Hello everyone!
As the title suggests, I have built an amp with the following configuration and I have a less than usual hum problem. I have tried the 2 ways of connecting the input ground and the symptoms are quite similar: If i plug in only one channel(doesn't matter witch one of them) there is no hum or noise whatsoever, but when I plug both I get a very loud hum. I have also tried to tie together the 2 star-grounds and the hum is lower but still very loud.
In my drawings there is no safety ground because where I live we have two-prong AC mains. However, there is AC coupling between each amplifier ground and chassis though 220 nF film cap.
What are your suggestions regarding this problem? Many thanks in advance!
As the title suggests, I have built an amp with the following configuration and I have a less than usual hum problem. I have tried the 2 ways of connecting the input ground and the symptoms are quite similar: If i plug in only one channel(doesn't matter witch one of them) there is no hum or noise whatsoever, but when I plug both I get a very loud hum. I have also tried to tie together the 2 star-grounds and the hum is lower but still very loud.
In my drawings there is no safety ground because where I live we have two-prong AC mains. However, there is AC coupling between each amplifier ground and chassis though 220 nF film cap.
What are your suggestions regarding this problem? Many thanks in advance!
Attachments
Hi kin,
1. Are your input RCAs isolated from the chassis ?
2. Does the PSU Ground connections (Gnd Left/Right) have connection to the Signal Gnd in on the PCB ?
3. How is your PSU Gnd's connected to chassis and PSU ?
If not allready done this way, try isolate the RCA from the chassis. If there is a Gnd Left/Right to Input Gnd connection on the PCB, try setup one again.
If there are no connections between Gnd's on the PCBs, connect the RCA Gnd's together, and connect this RCA Gnd to a common GND point between the big filter caps (as in setup two, but RCA
Gnd's should connect to a common Gnd beyween the PSU caps). This common Gnd point (now a star Ground should the be connected to chassis as chassis Gnd).
Always make sure that you only have ONE Gnd connection to each circuit, connector or chassis.
If you still got problems, try to post a photo here showing your wiring.
1. Are your input RCAs isolated from the chassis ?
2. Does the PSU Ground connections (Gnd Left/Right) have connection to the Signal Gnd in on the PCB ?
3. How is your PSU Gnd's connected to chassis and PSU ?
If not allready done this way, try isolate the RCA from the chassis. If there is a Gnd Left/Right to Input Gnd connection on the PCB, try setup one again.
If there are no connections between Gnd's on the PCBs, connect the RCA Gnd's together, and connect this RCA Gnd to a common GND point between the big filter caps (as in setup two, but RCA
Gnd's should connect to a common Gnd beyween the PSU caps). This common Gnd point (now a star Ground should the be connected to chassis as chassis Gnd).
Always make sure that you only have ONE Gnd connection to each circuit, connector or chassis.
If you still got problems, try to post a photo here showing your wiring.
Thanks everyone for the advice!
@mjf
I have tried connecting the two stargrounds together(with little improvement) but not to the chassis, I will give it a try.
@Jan Dupont
1. Yes
2. Yes, in fact the pcbs have a ground plane.
I don't think this is the problem though, because, as I said above, if I connect to the input RCA only one channel(that is the cable witch comes form the source, the internal wiring is not changed) there is no noise, problems occur only when i connect both channels. Also, with the source disconnected(that is both inputs floating), there is no noise.
So this makes me think that there is a difference between left and right ground and somehow hum appears into my signal.
3. No, because my mains is 2-prong.
Could this be the problem? I will try connecting starground to chassis and post the results. Speaking of this, I have seen in Self's book that he recommends connecting gnd to chassis in the vicinity of input RCA's, what are your thoughts on this?
I have also tried connecting a volume potentiometer, using both above setups. When in setup one, there is noise even at low levels. With setup two, at minimum there is no noise but it rises as I crank the volume up and at max there is the same noise as in the case when i had no volume pot. Maybe this info is useful.
@mjf
I have tried connecting the two stargrounds together(with little improvement) but not to the chassis, I will give it a try.
