I have this Chinese copy of the Brazilian design that is powering up to the point where the Power and Clip LED are on.
When I got it, someone had fitted two IRF640N's in the blown side and made things worse. This uses the outputs marked as H9.
I have read through as many threads as come up in search for IRS20957S. This is very nearly identical to the Sundown SFB-5000.
So far what I have done is.
Replaced the blown side outputs with IRFB4127. The other side still has the original H9 installed.
Replaced two defaced 6 leg drivers with ZXGD3005E6TA on the blown side. The other side are still original pair.
Replaced four 15V Zener Diodes, four 4148 and four 13 Ohm gate resistors on the blown side. The other side is original set.
Replaced both IRS20957S. Replaced two 10 ohm resistors also.
Replaced three TL074's
Replaced LM319
Replaced one LMR62014XMFE/NOPB.
Power supply is operational. I have 15V out of the lower supplies now. I have clean drive to Pin 3 of both IRS20957's but Pin 2 is ~3.8V.
I have watched the you tube video. This amp had more issues likely from previous tech fitting in 640's.
I have at one point removed the 4427 and the two diodes. I saw what looked like starts of drive signals on each IRS20957 output pins. I did have over 5V on Pin 2 this way.
Seeing that, I put it all together to get result of both LEDs lit.
What I am not sure of is the choice of using the IRFB4127 or the choice of using the IRFB4127 only on the one side.
Aside from making both sides match, I have replaced everything I have found defective, what am I missing?
When I got it, someone had fitted two IRF640N's in the blown side and made things worse. This uses the outputs marked as H9.
I have read through as many threads as come up in search for IRS20957S. This is very nearly identical to the Sundown SFB-5000.
So far what I have done is.
Replaced the blown side outputs with IRFB4127. The other side still has the original H9 installed.
Replaced two defaced 6 leg drivers with ZXGD3005E6TA on the blown side. The other side are still original pair.
Replaced four 15V Zener Diodes, four 4148 and four 13 Ohm gate resistors on the blown side. The other side is original set.
Replaced both IRS20957S. Replaced two 10 ohm resistors also.
Replaced three TL074's
Replaced LM319
Replaced one LMR62014XMFE/NOPB.
Power supply is operational. I have 15V out of the lower supplies now. I have clean drive to Pin 3 of both IRS20957's but Pin 2 is ~3.8V.
I have watched the you tube video. This amp had more issues likely from previous tech fitting in 640's.
I have at one point removed the 4427 and the two diodes. I saw what looked like starts of drive signals on each IRS20957 output pins. I did have over 5V on Pin 2 this way.
Seeing that, I put it all together to get result of both LEDs lit.
What I am not sure of is the choice of using the IRFB4127 or the choice of using the IRFB4127 only on the one side.
Aside from making both sides match, I have replaced everything I have found defective, what am I missing?
If no one else tries to help, post at least one photo of the board.
The 4427? The one that drives the PS FETs?
Two diodes?
Have you tried lifting the two pin 2 terminals to see if both go up to a level where the amp will produce audio?
Thanks for posting in a way that made it easy to read.
The 4427? The one that drives the PS FETs?
Two diodes?
Have you tried lifting the two pin 2 terminals to see if both go up to a level where the amp will produce audio?
Thanks for posting in a way that made it easy to read.
Last edited:
Yes, the 4427 that drives PS FETs in order to keep it from making full rail voltage for testing of output drive and the two diodes that are in line with both Pin 2's of the 20957S's so they are not driven low by the PIC.
Everything is in circuit now.
I have not lifted the pins as I thought taking the diodes out of circuit was essentially the same thing.
Two things I forgot to mention, this is advertised to drive 1 Ohm loads and the rail caps are 160V.
For perspective of the picture, it is the top side outputs that was blown and the bottom side, with exception of the IRS20957S, is original parts.
Everything is in circuit now.
I have not lifted the pins as I thought taking the diodes out of circuit was essentially the same thing.
Two things I forgot to mention, this is advertised to drive 1 Ohm loads and the rail caps are 160V.
For perspective of the picture, it is the top side outputs that was blown and the bottom side, with exception of the IRS20957S, is original parts.
Here is a pic of in process clean up.
That is where the IRF640's were placed and the 15V Zener burned the board and pads.
Maybe it is possible there is a connection I missed in reconstructing that area.
That is where the IRF640's were placed and the 15V Zener burned the board and pads.
Maybe it is possible there is a connection I missed in reconstructing that area.
