Lets have a look at the GND connections to the P101. You have connected signal GND to PGND though your star ground. In doing so you bypassed R19/C11 (the ground loop breaker) and it carries the feedback current to PGND.
Let us start again.
First lets take AndrewT advise and place all flow and returns together.
Starting with one P101.
V+, PGND, V- to PSU
signal GND, signal in to POT
speaker out, speaker return
Let us start again.
First lets take AndrewT advise and place all flow and returns together.
Starting with one P101.
V+, PGND, V- to PSU
signal GND, signal in to POT
speaker out, speaker return
Attachments
Last edited:
It's more probable than you get less hum with the setup I suggested than with the one of yours.
There is no need in local ground loop breakers since there are no connections that'd generate ground loops at all.
When you split main ground connections there will be ground loops through chassis of different magnitudes.
There is no need in local ground loop breakers since there are no connections that'd generate ground loops at all.
When you split main ground connections there will be ground loops through chassis of different magnitudes.
So far I have not connected anything to chassis!!!
The idea is to get each step of the build to work without hum.
Connect a CD player to the temporary input and test the output. Don't start on the next step until you are satisfied with the current step. This is not a paint by numbers design but a work in progress. Step two may be wrong, lets get step one done first.
The idea is to get each step of the build to work without hum.
Connect a CD player to the temporary input and test the output. Don't start on the next step until you are satisfied with the current step. This is not a paint by numbers design but a work in progress. Step two may be wrong, lets get step one done first.
Last edited:
The loop areas in post60 are atrocious.
Mark's version in post61 is far better.
Get all the pairs installed, each with low loop area.
Then add the minimum number of reference connections.
A power amp needs one reference connection from Input to Output.
A pre-amp of similar topology will need one reference connection from Input to Output.
The "end" as far as audio performance is concerned.
This will work on the test bench. Prove it works on the test bench.
If it's a multi-channel amp then add the second channel and prove it too works on the test bench !
Use post62
Then fold it all up into a package that fits inside the selected chassis.
Now Safety:
Connect Chassis to Mains Protective Earth (PE) permanently.
Connect Main Audio Ground to Chassis.
Mark's version in post61 is far better.
Get all the pairs installed, each with low loop area.
Then add the minimum number of reference connections.
A power amp needs one reference connection from Input to Output.
A pre-amp of similar topology will need one reference connection from Input to Output.
The "end" as far as audio performance is concerned.
This will work on the test bench. Prove it works on the test bench.
If it's a multi-channel amp then add the second channel and prove it too works on the test bench !
Use post62
YES !!! do this BEFORE attempting to install inside a Chassis/case. Get all the audio working properly.
Then fold it all up into a package that fits inside the selected chassis.
Now Safety:
Connect Chassis to Mains Protective Earth (PE) permanently.
Connect Main Audio Ground to Chassis.
Last edited:
Can't see the one you are reffering too. Please specify. We are here to help each other no to brag about anything.
There is a loop breaker in the board between P-ground and signal ground. Can't see other possible loops since everything is routed through just one path to the star.
Here are two large area loops. Others are formed around the PSU grounds and their supplies.
Attachments
Last edited:
If you wire the thing as above there are no ground loops. All components are refferenced to the same spot and just once each one. I've implemented this and works perfect.
All pot wires negative side are attached just at the pot ground pins and to the star. There won't develop any ground loops.
All pot wires negative side are attached just at the pot ground pins and to the star. There won't develop any ground loops.
If you wire the thing as above there are no ground loops. All components are refferenced to the same spot and just once each one. I've implemented this and works perfect.
All pot wires negative side are attached just at the pot ground pins and to the star. There won't develop any ground loops.
All pot wires negative side are attached just at the pot ground pins and to the star. There won't develop any ground loops.
You are not understanding what a LOOP is and not understanding what AREA is enclosed by the loop.
A loop with some area behaves as an aerial.
That aerial radiates energy and receives energy.
It's the energy transferred between LOOPs that creates interference.
A loop with some area behaves as an aerial.
That aerial radiates energy and receives energy.
It's the energy transferred between LOOPs that creates interference.
I really understand what you say, but in practice the setup above works. I tried it and found to be the best hum proof scheme.
