AMB Laboratories DIY Audio

[Tilpo]

Sorry for bothering you again. I sincerely hope this will be the last time.

First of all on the left channel the initial setup is fine. However when I read the voltage across R9 it sometimes jumps to 10~11V and then back to 4.5V again after a few seconds.
When I listened to the left channel I heard a lot of humming and no music. At one point I did get music to play but it was impossible to adjust the volume, and there was a lot of distortion. At times I also get high pitched screeching sounds (~4kHz).
Additionally the sound (whether humming, or music) is always in the right channel of the headphones, regardless of whether the input is right or left channel.

With the left channel R9 is fine, but R12 and R10 read 0mV. R11 is fine as well.
For this channel there is a lot of humming also, and the sound is again on the right channel of the headphones always.



I personally suspect both an issue with ground somewhere, and there is obviously something wrong with both channels as well. Can I please be pointed in the right direction?


Thanks.

[amb]

Check the voltages across D11 and D12. Are either of these more than a couple of volts? If so, the corresponding Q25 or Q26 is most likely blown.

[Tilpo]

Right channel (the one where R10 + R12 were misbehaving):
D11: 2.0V
D12: 1.15V

Left channel (the one where R9 misbehaved):
D11: 1.5V
D12: 1.2V


If it means anything: This all happened after I replaced all the JFET's, because some of them were blown; Both circuits were powered on with one or more blown JFET's.

[amb]

OK, those voltages look reasonable.
On your left board, examine it carefully for cold solder joints, which are the most likely cause of intermittent problems.
Also check your input and output jack wiring -- a problem there could cause the sound to only appear on one channel.

[Tilpo]

I have checked the entire left board for cold solder joints, and corrected every single joint that looks even only slightly suspicious. However, the problem of the sometimes high voltages persists. For now, I think it's best to assume that it's a measuring error, as it doesn't seem to happen very often.

First I want to focus on the input/output jack wiring. I'm quite sure the output wiring is fine, so the problem is probably with the input wiring.
I think it's some kind of grounding problem, since it's the same problem on both channels. It could also be a problem with the wiring on the pot, or a component failure caused by the blowing of the JFET's (since this happened on both boards). Since I can't seem to find the problem myself, I want to ask you whether what I'm currently doing is correct.

The resistances between right, left and signal ground are fine.
The input jacks are grounded to signal, and connected to the pot via a DPDT switch so that I can choose between two inputs. The pot is connected in the way shown in the datasheet, or at least my interpretation of that.
All of the wires connected to the chassis ground are done so at the same point. Chassis ground is then connected to the power receptacle. Also makes me wonder; can I connect the receptacle to a socket not grounded to earth? (not that I have done this)
All the signal grounds are connected to a ground terminal on the Sigma22. One of these terminals is connected to chassis ground via a ground loop breaker. This one consists of a 0.1uF capacitor and a 1 Ohm resistor connected in parallel.

[amb]

Can you post some clear photos of all this?
Btw, 1 ohm is probably too low for an effective ground loop breaker.

[Tilpo]

Would a too low resistance be able to cause a (loud) hum?

I've uploaded photos here. Tell me if you need any more photos.

[amb]

Looks like you did not adhere to the ground wiring scheme shown in the β22 website "wiring & ground" section. Change it to match the diagram for the single-chassis scheme first to see if it's better. The only differences between that and yours are your extra selector switches, but they shouldn't alter the ground wiring scheme.

As I said, a 1 ohm ground loop breaker is not effective, and may cause a ground loop itself. I recommend 10 ohms (or a little higher is ok too).

Other comments:

- Electrical tape for RCA jack isolation? Not reliable...
- A length of solder (?) as the ground wire from the RCA jack to the σ22 board? Other issues aside, that shouldn't even be there (according to the recommended wiring scheme).
- The wires for the power switch (containing AC mains) run too close to some of the other wires and may be inducing hum too.
- You may not be able to get rid of all hum (even when all wiring is done properly) due to the proximity of the (unshielded) power transformer to the amp boards and low-signal wiring.
- You have lots of metal shavings and debris in the case which you should vacuum out thoroughly. They are all potential short circuit hazards.

[Tilpo]

Thanks for your reply. I tried to match the scheme from your website quite carefully, and I don't really get the difference, but I will take a closer look very soon.

The electrical tape for isolation is only temporary. I couldn't think of a better way to isolate them from the chassis at the time.
The use of solder to ground the RCA jacks to signal ground was only done as a test. Initially they were not grounded at all.
Is there a way to insulate the AC mains wires such that they would not cause hum/less hum? I really do want to have a power switch on the front panel.
Is there a way to shield the power transformer effectively?
I kept putting of cleaning up the dirt inside the amp, but I will do it very soon.

[amb]

Thanks for your reply. I tried to match the scheme from your website quite carefully, and I don't really get the difference, but I will take a closer look very soon.

There are some glaring differences -- the website diagram does not tell you to run a wire from the RCA jack ground to the power supply. Instead, it calls for wiring the RCA jack ground to the volume pot and then to the amp boards (I would run separate ground wires for each channel for this purpose). But those appear to be missing in your build.

The electrical tape for isolation is only temporary. I couldn't think of a better way to isolate them from the chassis at the time.

Most RCA jacks should come with isolating washers, but if not, you should drill a somewhat oversized hole on the panel and then mount the jack to the panel using two non-conductive washers (one on each side), taking care to center the jack such that it doesn't touch the edge of the hole. Tighten the mounting nut really well so that the jack won't shift when you connect/disconnect cables.

Is there a way to insulate the AC mains wires such that they would not cause hum/less hum? I really do want to have a power switch on the front panel.

It's hard to shield AC, if that's you mean. The better solution is to run the wires away from the sensitive stuff. It's all about wire routing.

Is there a way to shield the power transformer effectively?

It's hard to do it effectively after the fact. A shielded transformer is constructed with shielding bands and/or other implements wrapped inside. You can try putting the transformer in a ferrous box (not aluminum) with mu-metal sheets lining inside, but chances are the magnetic field would still leak out. Rotating the transformer around sometimes helps a bit.