Ah Ok, I wasn't sure if it was fully or part shielded, mine (from some old gear I took to bits LOL) is metal to the backpanel. They are great though, any mains LC filter like that, makes a big difference.shielded
Ah Ok, I wasn't sure if it was fully or part shielded, mine (from some old gear I took to bits LOL) is metal to the backpanel. They are great though, any mains LC filter like that, makes a big difference.shielded
First, here are two Amplifier PCB captures; with and without the ground highlighted, the second for the silkscreen clarity. J5 is the PCB HQG connector and J2 is the speaker output connector. J1 is the input connector. J3 and J4 are +55 and -55 VDC respectively. In the casing, one PCB have it's -55 VDC in the bottom of the case and the other channel have it's -55 VDC near the top cover of the case.Do you have a picture of the internals of your amp from above with the wires labelled?
As Kleinhorn has noted: this earthing pattern is not ideal.Amplifier PCB
I am not an expert in pcb design and electronic, but in my opinion you have opend a door for groundloops on your pcb. The ground track in the middle of your amp pcb is not nessesary, to much crossings to other components. Left and right were enough, or only in the middle. Now the ground tracks are parallel to other ground tracks at some place.
Why is the servo so far away from the input ? The tracks are rather thin. But i dont know if this is the reason for the humming effect.
I have took one picture and marked the cabling on the right side of your housing. Perhaps is possible to take a way above the input net filter to the amp pcb.
No good idea to make the cabling in this, your, way...always keep distance to the net input. Keep away from the amp input or output. Seperate all cables to each other. Secondary wires, Output wires, input wires and so on. If not possible you can cross, no parallel wiring..i think you know...
Drilling is a good idea.
The left RCA wire is very near to the secondary supply wires, perhaps touching.
The second thing, easy to try, is to let one shield open at your RCA input.If your solution with the grounding to RCA from the star point grounding has made something better, follow the advice of Bonsai or mine and solder in a filter or resistor to cinch input and the shielding.
In a bad case RCA shielding becomes active, thats the reason why XLR is the better solution. Perhaps you have got different potential on the shielding. That may the reason for humming. One reason may be the equal grounding of the input and the powersupply. This problem does not occur in every case.
Perhaps you can make a measurement of current if you split of the RCA shield and switch in a multimeter.
The second thing, easy to try, is to let one shield open at your RCA input. You have already made a connection to your star ground, if i look at the last pictures. On the first picture this connection was not done.
You can set spacers on your floorplate to get distance from net wires and secondary voltage. Its not allowed to tie the primary net wires to the metal floorplate, to dangerous...
It is...Better with space through plastic spacers and cableties....Dont know about the canadian regulations...but thinking they are similar to the german regulations.
Because of your cabling colours i can't see if PE ist connected to your housing...for safety and EMI/EMV prevention PE should be connected.
Also a note about switching 50Vdc with a single contact - what is the rating of the relay? It may be necessary to use a double pole relay and put the contacts in series, so the switching voltage then drops to 25V per contact, which it's possible to actually switch off
Very few relays are rated to switch off 50Vdc as they turn into arc lamps: there's a cool thread on Diyaudio about making a nice MOSFET design to do this - as most solid state relays are more expensive than simply using MOSFETS, or a triac and fuse.
For this reason I'd also put fuses between the main PSU caps and the amplifier boards, 2-3AT, so if all goes horribly wrong it's another layer of protection for the woofer coils
Again, this is a totally new design... I am not ready to go in this path...Apologies for for so many comments , I have momentum... looking at the PSU schematic, my observations are:
The 2k bleeder resistors would be best replaced with current sources iMO, as they are much more efficient, and add a LED on each rail to show when active (for safety and convenience), flowing 1mA (bright enough).
My future design will surely include resistors in the snubbers. Although it is the first time I see resistors in series with the snubbers capacitors, but I admit it is my first Audio Power Amplifier Supply.ETA: Diode snubbers, best in series with 4.7R resistors, in my view.