@Jan Dupont
1. Yes
2. Yes, in fact the pcbs have a ground plane.
I don't think this is the problem though, because, as I said above, if I connect to the input RCA only one channel(that is the cable witch comes form the source, the internal wiring is not changed) there is no noise, problems occur only when i connect both channels. Also, with the source disconnected(that is both inputs floating), there is no noise.
So this makes me think that there is a difference between left and right ground and somehow hum appears into my signal.
3. No, because my mains is 2-prong.
Could this be the problem? I will try connecting starground to chassis and post the results. Speaking of this, I have seen in Self's book that he recommends connecting gnd to chassis in the vicinity of input RCA's, what are your thoughts on this?
I have also tried connecting a volume potentiometer, using both above setups. When in setup one, there is noise even at low levels. With setup two, at minimum there is no noise but it rises as I crank the volume up and at max there is the same noise as in the case when i had no volume pot. Maybe this info is useful.
to avoid unwanted ground loops by 2-ch amps in same cabinet, the royal way is always to use dual mono power supplies (i. e. two identical independend versions include two transformers) so as outsourced power supplies (i. e. additional enclosures).
Now the right place between left and right GND is the input GND. At same places I connect the metal enclosure for power amps (now there are optimized screen function).
In all the other cases I can't use the mentioned royal way und unfortunately I must choise an other GND connection approach. Residual hum effects don't disappear in all cases
-
Now the right place between left and right GND is the input GND. At same places I connect the metal enclosure for power amps (now there are optimized screen function).
In all the other cases I can't use the mentioned royal way und unfortunately I must choise an other GND connection approach. Residual hum effects don't disappear in all cases
-
Hi kin
Jan is quite right.
Take one interconnect and disconnect the shield wire at one end.
This will break the ground loop. There should be no hum.
Do remember that the normal grounded interconnect must not be unplugged otherwise you will get a loud hum from the lifted interconnect.
Regards
Mike
I am very grateful to you for your suggestions.
I have tried the other suggestions, that is connecting the two star-grounds together and then to the chassis. I as suspected, connecting it to the chassis made no difference. In fact, i tried connecting the chassis gnd at input, as per D self, also no difference. What is certain is that when I connect the 2 stargrounds together i get considerably lower hum, however not neglijable.
Regarding the ground loop formed at the source, i am aware of it, but that happens to every unbalanced amplifier as far as I can tell. I will try disconnecting one shield and report here the results(however I think this solution is somehow a compromise, as the loop formed by the input current of that channel will be considerably greater).
@ tiefbassuebertr
I know the absolute best solution would be two separate power supplies, and i think that the problem(or at least part of it) is the psu. Maybe converting it by using only two rectifier bridges and having only one star-ground would reduce the hum.
I have tried the other suggestions, that is connecting the two star-grounds together and then to the chassis. I as suspected, connecting it to the chassis made no difference. In fact, i tried connecting the chassis gnd at input, as per D self, also no difference. What is certain is that when I connect the 2 stargrounds together i get considerably lower hum, however not neglijable.
Regarding the ground loop formed at the source, i am aware of it, but that happens to every unbalanced amplifier as far as I can tell. I will try disconnecting one shield and report here the results(however I think this solution is somehow a compromise, as the loop formed by the input current of that channel will be considerably greater).
@ tiefbassuebertr
I know the absolute best solution would be two separate power supplies, and i think that the problem(or at least part of it) is the psu. Maybe converting it by using only two rectifier bridges and having only one star-ground would reduce the hum.
