Thanks for your answer!
I just tried, and yes, same hum with only one amp connected. But I need to decrease bias to 150mA to turn off hum.
Like pos and neg also GND is connected by soldered 1,5mm wire directly, so I have no idea how the LME could have a GND issue.
One more hint... the signal is SE-connected, so inv is connected to GND onboard.
I just tried, and yes, same hum with only one amp connected. But I need to decrease bias to 150mA to turn off hum.
Like pos and neg also GND is connected by soldered 1,5mm wire directly, so I have no idea how the LME could have a GND issue.
One more hint... the signal is SE-connected, so inv is connected to GND onboard.
Is this the small power amp PCB with the National driver chip squeezed into the middle and not enough room to fit a decent cooler?
If so there is a mistake in the input stage traces.
Fairly easy to modify the traces and solve the grounding error.
A link to a schematic and pcb would be helpful. some photos too...
If it starts to hum immediately after a specific bias point it is very likely that the amp starts to oscillate hence draws more current from the power supply. HUM HUM
Try to increase the Miller cap!
BR,Toni
Last edited:
I bias at 360 mA and have the green PCB. A tiny hum/noise when connected to a preamp (SE) or some other source. Also with only one channnel connected. Without connection silence.I didn't try less bias yet. I don't see oscillations other then a short one when shutting off without speakers. Until now never played full power to see what happens.
I have the green one, but I can't see any version labelling on it, not even the name of the amp.
OK, so seem to be different PCB's.
I made another discovery... Biasing (and hum) increases until 320mA, then it decreases again, although I keep turning the potentiometer.
I checked the PCB's few times, no failure regarding the BOM. I have no idea, what else could be responsible for that. I could have made one failure, but not the same on 4 boards...
I made another discovery... Biasing (and hum) increases until 320mA, then it decreases again, although I keep turning the potentiometer.
I checked the PCB's few times, no failure regarding the BOM. I have no idea, what else could be responsible for that. I could have made one failure, but not the same on 4 boards...
Are you using the same power supply for the LME chip and the output Transistors? If you are using the same power supply for both are you using 6 wires or 5 wires to connect to the power supply? I am using +/- 75 regulated for THE LME and +/- 65 unregulated for the outputs.
Did you try a Small value resistor 5-10ohm for the 0R resistor?
Did you try temporarily disconnecting the earth ground?
Did you try a Small value resistor 5-10ohm for the 0R resistor?
Did you try temporarily disconnecting the earth ground?
No difference between connected LME- and Output- GND and unconnected by 0R, I've not tried another resistor between both GND's
Also tried separate ps for the LME and common ps.
The transistor part is connected using + and - and speaker GND (as I use a star ground).
The LME part is connected using + and - to common / separate ps, GND to the star ground.
Another question:
Which ps are you using as regulated pc for the LME???
Also tried separate ps for the LME and common ps.
The transistor part is connected using + and - and speaker GND (as I use a star ground).
The LME part is connected using + and - to common / separate ps, GND to the star ground.
Another question:
Which ps are you using as regulated pc for the LME???
Such a pity I can't edit anymore after some minutes, sorry for "spamming" this thread!!!
No difference between connected LME- and Output- GND and unconnected by 0R, I've not tried another resistor between both GND's
Also tried separate ps for the LME and common ps.
The transistor part is connected using + and - and speaker GND (as I use a star ground).
The LME part is connected using + and - to common / separate ps, GND to the star ground.
When increasing bias above 320mA hum is getting louder an turning back again. keeping turning my speaker protection disconnects, and when the circuit is open again, quiescent current jumps to 600mA and more. Speakers are normal 4R speakers, no issue with that.
Between 90-100mA my the output signal begins to get a 10MHz signal, which increases with increasing the bias, like described above. Coming to 300mA the signal is growing anymore, but it's getting closer to the negative site, while positive amplitude is also drops off.
Man it drives me crazy!
Another question:
Which ps are you using as regulated pc for the LME???
No difference between connected LME- and Output- GND and unconnected by 0R, I've not tried another resistor between both GND's
Also tried separate ps for the LME and common ps.
The transistor part is connected using + and - and speaker GND (as I use a star ground).
