Paper clip heatsinking for LM317?

[tomb]

Very entertaining and useful, Tangent!

[Voodoochile]

Well tangent, I guess you are all caught up on other projects, eh?
That's pretty nice experimenting.

I was half joking when I suggested the epoxied penny. I don't think it's a good idea. But I did think it would be better than a paperclip. Although the contact area would be far from optimal, it's a lot better than a paperclip. And then the thought of the paperclip falling off, and landing somewhere around the legs of the reg...

Of course it's best to buy a proper heatsink, and for a little device like that, they are cheap. But I guess if I absolutely could not find one locally, and shipping was horrible, I'd make one somehow. Probably not out of clips though.

I once saw a fairly detailed article about some guy who made a heatsink for a graphics card out of the better part of a roll of pennies. It was just plain scary looking.

[tomb]

Originally Posted by Voodoochile

... snip ...
I once saw a fairly detailed article about some guy who made a heatsink for a graphics card out of the better part of a roll of pennies. It was just plain scary looking.

Not quite as scary as the thing that guy in your avatar built (Oppie).

[tangent]

Originally Posted by Voodoochile

Well tangent, I guess you are all caught up on other projects, eh?

Of course not. This is pure, unadulterated procrastination in action.

Oh, was that my outside voice? Um, I mean, I had constructed a scenario to evaluate the operation of my test equipment. Yes. That's the one.

I was half joking when I suggested the epoxied penny. I don't think it's a good idea.

It's going to come out better than nothing still, I'm sure of that. Just not quite as high up the rankings as in the original test, I think.

Although the contact area would be far from optimal, it's a lot better than a paperclip.

I think the reason the paperclip does so well is that it has a lot of surface area away from the regulator, in contact with the air. Most of the surface area of the penny is just mimicking the TO-220's tab, so no advantage.

And then the thought of the paperclip falling off, and landing somewhere around the legs of the reg...

Okay, so you find a copper paper clip, then you get a really hot iron, and....

Of course it's best to buy a proper heatsink

Yep. Even the little 21 cent Aavid is still beating out all the DIY hack-jobs.

[Voodoochile]

How about this one tangent. Take a penny, and place it on top of the silicon package, against the side of the tab facing the part number labelling. Mark and drill the hole there.

Then drill a second penny to match, and bolt the two of them so that they are more up in the air, with the gap between them formed by the tab rather than a nut.

Finally, tighten it to 65 foot-pounds.

I've got to admit that while the clip did not perform that well, I am surprised that it dropped the temp that much.

Have you considered some 16g solid wire to make your own copper clip? I suppose that would defeat the idea of using what you have on hand (eg, clip or penny).

Originally Posted by tomb

Not quite as scary as the thing that guy in your avatar built

No kidding. I'd not want to have that on my resume, brilliant as it was. Very interesting fellow, though.

[tomb]

Originally Posted by Tangent

I think the reason the paperclip does so well is that it has a lot of surface area away from the regulator, in contact with the air. Most of the surface area of the penny is just mimicking the TO-220's tab, so no advantage.

That is absolutely the key. I'm with Voodoochile on raising the pennies, though. Just rotating one up to the tab part of the regulator would get some interesting results, I think.

[ilikepie]

Ive used staples as a heatsink and it helps a little. If your gonna use pennies you can stack them zig zagish to increase surface area. Glueing them on is a pain though. I use these heatsinks for my computer because im too lazy to glue staples on all day, they come with some tape, and its cheap.

http://www.svc.com/crc-u01.html

[luvdunhill]

I'm looking to step up from the 24.4 C/W heatsink recommended on the TREAD by Tangent, and was thinking about using this one but mounted "the wrong way" (i.e. the fins sticking out away from the board).. this gets my thermal resistance down to 9.6, but I was wondering a couple things:

1.) Does this really matter since I will not be using a fan, so I will not have any air velocity (see the graphs for the part above)

2.) Is there another heat sink that would "fit" on the TREAD? The larger style used on the STEP doesn't look like it will even come close to fitting!!

Thanks!

