Heatsink on OPAMPs

[M23Bomber]

Hello All 

Did anyone notice a diference in the pedal/tone stability if a heat sink is glued to the DIP8 OPAMP such as the JRC4558 in the tube screammer?

Regards,
M.

[slacker]

In something like a tube screamer the opamp shouldn't be getting hot, so I can't see how adding a heat sink could have any effect.

[R.G.]

I've always been fond of painting on racing stripes and supergluing on tailfins.

[bluebunny]

LOL!   

Come on, R.G. - you just know that with a beautifully facetious remark like that, someone's just gonna have to do exactly that and post the results in the Pictures thread!  I, for one, can't wait.   

Actually, I feel the need to rustle up a screamer or similar just to try it myself...   

[boogietone]

The Cadillac pink ones will soon be the new, new old stock. 

[tubegeek]

EVERYONE knows that you have to use epoxy on the tailfins. Superglue is for the chin spoiler.

[M23Bomber]

Hello All,

At work, I found a few 8,5 mm x8,5 mm x 5mm heat sinks, I normally use them for other things, but since I like to  run most of my pedals at at different voltages, from 6v to 18v normally, because its an easy way of  getting different tones ( If someone wants to try this please remember to calculate the wattage of some of the resistors in the power supply especially, if you dont have resistors with the right wattage bad things might happen) I noticed that some opamps produce more heat than others. Some of them such as the TL082 gets really hot.To extend their useful life and preserve their tonal characteristics, I started adding heat sinks, and even on the oscilloscope  I see improvements when running them in different configurations than the standard ones.

NOTE FOR NON ENGINERS aka DIYers: Resistors have thermal noise and ,OPAMPS have lots of resistors inside, they also have TRANSISTORS, this means "Dynamic resistors". And overall heat is a awful thing in any electronic device.

Regards,
M.

[nick d]

          I've used all sorts of OpAmps  at +/- 12V in synth circuits , never known one even get lukewarm .
           
           Perhaps I've just been lucky .

[M23Bomber]

Hello all 

Let me define hot in some opamps as being a close tepid temperature when playing for a long time , not boiling hot or  200 celsius, I  just noticed that when heat sinks are attached some Opamps seem to be more accurate after large periods of time at higher than normal voltages, 9v, due to constant temperature transfer. Its just a detail of almost unnoticeable proportion but its there.

Regards,
M.

[R.G.]

I do in no way doubt that you are seeing (on the scope) and hearing something.

However, the wildly spurious claims that are rampant in the hifi market and becoming entrenched in the pedals market immediately raise the BS sensors of a lot of technically-inclined folks.

Temperature drifts and cross-chip temperature gradients are well-known issues in linear circuit chip layout circles (these are small circles, as you're no doubt aware). That being said, there are some questions to be answered.

1. You give advice to "NON ENGINEERS".  Are you an engineer, if so what kind, and if EE, what speciality within EE? Just so we can gauge your background, you know.
2. Some opamps do indeed produce more heat than others. The tradeoffs of bias current versus power supply voltage versus thermal drift versus output current capability is a field that's been being plowed since the days of the 709 opamp.

Fortunately, this doesn't need to be mysterious. The data sheets tend to list idle current per section for opamps, and as we all know, power equals current times voltage. So you can triple the internal dissipation of an opamp from its standby current by simply tripling the supply voltage from 6V to 18V - which may not make any difference at all. It depends on how big the internal currents are, and the internal drifts depend on how symmetrically the chip is laid out and how the thermal gradients across the chip affect the (in most cases) elaborate precautions that the chip layout guys went through to make sure that items with thermal drift are (a) compensated for and (b) affected equally by chip dissipation.