Here in Canada we used 120 VAC.The speaker relays as noted. The 0.22R resistors I'd also eliminate, but I'd add a small, seperate diode bridge + caps for the driver part of the power amplifier boards, so voltage sags from the main transistors do not drag down the rails of the driver section.
I would also add a soft-start relay on the 220Vac line, with a 56R 10W resistor of about 6 seconds, for a nice gentle startup.
This is the blessing, and the curse of DIY: everything can always be alterednew PCB
In a fault condition - one MOSFET welding itself shut, there may be 50V across the speaker...The relays doesn't switch the 50 VDC! They switch in and out the Speakers.
Input and General wiring.
Looking at your input wiring, it seems you have the signal + signal return going to the amp modules via the black-screened cables, and then you have joined the signal grounds and connected them to the star ground with the white wire. This means you have very large input circuit loop areas and two grounding points so highly likely a common impedance coupling problem as well. The signal grounds must go back to the amp modules and nowhere else.
I would not route the input cables under the base plate. Run them around the top of the amplifier above the PCB and right next to the chassis. You want to keep the loop area between the signal ground and the chassis as small as possible (leaving aside the loop areas on the PCB for now).
I agree with other comments that your mains wiring is too close to the amp module that appears on the RHS of the image (the left channel I think). You should try to reconfigure that.
That's the part I hate, cutting the pcb lol... It means I cannot think of it as finished. All those recent comments start to make me think of making a totally new design... $$$$,$$If you have not fitted an HBR, then you should do it - see the picture below and note very carefully how it is done. It is not just a resistor in series with the signal return. If you do it like that, you will make your problems worse. To fit it, you will need to cut some tracks on your PCB. I recommend a value of 2.2 to 4.7 Ohms.
PCB Layout.
The kind of layout you have used I call a 'folded' layout (the other way is with all the output transistors in a line). The important thing is to keep the V+, output rail and 0V tracks for the positive side and then the V-, output rail and 0V tracks on the -ve side very close together. Lay the V+ and V- tracks over the output rail track and keep the 0V (which you use for decoupling) right next to these. This way, the radiating loop areas for the output and decoupling are kept small and the result is lower noise and lower distortion (The distortion arises because the high output currents generate large mag fields that couple into the small signal areas of the amp). In your layout, you show a central ground track going to the input circuit. This is the correct way to do it, with the following caveats. This signal ground track must run under the feedback and gain setting resistors - again, to keep the loop arwea as small as possible. The gain setting resistor is return to the signal ground trace through the HBR and the input signal; ground connects to the other side (ie not the signal ground trace side) of the HBR. Again, keep the input circuit loop areas as small as possible. Do not connect any other components to this signal ground trace. If you put your decoupling capacitor returns, you will cause excess noise.
One other point is where you take the speaker return to on your PCB - locate it on the other side of the connection back to the PSU Module. if you put it on the same side as the PCB 0V return, you will create distortion through common impedance coupling.
I attached another picture at the bottom to hopefully clarify these things. I think before you look outside your amp, you have to get it absolutely silent on both channels.
YesSo would I hear any more difference making all these new design changes?!
I think this effect is some universal property of this universe / school / matrixi make an upload with the gerber files i am always too fast. By a second or third look i always find things to make something better.
The problem here is the space available onto the heat-sink surface. The PCB is much longer than it is high. So passing traces from left to right (front to rear of the case) is very limited by this height with all the components and others traces occupying the surface. Not to mention the mounting holes of the transistors and the PCB itself onto the heat-sink. Then one PCB rule design is to always stop and start from a part donuts, the least possible from an existing trace.As for symmetry, I don't see why that should not be welcomed in a PCB, the +ve and -ve sides are in general, symmetrical in many respects
Could just mean you have other problems in your system, which are masking the differences between amplifiers. Don't give up, is all.I don't hear any difference in the sound between my DIY and my Rotel Amplifiers,...