I also had a resistor first time when I powerd it up, but removed it afterwards. The sound is just the same, as well the lack of hum and noises. The only noise I have is when I power down the amp, due to the fact that the pot is very close to main power cord, switch and toroid.
Funny part: some days ago I moved my spekars around and they have hum while they're not connected to the amp
. It took me some time to realize that the passive filters where very close to a transformer.
Funny part: some days ago I moved my spekars around and they have hum while they're not connected to the amp
. It took me some time to realize that the passive filters where very close to a transformer.after upgrading my speakers i'm starting to hear a buzzing/humming from my gainclone as well.
i'd love to follow gainphile's tips (very well written btw!) however i'm using the lm3886 kit from chipamp.com, which iirc has signal ground, output ground, power ground and "chassis ground" all connected on the amp PCB.
i'm currently using the same size wire for all ground connections, and am wondering given that all the ground connections are already connected together on the amp PCB, would using different wire sizes be futile?
i'd love to follow gainphile's tips (very well written btw!) however i'm using the lm3886 kit from chipamp.com, which iirc has signal ground, output ground, power ground and "chassis ground" all connected on the amp PCB.
i'm currently using the same size wire for all ground connections, and am wondering given that all the ground connections are already connected together on the amp PCB, would using different wire sizes be futile?
For chipamp.com kit - bring two ground wires together (as thick as you can fit) from each of the PCBs - connect them to one side of the Elliot Sound ground breaker (10R||CAP||D1||D2) . Connect the other side of the ground breaker to the chassis ground point. Connect the mains ground to the same chassis ground point. You should not have any hum after this.
DECKY999 said:
thanks for the reply.
from which ground holes on the amp boards should i run the ground wire? at the moment I am running the ground cables from the "CHG" chassis ground holes to what i believe is a similar ground loop breaker (10R 5W resistor, 0.1uf cap, 35A 400V bridge rectifier) which then connects to the chassis/mains earth. i have verified that the zero volt line is raised above the chassis/earth by at least 10ohms.
so if i just thicken the cables coming from the CHG holes it should be fine? (at the moment i am using all cables of the same thickness) also am i correct in assuming that i need to use this thick cable all the way from CHG to the breaker?
btw here is a pic of my loop breaker:
An externally hosted image should be here but it was not working when we last tested it.
DECKY999 said:
ok i did this, upgraded the chassis ground wires to mains cable (largest size that i have that will fit in the PCB) and joined them as soon as possible to a thick subwoofer cable (about 4-5mm2) and then ran this to the loop breaker.
the noise is still there
is there an audio sample on the web somewhere of the humming? perhaps i'm tackling the wrong problem...
i managed to capture the sound of the humming/buzzing i'm experiencing, another piece of hi-fi equipment was malfunctioning which made the gc amplify the noise enough to record it. here's the sound that i hear:
http://temp.jw.id.au/gcbuzz.mp3
you can hear it at a distance of about 1m, and the volume control has absolutely no effect on it.
Hi,
that buzz is due to the high current charging pulses that flow from the secondary through the rectifier into the smoothing caps.
This is probably connected electrically due to a wiring error to the clean side of the amp supply or to the audio ground.
It may be EMI being radiated from the pulsing side and interfering with the signal side.
that buzz is due to the high current charging pulses that flow from the secondary through the rectifier into the smoothing caps.
This is probably connected electrically due to a wiring error to the clean side of the amp supply or to the audio ground.
It may be EMI being radiated from the pulsing side and interfering with the signal side.
AndrewT said:
that seems to correlate with another observation i made: when i turn the power off, the buzzing goes away immediately, even though music still plays for a few seconds. the power switch cuts off both the live and neutral connections, i.e. the whole 'dirty' side of the PSU, so i believe my problem is something power-related rather than ground-related.
here is a pic of my amp layout:
An externally hosted image should be here but it was not working when we last tested it.
with the exception of the ground loop breaker, i have followed the chipamp.com instructions exactly; so i do not believe there is a wiring error.
if it is EMI, how can i fix it without taking the PSU components out and putting them in a separate box?
also, would it help if i fitted snubbing capacitors to the rectifier diodes?
