the cold to see how it reacts to a temp change. It was probably a trace crack where the ground wire connects to the ps since that looked the most suspect and the output is not grounded otherwise. Is it normal not to ground aluminum heatsinks? The transistor packages themselves have plastic insulators between them and the heatsink.
Some are grounded, some are floating, some are at signal (speaker output) potential and some are at one of the rail voltages.
Most transistors have no internal insulation to the metal part of their case. As a result in a typical amplifier, they carry different voltages. This makes the isolation necessary.
For analog amplifiers, leaving the heat sink floating has no negative side effects. For digital (class D) amplifiers, it is mandatory to make a connection to the power ground to prevent the fast switching transients to cause EMI and signal integrity problems.
Since the heatsink is attached to the chassis via metal screws, it's not completely isolated but it surprised me not to see an actual ground connection from the output board to it's heat sink. The supply provides the ground line to the output boards and it also has a line going right next to this to the star point on the chassis so maybe there just isn't a need for one. Oh well.
Edit: I just though of the reason. Heat sinks get hot. Don't want that going to you PCB via a connection! Makes sense.
Back to topic. Seems to be ok for now but I will run it in for a while to make sure.
Edit: I just though of the reason. Heat sinks get hot. Don't want that going to you PCB via a connection! Makes sense.
Back to topic. Seems to be ok for now but I will run it in for a while to make sure.
Last edited:
When there is a connection to chassis at the supply then making another one at the amp would create a loop, which is often the cause for audible mains hum. For that reason making multiple ground connections is generally avoided.
Maybe a little unconventional but that once worked fine for me: get some thick litz cable (2.5mm^2 at least) and use it to make a strong low impedance path between the chassis of all signal sources and that of the amplifier, preferable a star starting at the amplifier. Can be done at the back, so nobody sees it 🙂 Luckily today I have an amp with differential inputs.
The amp has balanced inputs but the source does not have balanced outputs. 🙁 I'll make a nice thick ground cable. Thanks.
When the amp has balanced inputs you can try the following: connect + of source to + of amp, connect gnd of source to - of amp, connect a 10~47 ohm resistor from - of amp to gnd of amp. This way you create a loose gnd connection between the two, which the input buffer of the amp needs for biasing (if it is not ac coupled). You can still stiffen that with a thick extra wire, but this way the hum current does not flow through the signal cable's gnd wire. Another option could be using a differential XLR cable and converting it to cinch directly at the source's output jack.
- Status
- Not open for further replies.