I hope that you left the rectifier and cap wires twisted.
Next, I would take care of the wiring from IEC to transformer. Does a wire go from IEC to on/off switch? If so, that wire must follow the other wire from the IEC, to the transformer, and then follow the other transformer wire to the switch.
If the transformer primary wires that go to the IEC and the switch are no longer long-enough to twist, then disconnect one end of the IEC-to-switch wire and wrap it around them, from the IEC to the transformer and then from the transformer to the switch. If the wire isn't long-enough to wrap around and still go the distance, replace it.
I'm not sure that's how your switch wiring is set up. So it may go differently. But one way or another, the AC mains and transformer primary and any switch or fuse wiring must not have any separation between the AC input pair, anywhere.
I would take care of that, next, just to eliminate it from the equation, since it is definitely a problem.
And it worked for this guy:
http://www.diyaudio.com/forums/tubelab/243870-sse-question-about-hum-3.html#post3664612
Cheers,
Tom
Last edited:
The shielding is just icing on the cake. The main thing is keeping the signal and its ground against each other, always, so that there is no enclosed loop area to act as an antenna for the time-varying magnetic fields from the mains and rectifier loops.
Last edited:
They were never twisted in the first place, none of the wires in the amp ever were. But then none of the wires in My Krell KSA 200 were ever twisted either and that's utterly silent.
An externally hosted image should be here but it was not working when we last tested it.
The black and white wires you can see coming from the back panel are the transformer primary wires, they run up to the mounting block, via the front panel, and as you can see they are untwisted. And that white wire that disappears stage right off the lower edge of the image attaches to the front panel (pre/attenuator control board pcb) and that contact is a good 12cm away from where the black wires joins the same pcb.
The on/off switch is part of the front panel pcb, it's a tiny microswitch.
An externally hosted image should be here but it was not working when we last tested it.
Listening to the noise out of the speakers.
I can twist and untwist all I like it makes bugger all difference to the level and type of noise. But moving the signal input wiring and the pre-to-right channel signal wiring makes a huge difference.
I get buzz not hum, it's totally broadband noise.
In the quote below, you say you tried them twisted. Above, you say none of them ever were (twisted).
If you twisted them, in order to test them twisted, you should have left them twisted, because that is how they should have been in the first place. (I don't care what mistakes Krell made.)
If it doesn't sound too daunting, you really should make sure that none of the AC or rectified AC wiring has any significant enclosed loop area. It doesn't look like it would be TOO bad of a job. I would start with the AC Mains IEC input, transformer primary, and power switch wiring. if nothing else, it will eliminate one of the possible causes of the problems.
I can't quite see the AC wiring, well-enough, in your photos. If you copy the photos and paste them into MS Paint or some other graphics program, and somehow LABEL the wires and the connection points and the transformer, switch location, any fuses, etc (i.e. everything that the AC Mains and primary wiring goes to or from), then we can probably help you figure out what to twist with what, etc.
---
Full stop!
---
NOW it looks like you're saying you DON'T have hum, and that it's all broadband noise. Does (be careful!) moving the AC wiring have ANY effect on the noise, or whatever it is?
---
Have you tried Caig's De-Oxit on all of the IDC cable pins and sockets? Or, just replace the IDC assemblies entirely. IDC cables can be unreliable. Seriously, maybe you should do scope checks of the signals at each end of each IDC cable, by scoping something on each board that connects to the pins you're testing. Compare the scope view of the same signal/pin, on the boards, at both ends of the cable. Maybe move the cable while doing so. Start with any signal or signal ground pins.
If you find IDC problems that De-Oxit doesn't fix, you could consider making new cables. You don't have to use IDC-type flat cable. You could use some nice shielded multi-pair cable. It's easy to crimp the little pins or sleeves onto the wires and insert them into a connector housing. The trick is usually finding the right ones to use. Or, just buy new cables, or have them made.
---
How OLD is this amplifier? If you were going to replace parts, the first ones to suspect would be all of the electrolytic caps, NOT just those in the power supply.
---
Also, you have two supposedly-identical channels. You could easily take voltage and resistance measurements of corresponding points in each channel, relative to some common ground point, to make sure that the channels are, in fact, identical, or nearly identical. Any significant difference might point to a component or connector problem. (Resistances would be measured with power disconnected and caps drained.) You would probably need to take a photo of each board and assembly and number each measurement point, and then write down the measurements by number.
If you have not twisted all of the AC and rectified AC wiring, to test it, then you cannot say that it doesn't affect the noise.
If you are going to simply "decide" what the test results will be, instead of doing the testing, then there will be little hope of learning the source of the problem.
If moving the signal wiring affects the noise, then the source of the problem must be one of the following: a) the fields emitted by the untwisted AC wiring, or, b) some other fields, or, c) bad connections or broken wires in the signal wiring.
