Regulator heatsink

[pre65]

Hi-i am new to working with regulators and i would like some advice on whether the heatsink i have made for the regulator on my PT15 heater supply is good enough before i bolt it all together.

The reg is mounted on a purpose built heatsink which is mounted on a piece of alloy angle 6"long X 1.5 "X1.5" X1/16 thick.

I have lodged a photo in photobucket

http://i111.photobucket.com/albums/n146 ... ink002.jpg

[Darren]

Hi Philip, you need to look at the manufactures PDF files for that reg, the info should be there for you.

[Nick]

Well, I would guess the angle bracket will do as much (if not more) than the heatsink. I assume the regulator is isolated from teh metalwork?

It depends on what power you are getting the reg to provide. to know that, we need to know the voltage (average) thats supplying the reg. The resistors you have after it, and the voltage you intend to supply the valve with, and the current you expect at that voltage.

What TX are you using to supply the reg?

Without getting on the old soapbox, this is why breadboarding first makes life a lot simpiler, by the time you come to cutting metal, you will know how the regulator is going to run, and more to the point, you will know what the amp will sound like, so you will know if its worth the bother cutting all the metal, or if not, you may have found the changes needed to make it sound better.

If it gets to hot, it should just shutdown anyway. As long as the rest of the circuit is happy without the PT15 fil lit all that will happen is the light goes out.

[pre65]

Hi Nick-i am using the circuit you posted on the WD forum

9v toroid
LM338T regulator
1R5 6W sense resistor
6v (or 6.3V) filament in PT15
1.5A filament current
1K variable pot

[Nick]

Ok, 9v tX will in theory produce 12.6v, assuming you are using normal diodes, take a volt or so away for the bridge, that gives us 11.6v, you want 6.3v at 1.5A, so 1.5A through the 1R5 sense resistor will drop 2.25v (it will dissapate 3.4W so your 6W should be ok, but will get very HOT), so that leaves us with 6.3v after the resistor, that means we expect 8.55v after the reg and 11.6 before the reg.

So the reg is dropping 11.6 - 8.55 3.05v, thats maybe a little lower than we would like, as I think the low limit of the reg is 3v or so, so maybe it would be better with a 1R sense resistor. That would mean its dropping 1.5V, that leaves 7.8 after the reg, so the reg is dropping 3.8v

Assuming the 3.8v drop, that means the reg is dissapating 3.8 * 1.5 or 5.7w

Looking at the spec sheet, the chip to tab thermal resistance is 4 deg C/W, so at 6W, the chip will be at 24 deg above ambiant. The max temp is 125 deg (it will current limit after that), so that gives us 101 deg to play with. Assume the ambiant is 40 deg, we will need less than a 10 deg C/W heatsink.

Looking at the one you have used, it looks like a Rapid 36-0252, which is 7.8 degC/W, so without mounting to the ali, it should be sufficient, and we would expect it to run at
46 deg above ambiant, which will be warm, but within limits. With the extra mass of the ali, assuming the thermal contact between the heatsink and ali is good (otherwise it could make matters worst) it should be ok.

[pre65]

Hi Nick-thanks very much for those calculations.

In my head it sort of looked right !

I will add some 1R 12W resistors to my Bluebell order.

[Dave the bass]

I used a slightly different approach

1)Trundle through scrap Centrifuge PCB's to find a nice sized heat sink.
2)Remove said HS.
3)Cut in half.
4)Cut finger on burr.
5)File off burr.
6)Atatch to LM338
7)Power up, observe if it combusts? If no, place finger on HS, does flesh stay on the HS? If no run amp for 8hours (or convinient BH Weekend). Does LM338 go into thermal shutdown after 8 hours? If no job done

Bodger the Bass.

PS. Philip, PM me with the value resistors you need for your phono project and I'll try and sort you out (guv).

[Nick]

Might also be worth thinking about using schottky rectifiers if you are not allready, tjeir lower forward voltage drop will give the regulator more headroom, and will dissapate less heat than if you used (for example) a bridge rectifier.

[Nick]

I hope you don't think I did the calculations when I first built one, I used Dave's method, but I guessed you wanted a bit more than "it should be ok"

[pre65]

Hi Nick-i am using 4 UF5401 diodes to rectify the 9V AC.

Hi DTB-i have put together a Bluebell order for the remaining parts i need for the PT15 (resistors,caps,knobs) so i have added my Phono3s resistors to that order .

I should be ready to order in a day or so.

Thanks for the offer but i will wait for my order to come,PT15 is getting priority at the moment.

You just want to know what Phono3s sounds like,don't you ?

Perhaps DTB = Dave the BODGER !!

[Andrew]

I hope you don't think I did the calculations when I first built one, I used Dave's method, but I guessed you wanted a bit more than "it should be ok"

First time out I did the calcs, now I'm not being smug here because by the third, I'd got blase, changed the circuit and thought, ah it'll be OK. Needless to say, it wasn't and needed a bigger heatsink, but the interesting thing is that they can go into thermal oscillation, which confused the hell out of me, heat up, shutdown, cool down, switch on again, heat up, shutdown, cool down and so on.

cheers,

-- Andrew

[Dave the bass]

ah it'll be OK. Needless to say, it wasn't and needed a bigger heatsink, but the interesting but is they can go into thermal oscillation, which confused the hell out of me, heat up, shutdown, cool down, switch on again, heat up, shutdown, cool down and so on.

cheers,

-- Andrew

Ah y'see, your finger might have been out of calibration and giving you a false sense of security.

Top tip, don't use a finger with an open wound (like a burr-cut for example) to test heatsink capability, the whole project could go boob's-up if the goo in the infected finger causes false readings.

Wise words I think.

DTBodger

[Nick]

Andrew:

First time out I did the calcs

Oh, good, in that case, does my number make sense?

Philip:

Hi Nick-i am using 4 UF5401 diodes to rectify the 9V AC.

Ok, not schottky then, they have a forward drop of 1v, using something like STPS1545D (0.57v drop. 45v 15A) would mean you have 0.86v more to play with.

Whats the ESR of the cap you are using? you need to ensure the voltage into the regulator never drops below the voltage it need to regulate with the load and series reg, otherwise it will drop out of regulation, and you will get hum (with nasty harmonics) into the fillament.

Also, whats the VA rating of the transformer you are using, the voltage might be a bit over 9v is its not 100% loaded, which will help as well (though not with heat)

[pre65]

Hi Nick-the transformer says its 50VA with twin secondarys (i will connect both secondarys together in parallel)-one transformer for each PT15.

Cant measure the cap DCR cos its on a tagboard and wired up !

Just checked my parts box and i have 16 1N5820 which seem to be 20V 3A schottky diodes-will those be better ?

[Dave the bass]

Hi Philip,

Not DCR (caps don't have DCR unless they've gone faulty like a partial short or summat) but ESR 'effective series reactance'. I must admit I've never measured it myself. How do you do that Nick?

Data sheet info or summat?

DTB