I was thinking a pair of aluminum strips as heat bridges from the board to one flat back heat sink; that heat sink spanning across and shared by all 8 devices...
but the bodies of the devices will transmit bugger all of the heat, the heat is in the PCB, or in the center of the buffers, the epoxy body presents quite a bit more impedance, so while your idea will do something, it will be less effective than the already not so effective heatsink in the BOM. depends where you live i guess and whether you think heat will ever be an issue
Let's go crazy and use Peltier cooling! This would generate enough cooling on top to reach the back. Just put in a heatsink on top!
Geez... This looks like computer cooling... Maybe not a good idea!
Do
Geez... This looks like computer cooling... Maybe not a good idea!
Do
All of the smilies must be on holidays!
To put qusp's concern in perspective, we (I live 10km down the road from him) have days that are 35-40+degC in summer. Average daily temp on hottest days most summers is 32degC. A temp rise on the amplifier gets more significant with those ambient temps! 🙂
We do not have a Class A friendly environment.
To put qusp's concern in perspective, we (I live 10km down the road from him) have days that are 35-40+degC in summer. Average daily temp on hottest days most summers is 32degC. A temp rise on the amplifier gets more significant with those ambient temps! 🙂
We do not have a Class A friendly environment.
I think what Ed is suggesting (if I'm understanding you correctly Ed, correct me if I'm wrong) is that instead of machining down a thick based sink as HYPERTUNE is suggesting, you could instead obtain the same result by adding two aluminum bars of the correct height to an existing thin based heatsink to obtain the same general shape and result.
This would indeed work quite well, although as Ed correctly pointed out, it does add one more thermal interface between the added bar of aluminum and the heatsink proper.
Machining small simple bars might be easier than machining the entire heatsink, and since the actual half brick sink from DK is only about $4 it might save money as well.
I'm interested either way!
Cheers,
Owen
This would indeed work quite well, although as Ed correctly pointed out, it does add one more thermal interface between the added bar of aluminum and the heatsink proper.
Machining small simple bars might be easier than machining the entire heatsink, and since the actual half brick sink from DK is only about $4 it might save money as well.
I'm interested either way!
Cheers,
Owen
I'm able to get one rail to 15vdc. The 2nd rail is stuck at 1.258vdc no matter what adjustment I make. I've added the thermal barrier and verified that there's no short with the ground any longer. This is on the D21 side. The heatsink on that side is quite hot.
T.
T.
Can we see some updated pictures?
Have you measured voltage before and after regulator?
opc, I have the same understanding of Ed's proposal with the aluminium bars. I agree that it should work reasonably, though with extra interfaces it makes it one step more difficult to determine exactly how well it would work. Also, I'd be considering some way to secure the aluminium bar in place, I'm out of ideas on that at the moment. Perhaps the TIM would provide sufficient adhesive action to negate this concern and I'm just being paranoid.
Have you measured voltage before and after regulator?
opc, I have the same understanding of Ed's proposal with the aluminium bars. I agree that it should work reasonably, though with extra interfaces it makes it one step more difficult to determine exactly how well it would work. Also, I'd be considering some way to secure the aluminium bar in place, I'm out of ideas on that at the moment. Perhaps the TIM would provide sufficient adhesive action to negate this concern and I'm just being paranoid.
Last edited:
voltage before the regs is 22.29vdc on both sides. after one its 14.99vdc. the 2nd one is 1.258vdc. I tried replacing the pot but that didn't make a difference.
The pot i did replace though was stuck at 2kohms.
T.
The pot i did replace though was stuck at 2kohms.
T.
Last edited:
sounds like the regulator got killed... do you have a spare? Alternatively, you may have a short on the output side of the reg which is causing the reg to current limit giving you the voltage you're seeing and the hot heatsink.
If you measure output of the regualtor to GND with everything powered off, what do you get for resistance?
Without good pictures, it's going to be hard to tell which it is.
Regards,
Owen
If you measure output of the regualtor to GND with everything powered off, what do you get for resistance?
Without good pictures, it's going to be hard to tell which it is.
Regards,
Owen
0.2 ohms on the pin connected to R7.
