http://sound.westhost.com/project3a.htm
I built this amp about a year ago, i am very satisfied with it, but under certain conditions it generates a very loud distorted 50Hz tone.
It must be about a year ago i built it now, never really used it, found this forum a couple of months ago, but not until now i have had the time and will to do a few tests.
It consists of two mono amps using the same PSU (24-0-24 trafo with two 4700µF caps (one for + and one for -)).
When i ground both inputs (or have any resistance at all to ground) and short both input grounds (which is the case since i use a 2xCINCH to 3.5mm stereo plug cable) i get the following noise:
L: 50Hz. ~20mV positive spikes, about 1-2ms long.
R: 50Hz, ~20mV negative spikes, about 1-2ms long.
The noise is very audible, and is even higher when the bias current is correctly adjusted, the pots are at highest resistance now.
If i remove the short between the input grounds, and just connect the inputs to their respective ground, the following noise is present at the speaker terminals:
L: none at all
R: 50Hz, ~5mV positive spikes, not audible
Now, if i just short the input grounds and leave the inputs floating, both channels are clean.
This is done with a 8ohm speaker load.
I have measured all resistors, and they are OK.
Oh, i used the components in the schematics, except for the power transistors, which i replaced with motorola MJL3281 and MJL1302 instead, since the toshibas i purchased turned out to be counterfeits. (Don't buy 2SA1302/2SC3281 from www.jec.se)
I built this amp about a year ago, i am very satisfied with it, but under certain conditions it generates a very loud distorted 50Hz tone.
It must be about a year ago i built it now, never really used it, found this forum a couple of months ago, but not until now i have had the time and will to do a few tests.
It consists of two mono amps using the same PSU (24-0-24 trafo with two 4700µF caps (one for + and one for -)).
When i ground both inputs (or have any resistance at all to ground) and short both input grounds (which is the case since i use a 2xCINCH to 3.5mm stereo plug cable) i get the following noise:
L: 50Hz. ~20mV positive spikes, about 1-2ms long.
R: 50Hz, ~20mV negative spikes, about 1-2ms long.
The noise is very audible, and is even higher when the bias current is correctly adjusted, the pots are at highest resistance now.
If i remove the short between the input grounds, and just connect the inputs to their respective ground, the following noise is present at the speaker terminals:
L: none at all
R: 50Hz, ~5mV positive spikes, not audible
Now, if i just short the input grounds and leave the inputs floating, both channels are clean.
This is done with a 8ohm speaker load.
I have measured all resistors, and they are OK.
Oh, i used the components in the schematics, except for the power transistors, which i replaced with motorola MJL3281 and MJL1302 instead, since the toshibas i purchased turned out to be counterfeits. (Don't buy 2SA1302/2SC3281 from www.jec.se)
As long as i use the amps totally individually, it works great.
In that case, it IS a ground loop.
I had the same thing on my Leach Amp. I solved it by disconnecting the ground connection to the board from one of the input connectors and connected it to the other board's input ground, or something.
You have to isolate where the ground loop is originated and terminate it somehow.
Do you have pictures? It always helps to see what it looks like.
//magnus
Hm, weird thing is that the noise is still there even if i only connect one input ground thus breaking the ground loop (its just on one channel then though). I tried 10 ohm resistors in paralell, didn't help, havn't tried caps though.
Im fixing pictures as soon as i can figure out where the hell i put my digicam
Oh, and the PS is designed from Elliots guidelines, 300VA toroid with 2x4700µF caps for both amps.
Pics up: http://daviruz.homelinux.org/pics/amp/
This is just a test-setup chassis in case you wondered
Im fixing pictures as soon as i can figure out where the hell i put my digicam
Oh, and the PS is designed from Elliots guidelines, 300VA toroid with 2x4700µF caps for both amps.
Pics up: http://daviruz.homelinux.org/pics/amp/
This is just a test-setup chassis in case you wondered
Coarse ground..
The grounding scheme looks a little weird. It seems you run the speaker grounds individually back to the PCB and then to the cap center. This, AFAIK, means, among other things, that your input ground is shared with the speaker return ground and all of it would be causing a big ground loop.
