I'm having problems with a recently constructed 3886 amp. There's substantial hum when power is on, and even more if nothing is connected to to the inputs. Also, noise reduces dramatically when power is cut to the transformers, but the amp is still running (on power stored in the supply capacitors?).
I'm thinking the noise is from the transformers, and is making it's way into the signal path through the signal's ground somehow, but I'm just a beginner with this stuff.
What else could the problem be/ how can I fix it?
Thanks!
I'm thinking the noise is from the transformers, and is making it's way into the signal path through the signal's ground somehow, but I'm just a beginner with this stuff.
What else could the problem be/ how can I fix it?
Thanks!
Hi Kenneth,
Hum is one of those things that usually depends on exactly how everything is connected, and layed out.
So you would probably get better advice, sooner, if you could post a photo of your setup, and a schematic, and as much other relevant information as you can.
There are a lot of people here who will like to try to help you, if they are given enough information.
- Tom Gootee
http://www.fullnet.com/~tomg/index.html
Hum is one of those things that usually depends on exactly how everything is connected, and layed out.
So you would probably get better advice, sooner, if you could post a photo of your setup, and a schematic, and as much other relevant information as you can.
There are a lot of people here who will like to try to help you, if they are given enough information.
- Tom Gootee
http://www.fullnet.com/~tomg/index.html
I will put a picture or two up when I get a chance. I just wanted to see if my idea makes any sense. I'm at work right now, so the picture(s) won't be around for a while. Could the symptoms that I described point to anything? I understand if a picture is necessary for any concrete determinations of the problem, but I'm running low on things to do here (at work) and I could really use something to think about.
There are some good threads in this forum about solving hum and grounding problems, which are often related.
I don't know what kind of power supply topology you have used. So it's very difficult to make any specific statements, or even guesses, about it. But I can make a couple of general statements, that might give you some food for thought.
You should use what is called a "star" ground-return setup, in which all major ground current-return paths are run in separate conductors, which all meet at only ONE point (often the ground side of a main filter capacitor in the power supply). This is done because wires and PCB traces have parasitic resistance and capacitance, across which voltages are induced (back at the end of the conductor that's farther from the power supply, in the case of ground returns), by currents, which makes sharing of ground return paths "a bad thing".
Also, typically, the power supply and audio grounds, which usually are connected to each other only at the star ground point, are either not connected to the "earth ground" from the power cord, at all, or, are connected to it through a "disconnect network", which might include (all in parallel) a >= 10 Ohms resistor, a 0.1uF or so cap, and two opposite-facing diodes.
I don't know what kind of power supply topology you have used. So it's very difficult to make any specific statements, or even guesses, about it. But I can make a couple of general statements, that might give you some food for thought.
You should use what is called a "star" ground-return setup, in which all major ground current-return paths are run in separate conductors, which all meet at only ONE point (often the ground side of a main filter capacitor in the power supply). This is done because wires and PCB traces have parasitic resistance and capacitance, across which voltages are induced (back at the end of the conductor that's farther from the power supply, in the case of ground returns), by currents, which makes sharing of ground return paths "a bad thing".
Also, typically, the power supply and audio grounds, which usually are connected to each other only at the star ground point, are either not connected to the "earth ground" from the power cord, at all, or, are connected to it through a "disconnect network", which might include (all in parallel) a >= 10 Ohms resistor, a 0.1uF or so cap, and two opposite-facing diodes.
You might want to first check how your input connectors are grounded. If they are grounded to the case, that could be the problem.
Also, AC wiring should be kept well-away from signal and DC wiring (and components). If that's not possible, they should cross only at right angles to each other. It might also help to twist any pairs of wires you might have.
- Tom Gootee
http://www.fullnet.com/~tomg/index.html
Also, AC wiring should be kept well-away from signal and DC wiring (and components). If that's not possible, they should cross only at right angles to each other. It might also help to twist any pairs of wires you might have.
- Tom Gootee
http://www.fullnet.com/~tomg/index.html
gootee said:
I MEANT to type *** inductance*** (although there is also some capacitance between any two conductors, which depends on the materials and geometries involved).
- Tom Gootee
http://www.fullnet.com/~tomg/index.html
Hi there,
I was just looking at your thread and the picture and I can only see what looks like one bridge rectifier. Could you post a schematic of the power supply layout, I have found that some of the hum problems that "solve" themselves when you switch the power off and its running off the smoothing caps are related to the power supply topology. Also you will get more hum when there is nothing connected to the inputs.
