wintermute said:
Well, well...
Yes, you didn't ground the chassis without earth, did you?
That would solve your problem.
Earth causes ground loops, when you have more than one earthed component on your system.
But not only!
You are injecting noise on the signal.
But you can even have noise if you have only one earthed component, it that component is not the center of the system (the pre or integrated amp).
Basically, all is apparently well if you only have one source, but if you have more, problems will arise.
Now check it out how in 5 minutes I solved an old problem on a friend's system: he had hum, with a relatively high level, you could hear it even playing music, on the lower passages.
He had 4 sources connected to the amp.
Only one component was earthed: a Creek tuner.
I disconnected everything from the amp.
No signal. Volume up, silence.
I connected the tuner first (my main suspect). Still silence.
I went on connecting the sources one by one and testing.
At the third source connected, hummmmm...
Disconnected the earth on the tuner.
Connected all the four sources. Complete silence.
No, thanks.
Your amp must have a correctly rated fuse, you don't need earth.
carlosfm said:
Sorry Carlos I have to disagree on this point, a fuse is not going to save your life in a live chassis situation. It only takes a bit over 50mA to kill you, no fuse will ever blow as a result of the current flowing through a human being electrocuted!!! maybe if you have earth leakage detection circuits then you will be ok but without them (earth leakage, at least in countries like Australia with 240V mains) you will likely be toast!!
My new gainclone is very compact, is properly earthed to the mains, and has virtually no hum (only audible at full volume with my ear about 1cm from the speaker, and then I can barely hear it).
Unless you have a double insulated case, mains earth IS necessary (I'm not saying it's necessary to have no hum, but is necessary for safety reasons).
Tony.
I also agree with wintermute. Safety earth should be connected. But if you insist, use a loop breaker safety circuit. These are very effective according to what I have read... luckily I never had to find out if they do work as intended in case of a fault 
I'm in the process of assembling all the stuff back into the amp's case which is usually when the problems return :\ Up untill now whenever I seem to reach a buzz free condition, upon putting everything back in it's place, the buzz returns. If this time it's not acceptable I'll star doing a case for placing the transformer far away... This project is really giving me headaches
I'm in the process of assembling all the stuff back into the amp's case which is usually when the problems return :\ Up untill now whenever I seem to reach a buzz free condition, upon putting everything back in it's place, the buzz returns. If this time it's not acceptable I'll star doing a case for placing the transformer far away... This project is really giving me headaches
wintermute said:
You don't have 6 sources, do you?
I'm still trying to figure out how could you get an electrical shock with high voltage on the chassis of an amp like this.
Your '50ma' figure is based on what voltage?
We humans will only get an electrical shock at around 70V and up, and I don't have these voltages inside my amp.
PS: I understand your safety reasons, and I worry about that too, I just don't think it's necessary to earth this amp.
Hi Carlos,
I'm talking about fault conditions
Like if the insulation on the primary of the transformer is faulty and shorts to the chassis, or for what ever reason the input from the mains comes loose and touches the chassis. Not every day occurences, but they do happen! if this happens and you have 240V live on the chassis, and you are standing on a concrete floor (in bare feet) or happen to be touching something else that IS earthed, then you are in trouble!!!
You are correct on my gainclone I only have one input on my other amp which has problems I have 6 inputs and two outputs, but my noise problems are present regardless of whether anything is connected or not... even if I disconnect the pre from the power amp (it's an integrated) the noise remains.... different problem I think, possibly oscillation related.
Tony.
I'm talking about fault conditions
Like if the insulation on the primary of the transformer is faulty and shorts to the chassis, or for what ever reason the input from the mains comes loose and touches the chassis. Not every day occurences, but they do happen! if this happens and you have 240V live on the chassis, and you are standing on a concrete floor (in bare feet) or happen to be touching something else that IS earthed, then you are in trouble!!!