May be there's no single way of doing things. I'm transfering my real world experience, not just a theory.
The final arrangement on the chassis is highly important. So if this is really wrong it would probably affect the ground scheme performance.
I have an amp working this way with clear as distilled water sound, ZERO interference pick up or hum, even though it has 2 650VA toroids and 2 6V SMPS inside the enclosure.
Also implemented RF bypass caps before and after the volume pot and at the PS to keep that factor absolutely under control.
May be there's no single way of doing things. I'm transfering my real world experience, not just a theory.
The final arrangement on the chassis is highly important. So if this is really wrong it would probably affect the ground scheme performance.
I have an amp working this way with clear as distilled water sound, ZERO interference pick up or hum, even though it has 2 650VA toroids and 2 6V SMPS inside the enclosure.
Also implemented RF bypass caps before and after the volume pot and at the PS to keep that factor absolutely under control.
Alright, let me get back into the discussion here.
Since the layout suggested by Mark Whitney and AndrewT was exactly what I had prior to sextaafondo's suggestions (with the exception that the two amp grounds were not shorted together) and I'd also given connecting the speaker return directly to PSU (as specified by Rod) a try without seeing any improvement from it, I decided to try sextaafondo's scheme out on one channel. Please note that I had the same reservations about magnetically induced hum due to large loop areas put forward by Mark W. and AndrewT, but the fact that before I'd had both a 120hz buzz and some 60hz hum that I could not get rid of no matter what I tried made me decide to give the "total star ground" approach a try anyway.
I set everything up like in the diagram I posted 2 pages back (or rather the corrected one posted by sextaafondo a bit later) and the (quite positive, I'd say) result is that I can now fiddle with the output wires from the preamp and if I hold them just right, I get no hum and no 120hz buzz at all. I believe the buzz I get with the wires arranged in a "bad" way is being caused by the wires that come from the rectifier and pass right below/next to the pot. I'm going to try twisting them and running them further away from the pot and signal wires.
Also, It is possible to twist the ground return from the pot and the signal wires to it rather tightly, which should eliminate one of the loop areas pointed out by AndrewT and Mark W. I'm going to do all these things on one channel and see if it makes the amp reliably and completely quiet - having to rely on exact cable positioning isn't something I'm too happy with. I'll report back as soon as I have any results.
Since the layout suggested by Mark Whitney and AndrewT was exactly what I had prior to sextaafondo's suggestions (with the exception that the two amp grounds were not shorted together) and I'd also given connecting the speaker return directly to PSU (as specified by Rod) a try without seeing any improvement from it, I decided to try sextaafondo's scheme out on one channel. Please note that I had the same reservations about magnetically induced hum due to large loop areas put forward by Mark W. and AndrewT, but the fact that before I'd had both a 120hz buzz and some 60hz hum that I could not get rid of no matter what I tried made me decide to give the "total star ground" approach a try anyway.
I set everything up like in the diagram I posted 2 pages back (or rather the corrected one posted by sextaafondo a bit later) and the (quite positive, I'd say) result is that I can now fiddle with the output wires from the preamp and if I hold them just right, I get no hum and no 120hz buzz at all. I believe the buzz I get with the wires arranged in a "bad" way is being caused by the wires that come from the rectifier and pass right below/next to the pot. I'm going to try twisting them and running them further away from the pot and signal wires.
Also, It is possible to twist the ground return from the pot and the signal wires to it rather tightly, which should eliminate one of the loop areas pointed out by AndrewT and Mark W. I'm going to do all these things on one channel and see if it makes the amp reliably and completely quiet - having to rely on exact cable positioning isn't something I'm too happy with. I'll report back as soon as I have any results.
Last edited:
Alright, I twisted the wires I mentioned in my last post and it changed pretty much nothing about the situation - the small-signal wires (which are shielded) both to the pot and from the pot to the power amp are still extremely susceptible to noise pickup (which figures, considering the way they're run). And there seems to be another problem too, even with the power amp inputs shorted, I can still hear some hum (although it's so tiny and quiet that it might just be leftover ripple from the PSU). I don't know where this is coming from - I haven't been able to find a cause by jiggling the wires.