The biggest single dissipation item in most "jellybean" opamps these days is the load current. The output stage of an opamp can be biased, like a power amp, in A, AB, or B. The standby current reflects this. But if you hang a low impedance load on the output pin, the output currents can skyrocket, and like a discrete power amp, the internal dissipation can go up to about 40% of the delivered power for simple, non-malicious loads and signals, or equal to the output power for some ugly examples.  In this case, the chip dissipation skyrockets and all that stuff about minimizing gradients and compensating thermal offsets really starts to matter.

And there's a difference between internal heat and external heat. Heat sinks make equalizing the internal chip temperature and external temperatures easier. If there's significant internal chip dissipation, heat sinks help the heat escape from the outer surface after its tunneled through the insulating epoxy body.  Ooops. In most cases, the thermal resistance from chip to IC pin is lower than conducting across the epoxy body to a heat sink. But it depends on the epoxy and mounting technique - whether it's lead frame or bonding wires, etc. If there's not much internal dissipation, a heat sink helps the external air temps heat the insides better, so it can increase thermal drift too.

Finally, opamps are not alike under their black epoxy skins. Each chip is a unique effort in most instances. So the details of what power is generated, how the chip responds to loading, and how easily heat gets in and out are different from opamp to opamp.

3. There are some issues with the words you've used. It sets off some detectors sensitized by advertising artists from the hifi tweako world. This is infortunate, as you probably didn't intend to do that. The terms "more accurate" and "constant temperature transfer" set off my bells.

So - I apologize for my knee-jerk reactions.

Can you describe this "more accurate" and how you measured it?

[HOTTUBES]

LOL!   

Come on, R.G. - you just know that with a beautifully facetious remark like that, someone's just gonna have to do exactly that and post the results in the Pictures thread!  I, for one, can't wait.   

Actually, I feel the need to rustle up a screamer or similar just to try it myself...   

Sounds like a new TV show .... Pimp My OP amp !!   LOL !!!!!

[Seljer]

Doesn't the feedback magically compensate for most opamp temperature issues anyway?

[PRR]

I've known stuff that sounded better when HOT (*less* heat extraction).

Go for whatever sounds/looks better for you.

[M23Bomber]

Hello RG,

1. Aerospacial E. And EE, telecomunications. With a few acronyms in the mix.  anyway who cares?
2. Maybe in my lab our equipment is to sensitive for stompboxes, or no one else ever tried to analise. Which I am pretty sure about taking into consideration all the comments.
3. Im not British or American. Excuse me.

Regards ,
M.

[R.G.]

So cool - what measurements did you take and what did you find?

[DDD]

There's definitely something wrong on TS PCB if OpAmp needs a heatsink.

[R.G.]

I don't know about wrong - maybe just different.

I am intrigued that the OP has found ways to measure differences in opamp performance with and without a heat sink, as he says. I'd very much like to see that data.

[M23Bomber]

Hello all, 

Its all about distortion, and stability. Its something that can be noticed without measuring  equipment

M.

[electrosonic]

Its something that can be noticed without measuring  equipment

Is it measurable? Can you notice it with "measuring equipment"?

[R.G.]

Its something that can be noticed without measuring  equipment

Is it measurable? Can you notice it with "measuring equipment"?

Yeah - that's what fascinated me. We had:

I started adding heat sinks, and even on the oscilloscope  I see improvements when running them in different configurations than the standard ones.
...
NOTE FOR NON ENGINERS aka DIYers:
...
I  just noticed that when heat sinks are attached some Opamps seem to be more accurate after large periods of time at higher than normal voltages, 9v, due to constant temperature transfer. Its just a detail of almost unnoticeable proportion but its there.
...
Maybe in my lab our equipment is to sensitive for stompboxes, or no one else ever tried to analise. Which I am pretty sure about taking into consideration all the comments.

and I thought "Kewl! Someone who has found ways to find things that I was not able to back when I had access to huge rooms of measurement stuff."

There's no doubting that everything you do to an audio [insert whatever here] changes something, even if it's just your personal perception of it. But actually making and supporting that in a verifiable way has been remarkably little done over the last century.