Hi guys,
putting together 6 amps (active concept, 4x LME49830 and 2x LME49610, both from OPC, all asymmetric) I realized, that I get some hum from the speakers, when I connect the amps to the RME UCX.
Every amp is deadly silent, when unconnected, but connecting more then two amps together (to the UCX) it begins to hum.
Grounding is realized with a GND-star, so that all components are strictly connected together at one place (including amp-out)
I tried several combinations with grounding the UCX, disconnecting the Cinch-gnd, connecting resistors and capacitors, but had no solution yet.
My first issue is: Is there a recommended combination of connecting all components to the GND-star? Now I have following:
Chip (LME‘s)
LME49610 amps speaker „gnd„
Speakers-out „gnd„
LME49830 amps speaker „gnd„
Upper Level
--------------------------------
Lower Level
Linear/ Regulated PS „gnd“
Chip (LME’s) PS „gnd“
The limitation is, that all power supplis are on the lower level, all amps are on the upper level of the enclosure.
My second issue is, that the 49830-amps are connected to the gnd-star with one wire, speaker-cables are also wired to the star. Does the amps have a problem, when the (speakers-) gnd-reference is connected to the star, and not directly to the amp?
Ground-cables are thick and soldered, should reflect in less resistance than Cinch-to-UCX in any case. Another grounding cable from the UCX directly to the gnd-star has no influence to hum.
Another hint:
While the 49610-amp‘s power supply consists of two positive regulators (for positive and negative rail), the symmetric power supply for the mosfets consists of a positive and a negative regulator. Could this also be a point to think about?
I hope anybody has an idea and could help 😉
Regards
Stammheim
putting together 6 amps (active concept, 4x LME49830 and 2x LME49610, both from OPC, all asymmetric) I realized, that I get some hum from the speakers, when I connect the amps to the RME UCX.
Every amp is deadly silent, when unconnected, but connecting more then two amps together (to the UCX) it begins to hum.
Grounding is realized with a GND-star, so that all components are strictly connected together at one place (including amp-out)
I tried several combinations with grounding the UCX, disconnecting the Cinch-gnd, connecting resistors and capacitors, but had no solution yet.
My first issue is: Is there a recommended combination of connecting all components to the GND-star? Now I have following:
Chip (LME‘s)
LME49610 amps speaker „gnd„
Speakers-out „gnd„
LME49830 amps speaker „gnd„
Upper Level
--------------------------------
Lower Level
Linear/ Regulated PS „gnd“
Chip (LME’s) PS „gnd“
The limitation is, that all power supplis are on the lower level, all amps are on the upper level of the enclosure.
My second issue is, that the 49830-amps are connected to the gnd-star with one wire, speaker-cables are also wired to the star. Does the amps have a problem, when the (speakers-) gnd-reference is connected to the star, and not directly to the amp?
Ground-cables are thick and soldered, should reflect in less resistance than Cinch-to-UCX in any case. Another grounding cable from the UCX directly to the gnd-star has no influence to hum.
Another hint:
While the 49610-amp‘s power supply consists of two positive regulators (for positive and negative rail), the symmetric power supply for the mosfets consists of a positive and a negative regulator. Could this also be a point to think about?
I hope anybody has an idea and could help 😉
Regards
Stammheim
Last edited:
Some photos and a system schematic showing amps, PSU and inputs will make it easier to help you. Your problem is called ground loops. There are a few things that you can do to reduce the hum.
Hi Mark,
yes, but I never had so many problems, which I couldn't fight after several weeks of trying different combinations. That's also why I built in a strictly ground star. I'll draw something in the evening.
Regards
Stammheim
yes, but I never had so many problems, which I couldn't fight after several weeks of trying different combinations. That's also why I built in a strictly ground star. I'll draw something in the evening.
Regards
Stammheim
Hi. This has the potential to be very informative and interesting. The clue to your problem could well be that you built it "strictly ground star".
Have you read this?... http://hifisonix.com/wordpress/wp-c...re-up-a-Power-Amplifier_Updated-Autosaved.pdf
Have you read this?... http://hifisonix.com/wordpress/wp-c...re-up-a-Power-Amplifier_Updated-Autosaved.pdf
IMHO, There should be two star points (connected together). One for signal level connections and another for the high current connections. The single star point at present may be being polluted by the high current speaker returns.
Paul
Paul
Audio GND and safety earth connected at more than one place? You might need a ground loop braker to fix this (see Rod Elliot ESP pages).
Gesendet von meinem D5803 mit Tapatalk
Here we go...
The ground loop breaker didn't help btw. Hum got really loud!