600 ohms on the pin connected to C90
160 on the pinconnected to C96
600 ohms on the pin connected to C90
160 on the pinconnected to C96
Last edited:
I like the metal bar idea. Seems much simpler and less expensive than machining an entire heatsink. Stick some thermal paste in between the bars and the main heatsink and there shouldn't be too much loss in the interface.
BTW 32 degrees C, if my math is moderately correct, is about 90 degrees F? That's a pretty typical summer day in the Washington, DC area as well. Days over 100 degrees F are uncommon but will happen several times a summer. The solution is air conditioning 🙂
BTW 32 degrees C, if my math is moderately correct, is about 90 degrees F? That's a pretty typical summer day in the Washington, DC area as well. Days over 100 degrees F are uncommon but will happen several times a summer. The solution is air conditioning 🙂
well... it looks like your problem is on the pin connected to R7. That means:
- R5 might be shorted to the heatsink on the top of the board
- C87 was damaged and is shorting
- Both R7 and VR5 have failed short together (extremely unlikely)
- You have a solder bridge between any of those pads and GND.
Try removing C87 and see if that fixes the short. If not, reflow all those joints and inspect them very closely making sure there are no bridges.
Once you measure higher impedance there (something around the same as the other two pins, then you know you've found the short.
Regards,
Owen
- R5 might be shorted to the heatsink on the top of the board
- C87 was damaged and is shorting
- Both R7 and VR5 have failed short together (extremely unlikely)
- You have a solder bridge between any of those pads and GND.
Try removing C87 and see if that fixes the short. If not, reflow all those joints and inspect them very closely making sure there are no bridges.
Once you measure higher impedance there (something around the same as the other two pins, then you know you've found the short.
Regards,
Owen
it depends on what Ed was planning to bond it to, maybe i misunderstood, but i gathered he was talking about the bars contacting the buffer bodies, being that he said 2 bars and these have bugger all of the heat, so while the bars would do an ok job of getting rid of the heat they are in contact with; the way this would be set up, they wouldnt be in contact with all that much.
it'll probably be fine, its just that given the kind mates rates price HYPERTUNE is offering, i value my time more than that, let alone having a neater solution that can be housed any way you like.
it'll probably be fine, its just that given the kind mates rates price HYPERTUNE is offering, i value my time more than that, let alone having a neater solution that can be housed any way you like.
Looks like C87 was your problem then... hopefully you have a spare!
If not, you can actually run it without that cap for testing. The regulator will have worse ripple rejection, but it will work just fine otherwise. See if you can get 15V on the output now.
make sure you turn the pot all the way down first!
Regards,
Owen
If not, you can actually run it without that cap for testing. The regulator will have worse ripple rejection, but it will work just fine otherwise. See if you can get 15V on the output now.
make sure you turn the pot all the way down first!
Regards,
Owen
ha, you forget the last few years (well actually the end of last year was strangely mild, didnt get much above 35c, but its warming up now) it would not be unusual for there to be a couple of weeks in the 40s and one of those might contain up to 42-43c with 80% humidity!. aircon? nice for some, only in the lounge room. i live by the sea, so when the breeze is blowing its all good; but thats not always the case. then add the 6-8 channels of power amp is the plan (~about 700-800W total, about 40% class A) and a dac that puts out nearly 90W (thanks Owen!! well for 60W of that) and you start to get the picture
our worst summer days have death tolls.....
Last edited:
wow!! mate Owen these things are utterly indestructible!!
i had thought this amp might be a bit of a risky proposition with all those buffers in parallel like that and one going might mean the lot, but it seems that you would have to let off some sort of IED nearby to take them out!!
thats great to hear jazzm! how does she sound?
Ed: sounds like i misunderstood your solution, so the 2 bars are on the outside contacting the PCB not the sink and the sink contacts the buffer bodies? the former of course seems perfectly fine, the latter i'm not so sure about
i had thought this amp might be a bit of a risky proposition with all those buffers in parallel like that and one going might mean the lot, but it seems that you would have to let off some sort of IED nearby to take them out!!
thats great to hear jazzm! how does she sound?
Ed: sounds like i misunderstood your solution, so the 2 bars are on the outside contacting the PCB not the sink and the sink contacts the buffer bodies? the former of course seems perfectly fine, the latter i'm not so sure about
Last edited: