Andrew T.,
So...the mains safety earth...Right now, it's connected to the heat sink which is bolted to the chassis (=real good electrical conductivity with the chassis). Do you think I need to actually move the safety earth so that it is physically connected to the chassis? Personally I think there is enough conductivity with the chassis, but I appreciate opinions.
As for insulating the chip--am I good there? The way the edits look, I'm a bit confused. The amp functions beautifully save for the noise when no signal is attached (I haven't noticed any "destruction noises" while listening to it).
So...the mains safety earth...Right now, it's connected to the heat sink which is bolted to the chassis (=real good electrical conductivity with the chassis). Do you think I need to actually move the safety earth so that it is physically connected to the chassis? Personally I think there is enough conductivity with the chassis, but I appreciate opinions.
As for insulating the chip--am I good there? The way the edits look, I'm a bit confused. The amp functions beautifully save for the noise when no signal is attached (I haven't noticed any "destruction noises" while listening to it).
Hi J,
do you have tf versions?
Yes, you must move the safety earth to the chassis. And make it permanent.
Following on from BWRX on LIVE wires strewn about the insides. If you insist on running mains wires around like that then run double insulated two core for the flow and return and insulate the ends from interfering fingers/tools. It might not be you that opens it up one day!!!
do you have tf versions?
Yes, you must move the safety earth to the chassis. And make it permanent.
Following on from BWRX on LIVE wires strewn about the insides. If you insist on running mains wires around like that then run double insulated two core for the flow and return and insulate the ends from interfering fingers/tools. It might not be you that opens it up one day!!!
I just checked my order, and unless they sent me the wrong part, I have the TF--so I guess I'm good on that. I'll double check when I get home, though. I'm not sure I like the idea of destruction. I'll move that safety earth--I suppose I don't want to get electrocuted...and the strewn about live wires: I'll thake care of that too.
Now that we're safe, any ideas about noise?
Now that we're safe, any ideas about noise?
This might be helpful even though it is of a different design.
http://www.chipamp.com/docs/lm3886-manual.pdf
You can learn a lot here.... a must for anyone starting on amps.
http://sound.westhost.com/
Make sure the signal ground only goes to the amp PCB. You can separate the signal ground from the power ground by adding a 10R resistor in between.
If you haven't tried a ground wire from the star ground to the chassis earth, give it a go. You can also add a 10R 2W resistor that has 2 opposing diodes in parallel to the end of this wire like in the pic.
http://www.chipamp.com/docs/lm3886-manual.pdf
You can learn a lot here.... a must for anyone starting on amps.
http://sound.westhost.com/
Make sure the signal ground only goes to the amp PCB. You can separate the signal ground from the power ground by adding a 10R resistor in between.
If you haven't tried a ground wire from the star ground to the chassis earth, give it a go. You can also add a 10R 2W resistor that has 2 opposing diodes in parallel to the end of this wire like in the pic.
Attachments
The signal ground is to the PCB only, and through a 10 ohm resistor. I re-grounded the input shields and moved the live mains. I got cold feet about connecting the star point to earth. Could this really be causing my noise? I don't really want to do that unless it would solve my noise problem. I'm somewhat concerned about moving more current around on wires that go right next to the PCB where all the senstive stuff is. From what I understand, the way things are now there should be little or no current on the earth connection unless something goes wrong, and I'd kinda like to keep it that way. Is there a chance that making this connection would cause current on the earth conductor (substantially larger than the minimal current from the shields, and more frequent than current caused by some accident)? It seems like the smoothing capacitors would take care of any problems with noise at the star point. Anyhow, keep the advice coming, I really appreciate it!
Oh, and on a side note, where do I get those spiffy toroid mounting plates? I thought they would come with the 'roids, but they didn't, and I was way to impatient to go looking for some.
Thanks!!
Oh, and on a side note, where do I get those spiffy toroid mounting plates? I thought they would come with the 'roids, but they didn't, and I was way to impatient to go looking for some.
Thanks!!
[email]jumpus@mac.com[/email] said:
Hi Kenneth,
I don't quite see how your grounding is done. (A schematic would sure be nice...)
For a star ground, you'd probably want to take all grounds, *separately*, to the negative side of the main filter capacitor(s), or somewhere just off center, between them. --> i.e. If they all run together, in some length of wire, then you have ruined the star grounding. The center tap might be OK, if it were VERY close to the caps. But ideally you don't want the currents to share any length of wire.