I would check "c", quickly, since it should be easy to do, and then eliminate "a".
I've already reflowed all the IDC connector joints on the pcb's and replaced all the IDC cables with new ones. The amp is only five years old, runs in class Ab with low bias and minimal heat output, no components are anywhere near their rated life.
There's 10cm between the +/- screw fittings on each cap bank and there's no way to minimize this. There's also 15cm of open loop area between the two Ac input wires as they connect onto the back of the front panel.
There's no real loop area at the IEC, all the wires exit next to each other and the safety earth is connected to one of the screws for the IEC socket so has tiny area, maybe 2cm. I can twist these wires.
The long runs you can see from the 0v on the PSU caps back to the 0v transformer point on the terminal block have been trimmed to provide the shortest straight line from A-B, the photo pre-dates that change. This eliminated a good 20cm of 0v cap wiring.
I will twist the red/rectifier wires as best I can, they are solid core and diverge their paths quickly, so I'll be lucky to get two twist in each side.
I'll make these changes and take another pic later today.
There's 10cm between the +/- screw fittings on each cap bank and there's no way to minimize this. There's also 15cm of open loop area between the two Ac input wires as they connect onto the back of the front panel.
There's no real loop area at the IEC, all the wires exit next to each other and the safety earth is connected to one of the screws for the IEC socket so has tiny area, maybe 2cm. I can twist these wires.
The long runs you can see from the 0v on the PSU caps back to the 0v transformer point on the terminal block have been trimmed to provide the shortest straight line from A-B, the photo pre-dates that change. This eliminated a good 20cm of 0v cap wiring.
I will twist the red/rectifier wires as best I can, they are solid core and diverge their paths quickly, so I'll be lucky to get two twist in each side.
An externally hosted image should be here but it was not working when we last tested it.
I'll make these changes and take another pic later today.
To clarify point C.
"Moving the input signal and the right channel wires changes the noise."
The act of moving them, giving them a wiggle for example, does nothing. Placing them in different locations within the amp does. For example putting the right channel signal wiring anywhere near the two front panel mounted transformers induces good old 50hz 'hum'. (in addition to the brilliant buzz I get all the time).
I'll give all the IDC connectors a Squirt with Deoxit, but as I said all these are new anyway. I previously had a bad connector and the right channel was 3db down vs the left, it was a broken pin on the signal - of the balanced input, so I replaced them all. (The noise pre-dates this)
"Moving the input signal and the right channel wires changes the noise."
The act of moving them, giving them a wiggle for example, does nothing. Placing them in different locations within the amp does. For example putting the right channel signal wiring anywhere near the two front panel mounted transformers induces good old 50hz 'hum'. (in addition to the brilliant buzz I get all the time).
I'll give all the IDC connectors a Squirt with Deoxit, but as I said all these are new anyway. I previously had a bad connector and the right channel was 3db down vs the left, it was a broken pin on the signal - of the balanced input, so I replaced them all. (The noise pre-dates this)
Last edited:
measure it.
I gave your the different arrangements to measure in my earlier post.
Your ear cannot discriminate small changes.
So, the upshot of measuring last night, was zero.
There's no difference in the AC or DC at the speaker outs with the rca input, totally open or shorted to signal ground. AC is 15mv and DC 30mv on both channels.
With nothing connected to the RCA-in there's no hum, with my dac connected the speakers hum, with my dac connected and the psu plugged into my macbook pro, which is usb to the dac I get lots more noise.
I have the loan of a crimper tonight so i'll make up some wires to replace the incoming AC loom such that i can twist them all the way to the front panel switch.
There's no difference in the AC or DC at the speaker outs with the rca input, totally open or shorted to signal ground. AC is 15mv and DC 30mv on both channels.
With nothing connected to the RCA-in there's no hum, with my dac connected the speakers hum, with my dac connected and the psu plugged into my macbook pro, which is usb to the dac I get lots more noise.
I have the loan of a crimper tonight so i'll make up some wires to replace the incoming AC loom such that i can twist them all the way to the front panel switch.
Bingo. Getting hum only when something is connected to the inputs points to the enclosed loop area of the input signal and ground picking up the fields radiated by the mains or other wiring: connecting something closes the loop. The induced currents can then flow, and they induce voltages across all of the impedances, such as the input resistors.
Quiet with nothing on the inputs is a good sign.
Quiet with nothing on the inputs is a good sign.
Well that's interesting, doubling part of the black wire that runs to the switch from the IEC, so that it can loop back to follow the white wire for its entire length and thus close that 12cm loop actually makes things worse.
I'm going to try removing the loop and placing both wires such that they don't run past those two small block transformers on the front panel. I'm also going to replace all the wires from diodes to caps so that i can twist them more than is currently possible. I can rotate the caps in their mounts to bring the +/- closer in vertical alignment.
