Hi everyone,
I just recently re-wired and generally overhauled my homebuilt integrated amplifier due to the Lighter Note attenuator I had used in it having given out (can't recommend that design - the LDRs keep breaking all the time and Uriah Dailey refuses to publish the schematics). The power stage uses Rod Elliott's P101 and for the preamp I built opc's "The Wire". Everything is working perfectly now, except that I'm getting some noise that makes me think there's a ground loop. If you look at the picture below, you'll see that I made a dumb decision (didn't quite think it through when I chose what should go where) and separated star ground from the ground rail of the power supply - a very bad idea in hindsight. Also, grounding The Wire's PSU to the star ground makes no difference, but grounding the pot to the Wire PSU ground point has taken away some of the noise. In the pic, the pot is grounded to the gnd rail of the big PSU while the Wire PSU is grounded to the star ground. In the mean time, I've experimented a bit, but can't really say which gives less noise. My guess would be that grounding the pot directly to the Wire PSU should be better since there's less potential for ground loops that way, but as it's not something that's easily doable "on-the-fly" with the amp on and speakers connected, that remains a guess.
Another thing I did was instead of wiring the speaker signal returns (green wires twisted together w/ the red ones at the top left and right) directly to the star (or PSU) ground, I jury-rigged faston connectors to connect to the power ground lug on the P101 boards that had two wires attached: One to the speaker signal return and one to the star ground point. The reason for this is two-fold: It made it easier to twist the output cables, and it means less lugs on the star ground point which makes everything look a bit cleaner (which my amp is in dire need of, as you can see below...). I'm not sure if that's a bad or a good idea, I just wish Rod had included a separate speaker ground pad on the boards...
What's also not visible too well on the picture is that the connections from and to the pot are in shielded wire, and the shield is also connected to star ground (with those blue isolated faston tabs).
I should also mention that there must be two sources of noise: Part of the hum changes volume with the pot (which is wired after the preamp output) and is audible from about 5-10cm from the speaker with the volume on full blast, but there's a tiny bit left (audible only with my ear held directly to the speaker) even with the volume turned down all the way.
So I guess my question is: What would be a noise-free ground setup in my amp? I'm fearing that I'll have to move the star ground point to the large ground rail, but will that fix everything? And since the hum the pot produces is fixed only by grounding it to the Wire PSU in some way or other, should I just leave the Wire PSU floating and connect the pot only to that "ground"?
BTW, my source is a gamma-2 currently fed by a cheapo "usb charger" style SMPS, so no earth ground connection there.
And now the pic of my setup (you can see there's also a P39 in there to limit inrush current to the large transformer) - I hope nobody gets eye cancer.
On a side note, in the course of my experiments I also managed to blow one of my beloved Energy Point 1es' woofer (accidentally touched a V+ rail on one of the P101's with a ground lead).
Luckily I've found a replacement pair on ebay already. The old pair will probably be gutted and I'll find some fitting tweeters and woofers (or maybe just woofers as the tweeters are said to be very good for their price class, I'll head over to the speaker section later to get some advice on that too) to get them running again.
And if you've read up to this point, thanks a lot for your patience, and my apologies for the wall of text - I just wasn't sure what was relevant and what wasn't so I decided to mention everything I could think of.
Edit: Just thought of another question: If I do end up moving the star ground point to the P101 PSU ground (which I have a feeling will be inevitable if I want a dead quiet amp), that would mean that star ground is quite far away from the circuits and the power IEC socket and I'll have to pull some rather long wires - could that have any negative effects on sound quality/hum (assuming all other factors are fixed)?
I just recently re-wired and generally overhauled my homebuilt integrated amplifier due to the Lighter Note attenuator I had used in it having given out (can't recommend that design - the LDRs keep breaking all the time and Uriah Dailey refuses to publish the schematics). The power stage uses Rod Elliott's P101 and for the preamp I built opc's "The Wire". Everything is working perfectly now, except that I'm getting some noise that makes me think there's a ground loop. If you look at the picture below, you'll see that I made a dumb decision (didn't quite think it through when I chose what should go where) and separated star ground from the ground rail of the power supply - a very bad idea in hindsight. Also, grounding The Wire's PSU to the star ground makes no difference, but grounding the pot to the Wire PSU ground point has taken away some of the noise. In the pic, the pot is grounded to the gnd rail of the big PSU while the Wire PSU is grounded to the star ground. In the mean time, I've experimented a bit, but can't really say which gives less noise. My guess would be that grounding the pot directly to the Wire PSU should be better since there's less potential for ground loops that way, but as it's not something that's easily doable "on-the-fly" with the amp on and speakers connected, that remains a guess.