Edit: Scratch that last part. With earth ground connected and the power amp inputs shorted, it's dead quiet.
And with "good" positioning of the wires, I can get it to be almost dead quiet - there's a very low 60Hz hum that I can't seem to get rid of no matter what I do with the wires (must be coming from before the pot as it changes volume with the pot setting).
Edit: Scratch that last part. With earth ground connected and the power amp inputs shorted, it's dead quiet.
And with "good" positioning of the wires, I can get it to be almost dead quiet - there's a very low 60Hz hum that I can't seem to get rid of no matter what I do with the wires (must be coming from before the pot as it changes volume with the pot setting).
Last edited:
Yay, another triple-post!
I figured out how to make the small-signal wires less sensitive to noise pickup: Instead of grounding the power amp's input signal ground directly to star, I use the shield of the wire as the return connection and ground it at the pot - in other words, I'm using my shielded wire as a plain old coaxial cable. Makes for an awfully long return path, but since it drastically reduces the loop area formed by that wire and its return, the wire becomes MUCH less sensitive to EMI. Bad wire positioning that would have caused buzzing clearly audible from 2m away before now causes hardly anything, and good positioning makes the buzz (obviously only that which was picked up by the pot->P101 wire - I just turn the volume all the way down to make rest disappear) go away entirely.
I'm thinking of doing something similar with the wires going preamp->pot:
My idea was to use a shielded, twisted-pair cable to carry the two channels' "hot" connection in the twisted pair and use the shield of that cable to connect the pot's ground to star ground next to the preamp. This would be pretty elegant (only one cable needed from pot to preamp and grounding is all taken care of) and should have the same effect on the preamp->pot connection as grounding the power amp input ground to the pot had on pot->P101: All the signal current flowing to the pot is supplied by the preamp, so grounding the pot with that shield would give 0 total current inside the wire. Any opinions on this before I get started with it? I believe this setup, combined with running all small-signal wires along some better paths - I was thinking behind the right power amp module -, should get me pretty close to turning this amp into "Hum-Free Bogart" (I just had to bring up that pun somewhere, please forgive me).
So basically, I would implement sextaafondo's star ground topology with some additional tweaks to account for the loop area issues that were mentioned by Andrew and Mark. The only thing that's maybe not so elegant is that a bunch of signals would share their star ground connection at the pot, but I think that should not cause any (audible) problems - at least it doesn't right now from what I can tell.
I figured out how to make the small-signal wires less sensitive to noise pickup: Instead of grounding the power amp's input signal ground directly to star, I use the shield of the wire as the return connection and ground it at the pot - in other words, I'm using my shielded wire as a plain old coaxial cable. Makes for an awfully long return path, but since it drastically reduces the loop area formed by that wire and its return, the wire becomes MUCH less sensitive to EMI. Bad wire positioning that would have caused buzzing clearly audible from 2m away before now causes hardly anything, and good positioning makes the buzz (obviously only that which was picked up by the pot->P101 wire - I just turn the volume all the way down to make rest disappear) go away entirely.
I'm thinking of doing something similar with the wires going preamp->pot:
My idea was to use a shielded, twisted-pair cable to carry the two channels' "hot" connection in the twisted pair and use the shield of that cable to connect the pot's ground to star ground next to the preamp. This would be pretty elegant (only one cable needed from pot to preamp and grounding is all taken care of) and should have the same effect on the preamp->pot connection as grounding the power amp input ground to the pot had on pot->P101: All the signal current flowing to the pot is supplied by the preamp, so grounding the pot with that shield would give 0 total current inside the wire. Any opinions on this before I get started with it? I believe this setup, combined with running all small-signal wires along some better paths - I was thinking behind the right power amp module -, should get me pretty close to turning this amp into "Hum-Free Bogart" (I just had to bring up that pun somewhere, please forgive me).
So basically, I would implement sextaafondo's star ground topology with some additional tweaks to account for the loop area issues that were mentioned by Andrew and Mark. The only thing that's maybe not so elegant is that a bunch of signals would share their star ground connection at the pot, but I think that should not cause any (audible) problems - at least it doesn't right now from what I can tell.
Last edited:
- Status
- Not open for further replies.