I would probably do/try the following:
1. Put speaker return at the new star ground point which seems to fit at the cap center connections.
2. Return the PCB ground to the star point.
3. Connect chassis ground to power suppy ground via 10 Ohm resistor.
4. Return input grounds individually to the star point if needed.
Unless somebody tells me I am completely off then I think that would help.
/UrSv
The grounding scheme looks a little weird. It seems you run the speaker grounds individually back to the PCB and then to the cap center. This, AFAIK, means, among other things, that your input ground is shared with the speaker return ground and all of it would be causing a big ground loop.
I would probably do/try the following:
1. Put speaker return at the new star ground point which seems to fit at the cap center connections.
2. Return the PCB ground to the star point.
3. Connect chassis ground to power suppy ground via 10 Ohm resistor.
4. Return input grounds individually to the star point if needed.
Unless somebody tells me I am completely off then I think that would help.
/UrSv
Then there is definitely something fishy going on. Hard to say when I can't see up close.
Check everything. Component orientation, component values, solder bridges, solder joints etc...
What I can see that could be suggestions:
1.
Are the screws holding the PCBs really not touching the track on the bottom side? Lower fixing points look close. Measure and verify?
2.
Heatsink properly isolated from all transistors? Measure and verify?
3.
I have had occasional problems when not isolating the bridges from chassis. Try lifting the bridge from the chassis in case it is attached to metal. Don't ask me why.
4.
Oscillations? Check with a scope during the test.
Other than that I am out of ideas.
Good Luck/UrSv
Check everything. Component orientation, component values, solder bridges, solder joints etc...
What I can see that could be suggestions:
1.
Are the screws holding the PCBs really not touching the track on the bottom side? Lower fixing points look close. Measure and verify?
2.
Heatsink properly isolated from all transistors? Measure and verify?
3.
I have had occasional problems when not isolating the bridges from chassis. Try lifting the bridge from the chassis in case it is attached to metal. Don't ask me why.
4.
Oscillations? Check with a scope during the test.
Other than that I am out of ideas.
Good Luck/UrSv
Thanks for all the help so far.
The chassis is plastic, the screws are not touching the tracks, but even if they were it wouldn't be an issue.
The trannies are isolated from the heat sink, i just measured them (did this before first start up too ofcourse, i think things would have gotten a bit hot else.
I'm thinking of replacing all bc546's and all the caps on both boards, will borrow some desoldering stuff from school tomorrow i think. I have checked component orientation etc a million times.
The chassis is plastic, the screws are not touching the tracks, but even if they were it wouldn't be an issue.
The trannies are isolated from the heat sink, i just measured them (did this before first start up too ofcourse, i think things would have gotten a bit hot else
.I'm thinking of replacing all bc546's and all the caps on both boards, will borrow some desoldering stuff from school tomorrow i think. I have checked component orientation etc a million times.
No no, of course not! I wasn't commenting on the ground loop problem, just the cap size. The two don't have anything to do with each other.
paulb: 4700 may be OK for operation of a class-B or AB amp at low signal levels, but don't expect to get great performance out of it when you crank it up. Just remember your ripple will increase as you draw more current... and the P3A as I recall, is good for 60-100W/8R. If you start pushing that kind of voltage into a speaker load with an impedance dip, you'll find 4700 is not sufficient, especially if you want clean bass. At that kind of power output, your peak current draw from the PSU will rival that of a push-pull class-A unit, and furthermore will contain a lot more noise due to the class-B circuit effectively "rectifying" the current waveforms going back to the PSU. Bigger is better, even though you might get away with only 4700uF here.
hifiZen said:
I know. I was going to go for a second pair of caps and another bridge (thus making indivudual power supplies except for the toroid). However, when i started building it i could barely afford a single psu, and since i never got it working i never bought a second pair of caps.
Speaking of which, could splitting the psu acrually solve the problem? Would be an ugly fix though, something is appearantly wrong somewhere.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.