Concerning ground loops (for your additional audio projects in the future) I have found by google's sub search engine
http://books.google.com/
this these extracts from several audio books
http://books.google.com/books?id=r86J_aNNReEC&dq=ground+loops+audio
http://books.google.com/books?id=d7ft6F8ZUdcC&pg=RA6-PA529&dq=ground+loops+audio
http://books.google.com/books?id=rgAMRAW8bZYC&pg=PA59&dq=ground+loops+audio
http://books.google.com/books?id=o7pAJBGWVb4C&pg=PA405&dq=ground+loops+audio
http://books.google.com/books?id=nn5PToRqfIAC&pg=PA10&dq=ground+loops+audio
a good (German language) book in this case is this one
http://www.buecher-nach-isbn.info/3...r-Audio-Schaltungstechnik.-3-7723-7322-4.html
perhaps also released in English and other languages
http://books.google.com/
this these extracts from several audio books
http://books.google.com/books?id=r86J_aNNReEC&dq=ground+loops+audio
http://books.google.com/books?id=d7ft6F8ZUdcC&pg=RA6-PA529&dq=ground+loops+audio
http://books.google.com/books?id=rgAMRAW8bZYC&pg=PA59&dq=ground+loops+audio
http://books.google.com/books?id=o7pAJBGWVb4C&pg=PA405&dq=ground+loops+audio
http://books.google.com/books?id=nn5PToRqfIAC&pg=PA10&dq=ground+loops+audio
a good (German language) book in this case is this one
http://www.buecher-nach-isbn.info/3...r-Audio-Schaltungstechnik.-3-7723-7322-4.html
perhaps also released in English and other languages
Jan has pointed out your ground loop. At the source you will have the L & R grounds tied. Your loop is from the source ground, through one channel, back through its power supply, to the other channel power supply then back through that channel back to the source.
Place a wire joining the ground lugs of the L & R RCA input connectors of your amp, this will break the loop. Don't bother with the breaking of the shield wire.
Trust me, I discovered this the hard way with a setup just like yours (and is very common occurance if you read the chip amp forums...)
Cheers,
Chris
Place a wire joining the ground lugs of the L & R RCA input connectors of your amp, this will break the loop. Don't bother with the breaking of the shield wire.
Trust me, I discovered this the hard way with a setup just like yours (and is very common occurance if you read the chip amp forums...)
Cheers,
Chris
I'm back with more conclusions. I tried disconnecting one ground shield and the channel with the shield still connected had absolutely no noise but the channel without the shield had hum, slightly louder than before. I have also tested with the volume pot, and the channel with shield still attached - no problem, the one without had an increasingly louder hum as I turned the volume up(at minimum there was no noise). Also I tried connecting the "disconnected" channel shield through a 22 ohm resistor and there was no difference. The tests were made with the two star-grounds both connected and disconnected. I always get louder hum with them disconnected.
@tiefbassuebertr
Thank you for the references, I already have most of those books, I will try to buy the rest of them.
@AndrewT
Well, obviously when the two star-grounds are connected there is a short between them. When trey are disconnected, there is an open circuit (in fact that was the whole point with the 4 bridges arrangement, to have 2 independent PSU's for the two channels - for crosstalk's sake- but i also get much louder hum).
@ chrish
I already tried that and there is absolutely no effect.(that was a little bit strange also because I expected something to change). I have also tried to short them while one shield was disconnected, again with no effect.
Could you please point to me some of those threads in chip amp section, maybe I can find there some more info.
@tiefbassuebertr
Thank you for the references, I already have most of those books, I will try to buy the rest of them.
@AndrewT
Well, obviously when the two star-grounds are connected there is a short between them. When trey are disconnected, there is an open circuit (in fact that was the whole point with the 4 bridges arrangement, to have 2 independent PSU's for the two channels - for crosstalk's sake- but i also get much louder hum).
@ chrish
I already tried that and there is absolutely no effect.(that was a little bit strange also because I expected something to change). I have also tried to short them while one shield was disconnected, again with no effect.
Could you please point to me some of those threads in chip amp section, maybe I can find there some more info.
if you have no Safety Earth (protective earth wire) in your house wiring then you do not have a need for a Safety Earth connection inside your amplifier.
Try to insulate every component from the outside world. eg. Plastic buttons and switches, Plastic encased Speaker terminals, shrouded input sockets (XLR with a plastic surround), etc.
Do not connect the two channels, keep them completely independent of each other. You end up with a pair of true monoblocks inside a common case.
Now check that the two RCAs read open circuit.
