I plan on using the LT3080 to regulate a variable 1-36V 7A DC power supply. This regulator by itself will not provide more than an amp or so, so I will have to use an external series pass transistor per the recommendation in their datasheet:

LT3080 with series pass PNP

Now, I am sure this will work fine, however the recommended BJT is expensive, something like $7 a piece. Given I am planning to make a few of these regulator boards, I would like to keep the cost down by choosing a comparable part. This is where I need help.

On Mouser I see plenty of PNP BJTs that support 100v emitter-collector, and 10s of amps for 60 cents, but I don’t know if that will work. So what parameters should I be looking for in the datasheet that would be suitable in this application? It would be great if you can also explain why this part was chosen in the first place and why it is more expensive than other seemingly similar BJTs.

For further clarification, the LT3080 is set to always be a fixed value below the input voltage. This dropout voltage will be maximum 5V. This is to give enough headroom for regulation as well as to account for the voltage drop across the transistor.

What I’m planning to build is meant to suppress ripple and switching noise in a switch mode lab power supply with adjustable current. LT3080 is thus used as a post-regulator for cleaning up the supply rails.

Here's a simple schematic of what I'd like to build. enter image description here

and here's the full version I was playing with earlier with supposedly better ripple suppression.

enter image description here

I'm not experienced in designing circuits (software engineer by profession), so if I'm doing some obvious mistakes please let me know.

  • \$\begingroup\$ Comments are not for extended discussion; this conversation has been moved to chat. Any conclusions reached should be edited back into the question and/or any answer(s). \$\endgroup\$ – Dave Tweed Dec 26 '19 at 19:15

Power dissipation ,Current ,Voltage and Safe operating Area .On the Ap note circuit life is good because the input /output differential voltage is only 2V7 so power wasted is manageable with the Husky MJE4502 on a heatsink .You want to increase max current by 40% Which is not a show stopper .However your expected input /output differential will be more than 36V .This will call for many Power devices in parallel .Your Calculated Heatsink will have to be much less than 1 Kelvin per watt which as a rule of thumb it wont be easy .Consider reducing your input volts .Prebuck works well for me .Some large off the shelf linear lab power supplies have automatic relays that step the tapped transformer secondary windings to save heat by keeping input /output differential volts not too high .

|improve this answer|||||
  • \$\begingroup\$ Please see my comments for further clarification. Maximum voltage drop across the transistor will be 5V because the output is set to track the input. \$\endgroup\$ – darksky Dec 26 '19 at 1:04
  • \$\begingroup\$ Other than heat dissipation and voltage and current limits, is there anything else in the transistor that is relevant to this application? Like its bandwidth product, output capacitance, switching characteristics, or anything like that? \$\endgroup\$ – darksky Dec 26 '19 at 1:20
  • \$\begingroup\$ What I want to understand is whether the $7 you’d be paying for vs 60 cents is mainly for the package and heat dissipation capability. \$\endgroup\$ – darksky Dec 26 '19 at 1:29
  • 1
    \$\begingroup\$ @darksky At high voltages Safe operating area can limit things befor max power dissipation kicks in .Switching does not matter .Transition frequency should not be a problem because the circuit is designed for slow transistors . \$\endgroup\$ – Autistic Dec 26 '19 at 1:46

How to make a better linear power supply. (Merry Xmas)

enter image description here

Assuming you had a bank of matched NPN 2N2955(PNP)'s on a 200W heatsink and a 40V Op Amp that senses to 0V input (PNP type) with any voltage reference like 1.25V bandgap ref. with a log scale 1k Pot to fine tune the more sensitive high V range near 0 Ohms.

This is what I imagined it to look like. The dummy load is can be like the quasi NPN driver except 2N3055(NPN)'s on a 400W heatsink. As this design goes to 10A @ 40V.

The noise generator I added to test loop stability with an RC load. You will need a load cap to buffer RF load currents and will probably need a cap on the wiper to shunt a noisy pot. Note the sliders on right side of this SIMULATION adjust Vout and I load. Enjoy.

|improve this answer|||||

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.