Cheers,
Ian
I was just looking at your thread and the picture and I can only see what looks like one bridge rectifier. Could you post a schematic of the power supply layout, I have found that some of the hum problems that "solve" themselves when you switch the power off and its running off the smoothing caps are related to the power supply topology. Also you will get more hum when there is nothing connected to the inputs.
Cheers,
Ian
Hi,
looks like you are running two transformers with the secondaries in parallel. Not good.
You can run the secondaries of one transformer in parallel if they are identical. It is unwise to run different transformers in parallel.
There is no problem with running the primaries in parallel off the mains. We do it all the time when we plug a second unit into a mains socket.
What is that white wire connecting the chipamp heatsink to somewhere?
What have you used to insulate the chipamps from the heatsink?
How do you intend cooling that flat plate heatsink?
Since you have two transformers and two capacitor banks and two chipamps, then I suggest you buy another rectifier and build a true dual mono amplifier. KEEP EACH AMPLIFIER COMPLETELY SEPARATE from each other electrically, until you are ready to connect the chassis safety earth (where is it?) to the two amplifier audio grounds.
You can post upto 1000 by 1000 pixels resolution if the compressed file is under 100kb
looks like you are running two transformers with the secondaries in parallel. Not good.
You can run the secondaries of one transformer in parallel if they are identical. It is unwise to run different transformers in parallel.
There is no problem with running the primaries in parallel off the mains. We do it all the time when we plug a second unit into a mains socket.
What is that white wire connecting the chipamp heatsink to somewhere?
What have you used to insulate the chipamps from the heatsink?
How do you intend cooling that flat plate heatsink?
Since you have two transformers and two capacitor banks and two chipamps, then I suggest you buy another rectifier and build a true dual mono amplifier. KEEP EACH AMPLIFIER COMPLETELY SEPARATE from each other electrically, until you are ready to connect the chassis safety earth (where is it?) to the two amplifier audio grounds.
You can post upto 1000 by 1000 pixels resolution if the compressed file is under 100kb
To get quiet amp, you need to build using a few basic rules/guides.
Layout is so important and the best results I've had have the amp PCB's at the back and transformers and power supply at the front.
Make sure all your grounds come back to a single point star earth. That includes the transformer, amp PCB's, speaker cold, power supply. This star ground can then be attached to the mains earth on the chassis but sometimes it is quieter without the connection. Under no circumstances remove the chassis from the mains earth.
Wiring needs to be kept apart and runs twisted where possible. Keep signal wiring away from everything if possible and if not run at 90° to speaker wiring. All AC & DC wiring to be twisted and kept away from signal and speaker hot.
Keep the input RCA and speaker out away from each other.
One thing that can give you a hum that is not power amp related, is the body of the pot needs to be grounded in most instances.
The amp in the pic is not a Chip Amp, but notice the layout and the use of the star ground in the centre of the power supply. This amp is super quiet.
Layout is so important and the best results I've had have the amp PCB's at the back and transformers and power supply at the front.
Make sure all your grounds come back to a single point star earth. That includes the transformer, amp PCB's, speaker cold, power supply. This star ground can then be attached to the mains earth on the chassis but sometimes it is quieter without the connection. Under no circumstances remove the chassis from the mains earth.
Wiring needs to be kept apart and runs twisted where possible. Keep signal wiring away from everything if possible and if not run at 90° to speaker wiring. All AC & DC wiring to be twisted and kept away from signal and speaker hot.
Keep the input RCA and speaker out away from each other.
One thing that can give you a hum that is not power amp related, is the body of the pot needs to be grounded in most instances.
The amp in the pic is not a Chip Amp, but notice the layout and the use of the star ground in the centre of the power supply. This amp is super quiet.
Attachments
Ian:
I won't be able to get you a schematic until later (at work now). The power supply is pretty basic though--nothing too tricky, and I haven't been electrocuted yet, so I assume it's safe. What kind of things would I need to fix for a supply problem (like the one you suspect?). Power supply wiring: Neutral from mains (white coming in the back) goes to the primaries. The hot from mains (black from back) sneaks in and attaches to fuse holder. From there, hot (ol' blackie?) shoots across the chassis to the power switch at the front, then strolls over to the other connection for the primaries. The secondaries are soldered (gooed as I like to say) to the appropriate sides of the rectifier. As you brought up earlier, the rectifier is shared between the two transformers. From the rectifier, red and blue scoot over to the smooth(ing) capacitors. The earth from mains (green) meanders to the heatsink for case grounding.