I would have to disconnect the gnd within the UCX if I wanted to break the loop :\
An externally hosted image should be here but it was not working when we last tested it.
The ground loop breaker didn't help btw. Hum got really loud!
I would have to disconnect the gnd within the UCX if I wanted to break the loop :\
Last edited:
Here another pic, but connecting and disconnecting PE doesn't make difference.
An externally hosted image should be here but it was not working when we last tested it.
In my humble opinion I think connecting the gnd of five different power supplies on a single point to speak of star wiring is not at all good for mass loops. Take the case of a dual mono and you will see that the gnds of the power supplies are not brought to a point. Another point, the only connection that should be connected to the box is the earth connection (no ground).
Begin feeding your amplifier stages as it should and forget about your center ground point.
Begin feeding your amplifier stages as it should and forget about your center ground point.
You should be thinking about something like this. Where possible keep the grounds per PSU separate. Use a ground loop breaker for each PSU.
The more details you can give the better. What kind of cables are you using? how are the channels connected to the speaker protection? What are you using as a source?
The more details you can give the better. What kind of cables are you using? how are the channels connected to the speaker protection? What are you using as a source?
Attachments
Disconnecting PE does make a difference, because PE can save lives. A bit of hum will be irrelevant if you or someone else gets electrocuted. But I guess that was not your point.
It's weird that inserting ground loop breakers results in more hum, while disconnecting PE does not change the hum. From your drawing I'd suspect there is at least one loop involving the audio GND + PE, but I do not completely understand your drawing with respect to the GND and PE connections.
Is PE connected to audio GND in any way? Use a Ohm meter to test.
Do the speaker protection modules have a common GND for the different audio channels? This might cause ground loops. Check by removing the protection modules. The protection modules may also have their own power supplies, and have audio GND(s) connected to PE (yet another loop).
Your Cinch connections from UCX to the amps probably have audio GND connections. Add these to your drawing an you'll see ground loops. Do you really need the combined star ground for the UCX and your amps?
Photos of the whole setup would be useful.
Guys, many thanks for your thoughts, I'll try separating the GND, unfortunatelly not before sunday.
PE and GND is NOT connected, multimeter shows infinite resistance. The trafos are connected to PE with their screen winding.
The 6-CH speaker protection module has a separate 12V ps. Both, the ps and the protection module are connected to the gnd-star.
PE is mandatory, but of course I meant it has no influence to the hum.
Yes, objectively speaking the UCX (source) is making a ground loop, but is an interface not always causing a loop? Which connection is to be disconnected? I tried disconnecting the Cinch-gnd on both sides, while using R's, C's, diodes, connecting to the star-gnd etc...
The Cinch-cable is a HQ Cinch cable, while power and gnd-cables are massive connected 4mm cables.
PE and GND is NOT connected, multimeter shows infinite resistance. The trafos are connected to PE with their screen winding.
The 6-CH speaker protection module has a separate 12V ps. Both, the ps and the protection module are connected to the gnd-star.
PE is mandatory, but of course I meant it has no influence to the hum.
Yes, objectively speaking the UCX (source) is making a ground loop, but is an interface not always causing a loop? Which connection is to be disconnected? I tried disconnecting the Cinch-gnd on both sides, while using R's, C's, diodes, connecting to the star-gnd etc...
The Cinch-cable is a HQ Cinch cable, while power and gnd-cables are massive connected 4mm cables.
In my 4 channel set up with 2 stereo speaker protection circuits I had to use two separate 12V windings or I had hum. They are not connected to audio ground, the speaker pair pass through the protection circuit only
Multi-channel amplifiers are massively more difficult to get hum free than monoblocks.
The more you increase channels the more chance you have of creating interference gathering aerial loops. It's those multiple aerial loops that generate unwanted hum and other interferences.
Even two channel power amps are difficult to get right. Have you read D.Joffe and HBRR/HBRL?
The more you increase channels the more chance you have of creating interference gathering aerial loops. It's those multiple aerial loops that generate unwanted hum and other interferences.
Even two channel power amps are difficult to get right. Have you read D.Joffe and HBRR/HBRL?
This may help understand your grounding problems. See post
"http://www.diyaudio.com/forums/construction-tips/251996-ground-loops-8.html" #73
Duke🙂
"http://www.diyaudio.com/forums/construction-tips/251996-ground-loops-8.html" #73
Duke🙂
That is not the way to wire it up.
TonyT's diagram makes the classic mistake of using the charging circuit as the voltage reference. The pulses across that charging link corrupt every connection taken to that location.
After reading this, I took it as a slightly ambiguous way of saying "this is not how to do it", at least I hope the OP took it this way and didn't get confused and waste his time
- Status
- Not open for further replies.