Hi, back with my second my_ref, but now it's an integrated version. Really like the amp, still amazed it played better than my tube...
Problem is there's a lot of hum from both channels in the integrated one, probably a ground loop I cant find.
-hum increases with volume turned up
-hum stays when inputs are disconnected
-hum stays when selector is set to 3 of the not used positions
-when touching transformers, i hear a 'pop' (softly) through the speakers, also while touching mass of cinch/input
-dc offset is 3mV on right channel, 6mV on left channel.
-music sounds already rather good, even through 7 dollar test speakers
Is the mass from cinch to IN (myref gnd) and from volume to IN (on myref gnd) correct? Took that connection from the audiocreative integrated chipamp. DIY LM3886 Chipamp
See attached photos for drawing of the spaghetti mix of red wires. Still need to clean that up...
Problem is there's a lot of hum from both channels in the integrated one, probably a ground loop I cant find.
-hum increases with volume turned up
-hum stays when inputs are disconnected
-hum stays when selector is set to 3 of the not used positions
-when touching transformers, i hear a 'pop' (softly) through the speakers, also while touching mass of cinch/input
-dc offset is 3mV on right channel, 6mV on left channel.
-music sounds already rather good, even through 7 dollar test speakers
Is the mass from cinch to IN (myref gnd) and from volume to IN (on myref gnd) correct? Took that connection from the audiocreative integrated chipamp. DIY LM3886 Chipamp
See attached photos for drawing of the spaghetti mix of red wires. Still need to clean that up...
Attachments
Hi, back with my second my_ref, but now it's an integrated version. Really like the amp, still amazed it played better than my tube...
Problem is there's a lot of hum from both channels in the integrated one, probably a ground loop I cant find.
-hum increases with volume turned up
-hum stays when inputs are disconnected
-hum stays when selector is set to 3 of the not used positions
-when touching transformers, i hear a 'pop' (softly) through the speakers, also while touching mass of cinch/input
-dc offset is 3mV on right channel, 6mV on left channel.
-music sounds already rather good, even through 7 dollar test speakers
Is the mass from cinch to IN (myref gnd) and from volume to IN (on myref gnd) correct? Took that connection from the audiocreative integrated chipamp. DIY LM3886 Chipamp
See attached photos for drawing of the spaghetti mix of red wires. Still need to clean that up...
Unshielded low level signal wires so close to unshielded 230Vac wires and input plug is not a good idea...
I don't know if that could be enough for the hum you hear, but i'll fix that anyway.
Then try using a proper ground lift (as described in the tutorial) and connecting PGNDs to ground lift chassis GND.
Thnx Luca, I'll try that when i have those components.
Decoupled volumepot, selector and cinch from PCB, hum is still present.
Inputs are isolated from case and grounded.
Removed all the wires, selector and volume: Just the boards and output. Stil humming. I will remove the boards from the case and report back.
Wait a minute.. it is my favourite case: You did not connect any way the board PGND references to the case. I see the PE connected to the case but the boards are floating..
This is the typical situation forced around here: all heavy attention to the safety rules - - and zero attention to signal integrity..
Sorry it's not directed to You but the general attitude..
Ciao, George
This is the typical situation forced around here: all heavy attention to the safety rules - - and zero attention to signal integrity..
Sorry it's not directed to You but the general attitude..
Ciao, George
That is, to be very clear: these amps are high speed solid state devices. Easily go to the MHz region in their usable bandwith. They have to be treated as such. They are not Toobzz!
For any kind of sensitive high speed electronics : shielding is an all important element.
The case cannot be floating antenna receiver closely coupled to the sensitive parts (like high impedance input pins of LM3886 <> heatsink)
The case must be connected with low ! Inductance, short thick conductor to the metal case / shield & heatsink.
PE earth connection is for safety. It is in clear collision with signal integrity requirements, and would need extra care to resolve properly.
Ciao, George
For any kind of sensitive high speed electronics : shielding is an all important element.