You are correct on my gainclone I only have one input
on my other amp which has problems I have 6 inputs and two outputs, but my noise problems are present regardless of whether anything is connected or not... even if I disconnect the pre from the power amp (it's an integrated) the noise remains.... different problem I think, possibly oscillation related.Tony.
wintermute said:Hi Carlos,
I'm talking about fault conditions
Like if the insulation on the primary of the transformer is faulty and shorts to the chassis, or for what ever reason the input from the mains comes loose and touches the chassis. Not every day occurences, but they do happen! if this happens and you have 240V live on the chassis, and you are standing on a concrete floor (in bare feet) or happen to be touching something else that IS earthed, then you are in trouble!!!
I've never seen that happen, a short in the primary will blow the fuse immediately. And you should have one.
Also, I trust my connections with blind faith, because I solder all the wires (including mains), and isolate everything.
No, you won't see me tightening a screw on a mains wire, I even solder the mains plugs!
But let's not discuss this, because it's really no harm to be extra prudent, so if you want to earth the amp, by all means do it.
Fair enough... I think it is important for everyone to know that this is your oppinion, and not fact though! I guess we can discuss off line if you want to IMO It should be possible to have mains faults where the chassis is live and the fuse doesn't blow. This is why there are regulations on this sort of thing!!
Tony.
IMO It should be possible to have mains faults where the chassis is live and the fuse doesn't blow. This is why there are regulations on this sort of thing!! Tony.
wintermute, I don't know if we are talking the same thing, but the important point I made is that you should not connect the earth to the ground of the amp, you are polluting the signal, and you will also have hum, in some circumstances.
My external PSUs are earthed (earth connected to the metal chassis), but not connected to the PSU ground.
So, my amps are not earthed.
The PSUs are.
My external PSUs are earthed (earth connected to the metal chassis), but not connected to the PSU ground.
So, my amps are not earthed.
The PSUs are.
ok now I get it!!!!! you have an external case with the transformer and power supply in it, and this DOES have a connection to mains earth???
we are talking about two different things me thinks
edit: this is better, but I think there is still one possible fault although unlikely, and that is where a short occurs between primary and secondary winding of the transformer, allowing (potenially depending on PS circuit topology) the 0V line to be at mains potential (meaning that input and output sockets could become live).... I sent you an email just before this post
Tony.
carlosfm said:Likely?
No, very likely.
I've had problems with aligator clips many years ago, and I tell you what you must do when you buy a bunch of those cheap things: open them all and solder the wires, if you don't wanna have problems!
Some of them are so badly crimped that they only work by pure luck, at a certain angle.
You'll only have a little work once, then you'll have peace of mind.
Carlos, thanks for reminding me. I often do resolder the alligator clips and sometimes replace the wire as well.
I just double checked and all alligator clips are soldered (surprised
). BTW: I do have a "solder and heat shrink" phobia. 
Regards
Finally!!!!
After many hours experimenting with every possible combination I could think of, I've solved the problem.
Basically the problem was related to how I wired the input signal cable. I left the shield connected on one end only between input plug and amp board. This, of course, created a big current return loop area that picked up all the noise radiated from the transformer and injected it into the input of the amp. It was like having an antenna connected to the amp. The solution is very obvious, I just connected the shield on both ends and disconnected the wire that was going from input plug ground to star point. Another thing that helped was connecting the transformer center tap to the star ground and not to the beggining of the filter bank. This way the transformer stays inside the case and the noise is very little. It could be better but I suspect now the problem is within the pre-amp, so I'm working on that now.
Remember this: the cable from input plug to amp board should work as an extension of the interconnect cable. That is why it is so important to isolate the plug from the chassis in case you use unbalanced inputs and outpus.
This was a hard lesson to learn but I'll never have problems again because of a bad ground layout. Thank you to all who posted and I hope this can be useful to someone
After many hours experimenting with every possible combination I could think of, I've solved the problem.