The inputs' shields' grounds running to PCB signal ground through a small resistor (like 4.7 Ohms) is a good thing. BUT, you DON'T want the input jacks' grounds ("outer barrels" of the connectors) to be connected to the case. So you'd have to have the insulated type of jacks, or else mount them on a non-conducting material in a cutout.
Where do your speaker grounds go? Do they run separately (i.e. by themselves) all the way to the center tap (or, preferably, the gnd side of the caps)? Note that they should be insulated from the case, too, just like the inputs.
Your mains wiring could be causing hum, too. Remember that wires with changing currents will act like transmitting antennas, and can capacitively couple to other circuitry.
Anyway, I don't remember you saying anything about what happens when you short the inputs to ground. Is there hum, then?
I'd use an ohmeter and measure from your center tap or caps' gnd to the case, first, and verify that there's no connection, anywhere. Then, if you want to try connecting the center tap to the case through a 10 Ohm resistor, you can do that and see if it helps. If it does, you'll want to use a full "disconnect network", instead of just the resistor, which is usually two antiparallel diodes and a 0.1uF cap, all in parallel with the 10R.
- Tom Gootee
http://www.fullnet.com/~tomg/index.html
[email]jumpus@mac.com[/email] said:
If your input shielding is also your input ground, and you've connected it to the case, that might be the problem.
If the shields are not connected to the input signal or ground, they could still cause a problem, if you have them grounded at both ends. Only one end of each shield should be connected. Otherwise you probably create a ground loop.
- Tom Gootee
http://www.fullnet.com/~tomg/index.html
[email]jumpus@mac.com[/email] said:
No! NOT for the return currents, which is what the star ground is for. The NOISE is created back at the OTHER end of each ground return, where it does its harm. It's caused by the distributed resistance and inductance of the wires and traces, when currents flowing through them induce a voltage across them. So if a big, noisy return-current, heading for the star ground, shares a length of ground-return conductor that connects to anything that connects back to a "sensitive" point in your circuit, it will pollute that point with voltage noise. The amount of pollution will be proportional to the length of conductor that is shared, et al.
- Tom Gootee
http://www.fullnet.com/~tomg/index.html
Ok. I'm going to do my best to describe the grounding scheme. I don't really want to draw a schematic because that means I'd have to use the scanner which is an eternal hassle. If this doesn't work, though, let me know and I'll draw one up after work.
The power supply is connected identically to this:
http://www.sound.westhost.com/psu-wiring.htm
That leaves us with the chassis grounded to earth, and the ground created between the filter caps:
Connected to chassis ground point:
back pannel, RCA shielding (by one end as suggested)
Between the filter capacitors (on their respective ground sides), There is a HUGE wire (.3 ohms from end to end according to multimeter). Soldered to this:
Center tap of Transformer (each transformer gets its own wire)
Output grounds (through separate wires)
PCB ground (through separate wires)
The input grounds are connected to the PCB. From there, the input ground goes through a 10 ohm resistor to PCB ground (at a point VERY close to where the PCB ground goes to filter cap ground)
That's it for grounding. If you have questions about that, let me know--I would rather answer those than deal with that scanner, but like I said before, If you need a schematic I can send you one.
I was thinking last night about the one wire rule for each ground: The RCA shields go to a point on the chassis (seperately) then from that point, a single wire goes to the saftey earth. I don't think this is the source of the noise, though, because I disconnected the shields and the noise persisted without any audiable change. I don't think the extra wire to safety earth created a loope either: I checked with my multimeter.
I also checked that the input shielding is in no way connected to input ground (multimeter again)
I haven't tried shorting the inputs to ground. Which ground? earth? I'll try that after work.
And the center tap to case connection: would it just be--filter cap ground (star point/huge wire) to safety earth through a 10 ohm?
Thanks!!
The power supply is connected identically to this:
http://www.sound.westhost.com/psu-wiring.htm
That leaves us with the chassis grounded to earth, and the ground created between the filter caps:
Connected to chassis ground point:
back pannel, RCA shielding (by one end as suggested)
Between the filter capacitors (on their respective ground sides), There is a HUGE wire (.3 ohms from end to end according to multimeter). Soldered to this:
Center tap of Transformer (each transformer gets its own wire)
Output grounds (through separate wires)
PCB ground (through separate wires)
The input grounds are connected to the PCB. From there, the input ground goes through a 10 ohm resistor to PCB ground (at a point VERY close to where the PCB ground goes to filter cap ground)
That's it for grounding. If you have questions about that, let me know--I would rather answer those than deal with that scanner, but like I said before, If you need a schematic I can send you one.
