High Power TL432?

12V 200W

I am thinking about either Tarlington or Qarlington. Part Package Vbe Ib Ic Power TL431 SOT-23 0.05A 1/4W

Tarlington (3 stages)

2N2907A TO-92 5V 0.015A 0.8A 1W TIP147 TO-220 5V 0.5A 10A 125W 2N2955 TO-3 7V 7A 15A 115W

Qarlington (4 stages, 2x Darlington) MJD117 IPAK 5V 0.05A 2A 10W MJ11011 TO-3 5V 1A 20A 200W

On a sunny day (Thu, 25 Aug 2022 18:08:26 -0700 (PDT)) it happened Ed Lee snipped-for-privacy@gmail.com wrote in snipped-for-privacy@googlegroups.com:

Would not a switcher series regulator to stabilize voltage be more green and cheaper too, given sky rocketing cost of electrickety? What's the application?

The application circuits, so-called, in the data sheets are just a bunch simplified conceptual ideas you would never want to use as a finished circuit. They have no protection for all the kinds of things that can go wrong with the components or adverse external events the circuit may be subjected to, especially as they cause component failure. And it doesn't make any sense whatsoever to use these old fashioned linear circuits at relatively high power given the simplified thermal management, availability and low cost of self-protecting, nearly monolithic solutions using switching technology.

I knew a guy who had a nice little business putting app note circuits on little PC boards with barrier strips.

If he wants a shunt regulator, that has to dissipate power. Switchers don't.

The OP is talking about using through-hole parts, so I'm thinking this isn't a production design. ;)

Switching isn't always a win even in production--it can be difficult to keep spikes out of sensitive circuitry, for one thing. For another, there are some applications such as moderate-accuracy temperature control, where you can mount the pass device and heater together. That way you get 100% driver efficiency even with a linear control. (The drawback is that the thermal gradients tend to move around a lot during transients, as the dissipation moves between the heater and the pass transistor.)

For less sensitive and less specialized things, though, switching is a win.

We have a high-performance laser driver design that includes a Class H driver to run thermoelectrics--the switching part keeps the voltage drop in the linear part down to below half a volt--just enough to keep the BJT bridge out of saturation so that the active-filter action works.

Cheers

Phil Hobbs

The worst that can happen is that the circuit burn out as one-time fuse.

Sot-23 and Ipak are SMT. SO-3 is mainly for the high power capacity.

Think of it as a over voltage circuit breaker/protector. When it's needed, efficiency is not an issue.

Okay, well there are plenty of precision power clamp circuits in those voltage reference datasheet apps. Your dissipation should be transient. It helps a LOT if your clamp detection switch can signal a shutdown and make it happen. Dunno how you expect a Darlington to act as a clamp when it's only able to deliver current.

Probably this BOM: Tarlington (3 stages) Part______ Package Vbe Ib Ic Power TL431____ SOT-23__ --- --- 0.05A 1/4W PZT2907A SOT-223 5V 0.015A 0.8A 1W

2N6287___ TO-3___ 5V 0.5A 20A 160W

The problem with running bipolar with very low headroom is you lose input ripple rejection with frequency- not completely, there is some attenuation, but if you're hyper about noise then it's a big deal. Like I'm really telling you something you don't know...

The 160W Power Darlington (with big heat sink, perhaps with power resistor as well) should stop the voltage from going higher. Cannot shutdown. The power source should continue to work, just not going too high in voltage.

And just exactly how will it do that when your external power source is back biasing its BE junctions?

No, the TL432 is biasing PZT2907 and then 2N6287.

0.5% TL432. It's OV Shunt. Cannot be in series.

Guy Macon? He used to do that sort of stuff.

Cheers

Phil Hobbs

No, a friend who lived in Foster City. Long, sad story. Day trading. Nth mortgages. Cruises and fine dining. Divorce.

Yeah, you have to choose the right transistors, but some are quite good. It also helps to be able to run the bases from a higher supply--the bottom devices are allowed to saturate because the ground is quiet.

Our gizmo uses an asymmetric bridge--one side is a +- current conveyor, and the other side is a complementary emitter follower, so it can heat as well as cool--diode lasers generally are happiest around room temperature.

Cheers

Phil Hobbs

Well, that's MUCH better, a standard clamp configuration. You need to watch that IKA,off spec with your super large current gain. Pulse response fig 7-14, 7-15 no faster than 1/2 us- which ought to be a few orders of magnitude faster than the power supply overshoot you're protecting against.

For how many milliseconds will the power transistor be dissipating 200W?

No realistic heatsinking can handle that dissipation in a single TO-3 package - if the application calls for that power over any extended period you will need multiple paralleled power devices (with emitter ballast resistors to share the current).

piglet

several minutes.

Yes, there will be power resistors to share the heat. I pick TO-3 mainly for the higher currents: 20A for 2N6287 and 30A for MJ11011.