Try to insulate every component from the outside world. eg. Plastic buttons and switches, Plastic encased Speaker terminals, shrouded input sockets (XLR with a plastic surround), etc.
Do not connect the two channels, keep them completely independent of each other. You end up with a pair of true monoblocks inside a common case.
Now check that the two RCAs read open circuit.
Well, I have already made sure that everything inside is isolated form the chassis - safety is paramount.
As much as I would like to have the two channels completely separated, that is not possible since I am using a single toroidal transformer, as shown in the drawings. Anyways, if I keep the two star-grounds separated, I read open-circuit(well, not quite, actually I get 6-8 Megaohms but you get the idea) but the hum is horrible. Much better with them tied together(but still far from perfect considering that only one channel connected gives no hum at all). I really don't know what to try anymore. Btw, this is not the first layout I tried. Before that, I made a pcb with separate signal and power gnd(single layer) and I had even more hum(now i have double sided with gnd plane).
As much as I would like to have the two channels completely separated, that is not possible since I am using a single toroidal transformer, as shown in the drawings. Anyways, if I keep the two star-grounds separated, I read open-circuit(well, not quite, actually I get 6-8 Megaohms but you get the idea) but the hum is horrible. Much better with them tied together(but still far from perfect considering that only one channel connected gives no hum at all). I really don't know what to try anymore. Btw, this is not the first layout I tried. Before that, I made a pcb with separate signal and power gnd(single layer) and I had even more hum(now i have double sided with gnd plane).
Try to connect both amplifier channels to only one pair of rectifiers. Two pairs of rectifiers will not be matched good enough. They will create different potentials at the two star grounds. Equalizing currents will produce hum, when they flow through the only connection they have, the path through the RCA cable shields and source ground.
Yes, this is the only right thing without dual mono power supplies.pacificblue said:
Additional should be noted that also the filter capacitors for both channels are in use. GND from filter capacitors is at the same time the star GND for both channels.
In some cases there are additional resistor arround 10 ohms between the dirty GND (central star GND) and input GND on the voltage gain PCB.
But all this is an improvised solution and not the royal way. All commercial amplifier devices with such GND management (unfortunately the most) I don't like.
Thanks everyone for sticking with me!
@mjf
Yes I use a Roksan Kandy MK3 Cd-player. I have also tested it from a Nad cd-player, no difference. The interconnect cable is Luxmann JPX 2000 and I have tested with a viablue nfs-quattro again no difference. So this is not the problem.
@pacificblue
Indeed that's what I fear, that the culprit is the psu. It will be quite a fuss modifying it since I have to do a completely different pcb because I have 4 discrete ultra-fast bridges + soft-start and stand-by circuitry on the same board. But i'll certainly try it if I don't find any other solution.
@tiefbassuebertr
I must admit that I also was crazed with separation of dirty ground from signal gnd with 10 ohm rez, but that didn't work. And besides I'm sure that it's not the source of the problem because if only one channel is connected i get absolutely ZERO hum&buzz (tested on 96 db supravox speakers) only a very very faint "fsssss". The psu arrangement and interaction of gnd between channels seems guilty.
@mjf
Yes I use a Roksan Kandy MK3 Cd-player. I have also tested it from a Nad cd-player, no difference. The interconnect cable is Luxmann JPX 2000 and I have tested with a viablue nfs-quattro again no difference. So this is not the problem.
@pacificblue
Indeed that's what I fear, that the culprit is the psu. It will be quite a fuss modifying it since I have to do a completely different pcb because I have 4 discrete ultra-fast bridges + soft-start and stand-by circuitry on the same board. But i'll certainly try it if I don't find any other solution.
@tiefbassuebertr
I must admit that I also was crazed with separation of dirty ground from signal gnd with 10 ohm rez, but that didn't work. And besides I'm sure that it's not the source of the problem because if only one channel is connected i get absolutely ZERO hum&buzz (tested on 96 db supravox speakers) only a very very faint "fsssss". The psu arrangement and interaction of gnd between channels seems guilty.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.