Andrew T.:
The white wire grounds the case. That connection on the heat sink is the green wire from the mains (earth with any luck). Insulate the chipamps from the heat-sink? I put some thermal compound underneath to do the opposite (conduct the heat out of those things quick). Unless you're talking electrically--in which case nothing, I thought the chip's package material would be enough. The picture of the heat sink is a bit deceiving: It's a normal heat-sink, just mounted upside-down. I think it's big enough that it should cool well enough even though the fins are down. I'm not too worried about that, though. If I read right, the 3886 has some thermal protection mechanisms built in, so if those ever trip, I'll try remounting...it's just such a pain, though. As for the two-transformers-powering-one-supply-through-one-rectifier problem, could that cause the noise that I'm having?
If noise is not the problem with having the powersupply this way, what other problems would I experience? As I said, I'm unexperienced with this amplifier building thing, but my friend had no issues with wiring the supply this way (he helped me through the process, but he's gone out of town now, and I'm anxious to
get this noise problem solved), and he's quite experienced with this diy stuff. He thought it was a grounding issue, so I'd kinda like to check that out before re-doing the power supply. If that's what it takes though...
rabbitz:
Everything except the chassis ground (green wire/earth, to heatsink, white wire to chassis) is grounded to that fat fat wire running between the capacitors. At some point (you can see if you look close) the fat fat wire is attachedto the center tap of the toroids. Does this count as "star ground?" Might the noise go away if I run the earth connection (green from mains) to the star point(fat fat wire)? Does it look like my wiring is separated well enough? The signal ins gets kinda close to the output wires, but those are shielded (more on that below). I don't have any pot in that amplifier (full power all the time!! YEAH!). I figure volume can be managed by the input device until I get my preamp built (project #2 for the summer).
If you (in response to no one in particular) look real close, you can see the signal input jumping up over everything there. Those are the (other) white wires. Those are shielded, and I had them (shields) grounded to that earth connection on the chassis. The noise persisted. I disconnected the shield from there and the noise didn't change at all (that I could tell). An interesting note: the signal ground goes to the PCB, and then to the fat fat wire (star ground?) through a short (really short) trace and resistor. Would this cause a problem? My friend designed the board, so I'm a bit wary to modify anything there--I have confidence that he knew what he was doing when he designed it. Would there be an advantage to connecting the fat fat wire (my star point?) to the mains earth? How about connecting the signal ground to mains earth?
Hopefully I responded to all of the issues you guys brought up. If I didn't, let me know. Thanks for the advice, I really appreciate it!
I won't be able to get you a schematic until later (at work now). The power supply is pretty basic though--nothing too tricky, and I haven't been electrocuted yet, so I assume it's safe. What kind of things would I need to fix for a supply problem (like the one you suspect?). Power supply wiring: Neutral from mains (white coming in the back) goes to the primaries. The hot from mains (black from back) sneaks in and attaches to fuse holder. From there, hot (ol' blackie?) shoots across the chassis to the power switch at the front, then strolls over to the other connection for the primaries. The secondaries are soldered (gooed as I like to say) to the appropriate sides of the rectifier. As you brought up earlier, the rectifier is shared between the two transformers. From the rectifier, red and blue scoot over to the smooth(ing) capacitors. The earth from mains (green) meanders to the heatsink for case grounding.
Andrew T.:
The white wire grounds the case. That connection on the heat sink is the green wire from the mains (earth with any luck). Insulate the chipamps from the heat-sink? I put some thermal compound underneath to do the opposite (conduct the heat out of those things quick). Unless you're talking electrically--in which case nothing, I thought the chip's package material would be enough. The picture of the heat sink is a bit deceiving: It's a normal heat-sink, just mounted upside-down. I think it's big enough that it should cool well enough even though the fins are down. I'm not too worried about that, though. If I read right, the 3886 has some thermal protection mechanisms built in, so if those ever trip, I'll try remounting...it's just such a pain, though. As for the two-transformers-powering-one-supply-through-one-rectifier problem, could that cause the noise that I'm having?
If noise is not the problem with having the powersupply this way, what other problems would I experience? As I said, I'm unexperienced with this amplifier building thing, but my friend had no issues with wiring the supply this way (he helped me through the process, but he's gone out of town now, and I'm anxious to
get this noise problem solved), and he's quite experienced with this diy stuff. He thought it was a grounding issue, so I'd kinda like to check that out before re-doing the power supply. If that's what it takes though...
rabbitz:
Everything except the chassis ground (green wire/earth, to heatsink, white wire to chassis) is grounded to that fat fat wire running between the capacitors. At some point (you can see if you look close) the fat fat wire is attachedto the center tap of the toroids. Does this count as "star ground?" Might the noise go away if I run the earth connection (green from mains) to the star point(fat fat wire)? Does it look like my wiring is separated well enough? The signal ins gets kinda close to the output wires, but those are shielded (more on that below). I don't have any pot in that amplifier (full power all the time!! YEAH!). I figure volume can be managed by the input device until I get my preamp built (project #2 for the summer).