Is 5v enough on pin 2?
Aren't the two CSD terminals directly connected so one could affect the other?
On some of these amps, you can remove the supply resistor that feeds pin 6 of the 4427/33152 driver to shut it down.
Is 12v (without rail drive) enough for the amp to produce audio? Have you tried an external voltage source to get a bit more voltage on the rails? I don't know if this is necessary.
Aren't the two CSD terminals directly connected so one could affect the other?
On some of these amps, you can remove the supply resistor that feeds pin 6 of the 4427/33152 driver to shut it down.
Is 12v (without rail drive) enough for the amp to produce audio? Have you tried an external voltage source to get a bit more voltage on the rails? I don't know if this is necessary.
I am not really very familiar with these IRS20957S's, I thought I read somewhere in the forum that it needs to be at least 5V on pin 2 to start up. Maybe I misread that?
Yes, both Pin 2's are directly connected.
In fact, I forgot to mention it, but a 10 Ohm resistor connected to Pin 6 of the 4427 was out of tolerance and replaced.
12V was enough to produce output oscillation at the outputs in the video I watched. I was attempting to duplicate what I saw there.
I do not have output drive from the 20957's
Should I lift the CSD pins and power the amp up with 45V to 50V of rail? Thats about how much rail I get with around 10.5V or so on the B+ & GND
Yes, both Pin 2's are directly connected.
In fact, I forgot to mention it, but a 10 Ohm resistor connected to Pin 6 of the 4427 was out of tolerance and replaced.
12V was enough to produce output oscillation at the outputs in the video I watched. I was attempting to duplicate what I saw there.
I do not have output drive from the 20957's
Should I lift the CSD pins and power the amp up with 45V to 50V of rail? Thats about how much rail I get with around 10.5V or so on the B+ & GND
Try it without any additional rail first. These use the 62014 to produce supply voltage to the 20957 so rail may not matter.
Each of the ICs have internal protection that can affect the CSD so that's why I asked about lifting them.
According to the datasheet, I think they need ~60% of vcc. What do you have across pins 9 and 11? Approximately 13.5v?
Each of the ICs have internal protection that can affect the CSD so that's why I asked about lifting them.
According to the datasheet, I think they need ~60% of vcc. What do you have across pins 9 and 11? Approximately 13.5v?
The IRS 20957 needs to be 9 volts for the amp to start up without the clip light. If the output circuit is fine you will get 9 volts at start up on pin9. Make sure that the diode you replaced is put back in the correct orientation. This is how it should be put back. The 4148 anode side needs to go on the Gate pin of the fet. The zener diode Cathode is on Gate pin and the Anode is on the Source pin. If these diodes are not installed in the correct way then it causes the amp not to switch on. Also, check the rectifiers more so on the blown side of the outputs. I have seen alot of these amps with blown outputs with one rectifier blown.
I would suggest you put the amp in service mode and then check the output stage. When in service mode you do not have the high rail voltage and the amp will also be able to produce output. It is far much safer to work with 12v as rail voltage than with 150v. This way you also prevent damage to any components.
I would also suggest replacing all the output fets with irfb4115 even on the side that is not blown.
I did send you a PM explaining how to put this amp in service mode and some information on what to check.
Hope this information helps you
I would suggest you put the amp in service mode and then check the output stage. When in service mode you do not have the high rail voltage and the amp will also be able to produce output. It is far much safer to work with 12v as rail voltage than with 150v. This way you also prevent damage to any components.
I would also suggest replacing all the output fets with irfb4115 even on the side that is not blown.
I did send you a PM explaining how to put this amp in service mode and some information on what to check.
Hope this information helps you
Thank you for sharing this detailed information.
I have some things to check over including switching outputs to the IRFB4115.
I have some things to check over including switching outputs to the IRFB4115.
Are the irfb4115s definitive exact replacements for the original outputs?
One problem with repairing these amplifiers with defaced parts is the over-current protection circuits. They're programmed with resistors based on the RDSon of the output. If the RDSon of the replacement is not the same as the original, the over-current protection may engage too early or possibly too late.
One problem with repairing these amplifiers with defaced parts is the over-current protection circuits. They're programmed with resistors based on the RDSon of the output. If the RDSon of the replacement is not the same as the original, the over-current protection may engage too early or possibly too late.
An update for the thread.
With the amp in the condition as described above, I put the amp in diagnostic mode. It still powered up with both power and clip LED's lit which means there is still something to wrong to find.
