Page 1 of 1

#1 TL783C

Posted: Thu Jul 05, 2012 9:31 pm
by Paul Barker
It says in the data sheet that it can handle 180mA (a guess from the curves not a quoted number) at a volt drop of 125v, at a junction temperature of 125 Celcius. so that is 22.5 watts.

So what heat sink would be required for the TO220 package?

#2

Posted: Thu Jul 05, 2012 10:13 pm
by Nick
At a 25C ambient, I would say somewhere around 1C/W

3C/W from inside to the tab, at 22W, thats 66W, add 25C ambiant, thats 91, so that leaves 34C. 34/22 gives 1.5 C/W

#3

Posted: Fri Jul 06, 2012 7:49 am
by Paul Barker
Thank you, so that would be big.

I found this 1.21 C/W Heatsink for 220 packageat Farnel.

But I would rather be safe and go for 1 C/W.

Maybe if I boost it's performance with a £1 fanrun at half voltage for quietness?

The purpose is to use these chips as constant current power supplies to precede the shunt regulator. Apart from the possible sonic advantage, they have a practical advantage of less sensitivity to supply voltage variations.

I suppose the ideal would be to only aim to use 60v of headroom across the chip but put in place heatsink capability for the full 125v.

Obviously for a CCS anode load of the first VA stage of an amp the current is miniscule and the heatsink requirements of no consequence. It is the power supply application which I am eyeballing this chip.

As I said before it is a modern day Barrister to a power supply designer, i.e. offers great flexibility.

The other appeal about the chip is that it is no longer a poor man's CCS which the usual 3 legged VR chips would be compared to the efforts of the likes of Gary Pimm.

"Excellent performance specifications,
superior to those of most bipolar regulators, are
achieved through circuit design and advanced
layout techniques.
As a state-of-the-art regulator, the TL783
combines standard bipolar circuitry with
high-voltage double-diffused MOS transistors"

If it does tun out to have the "right" internal complexity and materials choices it could prove to sound good and save an awful lot of faffing around that most of us never achieve normally.

Time someone in the solid state wold amongst us evaluated it against such benchmarks as the Pimm stuff.

I can't do that, but I can make one and see if I like it. Not a lot to go on but things have to be pretty good before I like them.

#4

Posted: Fri Jul 06, 2012 8:13 am
by Paul Barker
OK I have done a more practical recalculation of power supply concept.

Using a cheaper and more practical sized 3.3 C/W heatsinkIt should tolerate up to 88 volts maximum, so if we aim for 44 volts as a design objective in the power supply hopefully we shouldn't see thermal cutout even with passive cooling.

44v is a reasonable amount of headroom to design a shunt regulator around.

People who don't know the way I work with shunt regulators may be mistified by this. But the problem with the shunt regulators I prefer to use based on Stephy Benches very early stuff involves the use of VR tubes so the supply voltage has to acomodate the inflexibility of the range of VR's available. Nomally one has to have a hand on the variac. One could just as easily use a valve series regulator preceding, but the CCS method may have sonic advantages and this chip is a done deal. (complexity all housed within itself, less work for me, I am geting lazy.)

#5

Posted: Fri Jul 06, 2012 8:15 am
by Paul Barker
Please feal free Mike H to poke fun at the massive turn around in Paul Barker's preferences!

#6

Posted: Fri Jul 06, 2012 2:07 pm
by colin.hepburn
If you are looking for heat sinks this company Conrad are good and the cheapest but shipping time maybe a problem coming from Australia take a look the 8FT Series but may not but worth it for one or two small sinks but if you are wanting big sinks for the like of the nelsion pass F5s this the place
http://www.conradheatsinks.com/products/channel.html

#7

Posted: Fri Jul 06, 2012 9:04 pm
by Paul Barker
Yes they do look good value.

#8

Posted: Sat Jul 07, 2012 12:39 pm
by Mike H
So can get decent heatsinks! (Farnell) 'Cept the big ones I might want are "out of stock" :roll:

Paul Barker wrote:Please feal free Mike H to poke fun at the massive turn around in Paul Barker's preferences!
Who moi :lol:

#9

Posted: Sat Jul 07, 2012 12:43 pm
by Mike H
That is an interesting chip actually