I would like to pick your brains about the diode heatsinking.
I'm building 4pr per side F5T monoblocks. I'm using the DIYstore boards and I'm toying with using these
on the diodes instead of mounting the diodes right on the main heatsink.
Would this method be better in preventing thermal runaway than having them fixed onto the main heatsink along with the mosfet? I'm planning to run the bias pretty high....
Note the dimension of the above are here http://www.digikey.com/product-detail/en/657-15ABP/345-1033-ND/340331
Any input would be appreciated.
I'm building 4pr per side F5T monoblocks. I'm using the DIYstore boards and I'm toying with using these
An externally hosted image should be here but it was not working when we last tested it.
on the diodes instead of mounting the diodes right on the main heatsink.
Would this method be better in preventing thermal runaway than having them fixed onto the main heatsink along with the mosfet? I'm planning to run the bias pretty high....
Note the dimension of the above are here http://www.digikey.com/product-detail/en/657-15ABP/345-1033-ND/340331
Any input would be appreciated.
Last edited:
Hey, I am planning the same thing. I prefer not to install the diodes onto the heat sink - which will get hot
. Regards,
Diodes conduct when the voltage across the resistor (parallel to it) increases. One way to reduce this phenomenon is to lower the resistance, thus reducing the voltage drop (for the same current), shifting the point where the diode conducts which means it gets less warm. This should be higher than the diode's own forward voltage drop specification, else it's pointless using them.
Diodes can usually take very high temperatures, I wouldn't worry. Those sinks look really, really small. Diodes when conducting will be putting out a bit of their own heat and could use the assistance of the main sinks, or at least, more sizeable ones than the ones in that picture. I wouldn't use those for anything more than a watt or three, maybe a voltage regulator doing light duty.
Diodes can usually take very high temperatures, I wouldn't worry. Those sinks look really, really small. Diodes when conducting will be putting out a bit of their own heat and could use the assistance of the main sinks, or at least, more sizeable ones than the ones in that picture. I wouldn't use those for anything more than a watt or three, maybe a voltage regulator doing light duty.
What bothers me is that the main sink is at relatively high temperature already (as compare to 20-25oC). By installing the diodes on the main sink, you are heating up the diode , not cooling it. I thought the purpose of cooling the diodes is to prevent any "thermally activation", causing more current to pass thru the MOSFET and eventually burn it.
Regards,
Yeah thats exactly what I was thinking. I am using 10x10 Heatsink USA heatsinks with 2 mosfets on them. I'm afraid that adding the diodes onto them will not "cool" the diodes per say. Although I do understand the perhaps the little ones that I show may not be enough? I'm hoping that at idle and when the amp is playing at regular volumes the diodes will not be conducting anyways.
I also thought (please correct me if I'm wrong ) that increasing the ambient diode temperature will also lower the point at which the diode will conduct?
Is this a wrong assumption?
Last edited:
No. You are right. Papa inlcuded a graph that descibes the current vs voltage drop relationship at different temperatures (for the Vishay diode) in the F5T article. For example, a voltage drop of 0.6 V at 25oC correspond to a current of 0.3 amp. At 150oC, same 0.6 v drop corresponds to a current of 5 amp. From the same graph, the diode starts to conduct (with a 0.1 amp current) at ~0.5 v at 25oC. At 150oC , it starts to conduct at just above 0.3 v.
By the way, your heat sink seems more than sufficient for 2 output deveices. Well, I guess it depends on how much bias you are going to use
. I am also planning to build the V3 in mono blocks, with 4 output devices on one 15" x 5" (2" high cooling fins) heatsink. Hope it is sufficient for relativley high bias.
Regards,
No. You are right. Papa inlcuded a graph that descibes the current vs voltage drop relationship at different temperatures (for the Vishay diode) in the F5T article. For example, a voltage drop of 0.6 V at 25oC correspond to a current of 0.3 amp. At 150oC, same 0.6 v drop corresponds to a current of 5 amp. From the same graph, the diode starts to conduct (with a 0.1 amp current) at ~0.5 v at 25oC. At 150oC , it starts to conduct at just above 0.3 v.
By the way, your heat sink seems more than sufficient for 2 output deveices. Well, I guess it depends on how much bias you are going to use
I am also planning to build the V3 in mono blocks, with 4 output devices on one 15" x 5" (2" high cooling fins) heatsink. Hope it is sufficient for relativley high bias.