The case cannot be floating antenna receiver closely coupled to the sensitive parts (like high impedance input pins of LM3886 <> heatsink)
The case must be connected with low ! Inductance, short thick conductor to the metal case / shield & heatsink.
PE earth connection is for safety. It is in clear collision with signal integrity requirements, and would need extra care to resolve properly.
Ciao, George
Hi George, thanks for the tips. Just to be sure, do you advise the same as luca (loop breaker, heatsink connected on zero volt Line)? This breaks the floating design Dario proposes, but listening to that hum is quite a challenge.
The heatsinks seem to be the source of the hum. Just with one channel installed, without anymore extra wires the hum remains.
Hopefully the stuff for the loopbreaker will be here at friday.
Last edited:
I'm afraid there could be a misunderstanding here..
As far as I remember Dario always told: connect PGND to the case.
Now, seriously: even without loop braker - you could do a test. Just connect PGND with and then without PE connection. Without is only a temporary, transitorial test, not stable solution.
Ciao, George
As far as I remember Dario always told: connect PGND to the case.
Now, seriously: even without loop braker - you could do a test. Just connect PGND with and then without PE connection. Without is only a temporary, transitorial test, not stable solution.
Ciao, George
You have two ground paths, I would remove the ground wire that goes from My_Ref boards input GND to the ground bus on RCA inputs.
Also shielding (the missing of) and not twisted wires coudl have a (big) role.
Wait! Did you short boards inputs during this test?
Unless connected to signal ground via a source or a preamp the amp will hum, it is absolutely normal.
The amp should be silent when connected to a source/preamp or with shorted inputs.
If with shorted inputs the amp is not silent you can:
- double check heatsink grounding
- rotate transformers until hum disappear
Last edited:
?
Then were would Myref boards (and pot) connect to input GNDs?
Maybe i didn't understand well what you wrote...
Ahum.. PGND ... IS the Ground!
It cannot be connected if not to itself.
Joking.
I will try to reassume:
On the PCB there had been designed a ground separation scheme.
Power Ground (PGND is part of it) and Signal Ground.
R11, the 1ohm resistor unifies these two ground domains.
All signal cabling should be connected to the Signal Ground.
The input power parts, the output load should be connected to the Power /ground domain. But One does not have to really think about it, because all the connectors are correctly assigned already.
PGND is a connection point, where one should connect the external schielding, in our case: the metal casing.
The schielding must be equipotential, that is all parts of it must have good conductivity between each other. The Heat Sink should be such a part of the metal case.
So yes, the heatsink should have a direct path to the PGND point on the boards.
On the contrary, all signal cables, input connectors should only be connected directly only to the Signal Ground. Which is the Ground pin of the two pin input terminal.
Each of the stereo channels should be connected to the appropriate input gnd only.
Not connected between each other.
So the input cinch connectors must not be directly unified between them, on the back panel.
I hope this helps a bit.
It cannot be connected if not to itself.
Joking.
I will try to reassume:
On the PCB there had been designed a ground separation scheme.
Power Ground (PGND is part of it) and Signal Ground.
R11, the 1ohm resistor unifies these two ground domains.
All signal cabling should be connected to the Signal Ground.
The input power parts, the output load should be connected to the Power /ground domain. But One does not have to really think about it, because all the connectors are correctly assigned already.
PGND is a connection point, where one should connect the external schielding, in our case: the metal casing.
The schielding must be equipotential, that is all parts of it must have good conductivity between each other. The Heat Sink should be such a part of the metal case.
So yes, the heatsink should have a direct path to the PGND point on the boards.
On the contrary, all signal cables, input connectors should only be connected directly only to the Signal Ground. Which is the Ground pin of the two pin input terminal.
Each of the stereo channels should be connected to the appropriate input gnd only.
Not connected between each other.
So the input cinch connectors must not be directly unified between them, on the back panel.
I hope this helps a bit.
Thank you! It seems i lost my head somewhere. Didn't short the input the second time i tested, now everything is silent. Seems the hum was also the result of impropper grounding of the heatsinks, the did not make contact with PGnd.
