I've got an Alpine MRV-F540 that was losing one channel so I set it aside for a few years and just now pulled it out to try to figure it out and learn more about amps.
There are four IRF3205 FETS mounted in one corner that have been getting very hot, and one in particular is really nasty looking. The FET heat sinks are all isolated from the aluminum heat sink with a tape strip, which I believe is correct though I don't know why. Thermally it's obviously not working out too well though.
The automotive version of this FET appears to have the same specs as the standard except for a lower continuous current drain Tc (75A vs 110A). Would there be any advantage in going with the automotive version? It's $2.73 each versus $1.45, so not a big cash outlay either way.
Also is there any advantage to using some of the new thermal pastes that have come to market for the PC gaming crowd, or is that already being done?
Thanks!
There are four IRF3205 FETS mounted in one corner that have been getting very hot, and one in particular is really nasty looking. The FET heat sinks are all isolated from the aluminum heat sink with a tape strip, which I believe is correct though I don't know why. Thermally it's obviously not working out too well though.
The automotive version of this FET appears to have the same specs as the standard except for a lower continuous current drain Tc (75A vs 110A). Would there be any advantage in going with the automotive version? It's $2.73 each versus $1.45, so not a big cash outlay either way.
Also is there any advantage to using some of the new thermal pastes that have come to market for the PC gaming crowd, or is that already being done?
Thanks!
Power supply FETs rarely get hot enough to show signs of heat stress unless there is another problem.
I use the standard IRF3205. I've never bothered with the au version.
You won't see any benefits from using the expensive pastes. Dow 340 works well.
The tape or some other electrical insulator must be used. Apply heatsink compound when reassembling the amp.
I use the standard IRF3205. I've never bothered with the au version.
You won't see any benefits from using the expensive pastes. Dow 340 works well.
The tape or some other electrical insulator must be used. Apply heatsink compound when reassembling the amp.
Thanks for the quick reply Perry. This stuff is really interesting to me though I'm just starting out.
Here's a pic of the four MOSFETs, the entire area near them is pretty smoky, but the one looks much worse than the others:
Here's a link to that pic as well so you can get a closer view: https://goo.gl/photos/vJ4xBqbN2rqkoxk49
Here's a pic of the four MOSFETs, the entire area near them is pretty smoky, but the one looks much worse than the others:
An externally hosted image should be here but it was not working when we last tested it.
Here's a link to that pic as well so you can get a closer view: https://goo.gl/photos/vJ4xBqbN2rqkoxk49
They have failed. You need to replace all 4. The gate resistors and the driver IC may have failed as well. Failure of the output transistors is often the cause of a power supply failure.
Well if that's the case then I'm finally going to have to learn how to deal with SMD soldering as the resistors tied to the pins labeled "G" on each FET are surface mount:
https://goo.gl/photos/orfkWUWHxaCQtks76
But the good news is that this amp was working right up to when I took it out of the car, so I expect the damage can't be all that deep. Perhaps I should replace the FETs and the cute little resistors and then see how it works.
The dodgy tweeter was probably just dirt in the crossover pots or something simple like that. I suppose those will need a cleaning while this is apart as well.
So I'll need four of the IRF3205's, four of the 10 ohm surface mount resistors, a tube of Dow 340, and some thermal tape stuff? Are there any caps I should replace as a rule while I'm at it? People seem to do that alot.
An externally hosted image should be here but it was not working when we last tested it.
https://goo.gl/photos/orfkWUWHxaCQtks76
But the good news is that this amp was working right up to when I took it out of the car, so I expect the damage can't be all that deep. Perhaps I should replace the FETs and the cute little resistors and then see how it works.
The dodgy tweeter was probably just dirt in the crossover pots or something simple like that. I suppose those will need a cleaning while this is apart as well.
So I'll need four of the IRF3205's, four of the 10 ohm surface mount resistors, a tube of Dow 340, and some thermal tape stuff? Are there any caps I should replace as a rule while I'm at it? People seem to do that alot.
I took a look at the schematic, and figured out why the amp could continue running even with one of the FETs toasted so badly. They're installed in parallel, sharing load. If you damage one you lose some current capacity I suppose (or the other three have to suffer more), but unless the damaged FET shorts closed the rest of the system can keep on running.
I checked the crispy one and it's indicating an open circuit between pins, so this may be starting to make some sense now.
I checked the crispy one and it's indicating an open circuit between pins, so this may be starting to make some sense now.
Common problems for audio amp ppi 5440 for repair it turns on but I get no sound and it doesn't go into protection mode
This morning I desoldered the FETs and with them out of the way I could check the surface mounted gate resistors, and Perry you were correct in that the gate resistor on the dead FET was damaged and acting as an open circuit. All of the others measure out to 10.4 ohms so at least I only need to mess with one SMD resistor really.
What's a good way to clean up the old flux and smoke residue on the PC board? Maybe a light solvent of some sort? I've got quite an assortment: contact cleaner, brake cleaner, carb cleaner, laquer thinner, mineral spirits, you name it - I'm sure it's in my garage somewhere.
I checked all of the FETs per the BCAE1 website method and the three non-crispy ones check out good and consistently 0.464-0.465. One of the legs and a chunk of silicon fell off the crispy critter FET as I was desoldering it. Pretty sad bugger. I'll be replacing all four anyway of course, but it's just good to know that the other three were still in good shape even though they were carrying more load.
The tape Alpine used under these was pinkish-whiteish and 0.0065" thick, looking more like a band-aid than Kapton tape. It must be one of the Silicon pad materials. DigiKey sells a variety of pre-cut options for the TO-220 size so I'm going to order a few and see what looks best as they are pretty cheap.
What's a good way to clean up the old flux and smoke residue on the PC board? Maybe a light solvent of some sort? I've got quite an assortment: contact cleaner, brake cleaner, carb cleaner, laquer thinner, mineral spirits, you name it - I'm sure it's in my garage somewhere.
I checked all of the FETs per the BCAE1 website method and the three non-crispy ones check out good and consistently 0.464-0.465. One of the legs and a chunk of silicon fell off the crispy critter FET as I was desoldering it. Pretty sad bugger. I'll be replacing all four anyway of course, but it's just good to know that the other three were still in good shape even though they were carrying more load.
The tape Alpine used under these was pinkish-whiteish and 0.0065" thick, looking more like a band-aid than Kapton tape. It must be one of the Silicon pad materials. DigiKey sells a variety of pre-cut options for the TO-220 size so I'm going to order a few and see what looks best as they are pretty cheap.
Last edited:
I'd use acetone and cotton swabs to clean the soot from the board.
Mica insulators with heatsink compound are very good.
Mica insulators with heatsink compound are very good.
Thanks Perry. No Acetone in the garage, but luckily I have teenage daughters.
Is it wise to replace all four of the surface mount gate resistors to try to keep them as similar as possible, or is it OK to do just the one?
Is it wise to replace all four of the surface mount gate resistors to try to keep them as similar as possible, or is it OK to do just the one?
If three are within tolerance, I'd, at least, replace the one that's out of tolerance and the resistor for the adjacent FET that's directly connected to the burned resistor.
Well I got it all back together and working again. The SMD resistors went fine for my first time, although one did measure higher than the others a touch so it replaced it with another and now they all measure 10.2 ohms.
All of the channels work fine on the bench with a sine wave, and channels 1-3 measure +/- 3 mV or less DC offset, right on spec. But channel 4 is about 22 mV or so, and that's the channel that was giving me trouble when I had it in the car.
Where's a good place to start in looking for a problem like that?
All of the channels work fine on the bench with a sine wave, and channels 1-3 measure +/- 3 mV or less DC offset, right on spec. But channel 4 is about 22 mV or so, and that's the channel that was giving me trouble when I had it in the car.
Where's a good place to start in looking for a problem like that?
I forgot to mention that I ended up using the silicone pads under the FETs, they were pretty easy to work with.
The channel 4 signal going into the crossover board is showing only 0.2 mA of DC offset, but coming out of the crossover board it's 10.2 mA. The other channels all measure about the same, as high as 12.5 mA.
After the crossover board there's an electrolytic which I guess is there to knock out the DC again, but access to that is on the other side of the board so I haven't tested downstream of that cap yet.
After the crossover board there's an electrolytic which I guess is there to knock out the DC again, but access to that is on the other side of the board so I haven't tested downstream of that cap yet.
- Status
- Not open for further replies.