Just let it idle on the bench for 5 or 10 minutes while monitoring the offset, then blow across the board. The offset will change. There isn't much you can do about it.
Just learn to live with it. Got it. Thank you Perry. I am glad I have passive crossovers in place.
I am back again. This crazy condensation is really killing my amp. I swiped my finger across the pcb board one day and it is wet. I was trying to avoid the damage but at some point I used it until it had a chance to fully dry out (it was ok initially but when I was taking turns it started to make usual popping noises until I have lost the mid bass in my left channel.
What I have found this time is:
R811 (R146 equivalent) - replaced
R819 (R111 equivalent) - replaced
Q811 (Q51 equivalent) - replaced. I ended up using smaller footprint part 863-BC856BWT1G SOT-323 instead of (SOT-23-3 BC856BLT1G) since I ran out of them. Previously smaller part was working ok until I replaced it for cosmetic reasons on channel 1 if I am not mistaken.
Q819 (Q52 equivalent BF422 transistor) - replaced
Q807 (Q63. equivalent BF423 transistor) - replaced. This is new failure comparing to my previous one. I did load BF422 by mistake, but ended up replacing it with BF423 once my order arrived.
When powered on the unit starts to smoke and blowing resistors and Q811 if I wait too long.
Doing diode check and Ohming out in the area I haven't found anything unusual. I am open to any ideas of what else to check for... Output transistors Ohmed out just fine (same as other working channels). This is a real mystery...
What I have found this time is:
R811 (R146 equivalent) - replaced
R819 (R111 equivalent) - replaced
Q811 (Q51 equivalent) - replaced. I ended up using smaller footprint part 863-BC856BWT1G SOT-323 instead of (SOT-23-3 BC856BLT1G) since I ran out of them. Previously smaller part was working ok until I replaced it for cosmetic reasons on channel 1 if I am not mistaken.
Q819 (Q52 equivalent BF422 transistor) - replaced
Q807 (Q63. equivalent BF423 transistor) - replaced. This is new failure comparing to my previous one. I did load BF422 by mistake, but ended up replacing it with BF423 once my order arrived.
When powered on the unit starts to smoke and blowing resistors and Q811 if I wait too long.
Doing diode check and Ohming out in the area I haven't found anything unusual. I am open to any ideas of what else to check for... Output transistors Ohmed out just fine (same as other working channels). This is a real mystery...
Last edited:
What value did you use for R146? What happens to the heating of Q811 if you increase the value of R146 or reduce the value of R148?
Hi Perry.
I have used the default 100 Ohm resistor for R811 (R146) equivalent. R819 (R111 equivalent) is also 100Ohm resistor. What values would you recommend to try for R146 and/or R148?
BF423 I ordered is Philips PNP 250V 500mA 60MHz TO-92 transistor.
I have used the default 100 Ohm resistor for R811 (R146) equivalent. R819 (R111 equivalent) is also 100Ohm resistor. What values would you recommend to try for R146 and/or R148?
BF423 I ordered is Philips PNP 250V 500mA 60MHz TO-92 transistor.
Try decreasing R148 to whatever the next value down is to see if it makes a difference. You can measure the voltage across R146 to see if that changes. Lower voltage across it should reduce the current through Q811. I don't know if this is going to help but it sounds like you've covered most all the bases.
Today I have found that Q810 had 2 broken legs likely from me bending it sideways to repair Q819. Swapped it with another BF422 (new) and powered the system back up. Same issue. R817, R819 burning with Q811 shorted.
I did short power on session and measured:
R817 20-22V
R819 20-22V
R811 12V.
On a good channel I have probed R124 (R819 equivalent) and it had 0.5V. Then I proceed to probe R122 (R817 equivalent) and it started to burn. The Q54 (Q811 equivalent) was shorted as well. I have induced the failure by probing R122 on a good channel... 🙁
Perhaps 863-BC856BWT1G SOT-323 are not suitable and I need to place the order for SOT-23 BC856BLT1G... Haven't had a chance to try other resistor nominals. Need to get to the office as I don't have other values of 0805 resistors.
I did short power on session and measured:
R817 20-22V
R819 20-22V
R811 12V.
On a good channel I have probed R124 (R819 equivalent) and it had 0.5V. Then I proceed to probe R122 (R817 equivalent) and it started to burn. The Q54 (Q811 equivalent) was shorted as well. I have induced the failure by probing R122 on a good channel... 🙁
Perhaps 863-BC856BWT1G SOT-323 are not suitable and I need to place the order for SOT-23 BC856BLT1G... Haven't had a chance to try other resistor nominals. Need to get to the office as I don't have other values of 0805 resistors.
Found another issue on the good channel (ch 2 front right). Q74 transistor was shorted (Q65 equivalent).
Powered on the unit and noticed 24-29V (potentially) higher on the output of the bad channel (ch 3, rear left).
Powered on the unit and noticed 24-29V (potentially) higher on the output of the bad channel (ch 3, rear left).
Perry, besides removing output mosfets, is there anything else I should check? I don't see any shorts on output transistors with the Ohm meter. It appears that output transistors are connected to +/- 27V rails. Could it be lifted ground in preamp stage causing the problem? I need to reconnect the amp back to check if the output voltage is tight to high or low rail.
Isn't this in only one channel?
Why are you getting condensation? Condensation isn't generally an issue because it requires that significantly warmer moist air pass over a cold amplifier. How is the amp getting that cold in an area where the air is warm?