[Nordic]

I was testing my valve pre one night when my wife asked me why the tube was glowing so brightly, which surprised me as it was an ecc88 and they don't realy glow like that... or none that I have seen next moment I got the hot electronics smell from the regulator board supplying the 6V heater.... I forgot to attach the heatsink... and it swung to full voltage output... luckily it didn't do any permanent damage to the valve or the regulator.

[tomb]

Originally Posted by luvdunhill

I'm looking to step up from the 24.4 C/W heatsink recommended on the TREAD by Tangent, and was thinking about using this one but mounted "the wrong way" (i.e. the fins sticking out away from the board).. this gets my thermal resistance down to 9.6, but I was wondering a couple things:

1.) Does this really matter since I will not be using a fan, so I will not have any air velocity (see the graphs for the part above)

2.) Is there another heat sink that would "fit" on the TREAD? The larger style used on the STEP doesn't look like it will even come close to fitting!!

Thanks!

Luvdunhill,

The TREAD and the TO-220 LM317 are more flexible than you think. The heat sink you reference will work fine - assuming, as you say, it will fit. Even so, that's not really an issue unless you're trying to pack all this into a mint tin.

You can bend the leads on the LM317 so that it lays flat, away from the board, or even mount it in reverse on the bottom of the TREAD board (make sure the polarity is correct). Almost any heat sink will fit in those scenarios. If not, there's nothing that stops you from drilling another hole in the heat sink so that you can bolt it where it does fit. If your case is metal, you can delete the heat sink altogether and bolt the LM317 directly to the case. The mass of the metal will cool it better than most heat sinks.

If anything, this thread highlights the almost limitless flexibility of applying a heat sink design.

[Garbz]

Originally Posted by Nordic

I was testing my valve pre one night when my wife asked me why the tube was glowing so brightly, which surprised me as it was an ecc88 and they don't realy glow like that... or none that I have seen next moment I got the hot electronics smell from the regulator board supplying the 6V heater.... I forgot to attach the heatsink... and it swung to full voltage output... luckily it didn't do any permanent damage to the valve or the regulator.

Your wife knows how bright a tube should glow? My girlfriends mostly have no idea what a tube is!

Btw Tangent as a matter of interest what is your background in statistics? It seems you know quite a bit about data samples and expectation.

[tangent]

Originally Posted by Garbz

as a matter of interest what is your background in statistics? It seems you know quite a bit about data samples and expectation.

Most of my technical training came from this book.

I don't actually remember where I learned what I based my expectation on. It just seems to me that since it's a natural, linear system, it's most likely to be Gaussian. I'm not ruling out the possibility that it inherently has another distribution, perhaps because of some quirk of the way the regulator works. Let's just say that if I improve my techniques and start getting proper Gaussian distributions, I'm not going to be surprised at all.

[Nordic]

Well i was assembling the thing on the floor.. she did see me testing subsections of it earlier... but man those heaters wire bright. Gave me alot of respect for how tuff the tubes are.

[Tomo]

Hey,

Just for a record LM317 should work fine well beyond 100C (temperature compensated). ... In essence, keeping it free in air is fine according to Tangent's results. (Will you right a short paper for me? So I can refer to it.)

Nordic, you said your ECC88 glow hot. How are you powering the heaters? If you are using the LM317 current source like in the original YAHA amp, you may have wired it upside down. Check the wiring anyway. The glow like that comes from the tube heater getting massive current. This will fuse your heater coils.

I had a similar problem and I had it running for a week. Now I have Ei ECC88 with heater voltage of 1.8v(!!). This put approximately 12V across my LM317. This is no-no because voltage across the regulator is directly related to dissipated heat. I had my LM317 bolted to the aluminum chasis so the entire amp was hot to touch. I had to chuck the Ei tube out.

T

[Nordic]

No, the circuit was running fine, I think it was some form of thermal shutdown in the regulator that kicked in...when I let it cool down and put the sink back on it worked just fine...the heaters are just glowing faintly orange, I think it is about 365mA, but its been a while since I build it, so I'm not 100% sure. Works beautifully and is in daily use from about 8in the morning till 1 at night. I think I had a current limiting resistor in parallel just before the tube and I basicaly adjusted the regulator with the tube pluged in and dailed it to 6V.