Another thing I did was instead of wiring the speaker signal returns (green wires twisted together w/ the red ones at the top left and right) directly to the star (or PSU) ground, I jury-rigged faston connectors to connect to the power ground lug on the P101 boards that had two wires attached: One to the speaker signal return and one to the star ground point. The reason for this is two-fold: It made it easier to twist the output cables, and it means less lugs on the star ground point which makes everything look a bit cleaner (which my amp is in dire need of, as you can see below...). I'm not sure if that's a bad or a good idea, I just wish Rod had included a separate speaker ground pad on the boards...
What's also not visible too well on the picture is that the connections from and to the pot are in shielded wire, and the shield is also connected to star ground (with those blue isolated faston tabs).
I should also mention that there must be two sources of noise: Part of the hum changes volume with the pot (which is wired after the preamp output) and is audible from about 5-10cm from the speaker with the volume on full blast, but there's a tiny bit left (audible only with my ear held directly to the speaker) even with the volume turned down all the way.
So I guess my question is: What would be a noise-free ground setup in my amp? I'm fearing that I'll have to move the star ground point to the large ground rail, but will that fix everything? And since the hum the pot produces is fixed only by grounding it to the Wire PSU in some way or other, should I just leave the Wire PSU floating and connect the pot only to that "ground"?
BTW, my source is a gamma-2 currently fed by a cheapo "usb charger" style SMPS, so no earth ground connection there.
And now the pic of my setup (you can see there's also a P39 in there to limit inrush current to the large transformer) - I hope nobody gets eye cancer.
On a side note, in the course of my experiments I also managed to blow one of my beloved Energy Point 1es' woofer (accidentally touched a V+ rail on one of the P101's with a ground lead).
And if you've read up to this point, thanks a lot for your patience, and my apologies for the wall of text - I just wasn't sure what was relevant and what wasn't so I decided to mention everything I could think of.
Edit: Just thought of another question: If I do end up moving the star ground point to the P101 PSU ground (which I have a feeling will be inevitable if I want a dead quiet amp), that would mean that star ground is quite far away from the circuits and the power IEC socket and I'll have to pull some rather long wires - could that have any negative effects on sound quality/hum (assuming all other factors are fixed)?
Last edited:
That's the earth safety ground from the IEC inlet (hence the choice of wire color 😉 )
Edit: I just had another idea: What about moving star ground to the P101 PSU ground, but not connecting it to earth ground and having it completely isolated from the chassis (which would of course still be grounded to safety earth - it's not like I wanna get electrocuted)? That would leave everything involved with the signal floating and galvanically completely separated from earth and should eliminate any potential for ground loops with other equipment, and given the fact that I'd use a star ground scheme, there also shouldn't be any ground loops left in the amp itself... Good idea or bad idea? I'd also keep the "Wire" ground isolated from the P101 ground. i'm still not sure about the speaker return though - is it OK to have it connected like I do?
Last edited:
Connect the Earth safety ground to the chassis near the IEC inlet not the star ground. Your star ground is near the input jacks, just where it should be. If the inputs are RCA unbalanced interconnects, then they don't float.
Ok, but doesn't that leave some ground loop potential with the PSU ground not being the star ground point?
And what exactly would be the issue with keeping everything signal-related (that is, the P101 PSU ground as well as the ground point from the Wire preamp, which I'd only connect to the pot) floating and only connecting the chassis to safety earth? That should remove the inputs from the equation altogether, no?
Sorry for all the dumb questions, but grounding is still a pretty confusing issue to me...