At that point I removed the IRFB4127's I installed, and the other outputs marked H9. With those out of the circuit I probed various components to see what, if any measured different from another. I really didn't find anything glaring.
I swapped out the last two defaced buffers so that both sides are now the same and installed IRFB4115's. With the amp still in diagnostic mode, it powered up without the clip light. On Pin 2 of both 20957's there is 10V, and both have high & low side output.
Yes, this amp will produce audio at the speaker terminal at this low voltage.
I have been pulled away from this repair for a bit, I have to reassemble it and test at its normal voltages yet, but I'm optimistic that the fault has been found.
As to whether or not these outputs are the definitive replacement, I don't know. There are now quite a few threads discussing this design some die with little info while others suggest this output or that output. I tried to decipher best I could and selected the 4127's for this one.
I know in my area; this design of amp is flooding the streets and there are tons of blown ones around here to fix. I have declined these before but if it just comes down to gut it and replace everything I may start doing more.
With the amp in the condition as described above, I put the amp in diagnostic mode. It still powered up with both power and clip LED's lit which means there is still something to wrong to find.
At that point I removed the IRFB4127's I installed, and the other outputs marked H9. With those out of the circuit I probed various components to see what, if any measured different from another. I really didn't find anything glaring.
I swapped out the last two defaced buffers so that both sides are now the same and installed IRFB4115's. With the amp still in diagnostic mode, it powered up without the clip light. On Pin 2 of both 20957's there is 10V, and both have high & low side output.
Yes, this amp will produce audio at the speaker terminal at this low voltage.
I have been pulled away from this repair for a bit, I have to reassemble it and test at its normal voltages yet, but I'm optimistic that the fault has been found.
As to whether or not these outputs are the definitive replacement, I don't know. There are now quite a few threads discussing this design some die with little info while others suggest this output or that output. I tried to decipher best I could and selected the 4127's for this one.
I know in my area; this design of amp is flooding the streets and there are tons of blown ones around here to fix. I have declined these before but if it just comes down to gut it and replace everything I may start doing more.
If you're going to properly repair these with replacements that may not be the originals, you need to recalculate the over-current programming resistors for the high and low side and use the resistors appropriate for your output FETs.
After choosing and installing the resistors, you need to drive the amp to full power into the lowest rated load to confirm that it doesn't shut down prematurely. Then you need to test the amp by short-circuiting the output to ensure that the protection circuit is functioning properly. There's a risk that you will have to adjust the value if shorting the output causes the outputs to fail but you owe it to the owner to get it right. If there are others to repair, you can recoup the testing costs when repairing those.
I would have preferred you install the outputs half original, half new to compare the gate signals of each as well as the voltage across the gate resistor. It would also be good to compare the heating of the drive for the two halves of the amp.
Unknown outputs with this type of circuit makes a huge amount of work to get it right. Virtually no one does.
After choosing and installing the resistors, you need to drive the amp to full power into the lowest rated load to confirm that it doesn't shut down prematurely. Then you need to test the amp by short-circuiting the output to ensure that the protection circuit is functioning properly. There's a risk that you will have to adjust the value if shorting the output causes the outputs to fail but you owe it to the owner to get it right. If there are others to repair, you can recoup the testing costs when repairing those.
I would have preferred you install the outputs half original, half new to compare the gate signals of each as well as the voltage across the gate resistor. It would also be good to compare the heating of the drive for the two halves of the amp.
Unknown outputs with this type of circuit makes a huge amount of work to get it right. Virtually no one does.
The IRFB4115, 4321, and 4227 all work just fine in these amplifiers. There is no need to change any resistors for the protection circuit. When I did get these clone amps for the very first time I did compare them to the original SD amps and found that the protection circuit was the same. if one looks at the outputs of the SD amps(older versions) and these clone amps you will find that the only difference is the SD amps do not have the 4148 and they use a 1r resistor to the gate of the fets. On the clone amps, you will find that some use a 15r or 13r resistor on the gate and the 4148 diode. Also on the drive circuit itself, you will find that they use 2 different value resistors which SD calls them the RX2 and RX4.
'
'
If the FET has a different RDSon, the protection circuit resistors would be different. They may produce audio with any of those FETs but that doesn't mean that the protection circuit is properly calibrated to give proper protection. Even if the original resistors match other models, with all of the failures of the outputs, it could be that the manufacturer got it wrong, either by incompetence or by changing the FETs without consulting the circuit designer. Proper testing could produce a more reliable amplifier.