Basically the problem was related to how I wired the input signal cable. I left the shield connected on one end only between input plug and amp board. This, of course, created a big current return loop area that picked up all the noise radiated from the transformer and injected it into the input of the amp. It was like having an antenna connected to the amp. The solution is very obvious, I just connected the shield on both ends and disconnected the wire that was going from input plug ground to star point. Another thing that helped was connecting the transformer center tap to the star ground and not to the beggining of the filter bank. This way the transformer stays inside the case and the noise is very little. It could be better but I suspect now the problem is within the pre-amp, so I'm working on that now.
Remember this: the cable from input plug to amp board should work as an extension of the interconnect cable. That is why it is so important to isolate the plug from the chassis in case you use unbalanced inputs and outpus.
This was a hard lesson to learn but I'll never have problems again because of a bad ground layout. Thank you to all who posted and I hope this can be useful to someone
Devius said:
Who would have guessed you have done otherwise?
Well, it's solved.
You can try this, it may be even better on such a tight layout and a nearby trafo:
http://www.diyaudio.com/forums/showthread.php?postid=783494#post783494
I said it at least one time if I remember correctly, that I had connected the input signal ground to the star point. And it showed on the ugly layouts I've posted.
I don't think this point is stressed enough. It should be one of the first things to check whenever someone has a buzz problem. Hum is different.
BTW, what is the relevance of the post you indicated? I could only see a discussion about cables and a valve amp with lots of space to spare and everything very neetly arranged.
Devius said:
We were talking about shielded interconnects, and that member was talking about crosstalk with a dual shielded cable.
The semi-balanced arrangement I recommended isolates better not only from 'external' interferences but also from the several channels, even if the cables run close to each other.
Seems like a good solution for your layout, that's it.
Oh I see!! I thought the link was because of the physical disposition of elements inside the chassis and so I didn't pay much attention to the talk about cables. Your suggestion seems to be an interesting arrangement I'll give it a try when I can get some microphone cable. I do have 4 wire + shield but it's a bit thick and as so it's unusable on my tight layout.
I'll give it a try when I can get some microphone cable. I do have 4 wire + shield but it's a bit thick and as so it's unusable on my tight layout.Final conclusions after rewiring most of the pre-amp:
- The preamp itself wasn't responsable for any of the noise present at the output of the power amp. The headphones output, that is connected to the last amplifying stage, is absolutely silent. When I short the output of the pre-amp to ground the power amp still buzzes.
- If I short the input of the amps directly on the board to ground the buzz disappears, which means the noise is being injected into the input of the power amp by the signal cable. I could try CarlosFM suggestion of a semi-balanced connection, but it would be extremely troublesome in this case.
- The one thing that reduces noise to acceptable figures is putting a wire between the input plug ground and the star point. This only works on channels 3,4,5 and 6. Channels 1 and 2 are still noisy but as someone said, the noise is probably induced into the input of these amps because the wires connecting to psu, speaker out and signal all cross the transformer.
I will not try to do anything else for the time being. Maybe in the future if there is the need to use the noisy channels I'll make a new, more spacious case.
- The preamp itself wasn't responsable for any of the noise present at the output of the power amp. The headphones output, that is connected to the last amplifying stage, is absolutely silent. When I short the output of the pre-amp to ground the power amp still buzzes.
- If I short the input of the amps directly on the board to ground the buzz disappears, which means the noise is being injected into the input of the power amp by the signal cable. I could try CarlosFM suggestion of a semi-balanced connection, but it would be extremely troublesome in this case.
- The one thing that reduces noise to acceptable figures is putting a wire between the input plug ground and the star point. This only works on channels 3,4,5 and 6. Channels 1 and 2 are still noisy but as someone said, the noise is probably induced into the input of these amps because the wires connecting to psu, speaker out and signal all cross the transformer.
I will not try to do anything else for the time being. Maybe in the future if there is the need to use the noisy channels I'll make a new, more spacious case.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.