Regards,
The heatsinks should be enough, however I am running 50V rails its going to be 4pr /channel so 4 of those heatsinks per channel and I plan to try biasing high... close to 6A draw.
Thats why I'm worried about the diodes.... I need them cold.
Wow!! I think you are the first guy (correct me If I am wrong, anyone) in this thread that plans to run such a high rail voltage and bias
"Go big or go Home!!"
Regards,
Why do you need them cold? I want mine to stay at a constant 62C, as designed for with the targeted bias current. While being on the main heat sink with the FETs, the diodes have very little power contribution, and musical transients should not be an issue. Constant overload protection should be addressed with other circuitry (e.g. like suggested in the F5T Article)
-- xLoff
I would like ask the question - why keep them warm at 62oC? Why not just let the diodes standing in free air and take on whatever the ambient temperature is. Once the amp has warm up, the ambient temperature (inside the amp) should be stable.
Sorry, I do not quite undestand what do you mean by constant overload, as related to the temperature of the diodes.
For thermal protection, one can install one of those thermal switch (off at a preset temp e.g. 65oC) on the heat sink (close to the MOSFET) and connect to the main power supply. It will cut off the power supply at ~ 65oC and switch back on when the heat sink temperature has dropped by about 15oC lower.
Regards,
+/-50V and 6A bias on 4 pairs? have you calculated this?
6A/4pairs=1.5A pr device.
50V*1.5A=75W dissipation pr device. thats 150W pr sink. that is 600W dissipation pr ch.
and NO, your sinks can NOT handle that. far from it.
to stay within 25C above ambient, you need sinks of:
25/150=0.15C/W or less.
and your outputs will fry after a short time anyhow.
also. 1.5A thru a 0.5ohm resistance gives you a voltage drop of:
1.5*0.5=0.75V voltage drop. and you have bias runaway long before this.
this is the recipe to kill the amp 5 times over.
Last edited:
audiosan,
Sorry you're right, its been almost a year since I did the calculations. And its probably somewhere on this thread, in my notes somewhere. I'm using 0.25Ohm resistance and I calculated somewhere at just under a 0.6V voltage drop, pushing out roughly 75W per heatsink.
However,
6A/4pairs=1.5A pr device.
50V*1.5A=75W dissipation pr device. thats 150W pr sink. that is 600W dissipation pr ch.
Are you sure about this? Where did you get 150W/sink? from your pair calculation, each pair dissapates 75W not 150W....
Sorry you're right, its been almost a year since I did the calculations. And its probably somewhere on this thread, in my notes somewhere. I'm using 0.25Ohm resistance and I calculated somewhere at just under a 0.6V voltage drop, pushing out roughly 75W per heatsink.
However,
6A/4pairs=1.5A pr device.
50V*1.5A=75W dissipation pr device. thats 150W pr sink. that is 600W dissipation pr ch.
Are you sure about this? Where did you get 150W/sink? from your pair calculation, each pair dissapates 75W not 150W....
There will be some amount of drift this way, as ambient inside a well-ventilated case cools faster than sink.
More than that, you would want better heatsinks. If you do want to separate from main sink you can, but ideally not with LM317 regulator sinks. Those are terribly marginal and I can tell you the instantaneous peaks through the diodes will get it very hot.
Stable but high temperature is much better than peaky behaviour of components.
Good Luck
OK perhaps then we have a misunderstanding.
6A total draw over 4pairs which means there are 8 devices.
Yes each pair of devices draw 1.5A. I don't see why 6A/4pair =1.5A per device. Shouldn't this be per pair? 6A/8device = 0.75A per device
Which by your calculation output 75W per pair of devices.
So each device only dissipates 37.5W. So the heatsink with only 2 devices would need to dissipate 75W.
So over 8 devices and 4 heatsinks there should be 300W that the heatsinks need to dissipate.
Does that make sense?
Sangram,
Ok from your point then, if I want to separately heatsink the diodes, what size would I need them to be to be suffcient and cooler than the main sinks?
Last edited:
@dazed2, This is an educated guess at best - but assume vf of 0.8V per diode with current of 10A peak from the MUR3020 datasheet, that's 8W per diode. I would say about 5C/W will be as low as one should go (per diode) given the highish ambient. A 3x2 well-finned sink with a 3-4mm base should be fine, or you could attach them all on to a aluminum bar and fix that to a larger sink. 4 diodes should work well with a 1 or 2C/W sink.