Hi,
try reducing the radiating area of the primary side and of the charging side.
twist the mains input pair together.
twist the twin secondaries together.
Twist the rectified outputs together.
Finally slacken the transformer fixing bolt and rotate the transformer in small steps and listen/measure, the effect of each new position, at the output.
try reducing the radiating area of the primary side and of the charging side.
twist the mains input pair together.
twist the twin secondaries together.
Twist the rectified outputs together.
Finally slacken the transformer fixing bolt and rotate the transformer in small steps and listen/measure, the effect of each new position, at the output.
AndrewT said:
just spent the whole evening trying to debug my buzzing problem, starting with your helpful suggestion.
1) tried twisting the primary and secondaries together: no difference.
2) wasn't convenient to twist the rectified outputs together, so i just unbolted the power supply board and held it as far away from everything else as possible: no difference.
3) tried zip-tying the rectified output wires, no difference. tried moving them around as far as they would go while the amp was running, no difference.
4) unbolted the transformer a bit and rotated it every which way, no difference.
5) unbolted the transformer and stood it up vertically outside the case: NOISE IS GONE! something doesn't like being too close to the transformer!
6) put the transformer back in, and moved the loop breaker away from the transformer all the way to the other end of the case: buzz is still there.
7) popped the IEC socket out and held it as far away from the case as possible: no difference (buzz still there.)
8) tried standing the transformer on 1cm (3/8") tall rubber feet inside the case: no difference.
9) stood the transformer vertically inside the case: buzz is still there. if i move it away from the heatsink, buzz gets quieter; if i move it towards the heatsink, buzz gets louder. moving it towards/away from the power supply board makes no difference to the buzz.
10) unbolted the heatsink and elevated it using (non-conductive) foam pads: buzz still there.
11) constructed a metal shield out of aluminum sheet and placed it between the transformer and the heatsink, ensuring that the shield was electrically connected to ground: buzz still there.
i think that's all i tried. the noise seems to be inversely proportional to the distance between the transformer and the amp boards / heatsink.
is there anything else that i can try??
Try different transformer - joking!
You said when you trip the switch the noise goes away but the music is there while the caps are discharging.
OK try puting larger metal disk on top of the transfomer. Twist the primary wires and the socondary wires, twist +V-GND and -V-GND (output from your PSU) and move them away from the transformer. Consider using a shielded twisted pair for the voltage output. Maybe also turn your PSU PCB for 90deg so the output side is facing back of the case rather then the heatsink. By the way - is the noise equal level on both L and R? Check if your transforemer fixing screw is making good contact to the chassis.
- joking!You said when you trip the switch the noise goes away but the music is there while the caps are discharging.
OK try puting larger metal disk on top of the transfomer. Twist the primary wires and the socondary wires, twist +V-GND and -V-GND (output from your PSU) and move them away from the transformer. Consider using a shielded twisted pair for the voltage output. Maybe also turn your PSU PCB for 90deg so the output side is facing back of the case rather then the heatsink. By the way - is the noise equal level on both L and R? Check if your transforemer fixing screw is making good contact to the chassis.
DECKY999 said:Try different transformer
actually not an unreasonable recommendation... been doing a lot of searching on transformer hum and apparently lower quality units hum more than higher quality ones.
DECKY999 said:
correct, hence i believe the problem lies somewhere before the power caps.
DECKY999 said:
when i had the wires twisted, i tried moving them around but it made no difference to the buzz. I tried this with the wires untwisted as well, no difference. currently the wires are untwisted and as physically far away from the transformer as possible, no difference.
i have tried turning the PSU PCB 90 degrees every which way, it made no difference.
the noise is equal and the same on both left and right channels. the power cables going the left channel's amp PCB are about 3x closer to the transformer than the right channel's ones, so given that the noise is of the same volume in both channels, and that the noise does not change at all when moving these cables around, leads me to believe that these cables are not the culprit and shielding/twisting will not help.
the transformer screw is making good contact, and i tested with the screw in place as well as not in place; no difference.
in my searching i found a schematic for a dc blocker (http://diyparadise.com/dablok.html) and i bought the parts today and intend to try it out tonight. i really hope it works, otherwise i might be faced with having to build a separate box for the power supply (and keeping that box far away from the amp!)
Lets get back to basics.
