Hi Folk,
I am much depress on my just finished class A 50W Power AMP. Because its front and mid stage mounting on the same PCB and checked no problem. However, connect it with output stage with 6 pairs of MOSFET (Hitachi J56 / K176) per channel then the AMP have small "giisss" noise (not from speaker but it happen it some component). I think it should be oscillation by it self.
Just want to know how to solve ?
CK
I am much depress on my just finished class A 50W Power AMP. Because its front and mid stage mounting on the same PCB and checked no problem. However, connect it with output stage with 6 pairs of MOSFET (Hitachi J56 / K176) per channel then the AMP have small "giisss" noise (not from speaker but it happen it some component). I think it should be oscillation by it self.
Just want to know how to solve ?
CK
I have a 100MHz scope and signal generator. How to find where or which thing incur the oscillation.
Just find some circuit such output stage FET drain-gate and gate-source parallel with a pico value of cap. What it the purpose of it ?
Just find some circuit such output stage FET drain-gate and gate-source parallel with a pico value of cap. What it the purpose of it ?
Hello...
Could be a cold solder joint making a noise. Maybe a schematic
and some idea of localization of noise(I have heard resistors ring).
go back to step one of the build and double check placement and polarity of parts and all solder joints. check the basics for DC offset and bias.Look at the output on a scope and trace it backwards .
keep us posted.
Regards, Elwood
Could be a cold solder joint making a noise. Maybe a schematic
and some idea of localization of noise(I have heard resistors ring).
go back to step one of the build and double check placement and polarity of parts and all solder joints. check the basics for DC offset and bias.Look at the output on a scope and trace it backwards .
keep us posted.
Regards, Elwood
output inductors "sing" as well. it's a magnetic effect when large currents flow through the inductor. if this is happening with no signal applied, it could be oscillation. output inductor "singing" is usually heard while driving a signal into a dummy load.
Leave one channel measuring the oscillating output, then use the other to poke around certain areas of the circuit where such an oscillation might originate. If the oscillation is very fast frequency, the second scope probe may change it in some way via the internal components and can help point you toward the source of the problem.
I have found sometimes too small gate resistors with/or no gate Zobel and poor layout can cause a mosfet amp to become a RF transmitter.
I have set fire to the 10 Ohm output Zobel resistor through a small series 100nf cap a couple of times.(was using cheap carbon resistor🙄 )
I have found sometimes too small gate resistors with/or no gate Zobel and poor layout can cause a mosfet amp to become a RF transmitter.
I have set fire to the 10 Ohm output Zobel resistor through a small series 100nf cap a couple of times.(was using cheap carbon resistor🙄 )
Do you have gate resistors on each mosfet? Value? NJormally it's the N-channel device that oscillation and also above a certain current level, easy seen with an oscilloscope. The oscillation is local, only the mosfet oscillates.
I do not use inductor at end of speaker output. It seems a resistor generate this sound. Most likely some of it is wire wound type.
I know Zobel is a type of filter. Not sure it will usually exist on the speaker output between + and ground. And Zobel is around 5 ohm in series with 0.1 u cap. But I am not sure why it exist and sometimes no need apply on some AMP.
I use 470 ohm in series with the gate of output FET. Suppose my AMP is self-oscillation then power on it again to trace the problem. Due to the fault finding can spend much time. Will it damage to the AMP ? How can I do ?
I know Zobel is a type of filter. Not sure it will usually exist on the speaker output between + and ground. And Zobel is around 5 ohm in series with 0.1 u cap. But I am not sure why it exist and sometimes no need apply on some AMP.
I use 470 ohm in series with the gate of output FET. Suppose my AMP is self-oscillation then power on it again to trace the problem. Due to the fault finding can spend much time. Will it damage to the AMP ? How can I do ?
Zobel is a filter that keeps RF from entering the feedback loop via the speaker wires and output node, assuming you take your feedback from the output. It is used, and sometimes with an output inductor || small resistor for stability. If your amp has a global feedback loop taken from the output you should at least have a Zobel filter at the output to help separate the noisy RF world from your little circuit.🙂
Seems I do not have this filter .
What the parallel value of this coil and resistor be used ? If I make it by myself how many turns of coated copper and its diameter be used?
Does it connect in series to the path of SPK (+) O/P ? Or in parallel ?
One more question about my AMP where the schematic shown that at the end of SPK output have (5 ohm in series with 0.1 u CAP) thing paralleled to SPK (+) and SPK (GND). What is their purpose ?
What the parallel value of this coil and resistor be used ? If I make it by myself how many turns of coated copper and its diameter be used?
Does it connect in series to the path of SPK (+) O/P ? Or in parallel ?
One more question about my AMP where the schematic shown that at the end of SPK output have (5 ohm in series with 0.1 u CAP) thing paralleled to SPK (+) and SPK (GND). What is their purpose ?
Report my test result...
Seems the main problem is the O/P FET incur the oscillation. I had tried a Zobel filter but no function to the problem. After that I just only parallel 30n cap to FET's drain-gate and 270p cap to gate-source. (I observed some Japanese' design have this stuff) Lastly the whole circuit turn quiet.
Anyway, thanks to all of your valuable comments 😛
CK
Seems the main problem is the O/P FET incur the oscillation. I had tried a Zobel filter but no function to the problem. After that I just only parallel 30n cap to FET's drain-gate and 270p cap to gate-source. (I observed some Japanese' design have this stuff) Lastly the whole circuit turn quiet.
Anyway, thanks to all of your valuable comments 😛
CK
- Status
- Not open for further replies.