Questions about mounting of power transistor on a heat sink

Status
Not open for further replies.

Thread Starter

Peter Pan

Joined Mar 24, 2005
122
Hello All,

It's for first time now when I need to mount power IRF840 transistor on heat sink. I am not absolutely sure how to do it and have a couple of questions:

1. I was once advised that a heat sink with dimensions 300 x 100 x 26 mm (as shown on picture attached) would be good enough. Whether it's really good enough, in your opinion?

2. I am not sure how to mount IRF840 on a heat sink. IRF840 is produced in TO220AB package (also shown on a picture attached) and its drain contacts heat sink thermally and electrically as well (IRF840 is tighten to heat sink by bolt). So I will need to isolate whole heat sink from all other parts (case, for example) - is my understanding correct?

Thank you.
 

Attachments

millwood

Joined Dec 31, 1969
0
the right answer is always "it depends".

1) the heatsink is pretty sizeable, and I would say that it is good easily for 20-30w of heat at room temperature with good ventilation.

the geometry is not optimal in the sense that you get the best heatsinking by installing the fins vertically, to create a chimney effect.

it is hard to do that with a skinny heatsink like yours but that's not a problem if you aren't pushing the envelope here.

2) you don't have to insulate the transistors from the heatsink. insulation makes installing better, and avoids turning the heatsink into a giant antenna - important for high speed circuitry. However, insulation increases thermal resistance and worsens heat dissipation.

I typically insulate my transistors from their heatsinks.


insulation can be done via many methods. mica is for example a good and not so expensive material. I have also used film or scotch tapes with good success.

to220 transistors you will also need to isolate the metal tab. that can be done with nylon caps, or a short plastic sleeve made out of wires of the right size.
 

Thread Starter

Peter Pan

Joined Mar 24, 2005
122
the right answer is always "it depends".

the heatsink is pretty sizeable, and I would say that it is good easily for 20-30w of heat at room temperature with good ventilation.
Thank you for good advice!

I meant to use IRF840 in order to amplify very short (10-100 uS) RF pulses to 60W level and ... I did not mean to use ventilation at all - would this sink be good enough in the case without ventilation?
 

millwood

Joined Dec 31, 1969
0
it depends on how fast you can charge up and discharge the gate. 10us isn't that fast for this mosfet but you do need to have a good gate driver to do that.

as to how much heat it will dissipate (assuming you can charge up / down the gate fast), it depends on the current going through the channel when it is in conduction, and how deeply you can dirve it into saturation.
 

Thread Starter

Peter Pan

Joined Mar 24, 2005
122
to220 transistors you will also need to isolate the metal tab. that can be done with nylon caps, or a short plastic sleeve made out of wires of the right size.
Well, I am not quite sure how to get it (English is not my native tongue). Do you mean that drain of my TO220 has to be tighten to heat sink by something like nylon bolt? (when I want to isolate it electrically)
 

millwood

Joined Dec 31, 1969
0
you will need to fasten the metal tab to the heatsink with a screw. so you need to a) isolate the metal tab from the heatsink; and then you need to isolate the metal screw from metal tab.
 

Thread Starter

Peter Pan

Joined Mar 24, 2005
122
and then you need to isolate the metal screw from metal tab.
Thank you.
How exactly to isolate the metal screw from metal tab? (I had seen some power circuits recently and none of them had metal screw isolated from heat sink and metal tab - so I have no idea how to do it).
 

bertus

Joined Apr 5, 2008
22,276
Hello Vladimir,

Still busy with the RF power amplifier?
Mounting a TO220 transistor to a heatsink is not that difficult.
I have attached the parts needed.
There is a ring with a small standup rill.
This standup rill goes into the hole of the transistor.
The screww fits through the hole.
The plate (silicon in my case) goes between the transistor and the heatsink.
This plate is also available in mica (this is perhaps better for RF).
The screw can be fixed in a tapped hole in the heatsink or with a nut on the other side of the heatsink.







Greetings,
Bertus
 

Attachments

Thread Starter

Peter Pan

Joined Mar 24, 2005
122
Hello Vladimir,

Still busy with the RF power amplifier?
Hi Bertus,

Yes, I am back to that NMR spectrometer design!

I have got now an idea how to mount that.

What you think - whether that heat sink (described in the beginning of this thread) is good enough for the 60W amplifier?
 

bertus

Joined Apr 5, 2008
22,276
Hello Vladimir,

Do you know the heat capacity of the heatsink?
It should be given in °C / Watt.
When you multiply this number by the amount of watts you want to dissipate,
you will get the heatrise in degrees celcius of the heatsink.

Here is a page about heatsinks :
http://sound.westhost.com/heatsinks.htm

Greetings,
Bertus
 

Thread Starter

Peter Pan

Joined Mar 24, 2005
122
Hello Vladimir,

Do you know the heat capacity of the heatsink?
It should be given in °C / Watt.
When you multiply this number by the amount of watts you want to dissipate,
you will get the heatrise in degrees celcius of the heatsink.

Here is a page about heatsinks :
http://sound.westhost.com/heatsinks.htm

Greetings,
Bertus
This heat sink has heat capacity 5.3 degree/Watt/inch. Thanks to you and beenthere for the useful links.
 

millwood

Joined Dec 31, 1969
0
This heat sink has heat capacity 5.3 degree/Watt/inch. Thanks to you and beenthere for the useful links.
that's the thermal resistance of the heatsink, not its capacity.

you will then factor in other thermal resistance, like junction to case (for the transistor), and thermal resistance through the insulation, to calculation the temperature increase on the die for a given power dissipation figure.

how much heat the transistor dissipates depends on how efficient the amp is. how efficient the amp is depends on many other factors.
 

bertus

Joined Apr 5, 2008
22,276
Hello Vladimir,

The heatsink is 300 mm.
So the heatcapacity is (25.4 / 300) * 5.3 = 0.448 °C / Watt

At 60 Watts dissipation the heatsink will rise 60 * 0.448 = 27 °C above the surrounding temperature.
So when the surrounding temperature is 20 °C, the heatsink will become 47 °C.

Greetings,
Bertus
 

millwood

Joined Dec 31, 1969
0
the general rule of thumb is that a device can dissipate about 1/3 of its rated dissipation. for to220 devices, you want to keep the dissipation below 20 - 30w.
 

Thread Starter

Peter Pan

Joined Mar 24, 2005
122
Hello Vladimir,

The heatsink is 300 mm.
So the heatcapacity is (25.4 / 300) * 5.3 = 0.448 °C / Watt

At 60 Watts dissipation the heatsink will rise 60 * 0.448 = 27 °C above the surrounding temperature.
So when the surrounding temperature is 20 °C, the heatsink will become 47 °C.

Greetings,
Bertus
Thank you! 47 degrees is well within temperature limit allowed for IRF840. :)
 

millwood

Joined Dec 31, 1969
0
Thank you! 47 degrees is well within temperature limit allowed for IRF840. :)
it may or may not.

a big factor would be that 60w figure: what does it mean? if it is the output from the amp, then you have to factor in the amp's efficiency.

if it is power dissipated by the output devices, you will have to factor in various thermal resistance. the junction to case, with greased caseback, has a thermal resistance of 1.5c/w.

that brings the total thermal resistance to about 2c/w. to dissipated 60w over 20c ambient, you are talking about 20c+60w*2c/w=140c.

for a device rated 150c junction, you don't have much room for error.
 
Status
Not open for further replies.
Top