The LM packages with metal tabs have -ve DC on the smoothed side of the PSU.
Something on the heatsink is acting as an inductive coupler for the AC field from the transformer - a steel bolt or screw perhaps.?.. If so, this will induce an AC voltage in the heatsink, that will in turn modulate the smoothed -ve rail on the PSU. That produces HUM.
As you are trying to block an alternating magnetic field, I personally would choose a ferro magnetic material (thin sheet steel) to screen between the TX and the heatsink.
What is your earthing like - have you decoupled signal earth from PSU earth using the classic parallel cap (typically a few millifarads) and a resistor (typically 100 ohms IIRC).
Just a few thoughts.
Owen
The LM packages with metal tabs have -ve DC on the smoothed side of the PSU.
Something on the heatsink is acting as an inductive coupler for the AC field from the transformer - a steel bolt or screw perhaps.?.. If so, this will induce an AC voltage in the heatsink, that will in turn modulate the smoothed -ve rail on the PSU. That produces HUM.
As you are trying to block an alternating magnetic field, I personally would choose a ferro magnetic material (thin sheet steel) to screen between the TX and the heatsink.
What is your earthing like - have you decoupled signal earth from PSU earth using the classic parallel cap (typically a few millifarads) and a resistor (typically 100 ohms IIRC).
Just a few thoughts.
Owen
hi owen, thanks for your reply, more than happy to get back to basics!
the lm3886 chips i have are fully insulated (lm3886tf), i don't think they have metal tabs as such. don't have a pic of the back, but here's a pic of the front:
the heatsink is the following unit: (link) and has an extension coming out in the middle of the heatsink onto which i've mounted the chips. the heatsink is secured via 3 zinc-plated bolts to a couple of metal brackets which are bolted to the bottom of the case. i've tried unbolting the heatsink and sitting it on foam pads but that didn't reduce the noise at all.
i'm not sure what a ferro-magnetic material is, but i assume steel is and aluminum isn't? i've already tried aluminum and it didn't work at all, but i can probably find some steel scrap and give that a shot.
earthing is done using the ground loop breaker described on this page, the amp's signal/output/power ground all connect to the zero volt line which is only connected to the chassis/earth through this loop breaker. i don't think my problem is an earthing one however, as the noise is inversely proportional to the distance between the tx and heatsink, and the noise goes away as soon as power is cut (but audio still plays off the remaining charge in the power caps.)
it makes a lot of sense for the heatsink to be involved somehow; the tx is about 50% closer to the left channel's amp board than the right, but the noise is the same volume in both channels. *rummages around for random sheet steel*
the lm3886 chips i have are fully insulated (lm3886tf), i don't think they have metal tabs as such. don't have a pic of the back, but here's a pic of the front:
An externally hosted image should be here but it was not working when we last tested it.
the heatsink is the following unit: (link) and has an extension coming out in the middle of the heatsink onto which i've mounted the chips. the heatsink is secured via 3 zinc-plated bolts to a couple of metal brackets which are bolted to the bottom of the case. i've tried unbolting the heatsink and sitting it on foam pads but that didn't reduce the noise at all.
i'm not sure what a ferro-magnetic material is, but i assume steel is and aluminum isn't? i've already tried aluminum and it didn't work at all, but i can probably find some steel scrap and give that a shot.
earthing is done using the ground loop breaker described on this page, the amp's signal/output/power ground all connect to the zero volt line which is only connected to the chassis/earth through this loop breaker. i don't think my problem is an earthing one however, as the noise is inversely proportional to the distance between the tx and heatsink, and the noise goes away as soon as power is cut (but audio still plays off the remaining charge in the power caps.)
it makes a lot of sense for the heatsink to be involved somehow; the tx is about 50% closer to the left channel's amp board than the right, but the noise is the same volume in both channels. *rummages around for random sheet steel*
no,chylld said:
the audio ground should be separate from the zero volt line.
Then connect the zero volt line to the audio ground.
This avoids the charging currents flowing around the secondary/rectifier/smoothing caps circuit from contaminating the audio ground.
That is also the reason for twisting the charging circuit wires, to minimise the effect of these current pulses by reducing the radiating area between the flow and return wires.