I was thinking last night about the one wire rule for each ground: The RCA shields go to a point on the chassis (seperately) then from that point, a single wire goes to the saftey earth. I don't think this is the source of the noise, though, because I disconnected the shields and the noise persisted without any audiable change. I don't think the extra wire to safety earth created a loope either: I checked with my multimeter.
I also checked that the input shielding is in no way connected to input ground (multimeter again)
I haven't tried shorting the inputs to ground. Which ground? earth? I'll try that after work.
And the center tap to case connection: would it just be--filter cap ground (star point/huge wire) to safety earth through a 10 ohm?
Thanks!!
Hi Kenneth,
Thanks for the additional information.
About your 0.3-Ohm wire. It's probably much lower than that (we can hope). FYI: You can try shorting the meter leads together, and then subtracting that resistance measurement from every other measurement.
The inputs can be shorted to their own grounds, which I have understood to not be the same as their shields' ground (the case/earth). You do have two conductors plus a shield for each input, right? If you don't have shorting plugs made for the input connectors, maybe you could clip on a jumper from where each input connects to the PCB, with the other ends of the jumpers clipped to the star ground point. Shorting the inputs is usually one of the first tests you would perform.
Regarding connecting the star ground to the case/earth ground: Yes, you can try just touching a 10-Ohm R between them. If it's more convenient, and you don't measure any voltage between them, you can just use a wire, to test it. But note that, usually, connecting the star ground to earth ground makes hum worse, not better. As far as I know, it's usually only done for safety purposes, although I could be wrong.
One other thing you MIGHT want to try: This just now occurred to me. And I've never tried it. So it might be totally bogus. But, on the chance that the hum is radiated from your mains wiring, or something like that, and is somehow getting into your signal circuitry, maybe you could put some very thin sheet metal, e.g. aluminum foil, inside of a plastic bag (like a flat multilayer folded piece of aluminum foil in a thick, freezer-type one-gallon Ziploc bag, maybe; or two thick bags, to be safer), and experiment with using that to try to shield various things, to see if you can detect any difference. It might help to run a wire into the bag, and clip it to the foil, and clip the other end to the earth ground. Be careful, though! It sounds like it could be dangerous. Keep one hand in your pocket. Seriously. (Note, though, that while this might work to shield electric fields, it won't work well to shield magnetic fields.)
- Tom Gootee
http://www.fullnet.com/~tomg/index.html
Thanks for the additional information.
About your 0.3-Ohm wire. It's probably much lower than that (we can hope). FYI: You can try shorting the meter leads together, and then subtracting that resistance measurement from every other measurement.
The inputs can be shorted to their own grounds, which I have understood to not be the same as their shields' ground (the case/earth). You do have two conductors plus a shield for each input, right? If you don't have shorting plugs made for the input connectors, maybe you could clip on a jumper from where each input connects to the PCB, with the other ends of the jumpers clipped to the star ground point. Shorting the inputs is usually one of the first tests you would perform.
Regarding connecting the star ground to the case/earth ground: Yes, you can try just touching a 10-Ohm R between them. If it's more convenient, and you don't measure any voltage between them, you can just use a wire, to test it. But note that, usually, connecting the star ground to earth ground makes hum worse, not better. As far as I know, it's usually only done for safety purposes, although I could be wrong.
One other thing you MIGHT want to try: This just now occurred to me. And I've never tried it. So it might be totally bogus. But, on the chance that the hum is radiated from your mains wiring, or something like that, and is somehow getting into your signal circuitry, maybe you could put some very thin sheet metal, e.g. aluminum foil, inside of a plastic bag (like a flat multilayer folded piece of aluminum foil in a thick, freezer-type one-gallon Ziploc bag, maybe; or two thick bags, to be safer), and experiment with using that to try to shield various things, to see if you can detect any difference. It might help to run a wire into the bag, and clip it to the foil, and clip the other end to the earth ground. Be careful, though! It sounds like it could be dangerous. Keep one hand in your pocket. Seriously. (Note, though, that while this might work to shield electric fields, it won't work well to shield magnetic fields.)
- Tom Gootee
http://www.fullnet.com/~tomg/index.html
Tom-
Thanks for the advice. I was actually thinking the same thing (about shielding the mains, or the amp PCBs) last night. I'm about 90% sure that my long mains wire is causing the noise due to the nature of the hum (typical electrocution soundeffect). The odd thing is that the noise doesn't go away completely when the power is shut off<while the amp is still running off the smoothing caps>. It seems like if there were no current going through the live mains (suspect #1) due to it's circuit being open, there should be no more power supply noise. But that isn't the case. That same pulsating (probably ~60 Hz, but I can't count that fast) sound comes through the speakers, just at a substantially lower volume.
