Bonehead base current question

[Viitalahde]

Here's a power supply supply circuit using two TIP2955/3055 pass transistors per rail to boost output current.

http://sound.westhost.com/p30c-f2.gif

If I paralleled four transistors (TO3 2N3055/MJ2955's in my case), would the 33r resistor that supplies the base current need to be scaled too?

 

[ebartlet]

I'm no expert, but I think that the base current is neglagable, so the voltage drop across a 33 ohm resistor would not change very much at all.

 

[bcarso]

How much output current do you expect?

The betas on 3055/2955 are not all that high especially at high currents, so it will be worth a look at what's needed out of the three-term and the consequent dissipation. Right now I suspect the 33 ohm parts are mainly to assure the quiescent current of the three-term regs is healthy, around 20mA, and to keep the R*Cinput time constant small.

 

[Viitalahde]

Well, I'm expecting about 5-6 amps but I want to do a serious overkill to keep things cooler since I happen to have a few nice big heatsinks that can take a few extra TO3's.. I've understood a pair would be good for about 10 amps or so, but will require good cooling.

 

[bcarso]

There are a lot of things involved in determing the required base current and base voltage when currents get large. Check out the On Semi datasheet here:

http://www.onsemi.com/pub/Collateral/TIP3055-D.PDF

 

[Boswell]

You can try leaving the 33Rs as they are. The power transistors are acting as current sources, and the regulator is supposed to deal with whatever is left over.

The circuit acts so that not more than about 20mA goes through the regulator. However, as more current is diverted into the transistor bases, there is less left over for the regulator to feed on, so you may have to consider dropping the 33Rs to 27R.

The circuit looks OK for steady loads, but I would be careful about driving anything that was sensitive to overvolting if the load current could drop suddenly.

 

[Viitalahde]

Thanks guys, you've been helpful.

The load the supply sees is going to be steady (a console) and I don't see any reason why it would change, unless it blew up of course. :razz: In this PSU, I'm considering adding a fuse after the regulator and before the collector junction to protect the regulator in case of failure..

For overvoltage protection, I'm probably going to use a TVS.

 

[Boswell]

I think the fuse in that position is a bad idea. Consider what the bypass transistors will be doing if enough current is passing through the 33R and regulator to make the fuse blow - they'll be trying to clamp the output rails to the raw input rails. If you want to put fuses in, put them in the raw input rails before the transistors.

 

[Viitalahde]

[quote author="Boswell"]they'll be trying to clamp the output rails to the raw input rails.[/quote]

:?

You're right, didn't think of it that way. Thanks for bringing that up.

I'm putting the fuses in various places:

1) A big one at the mains (the last resort)
2) 5A or so after the raw DC capacitors
3) 5A or so at main rails

That combined with soft-start circuit (a 2x10A relay shorting a 68R/50w resistor located before the 500VA toroid in 100-200mS) and overvoltage protection at outputs (TVS's and probably a crowbar circuit too) should make up a safe supply - both for people and the console.

I was thinking of using two MJ15003/15004 transistors per side since they seem to be a whole lot more powerful than the 2955/3055's, but I'm doing this with four 2955/3055 per side afterall - will be both cheaper and the load will be shared better.

The final calculation for rail currents say a bit less than 4 amps per rail, so this configuration shouldn't even heat up too much with the big boy heatsinks I've got..