Can I make transistor heat sink out of aluminum foil?

Thread Starter

tjohnson

Joined Dec 23, 2014
611
I don't have any heat sinks currently, so I can't make circuits that require them, like this one:

I read that aluminum is a good conductor of heat, so I'm wondering if it would be possible to construct a functional heat sink for transistor Q2 out of aluminum foil?

Also, must Q2 be a high-power transistor, or would a low-power one (like a 2N2222 or 2N3904) work?
 

KL7AJ

Joined Nov 4, 2008
2,229
I don't have any heat sinks currently, so I can't make circuits that require them, like this one:

I read that aluminum is a good conductor of heat, so I'm wondering if it would be possible to construct a functional heat sink for transistor Q2 out of aluminum foil?

Also, must Q2 be a high-power transistor, or would a low-power one (like a 2N2222 or 2N3904) work?
Absolutely! You might want to stack a few layers together for some stiffness.
 

Thread Starter

tjohnson

Joined Dec 23, 2014
611
Absolutely! You might want to stack a few layers together for some stiffness.
Thanks! I suppose I'd need to prevent the foil from touching any of the transistor legs, or else I could end up with a short in my circuit, right?

Also, what is the answer to my second question regarding the transistor type?
 

MaxHeadRoom

Joined Jul 18, 2013
28,619
By foil you mean the rolls of Alcan cooking foil? It may be wiser to use the heavier gauge product used in foil pans etc that is less prone to creasing.
Also use some thermal compound.
Max.
 

Thread Starter

tjohnson

Joined Dec 23, 2014
611
By foil you mean the rolls of Alcan cooking foil? It may be wiser to use the heavier gauge product used in foil pans etc that is less prone to creasing.
Also use some thermal compound.
Max.
Do you mean that using a wad of cooking foil without thermal compound wouldn't be very effective? I assume by thermal compound you mean silicon gel or thermal grease.
 

Thread Starter

tjohnson

Joined Dec 23, 2014
611
Nobody (with good sense) designs in a high power transistor when a cheap one will work.
Then it sounds like if I want to improve the quality of my audio amplifier, it would be better to buy a cheap IC like the LM386 than to continue using transistors in my circuit.
 

WBahn

Joined Mar 31, 2012
29,979
I don't have any heat sinks currently, so I can't make circuits that require them, like this one:

I read that aluminum is a good conductor of heat, so I'm wondering if it would be possible to construct a functional heat sink for transistor Q2 out of aluminum foil?
The challenge will be getting a good thermal mating to the transistor case. Some cases styles will make this fairly easy and others will be very difficult. Many years ago I needed to make a heatsink for a TO-220 case and I cut a bunch (a dozen, I think) of 2" x 2" squares of aluminum foil and then mounted a TO-220 along the edge using a dead TO-220 as a backing plate to sandwich the layers. I didn't have any thermal compound, so I didn't us any. I bent the foil so that is spread out making a radial walls around the transistor and it worked very well.

Also, must Q2 be a high-power transistor, or would a low-power one (like a 2N2222 or 2N3904) work?
[/QUOTE]

Depends on how much power it is going to have to handle. For driving a speaker with any kind of volume, you will probably want a higher power transistor.
 

WBahn

Joined Mar 31, 2012
29,979
Then it sounds like if I want to improve the quality of my audio amplifier, it would be better to get an IC like the LM386 than to continue using transistors in my circuit.
I don't think that's what he is saying. I think he is saying that the designer of that circuit used a 3904 in the first stage but then didn't use one in the output stage. That probably was not an oversight because if a 3904 would have worked acceptably there, they probably would have used one there. That's not to say that they didn't just copy part or all of that circuit from someplace else and you could easily imagine copying one part of the schematic from one source and THAT person could have used any of a number of transistors, including a 3904, and just happen to choose the TIP31. Then this person snags the output circuit and uses it as is and is unaware that they could actually use a 3904 in both places. But if you look at the spec sheets for both, you will see that this is probably unlikely.
 

Thread Starter

tjohnson

Joined Dec 23, 2014
611
I don't think that's what he is saying. I think he is saying that the designer of that circuit used a 3904 in the first stage but then didn't use one in the output stage. That probably was not an oversight because if a 3904 would have worked acceptably there, they probably would have used one there. That's not to say that they didn't just copy part or all of that circuit from someplace else and you could easily imagine copying one part of the schematic from one source and THAT person could have used any of a number of transistors, including a 3904, and just happen to choose the TIP31. Then this person snags the output circuit and uses it as is and is unaware that they could actually use a 3904 in both places. But if you look at the spec sheets for both, you will see that this is probably unlikely.
Right, I understood #12 wasn't saying to use an IC. I just meant that based on what he said, I think I would rather buy an IC than a higher power transistor and thus avoid needing to improvise a heat sink.
 

WBahn

Joined Mar 31, 2012
29,979
As a really, really, really rough estimate, you have an 8Ω speaker and a 12V supply. Worst case power in the transistor will be when it is dropping half the voltage, 6V, which would result in about 0.75A through the speaker (this is assuming that the speaker can be treated as an 8Ω DC load -- for an AC load we pick up some leeway). So that's 4W of power and 750 mA.
 

WBahn

Joined Mar 31, 2012
29,979
Right, I understood #12 wasn't saying to use an IC. I just meant that based on what he said, I think I would rather buy an IC than a higher power transistor and thus avoid needing to improvise a heat sink.
What makes you think you won't need to heatsink the LM386 if you needed to heatsink the TIP31? It's a linear system, so you are going to have to vent that heat in either case.
 

#12

Joined Nov 30, 2010
18,224
Why so much resistance to heat sinks? Can't you walk outside and find a dead soda-pop can?
(Shortbus beat me to that one! :p )
Any old piece of metal will work (better than nothing) if it's flat.
 

Thread Starter

tjohnson

Joined Dec 23, 2014
611
What makes you think you won't need to heatsink the LM386 if you needed to heatsink the TIP31? It's a linear system, so you are going to have to vent that heat in either case.
Because I got this audio amplifier circuit to work without a heat sink:

so I assumed that if the circuit is designed well enough, it won't require a heat sink (and the circuit in an IC should be very well designed).

@WBahn: Your estimate in your previous post was very close, but my power supply is 9V. See this thread for more background details.

Why so much resistance to heat sinks? Can't you walk outside and find a dead soda-pop can?
(Shortbus beat me to that one! :p )
Any old piece of metal will work (better than nothing) if it's flat.
Because I didn't know they were so simple to make, and thought they might require thermal paste. Sorry for my ignorance.
 

#12

Joined Nov 30, 2010
18,224
Thermal paste always helps, but so does painting a car. Both will work better if they are used properly.
 

Thread Starter

tjohnson

Joined Dec 23, 2014
611
So would I need a heat sink with an LM386? As I mentioned in this post, sometimes just changing the value of a resistor and adding a capacitor can eliminate the need for a heat sink in a circuit.

Also, what are the main differences between an LM380 and LM386, other than that the LM380 appears to have a fixed 34 dB gain, whereas the gain of the LM386 can be adjusted from 20-200 dB?
 
Last edited:

Thread Starter

tjohnson

Joined Dec 23, 2014
611
I'm confused. Isn't this also a linear system:

It does not require a heat sink, although it did when the value of R3 was 1kΩ and there was no capacitor wired in series with the speaker.
 
Top