The Hum Breaking Resistor is placed between the input+feedback and the power ground. The Ground Loop Breaker goes between the power ground and chassis+PE. The principle is the same, in that it reduces the amount of current and moves the Voltage to where it can do the least damage. I Place two back to back diodes parallel to the HBR to protect it from too high a voltage/current. Using 1 ohm for the HBR and 10 ohms for the GLB works for me.
Current amp on the bench is >0.001% THD, so almost there.
I've been surprised at how tolerant THD has been of my sprawling open layouts all connected with alligator jumper twisted pairs.
The cap is just addressing one small problem, the rest is not much different than Self's layout. I don't know how much better or worse it is, hence the discussion. An isolation transformer and RCD breaker would make a lot of this unnecessary.
THANK YOU.
The terminology can be daunting for a new guy. I have that article. Clearly did not understand as much as I thought I did....
See slide 56 in the first presentation in post#1 on how to do it. Note how the connection to earth is part of the T off - it is not taken from the CT if you are using a single split secondary.
Should we not be making the direct-to-safety-earth connection at the RCA input, a-la Self, for exactly that reason - safety? With nothing physically connected, the RCA ground barrel is an exposed piece of metal, so surely that should be connected directly to earth?
At the risk of annoying everyone by complicating things and/or getting too far away from the original subject of the thread, I have a couple of thoughts/questions:
1. What about building the amplifier to "double-insulated" standard, so there is no safety earth requirement? If I understand correctly, this would solve some, but not all the problems. Does anyone have a reference on how to do this?
2. I feel the correct conclusion to draw from all this is that unbalanced connections are a horrific idea and should be avoided at all costs. What I'd prefer to see is more discussion on how to make amplifiers with balanced inputs and very high common-mode input impedance, in the lowest-distortion way possible. I believe Bob mentioned that his second edition should have a bit more on this. I'd also like to recommend Bruno's excellent The G Word, or How to Get Your Audio off the Ground article.
At the risk of annoying everyone by complicating things and/or getting too far away from the original subject of the thread, I have a couple of thoughts/questions:
1. What about building the amplifier to "double-insulated" standard, so there is no safety earth requirement? If I understand correctly, this would solve some, but not all the problems. Does anyone have a reference on how to do this?
2. I feel the correct conclusion to draw from all this is that unbalanced connections are a horrific idea and should be avoided at all costs. What I'd prefer to see is more discussion on how to make amplifiers with balanced inputs and very high common-mode input impedance, in the lowest-distortion way possible. I believe Bob mentioned that his second edition should have a bit more on this. I'd also like to recommend Bruno's excellent The G Word, or How to Get Your Audio off the Ground article.
^ Have a look at this thread The dozens schemes to wire an amp... particularly the posts by ilimzn
As far as I'm aware there is no guide for building a double insulated power amp. The transformers are unobtanium or too expensive, people want a metal chassis, etc. And DIYers can't dodge and weave through the various regulations.
Yes, balanced connections cause a lot of problems. They are not going away. A good unbalanced implementation is still the most cost effective target for the DIYer.
Yes, balanced connections cause a lot of problems. They are not going away. A good unbalanced implementation is still the most cost effective target for the DIYer.
That was a big disaster when I tried that 40 years go. Had to go out a buy some extruded plastic washers to isolate the RCA input connectors. I've done it this way ever since. Never use chassis for signal ground.
I think it depends on rail voltages. There is an acceptable upper limit to the voltage which may appear on exposed metal. If your rails are low enough it's not an issue.
Speaker terminals are usually exposed and yet I don't hear anyone warning about the voltage on those.
Earthed RCA barrels are incompatible with hum breakers as you cannot safely earth a chassis through a 10ohm resistor, at least for source-borne ground noise (maybe with a parallel diode bridge depending on who you ask). If you omit the hum breaker in my layout you can earth at the barrels.
If you isolate the barrels and connect earth to just chassis and the barrels, speaker 0V output fault current has no fault current path (10ohm resistor doesn't count), so the speaker outputs become dangerous. But with amps with 80V rails, the speaker output can have up to 160Vpp AC anyways (that doesn't seem to stop builders or manufacturers though).
Last edited:
I don't believe he's recommending a connection at the input connector to the chassis. He's still recommending isolating washers, but then run a ground wire from the screen at the input connector to the mains / chassis ground.
Yes, looking at his diagram, I think that’s what he was doing.
Re davada’s experience, If you allow the input barrel to make electrical contact with the chassis, and you then connect the safety ground(earth) and the amplifier 0V to different places on the chassis you will get massive earth loops. You can only have one ground point to the chassis.
So if you connect all three to the barrel, you won’t have any ground loops, so theoretically it will work provided you’ve taken care of the other stuff.
Let me make it clear, I have not tried this - one for the future though!
1. The issue here is that to claim something is double insulated means it has to meet very high standards since if it failed catastrophically (eg live off internally touching the metal chassis) someone’s life could be in danger. I understand that commercial gear making this claim are generally tested by independent labs. Would I trust a DIY amp that claimed DI? No. Would I offer anyone a DIY amp claiming it was DI if they followed the build instructions? No. It’s different if you are building stuff in volume - but one off’s are a real problem IMV
It’s just much, much safer to earth the thing
2. You are right, but as keantoken notes, there’s so much stuff out there that’s still unbalanced that you can’t ignore it. Besides, if we ditched it, half of the stuff on this thread would be unnecessary
This is what Self says: A better approach is to put the audio-chassis ground connection at the input connector, so in Figure18.1 , ground-loop currents must flow through A – B to the protected earth at B, and then to mains ground via B – C. They cannot flow through the audio path E – F. This topology is very resistant to ground loops, even with an unbalanced input; the limitation on system performance in the presence of a ground loop is now determined by the voltage drop in the input cable ground, which is outside the control of the amplifier designer.
See Pages 487,488
See Pages 487,488
Am I correct that an earthed appliance must be designed as if any exposed surface may come in direct contact with the mains through a fault?
If so then the vast majority of DIY power amps must be considered unsafe (and not discussable on DIYAudio), as well as any implementation of a hum breaking resistor.
Surely there must be a resolution to the cognitive dissonance?
If so then the vast majority of DIY power amps must be considered unsafe (and not discussable on DIYAudio), as well as any implementation of a hum breaking resistor.
Surely there must be a resolution to the cognitive dissonance?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.