Suggestions for heatsink regulators?

[m00dawg]

In my current SID design, I am still using the 7805 regulator for the CORE module and I'm trying to figure out ways to cool the thing down. I may eventually put the 7805 on my main power board so I can stick a gigantic heatsink on it (to at least power the main CORE module since it's driving all the LEDs and such, or will be once I"m done anyway). Until that time, however, does anyone have any suggestions on a small heatsink I can put on it? I'm using Smash's CORE board if that helps.

Thanks a ton!

[stryd_one]

There are a squillion available, just search your favourite store for TO220 heatsinks.

They shouldn't get all that hot really, unless you're overpowering them significantly.... What voltage are you feeding it?

[m00dawg]

Thanks Stryd! I did indeed find quite a selection at my electronics dealer. I think I'll buy a few of these at least to have as well as look at some heatsinks for the SIDs (saw a few threads related to that while trying to find an answer to my regulator heatsinks).

As for the power, I'm supplying CORE with 9V (regulated). I'm thinking of using a multi-tap transformer at some point, but for now I'm stepping the voltage down using regulators. Now, of course, the regs on my power board did get hot, so I put them under a large heatsink I grabbed from a busted motherboard. So they, plus the 7809's on the SIDs (I'm supplying them with a regulated 12VDC by the way) are good.

The 7805 on CORE, however, has gotten up around 60C after about an hour. Definitely too hot to comfortably touch. I have an LCD and a few control surface items, plus the +5V from CORE is also going two SIDs (I figured that was easier to do for now then using yet another regulator at least for now).

I made a pretty ugly attempt at mounting a fan, and while it helps cool the big heatsink, it doesn't help much for the 7805 on CORE, which is too far out of the way to get any wind I think.

[lylehaze]

All the positive regulators (78xx) have the tab grounded, unless you are "stacking" regulators. If you are using a metal case, you may be able to use your case as a heat sink.

To do this, I usually insert the regulator from the backside of the board, and I don't solder it until the board and regulator are mounted firmly. A bit of heatsink compound doesn't hurt either.

If the case is painted, Sandpaper will clean up a spot for the regulator.

DO NOT do this for negative regulators unless you have an insulating sheet and screw, as the tab of those is NOT grounded.

Have Fun,

LyleHaze

[stryd_one]

I never liked the idea of having the circuit exposed to the external world....Turning the case into a big ground plane just doesn't sit right with me... But it'll sure cool it down!

moody your voltages are totally normal, it must be drowing really high current. Have you considered a different VReg designed for higher loads? There are 2A and 3A 5V regs out there...

[m00dawg]

Thought cross my mind, but if I did that, I'd probably move the regulator to the power board itself, at least for the "main" CORE module. I still might do that, though. Trying to figure out the best way to approach the whole thing :)

Unfortunately, I don't have a metal case - at least not yet, so I think the sink is the best way to go.

[doc]

I never liked the idea of having the circuit exposed to the external world....Turning the case into a big ground plane just doesn't sit right with me... But it'll sure cool it down!

what?

In practise it is a good solution to groud (shield) a metal case. Even with audio Gear you have to ground a metal chassis at one point. And if this point is the back of your (positive) voltage regulator.... why not?  ;D

greets

Doc

[stryd_one]

In practise it is a good solution to groud (shield) a metal case. .....you have to ground a metal chassis at one point.

Why is that? (not arguing - hoping for a lesson :) )

[lylehaze]

When a metal case is grounded, it becomes a "shield" to prevent (or reduce) electrical noise from coming in.

It works in a way similar to the shield on a microphone cable.

Please correct me if I am wrong.

Lyle

[stryd_one]

Well yeh it'll help stop some EMF of course... but what if you apply power to it.... (say, by touching it with a pin from a phantom-powered mic cable that shares a ground with the device)

I see the value in shielding, but not in exposing the shield to the outside world, that's the bit I'm stuck on...

[lylehaze]

Phantom power..

