Please, look at the data sheet of the power transistor, to know the Collector/emitter saturation voltage, and add 2-4V for regulation margin. Better, tune-it in real life as indicated.
Please, as it is really out of topic, do-it by yourself or ask somewhere else.
Oscillations
Hi,
This is on topic, I promise!
My 2SJ162/2SK1058 arrived in the mail today. They look pretty and sound good.
But I'm dealing with an oscillation problem. I don't have a scope and I use the AM radio trick. The amp is oscillating (broadcasting?) at the begining of the AM band. I ear a whistle. The cause seems to be a long input cable. I use a 5m cable to connect the audio output of the pc motherboard to the amplifier. I'm not using an input coupling cap. Raising the VAS compensation seems to have no beneficial effect. I went up to 68pF. Soldering 100nF polyester caps between rails and ground (hopping the high ESR of polyester would damp the oscillation) also has no effect. What seems to have some effect is raising the input filter. I replaced the 20pF cap by a 680pF and the whistle became weaker but still there. What seems to kill the oscillation is using the 680pF cap and shorting the rails cable (from reservoir caps to amp board) to a minimum. But I need longer rails wires. I can't have the capacitor bank on top of the amp board!
Can you please help me on killing the oscillation? Thank you.
Hi,
This is on topic, I promise!
My 2SJ162/2SK1058 arrived in the mail today. They look pretty and sound good.
But I'm dealing with an oscillation problem. I don't have a scope and I use the AM radio trick. The amp is oscillating (broadcasting?) at the begining of the AM band. I ear a whistle. The cause seems to be a long input cable. I use a 5m cable to connect the audio output of the pc motherboard to the amplifier. I'm not using an input coupling cap. Raising the VAS compensation seems to have no beneficial effect. I went up to 68pF. Soldering 100nF polyester caps between rails and ground (hopping the high ESR of polyester would damp the oscillation) also has no effect. What seems to have some effect is raising the input filter. I replaced the 20pF cap by a 680pF and the whistle became weaker but still there. What seems to kill the oscillation is using the 680pF cap and shorting the rails cable (from reservoir caps to amp board) to a minimum. But I need longer rails wires. I can't have the capacitor bank on top of the amp board!
Can you please help me on killing the oscillation? Thank you.
Try a 3.3k resistor in series with the 20pF cap. Some amps like to oscillate with inductive input, and this will keep you from regressing the amp by trying to stabilize it with the miller compensation when that is not really the problem. Of course if it's not an input oscillation problem, this may not be the correct solution.
I'm sorry Keantoken, it didn't work
I tried the 20pF in series with 3.3k and didn't make much of a diference.
I may have two issues at play: the lenght of the input cable (wich is really long) and the lenght of the power rails. If I disconnect the input cable the oscillation disappear. About the power rails: they are at bare minimum lenght. If I increase the lenght of the psu wires the oscillation gets worse.
I also tried a 1nF capacitor - which reduced the intensity of the whistle even further - but I don't want to lower the corner frequency of the input filter more.
Thanks any way for your help, keantoken.
Last edited:
The source is fine. It doesn't oscillate with others amplifiers. The frequency also changed but I wasn't paying much attention to the dial on the radio.
It's not a single whistle. It's several whistles at different frequencies. The stronger is at the beggining of the band.
Hi paulo. Check the followin.
Do vas trannies share a common heatsink? If yes then the metal must be grounded. Same goes if the FETs share a common hs or even use different hs per device. Is there a zobel at the output? Check if it's soldered correctly. Shorten the psu wire as much as possible. Are the onboard decouple caps as close as possible to the FETs. Are the FETs directly on the board? If not then 100ohm stoppers at both ends of gate wire. Are the caps parallel to the zeners okay? Is the input 1k+100pF RC filter correctly installed(first the 1k then the parallel 100pF and 10k to ground, right from the joined input bases)? Is your 5m signal cable well shielded, it may be acting as a highly sensitive antenna. Shorten the cable to half meter and see if it helps.
shaan
Do vas trannies share a common heatsink? If yes then the metal must be grounded. Same goes if the FETs share a common hs or even use different hs per device. Is there a zobel at the output? Check if it's soldered correctly. Shorten the psu wire as much as possible. Are the onboard decouple caps as close as possible to the FETs. Are the FETs directly on the board? If not then 100ohm stoppers at both ends of gate wire. Are the caps parallel to the zeners okay? Is the input 1k+100pF RC filter correctly installed(first the 1k then the parallel 100pF and 10k to ground, right from the joined input bases)? Is your 5m signal cable well shielded, it may be acting as a highly sensitive antenna. Shorten the cable to half meter and see if it helps.
shaan
Let me share this interesting phenomena that's common to the CFB amps I build.
Whenever I touch the input pin while wearing my slippers the amp comes alive and a lot of AM stations can be "tuned" just by varying my grip on the input pin. The signal is quiet clear and has better fidelity than commercial AM tuners. The show stops if I touch the floor or the input ground. Funny anomaly.
Whenever I touch the input pin while wearing my slippers the amp comes alive and a lot of AM stations can be "tuned" just by varying my grip on the input pin. The signal is quiet clear and has better fidelity than commercial AM tuners. The show stops if I touch the floor or the input ground. Funny anomaly.
good afternoon mr.shann, mr.LC.saya want to assemble a klass amp ssa that, I read the article from the power amp bonsai if no one power amp pack assembly results bonsai nx and sx type .... I've assemble and mr.nafiri ssa power amps ... melodious voice, pleasing to the ear and pain in telinga.apakah mr.shaan, lc is assembling a klass amp?
Seems that now the oscillations are under control.
I soldered another 100 ohms at the gate of the mosfets (they're wired, not soldered directely to the board for now) and twisted the PSU cables.
The input cable is of very bad quality. It's not shielded and has been bitten and chewed many times by my cat. It's been soldered that many times after the bites. My cat bites every cable he can find.
The filter capacitor is at 330pF now with the 1K paralleled resitor to ground. Why did you choose only 20pF?
I still ear some noises from the radio but now it's not a whistle, it's noise. When I connect the input cable the noise increases. Is this an oscillation too or it's normal?
Thank you again for your help guys.
I soldered another 100 ohms at the gate of the mosfets (they're wired, not soldered directely to the board for now) and twisted the PSU cables.
The input cable is of very bad quality. It's not shielded and has been bitten and chewed many times by my cat. It's been soldered that many times after the bites. My cat bites every cable he can find.
The filter capacitor is at 330pF now with the 1K paralleled resitor to ground. Why did you choose only 20pF?
I still ear some noises from the radio but now it's not a whistle, it's noise. When I connect the input cable the noise increases. Is this an oscillation too or it's normal?
Thank you again for your help guys.
This is why I use an RC snubber across the input. Ideally the input would be buffered so this wouldn't happen. Just like an amplifier may anomalously oscillate into specific loades, it may also oscillate with specific sources. Oscillation may cause DC offset or otherwise destroy the speakers.
This had been said to you previously.Sorry about that. What I meant is:
The 1K resitor is in series with the input. The 330pF cap goes to ground and it's paralleled with a 10K resistor. Now it's good
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.