I'm going to try removing the loop and placing both wires such that they don't run past those two small block transformers on the front panel. I'm also going to replace all the wires from diodes to caps so that i can twist them more than is currently possible. I can rotate the caps in their mounts to bring the +/- closer in vertical alignment.
Sorry guys I ahve been a bit busy lately to work on the amp but it looks like there is some really good information here.
You can see my wiring on the pictures in post #9.
My AC mains go to the switch together and than from the switch to the trafo and they are packed together with zipties.
My secondaries are also ziptied together on the side of the trafo and go to the rectifier where they split. Than I have two black wires which are the rectified voltage that go to the BIG CAPS. They are by themselves.
It made no difference to the channel I was measuring. I disconnected one of the channels (the one that has the louder buzzing) and the untouched channel stil had the same white/broadband noise and slight humming. I do have to get about 5-6 inches away to hear the noise/hum.
I on the other hand get MORE humm and buzzing when something is connected :/ and one channel is worse than the other.
I am not sure what you guys mean by connectin a snubber to my secondary? I have only seen snuber circuits on in series ground connections.
Are you describing to wire them in series with my secondary to the rectifier?
Here are my options I guess.
1) I ahve a faulty component in one of the boards, hence the louder hum/buzzing in one of the channels. So I should test the MOSFETs?
2) Get my wiring checked out to the traffo?
3) Properly mount the trafo so it doesn't vibrate?
4) Rewire my RCA inputs. Use a twisted pear shielded cable to PCB, star ground to the input resistor on the PCB, star ground to the RCA shell?
5) Snubber circuit? I don't really wan't to use it and I probably wont have the parts.
6) Check my AC cable connection. Looks a little bunched up at the input. It might have kinked over the last.... 23 years! Also this amp does not have an IEC connector and it straight AC wiring :/
I probably won't get to do this untill next week.
Not in practice hehe. The wirging to the switch is the way you describe it but the wires are not twisted a lot but are packed together with zip wires.
You can see my wiring on the pictures in post #9.
My AC mains go to the switch together and than from the switch to the trafo and they are packed together with zipties.
My secondaries are also ziptied together on the side of the trafo and go to the rectifier where they split. Than I have two black wires which are the rectified voltage that go to the BIG CAPS. They are by themselves.
It made no difference to the channel I was measuring. I disconnected one of the channels (the one that has the louder buzzing) and the untouched channel stil had the same white/broadband noise and slight humming. I do have to get about 5-6 inches away to hear the noise/hum.
I on the other hand get MORE humm and buzzing when something is connected :/ and one channel is worse than the other.
I am not sure what you guys mean by connectin a snubber to my secondary? I have only seen snuber circuits on in series ground connections.
Are you describing to wire them in series with my secondary to the rectifier?
Here are my options I guess.
1) I ahve a faulty component in one of the boards, hence the louder hum/buzzing in one of the channels. So I should test the MOSFETs?
2) Get my wiring checked out to the traffo?
3) Properly mount the trafo so it doesn't vibrate?
4) Rewire my RCA inputs. Use a twisted pear shielded cable to PCB, star ground to the input resistor on the PCB, star ground to the RCA shell?
5) Snubber circuit? I don't really wan't to use it and I probably wont have the parts.
6) Check my AC cable connection. Looks a little bunched up at the input. It might have kinked over the last.... 23 years! Also this amp does not have an IEC connector and it straight AC wiring :/
I probably won't get to do this untill next week.
Good thread.
Off-topic comment here...
I previously had the (debatable) opinion that most power amps had little impact on sound quality. But I did evaluate all my amp purchases.
First purchase was the B&K ST-140.
Years later, I upgraded to a pair of Muse 175 monoblocks.
Then to my surprise, I realized both amps used the same output transistors.
Funny.
...or just coincidence ?
I still use all of them (no hum), and I'm contemplating replacing the (now) 20yo PSU caps in the Muse. From my best guess...replacement can caps would be.......
CGS103U100W4C
$58 x 8 caps = about $500
I think I'll buy one first, to make sure they will fit.
Off-topic comment here...
I previously had the (debatable) opinion that most power amps had little impact on sound quality. But I did evaluate all my amp purchases.
First purchase was the B&K ST-140.
Years later, I upgraded to a pair of Muse 175 monoblocks.
Then to my surprise, I realized both amps used the same output transistors.
Funny.
...or just coincidence ?
I still use all of them (no hum), and I'm contemplating replacing the (now) 20yo PSU caps in the Muse. From my best guess...replacement can caps would be.......
CGS103U100W4C
$58 x 8 caps = about $500
I think I'll buy one first, to make sure they will fit.
Attachments
![20160211_151747[1].jpg](/community/data/attachments/493/493220-27e3400e9e347c1884d3515ab97849e5.jpg)
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.