Last edited:
Alright, so I've been trying to fix this hum with no success so far. I have now moved the star ground point to the P101 ground rail and left that disconnected from chassis/earth ground. The shields as well as the volume pot ground cable are all connected to the preamp's ground point (I think that's the right way to do it? Only the small signal cables to and from the volume pot are shielded). One issue I've noticed is that the preamp PCBs (both the amplifier board and the PSU board) have plated-through mounting holes that are connected to the ground plane, so the preamp ground is still connected to chassis/earth ground right now. I'm gona get some nylon screws and nuts to isolate the preamp circuit from chassis as well - perhaps that will help. It's also worth noting that I'm noticing two sources of hum/buzz - one of them stays the same no matter the volume pot setting while the other changes its volume with the pot setting. They both sound the same (bad ground/ground loop noise), but they're slightly out of phase - otherwise I probably wouldn't even have noticed the second noise. This is really starting to confuse me...
I can't see your pic.
It says:
"Click the image to open in full size"
But nothing opens the arrow just goes round and round.
Could you attach a sketch of the wiring you have used to connect the modules together, from Input socket to Output socket?
It says:
"Click the image to open in full size"
But nothing opens the arrow just goes round and round.
Could you attach a sketch of the wiring you have used to connect the modules together, from Input socket to Output socket?
I build up a prototype on the bench with out any case/chassis
The mains comes in at the left end to the transformer.
A twisted triplet is run from transformer (left) to (right) rectifier.
A twisted triplet is run from rectifier (left) to (right) smoothing caps.
A twisted triplet is run from smoothing caps (left) to (right)amp PCB.
This strung out assembly takes up about 1m (~3') of bench length.
The signal pair come out of the amp to a temporary Phono socket. I can connect a signal source, or a dummy zero ohms here.
The speaker pair come out on the right. I connect a temporary L||R . This is where I can attach scope probes or Multimeter.
There are no connections to Mains PE anywhere on the prototype.
The ONLY PE connections are in the scope, the signal source and the bench top DMM.
I usually measure 0.0mVac and <0.5mVdc at the output in this "open" condition, even with compact fluorescent lamps ON and soldering iron ON. The soldering iron has terrible HF interference.
You should be able to replicate that 0.0mVac result on your set up.
BTW,
that 0.0mVac reading on the DMM means <0.05mVac. It does not mean zero mVac
The mains comes in at the left end to the transformer.
A twisted triplet is run from transformer (left) to (right) rectifier.
A twisted triplet is run from rectifier (left) to (right) smoothing caps.
A twisted triplet is run from smoothing caps (left) to (right)amp PCB.
This strung out assembly takes up about 1m (~3') of bench length.
The signal pair come out of the amp to a temporary Phono socket. I can connect a signal source, or a dummy zero ohms here.
The speaker pair come out on the right. I connect a temporary L||R . This is where I can attach scope probes or Multimeter.
There are no connections to Mains PE anywhere on the prototype.
The ONLY PE connections are in the scope, the signal source and the bench top DMM.
I usually measure 0.0mVac and <0.5mVdc at the output in this "open" condition, even with compact fluorescent lamps ON and soldering iron ON. The soldering iron has terrible HF interference.
You should be able to replicate that 0.0mVac result on your set up.
BTW,
that 0.0mVac reading on the DMM means <0.05mVac. It does not mean zero mVac
Last edited:
How do you know the interference signals are out of phase?
Are you using a two channel scope to see the phase difference?
Andrew,
To see the pic you're gonna have to take imgur.com off the banned URLs list of your AV software. It really doesn't belong there in the first place, imgur is a completely benign image host. I'll upload the pictures to diyAudio when I get time to resize them.
I figured the two hums were out of phase because when I turned the volume up, it just sounded like the original hum was overlaid with itself, but slightly delayed.
Also, I just tried the amp with the chassis not connected to safety earth, and that makes the hum that doesn't change with the volume control disappear, so I guess isolating the preamp from chassis will fix that part of the problem.