There are two possibilities:

If the case is not (yet) connected to system ground, which is likely if it's not yet plugged in while you fumble around with the MIC cable:

No problem. As there is no return path, there is no circuit.

If the case is already connected to ground, probably through some other cable already connected:

If you short Phantom power to the case, Phantom power will momenetarily fall to zero volts on that one lead. The phantom power supply and all other devices powering off of it should be unaffected, as each lead has its own "pullup resistor" to V+. Current through the short circuit point is limited to voltage (approx 48?) divided by resistance of the pullup for that wire. ref http://users.otenet.gr/~ATHSAM/Audio_mixer_6_ch_ENG.htm

The phantom power resistors are 6.8k to 10k.

Worst case would be 48V across 6.8K resulting in 0.00706 amps of current. Hopefully your PS can tolerate that for short periods.

If the same short happened and the case was NOT grounded, the "best available" path to ground might be through the operator. I recall a news story about a baptism ceremony a few years back in Texas. Standing in a pool of water, reaching up to a hanging Mic above. The problem was a short, the ceremony was even shorter.

I don't claim to be an expert. I'm just posting my understanding of the circuit. I invite others to correct me or discuss this further.

Lyle

[/tilted/]

Having a metallic chassis gives you a type of Faraday cage, which will help to keep wide band noise out of your gear. (not just RFI, but pretty much the whole spectrum) - ever wondered why your mobile phone stops working when you get into a lift(/elevator)?. Connecting your chassis to ground increases the effect greatly.

Phantom power (as of course just one example) has been designed to be impossible to short out outside of lab conditions.

- Think about it. All the inputs to your console, which could ever carry phantom power, are female sockets at the desk. All devices which use phantom have a male socket. This means that the free end of a phantom powered cable will always be a female end. You can't touch the phantom pins while they are active. Of course, you could try extra hard and get a male to male cable...

[lylehaze]

/tilted/

You are correct, of course.

But don't even think of the word "impossible" as long as the cable is laying on the ground. Forklifts, sport shoes with metal cleats, heavy boxes with sharp corners, performers with spiked hair, fools with staple guns, the list could go on forever.

Short circuits _MAY_ be impossible under lab conditions, but out in the real world, there are more possibilities than I can count.

Be prepared for anything, be suprised by nothing, always keep a roll of duct tape handy. :-)

LyleHaze

[stryd_one]

Heheh okay I think you've been distracted by the example. Forget phantom power, I was just trying to think of some random way that the circuit could get messed with. Point is, I've always felt uncomfortable with exposing the circuit to the outside world any more than necessary... Bad things happen outside the box ;)

[/tilted/]

lh: of course. I was talking more about problems due to "operator error", not so much the "build something foolproof, and nature supplies a greater fool" effect. Impossible is really not the word here. Perhaps it's more like "If you're not trying to break it, you won't..."

s1: I've always been somewhat adverse to supplying the outside world with any signal if it can be helped. My point is that if your gonna give the outside world any part of your circuit, then giving it a grounded chassis is definitely your best and safest option for most applications. (It is certainly better than giving it Sclk, or Mains Active for example...  ;D )

Of course, there are always exceptions... for example if the end user is likely to be exposed to high voltage supplies while using the equipment (some industrial applications) then you'll want your chassis to float, else it will potentially supply a handy return path for said high voltage...

Some audio folk prefer their equipment to be un-grounded, and this is achievable, but should always be done from the early design stages, not as an after-market modification (unless you are a skilled engineer). I've lost count of the times I've been asked to cut the earth pin off some piece of vintage gear. (for vintage, read "old!")... I've made a number of enemies that way, as I always refuse.

In terms of inadvertently applying power to a sheilded case, this is why power supplies have fuses.  ;D

[cimo]

Discussing about whether a chassis should be grounded or not, or about the size of a voltage regulator heatsink, that s the kind of folk i d like to hang out with...

[stryd_one]

Well said tilt. Especially the "cut the ground pin" thing. I hate that!

Cimo: you're a gee-eek, neener neener neeener!