Is there any orientation of the transformer that achieves both close proximity and lowest buzz level?
BTW,
the transformer leads should all be swapped to flexible cable/cord and not solid core to allow for this experimentation.
i would separate the audio ground if i could, however since the audio ground and power ground are connected together on the amp pcb (chipamp.com lm3886 kit) this is not an option.
i tried twisting the wires and it had no effect.
i tried rotating the tx around the y-axis (i.e. spinning it around on the spot) which had no effect. i also tried standing it up on its edge with no effect. of all the things that i tried, only the distance between the tx and the heatsink showed any correlation to the buzz.
i tried twisting the wires and it had no effect.
i tried rotating the tx around the y-axis (i.e. spinning it around on the spot) which had no effect. i also tried standing it up on its edge with no effect. of all the things that i tried, only the distance between the tx and the heatsink showed any correlation to the buzz.
ok it's almost 2am here, and i've spent the last few hours trying to debug the problem again. i figure it's silly working with mains power when you're tired and everyone else is asleep, so i stopped.
i found that the aluminium sheet i used as a shield before wasn't ferromagnetic, so i went and found some steel and constructed a wall around the transformer, covering >95% of the line of sight to the heatsink. it extended from the bottom to the top of the case, and i verified that a) the steel was ferromagnetic, and b) the steel was grounded to the case well. result: noise is still there.
so i tried twisting the rectified output wires since i realised i can do it in place, except for a small parallel non-twisted bit in the middle, but >90% is twisted. noise is still there. (leaving it twisted cos it looks pretty)
so then i tried the dc blocker. used 2 diodes to account for up to 1.4V of DC, soldered everything together, heatshrinked it all and connected it up: yep, you guessed it, noise still there.
so i figure the loop breaker is somehow making things go bad (i wasn't thinking clearly at the time) so i uninstalled that and tested it again: noise still there. the loop breaker actually works to eliminate computer noise i get through the mains ground, so i'm glad i don't have to throw that one away!
so now i'm back to square one. guys i really appreciate all the advice so far, and i'm honestly giving this my best shot. a solution to my transformer <-> heatsink hum cannot come any sooner!!!
i found that the aluminium sheet i used as a shield before wasn't ferromagnetic, so i went and found some steel and constructed a wall around the transformer, covering >95% of the line of sight to the heatsink. it extended from the bottom to the top of the case, and i verified that a) the steel was ferromagnetic, and b) the steel was grounded to the case well. result: noise is still there.
so i tried twisting the rectified output wires since i realised i can do it in place, except for a small parallel non-twisted bit in the middle, but >90% is twisted. noise is still there. (leaving it twisted cos it looks pretty)
so then i tried the dc blocker. used 2 diodes to account for up to 1.4V of DC, soldered everything together, heatshrinked it all and connected it up: yep, you guessed it, noise still there.
so i figure the loop breaker is somehow making things go bad (i wasn't thinking clearly at the time) so i uninstalled that and tested it again: noise still there. the loop breaker actually works to eliminate computer noise i get through the mains ground, so i'm glad i don't have to throw that one away!
so now i'm back to square one. guys i really appreciate all the advice so far, and i'm honestly giving this my best shot. a solution to my transformer <-> heatsink hum cannot come any sooner!!!
Could it be that the trick does not only lie in the distance between transformer and heatsink, but also in the fact that the transformer is removed from the amp case that looks very ferromagnetic.
Does it reduce the hum, when you lift the transformer up even without separating it further from the heatsink? I. e. put a sufficiently thick piece of wood or plastic below it, and so separate the transformer from the bottom and the sides? That would indicate that your case is the electromagnetic conductor.
Solutions could then be to move the transformer to a separate housing outside of the amplifier. Or to build a small shielding housing around the transformer, and mount that with as much separation as possible on all sides into your amplifier case.
Does it reduce the hum, when you lift the transformer up even without separating it further from the heatsink? I. e. put a sufficiently thick piece of wood or plastic below it, and so separate the transformer from the bottom and the sides? That would indicate that your case is the electromagnetic conductor.
Solutions could then be to move the transformer to a separate housing outside of the amplifier. Or to build a small shielding housing around the transformer, and mount that with as much separation as possible on all sides into your amplifier case.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.