I've also found the source of the second hum: it's the small signal cable running from the preamp to the pot. This cable is shielded, so I have no idea why it would pick up so much hum, but the source appears to be the power cables running to the amp. Those aren't properly braided/twisted yet - I only twisted the +/- lines but didn't include the ground line. I'll fix that next - together with the speaker returns which I'll pull directly to the PSU as they're supposed to be. That should be enough for a completely silent amp, I hope. I can already make it silent by holding the signal and power cables away from each other.
What's confusing me is why the signal cable would pick up so much interference from those power cables, considering that it's shielded... I did choose the right ground point for the shielding in the preamp ground, right?
Here's how it looks now - I managed to make it even uglier. I just couldn't figure out how to connect all those shields neatly to the preamp ground.
To see the pic you're gonna have to take imgur.com off the banned URLs list of your AV software. It really doesn't belong there in the first place, imgur is a completely benign image host. I'll upload the pictures to diyAudio when I get time to resize them.
I figured the two hums were out of phase because when I turned the volume up, it just sounded like the original hum was overlaid with itself, but slightly delayed.
Also, I just tried the amp with the chassis not connected to safety earth, and that makes the hum that doesn't change with the volume control disappear, so I guess isolating the preamp from chassis will fix that part of the problem.
I've also found the source of the second hum: it's the small signal cable running from the preamp to the pot. This cable is shielded, so I have no idea why it would pick up so much hum, but the source appears to be the power cables running to the amp. Those aren't properly braided/twisted yet - I only twisted the +/- lines but didn't include the ground line. I'll fix that next - together with the speaker returns which I'll pull directly to the PSU as they're supposed to be. That should be enough for a completely silent amp, I hope. I can already make it silent by holding the signal and power cables away from each other.
What's confusing me is why the signal cable would pick up so much interference from those power cables, considering that it's shielded... I did choose the right ground point for the shielding in the preamp ground, right?
Here's how it looks now - I managed to make it even uglier. I just couldn't figure out how to connect all those shields neatly to the preamp ground.
Last edited:
Small correction... When I wrote
Stupid edit time limit...
I meant to say that that's where the signal cable is picking up the hum.
Stupid edit time limit...
you may have more than 1 type of hum. 60/50 hz stray signal pickup, 100/120 hz power supply pickup and or ground loops.
One mistake which I make quite often is forgetting to ground the shaft or case of the volume pot.
Why did you put the volume control after the wire headphone amp? It has a very low output impedeance and very low output noise level.
I also use the SE to SE version of the wire as a preamp. Because of some low output sources, I changed mine to a gain of 4x. I use the wire as DC coupled. No input or output capacitors. All input sources to the preamp are AC coupled and the power amplifier is AC coupled input. I use the wire LME49830 with lateral mostfets from OPC as power amps. With a stepped 10k ohm volume control at the input of the wire preamp there is no audible noise at the speakers.
One mistake which I make quite often is forgetting to ground the shaft or case of the volume pot.
Why did you put the volume control after the wire headphone amp? It has a very low output impedeance and very low output noise level.
I also use the SE to SE version of the wire as a preamp. Because of some low output sources, I changed mine to a gain of 4x. I use the wire as DC coupled. No input or output capacitors. All input sources to the preamp are AC coupled and the power amplifier is AC coupled input. I use the wire LME49830 with lateral mostfets from OPC as power amps. With a stepped 10k ohm volume control at the input of the wire preamp there is no audible noise at the speakers.
My volume pot case is grounded (you can see the small green wire running to the shield ground point on the picture). I put the volume control after the preamp because it seemed more logical to me - something about the preamp being able to use its full dynamic range? Not sure if that has any merit but I read somewhere that after the preamp is better. Also, there's less wiring going through the chassis this way.
Either way, I've got the problem pretty much figured out - see my last post. I'm just curious why that shielded cable would pick up so much noise.
Last edited:
it appears that you have only one input. Are you using the wire to run headphones or as a buffer to the power amp?. How about if the input z of the amplifier is decent > 10k or so bypass the wire and have the input go directly to the volume control and then to the power amp.
Yes, I could do that - but there would still be long small-signal wires running through the whole case and would take away the benefit of being able to change 2 resistors for more gain in case I ever want to get really loud or have a weak source.
The input impedance of the P101 is definitely very high as I'm driving it from a 50k ohm pot.