I'll go ahead and try shorting the inputs the way you described.
On a side note, I'm assuming you build amps to some extent. How close do you need to get to the speakers to hear any noise with no signal and amp at full power? I always see people claiming that their amp makes no noise at full power with their head in the driver, but that doesn't seem legitimate. I have never experienced an amplifier that is 100% silent at full power (with no signal). The best one I've ever been in the presence of had an audiable hiss (whisper status) from the tweeters at full volume with no signal. Your ear had to be within ~2 inches of the driver to notice it, but there was noise. I'd be satisfied if the noise was reduced to that level, or even a bit louder. As it stands, I can just hear the humming from about 1.5-2 feet away.
Anyhow, thanks again!
Thanks for the advice. I was actually thinking the same thing (about shielding the mains, or the amp PCBs) last night. I'm about 90% sure that my long mains wire is causing the noise due to the nature of the hum (typical electrocution soundeffect). The odd thing is that the noise doesn't go away completely when the power is shut off<while the amp is still running off the smoothing caps>. It seems like if there were no current going through the live mains (suspect #1) due to it's circuit being open, there should be no more power supply noise. But that isn't the case. That same pulsating (probably ~60 Hz, but I can't count that fast) sound comes through the speakers, just at a substantially lower volume.
I'll go ahead and try shorting the inputs the way you described.
On a side note, I'm assuming you build amps to some extent. How close do you need to get to the speakers to hear any noise with no signal and amp at full power? I always see people claiming that their amp makes no noise at full power with their head in the driver, but that doesn't seem legitimate. I have never experienced an amplifier that is 100% silent at full power (with no signal). The best one I've ever been in the presence of had an audiable hiss (whisper status) from the tweeters at full volume with no signal. Your ear had to be within ~2 inches of the driver to notice it, but there was noise. I'd be satisfied if the noise was reduced to that level, or even a bit louder. As it stands, I can just hear the humming from about 1.5-2 feet away.
Anyhow, thanks again!
I just did the input-short-to-ground test. The noise was reduced dramatically. If I got close I could still hear the same humming, but it was quieter than ever before. Also, when I turned it off, and while it was running off of the smoothers the noise was down to an extremely faint hiss.
How do I make it that way all the time?
How do I make it that way all the time?
I discovered something just now: I think the noise is due to the lights that I have in the same room as the amp. I never noticed before, because it was always dark (lights on at night) when I heard the noise. But today, with the lights off, the noise was gone. Just a *slight* hum, and a more predominant hiss.
Is there any way to get rid of noise in an amplifier caused by something else on the same circuit?
The massive noise when the inputs are disconnected is still there...that could get annoying. How can I get rid of that?
How about the hiss--is there any way to reduce that?
Thanks.
Is there any way to get rid of noise in an amplifier caused by something else on the same circuit?
The massive noise when the inputs are disconnected is still there...that could get annoying. How can I get rid of that?
How about the hiss--is there any way to reduce that?
Thanks.
Hi,
that "lights" noise could be air borne or coming through the mains wiring.
Buy a mains input filter and wire that into your amp. If that is a success, a lower current version could/should be added to each source component that suffers the same problem.
If these don't improve the noise problem then it's air borne and full shielding should give a cure. That means metal cases for all equipment and twisted pair cable inside a shield grounded at one end only.
If that fails buy candles as a cheap skate solution or invest in gas lighting. I hear it was popular as an audio upgrade some time ago.
that "lights" noise could be air borne or coming through the mains wiring.
Buy a mains input filter and wire that into your amp. If that is a success, a lower current version could/should be added to each source component that suffers the same problem.
If these don't improve the noise problem then it's air borne and full shielding should give a cure. That means metal cases for all equipment and twisted pair cable inside a shield grounded at one end only.
If that fails buy candles as a cheap skate solution or invest in gas lighting. I hear it was popular as an audio upgrade some time ago.
Thanks for helping me trouble shoot this. It's a shame there's no easy fix. Oh well. I'll have to tinker with some things to see how much of the noise I can eliminate. Maybe I'll go with the audiophile lighting...
Just out of curiosity, does anybody know what the 'hiss' noise comes from?
One more thing: Why is there so much noise when the inputs are not connected to anything?
Thanks again!
Just out of curiosity, does anybody know what the 'hiss' noise comes from?
One more thing: Why is there so much noise when the inputs are not connected to anything?
Thanks again!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.