If you (in response to no one in particular) look real close, you can see the signal input jumping up over everything there. Those are the (other) white wires. Those are shielded, and I had them (shields) grounded to that earth connection on the chassis. The noise persisted. I disconnected the shield from there and the noise didn't change at all (that I could tell). An interesting note: the signal ground goes to the PCB, and then to the fat fat wire (star ground?) through a short (really short) trace and resistor. Would this cause a problem? My friend designed the board, so I'm a bit wary to modify anything there--I have confidence that he knew what he was doing when he designed it. Would there be an advantage to connecting the fat fat wire (my star point?) to the mains earth? How about connecting the signal ground to mains earth?
Hopefully I responded to all of the issues you guys brought up. If I didn't, let me know. Thanks for the advice, I really appreciate it!
[email]jumpus@mac.com[/email] said:
Hi Kenneth. You should try to keep the mains wiring as short as possible and as far away as possible from other wiring. I particularly don't like how you have the mains hot wire going from the fuse holder at the back of the chassis to the switch at the front and then back to transformers. It would have been better to place the switch on the back next to the fuse holder. There is a good reason they make IEC jacks with integrated fuse holders other than for convenience - safety!
What kind of transformers are those? Can you tell us the wire color code used for those transformers?
You are using chips with an insulated package so no other insulator is necessary.
I ran the mains wiring around like that so I could have a power switch at the front. I don't really want to move it back to the back since I already drilled a hole in the faceplate for it. Do you think it could really pose enough of a threat that I should relocate the switch? What about noise--shouldn't all of the friendly capacitors take care of that. I think I might move that wire to the other side of the 'roids to keep it away from things, but I'd really like to avoid moving the switch, and unless you think there is a serious threat to me or anyone who might operate the amplifier from the outside, I'll probably leave it the way it is.
The transformers: I got them from Apex Jr. and followed the wiring guide there:
"2 X 25V... For 115Vac Input - Tie the 2 Greys and brown together along with tying the 2 Violets and Blue together
OUTPUT- Black is the center tap and each side Red and Orange"
Do you think it's a problem that I'm only using one rectifier?
The transformers: I got them from Apex Jr. and followed the wiring guide there:
"2 X 25V... For 115Vac Input - Tie the 2 Greys and brown together along with tying the 2 Violets and Blue together
OUTPUT- Black is the center tap and each side Red and Orange"
Do you think it's a problem that I'm only using one rectifier?
If you don't want to relocate the switch then you should at least move the mains wire to the other side of the transformers (as you mentioned).
Since your transformers are identical, there shouldn't be any problem with the way you currently have the supply wired up. If it were my amp, I would use one transformer per chip to make your amp dual mono (Andrew mentioned this as well). To do this you will only need to add another bridge rectifier.
Since your transformers are identical, there shouldn't be any problem with the way you currently have the supply wired up. If it were my amp, I would use one transformer per chip to make your amp dual mono (Andrew mentioned this as well). To do this you will only need to add another bridge rectifier.
Hi Jumpus,
I see the green wire now.
Connect that mains earth wire direct to chassis and NEVER remove it even for maintenance.
You can add extra nuts above it for any additional connections to the mains safety earth.
The chipamps MUST be electrically isolated from the heatsink.
They must also be thermally connected to the heatsink. That means you MUST use a thermally conductive electrical isolator between the chipamp and the heatsink. They come in many types and prices.
The bolt fixing the chipamp must also be isolated.
edit,
If BWRX has correctly identified 3886tf then he is right that you have electrical isolation.
Ignore the following comment.
If you have been lucky the chipamp protection will have saved them from destruction.
I see the green wire now.
Connect that mains earth wire direct to chassis and NEVER remove it even for maintenance.
You can add extra nuts above it for any additional connections to the mains safety earth.
The chipamps MUST be electrically isolated from the heatsink.
They must also be thermally connected to the heatsink. That means you MUST use a thermally conductive electrical isolator between the chipamp and the heatsink. They come in many types and prices.
The bolt fixing the chipamp must also be isolated.
edit,
If BWRX has correctly identified 3886tf then he is right that you have electrical isolation.
Ignore the following comment.
If you have been lucky the chipamp protection will have saved them from destruction.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.