The LME part is connected using + and - to common / separate ps, GND to the star ground.
When increasing bias above 320mA hum is getting louder an turning back again. keeping turning my speaker protection disconnects, and when the circuit is open again, quiescent current jumps to 600mA and more. Speakers are normal 4R speakers, no issue with that.
Between 90-100mA my the output signal begins to get a 10MHz signal, which increases with increasing the bias, like described above. Coming to 300mA the signal is growing anymore, but it's getting closer to the negative site, while positive amplitude is also drops off.
Man it drives me crazy!
Another question:
Which ps are you using as regulated pc for the LME???
Last edited:
Second and last try to help: Increase the Miller cap and check the bias/hum problem again! Same results? Better?
I attach the sch that was issued for my green PCB.
and attach the sch dated 13/09/2011. This might be your white PCB.
There are component value changes and three more major changes.
R54 is removed.
The second Comp cap C80 is added.
The mute Zener is removed.
If your PCB has all these locations you may have the same PCB.
Attachments
Last edited:
Right, that's the sch of my amp:
https://drive.google.com/file/d/0B3MLwB9mdXl2c0dEbldWTk1DMlU/view
It's SE connected btw.
The Miller cap? Are you talking about the old sch?
https://drive.google.com/file/d/0B3MLwB9mdXl2c0dEbldWTk1DMlU/view
It's SE connected btw.
The Miller cap? Are you talking about the old sch?
In the old green PCB, R49 is left empty and C80 has a shorting link. Leaving only C79 as comp cap.
Your Google drive is the old green PCB.
Did the 2011 sch come out after the white PCBs were issued?
The error between the PCB and the sch was around R46, R46, R54 and the traces to the signal return.
Because the traces are built into the multilayer PCB, it took a little detective work to come up with what in the end was a very simple modification to correct the error.
I used SE but changed the resistor values to make them more conventional for that style of amp. I did not want a 5k7 input impedance that necessitates an enormous capacitor to get F-3dB down to 1Hz to 2Hz.
Your Google drive is the old green PCB.
Did the 2011 sch come out after the white PCBs were issued?
The error between the PCB and the sch was around R46, R46, R54 and the traces to the signal return.
Because the traces are built into the multilayer PCB, it took a little detective work to come up with what in the end was a very simple modification to correct the error.
I used SE but changed the resistor values to make them more conventional for that style of amp. I did not want a 5k7 input impedance that necessitates an enormous capacitor to get F-3dB down to 1Hz to 2Hz.
Last edited:
Well, that's the actual sch, which is available on OPC's all boards-site.
I wonder, if the error would still be existing in the newer white PCB's. But anyhow, did you have a similar behaviour before modifying your boards?
On the other hand I'm not the only one with the white boards, a present error should be known.
I wonder, if the error would still be existing in the newer white PCB's. But anyhow, did you have a similar behaviour before modifying your boards?
On the other hand I'm not the only one with the white boards, a present error should be known.
OPC never accepted that there was an error.
So I suuspect other buyers were not brave enough to argue.
I was not aware of this change in behaviour as bias was increased. But I can't remember what bias value I used. It is only a single pair and can't stand a lot of heat.
I used alfet ALF16N20W/ALF16P20W
I did add a Zobel between the two main decoupling capacitors. And added the 4 protection diodes that I recommend to others.
So I suuspect other buyers were not brave enough to argue.
I was not aware of this change in behaviour as bias was increased. But I can't remember what bias value I used. It is only a single pair and can't stand a lot of heat.
I used alfet ALF16N20W/ALF16P20W
I did add a Zobel between the two main decoupling capacitors. And added the 4 protection diodes that I recommend to others.
While we're here Andrew, I've evidently missed the whole part about the circuit and component value changes.
Can you explain here why they were changed and why the extra compensation parts were used after originally the decision to not use them was made with good confidence? Is that 10uF part of the compensation or related to bias adjustment?
And the gate resistors dropped and to the same value rather than slight different values?
Thanks.
Can you explain here why they were changed and why the extra compensation parts were used after originally the decision to not use them was made with good confidence? Is that 10uF part of the compensation or related to bias adjustment?
And the gate resistors dropped and to the same value rather than slight different values?
Thanks.