And to clarify this again: the wire is the preamp here. It "feeds" the pot and from there the signal goes into the power amps to make an integrated amplifier. I realize that the unity gain preamp is pretty much useless here, but as I said, I like having the option of a higher gain should I need it.
And what makes you think there's only one input? There are 2 RCA jacks in the back of the chassis with red/green twisted wires going into the wire preamp. I'm only using/testing one channel at a time though because I blew one of my speakers.
(Unles of course you were talking about one stereo input, which is the case)
The input impedance of the P101 is definitely very high as I'm driving it from a 50k ohm pot.
And to clarify this again: the wire is the preamp here. It "feeds" the pot and from there the signal goes into the power amps to make an integrated amplifier. I realize that the unity gain preamp is pretty much useless here, but as I said, I like having the option of a higher gain should I need it.
And what makes you think there's only one input? There are 2 RCA jacks in the back of the chassis with red/green twisted wires going into the wire preamp. I'm only using/testing one channel at a time though because I blew one of my speakers.
(Unles of course you were talking about one stereo input, which is the case)
Last edited:
No.
The speaker Return goes with the speaker Flow back to the amplifier and then close couples with the Power Ground wire that goes back to the Main Audio Ground.
The PSU has a twisted triplet that runs to the amplifier. That triplet includes ONE wire that is PSU Zero Volts. That ONE wire goes from PSU to PCB Power Ground.
I think you need to learn how to assemble a hum free amplifier. You need to "see" what a hum free assembly looks like. Where the necessary wire pairs go to and from.
When you have that, you can add on the pre-amp. When that is hum free, THEN and only then do you try to fold up the long strung out temporary assembly into a shape that fits inside the Chassis.
The important part of that "folding" is that ALL the wire pairs must not be changed in any way. You can make them shorter wire pairs, but they must connect source to receiver in exactly the same way.
Last edited:
Rod Elliott, the designer of the amp, specifies that the speaker returns should go directly to PSU ground. I've always thought it would make more sense to have the speaker return go to the amp so that the sum of currents flowing through the wires from PSU to amp is zero, but Rod is way more knowledgeable and experienced than me so I figured there's probably a good reason for it.
Anyway, if you read the last post again, I've already managed to get it completely hum free, now I just need to make it permanent and safe (no ground lift). My only point of concern is shielded wire picking up audible hum - that shouldn't happen, no?
Last edited:
shield plus one core when properly connected should attenuate most hum and noise.
But connect it wrongly and it can pick up so much noise that one would think one were using unshielded wire.
A twisted two core is just as good as a coax and less likely to be wired wrongly. This is what I use for all my internal wiring that is not RF, i.e. all audio.
A shielded twisted two core is better again.
A shielded star quad is best of all at rejecting interference, if the Source can drive the extra capacitance.
But connect it wrongly and it can pick up so much noise that one would think one were using unshielded wire.
A twisted two core is just as good as a coax and less likely to be wired wrongly. This is what I use for all my internal wiring that is not RF, i.e. all audio.
A shielded twisted two core is better again.
A shielded star quad is best of all at rejecting interference, if the Source can drive the extra capacitance.
Well, my shields are connected to preamp ground on one side only. I'm using two-conductor wire and the two conductors are individually shielded. The shields are all grounded at the same point, and I twisted the shields of two conductors that belong together (i.e. a signal/ground pair) together to connect them to the ground point. Being individually shielded, they're not twisted though. I was under the impression that the shielding should more than make up for that, but perhaps I was mistaken?
WRONG !!!!
using a coaxial as a signal connection requires the core to carry the Signal Flow and also requires the shield to carry the Signal Return.
If you connect the shield at one end only then you have broken the Return route for the Signal. The signal has to find another route to RETURN to Source.
There is your interference. The signal Flow and Return are no longer close coupled !!!!!!!
This is precisely where using a twisted pair instead of a coax avoids breaking the very important return route.
No Builder would connect the Source to the Receiver with a twisted pair and then break one of the twisted pair wires to mistakenly try to "cure" a hum problem
Last edited:
- Status
- Not open for further replies.