Why are you getting condensation? Condensation isn't generally an issue because it requires that significantly warmer moist air pass over a cold amplifier. How is the amp getting that cold in an area where the air is warm?
Perry, yes this is only one channel. I have tested channel 2 that I introduced the failure and it had 0.18V dc offset after the repair. Both amps X1001 and X604 are mounted in the trunk of my car. It gets really cold at night and when I open the trunk I see moisture all over the place. Typically when I was hearing popping noises I knew there is moisture buildup. The alpine headunit I am using doesn't even have off state (ILX650). That is if I am driving on the freeway and the issue happens I cannot even disable the amp (need to install the switch for amp remote). I have the amps covered with sun visor so there are no water droplets from the metal frame that is above the amps. Previously water will get inside through vent holes (condensation). The amplifier case is wet and the boards are wet as well on some days. Initially I thought I had a water leak in the trunk, but the amp is wet even when its covered and so are other metal parts in the trunk (Pacific North West).
Just checked only channel 3 (rear left) has the output tight at -29V (rail voltage?).
Just checked only channel 3 (rear left) has the output tight at -29V (rail voltage?).
Last edited:
I couldn't find anything wrong with the amp probing around. Decided to remove the Q811 (it was cooked since the paint with the part number has vanished away) and saw 2.6V on the output of that channel which likely suggests the issue is still with the biasing circuit. Next I am going to order BC856BLT1G transistors and report back my findings...
PS: I was thinking the pinout on BF422 and BF423 I have purchased could have been wrong, but the diode mode check in circuit proves otherwise.
PS: I was thinking the pinout on BF422 and BF423 I have purchased could have been wrong, but the diode mode check in circuit proves otherwise.
I received the correct parts BC856BLT1G but it made no difference (as expected as the smaller parts work just fine in the other channel). Tried 150 Ohm resistor in R146 spot instead of 100 Ohm and the output is still at negative rail voltage (-29V).
Not sure what's going on with it, but likely going to do the following:
1) replace Q65 and Q64 (equivalents)
2) if 1) doesn't help replace Q39 and Q40 (equivalents)
Open to suggestions/ideas. Thanks.
Not sure what's going on with it, but likely going to do the following:
1) replace Q65 and Q64 (equivalents)
2) if 1) doesn't help replace Q39 and Q40 (equivalents)
Open to suggestions/ideas. Thanks.
If that doesn't work and if you REALLY want to repair the amp, you may be left with swapping parts from a working channel to the bad channel to see if any of the problems follow the moved parts. It's sort of a last resort.
I have done so far:
1) replaced Q819 (Q40 equivalent)
2) replaced Q803 (Q65 equivalent)
Checked with external power supply zener diodes as well as TL431. No trouble found.
Measured multiple points on the board. Seems like there are errors in the schematic (R107 is connected to -50V, not +50). There may be more, need to sit down and figure it out.
Attached is the diagram showing voltages on Good (in black) and Bad (in red) channels. I was powering amp for 1 second or less to take readings but Q811 (Q51) died. So I removed it and continued on. All measurements in red with 2 column in front of them were done with Q811 removed. I had negative probe attached to the heatsink (chassis). There is 275Ohms between actual GND and heatsink. Note that Left and right channels (front and left) are inverted with respect to each other. On left channels GND point is common to - output and on right channels GND point is common to + output. I was using channel 1 as my reference (front left) and the bad channel is the read left.
1) replaced Q819 (Q40 equivalent)
2) replaced Q803 (Q65 equivalent)
Checked with external power supply zener diodes as well as TL431. No trouble found.
Measured multiple points on the board. Seems like there are errors in the schematic (R107 is connected to -50V, not +50). There may be more, need to sit down and figure it out.
Attached is the diagram showing voltages on Good (in black) and Bad (in red) channels. I was powering amp for 1 second or less to take readings but Q811 (Q51) died. So I removed it and continued on. All measurements in red with 2 column in front of them were done with Q811 removed. I had negative probe attached to the heatsink (chassis). There is 275Ohms between actual GND and heatsink. Note that Left and right channels (front and left) are inverted with respect to each other. On left channels GND point is common to - output and on right channels GND point is common to + output. I was using channel 1 as my reference (front left) and the bad channel is the read left.
A few weeks ago I thought that the problem is likely different pinout on the replacement transistors but for some reason I have convinced myself it wasn't the case when I ran the diode test...
Now I am working on deriving the schematic (as things don't add up with the one you have shared) but I was going off Philips bf423 datasheet and it appears that the output bias resistors (22 Ohms) are connected to the base of Q40 which makes no sense. So I pulled up another BF423 schematic for OnSemi part and sure enough the pinout is different. I hope that is my problem, but need to investigate further... 🙂
I have replaced 2 BF423 transistors thus far:
1) Q807 (Q63)
2) Q819 (Q40)
Now I am working on deriving the schematic (as things don't add up with the one you have shared) but I was going off Philips bf423 datasheet and it appears that the output bias resistors (22 Ohms) are connected to the base of Q40 which makes no sense. So I pulled up another BF423 schematic for OnSemi part and sure enough the pinout is different. I hope that is my problem, but need to investigate further... 🙂
I have replaced 2 BF423 transistors thus far:
1) Q807 (Q63)
2) Q819 (Q40)
Last edited:
Do you have any TO-220 transistors (hard wired with ribbon cable to the board) that you could use in place of Q51 to get numbers with the transistor in the circuit in that location?
What about in anything larger than the TO-92 that you could clamp to a heatsink (piece of aluminum stock...)?