DIY simple smps lab psu

[newbie666]

I want to build a simple psu to test my audio circuits. Requirements are +-15V@0.8A (not adjustable) and I'd like to be able to limit the current so that I avoid releasing the magic smoke in case something goes wrong.

My current plan is to purchase a meanwell smps module: PD-2515 (datasheet https://www.mouser.com/ds/2/260/r330_3-4354.pdf) and stick current limiting after it.

Now what I am trying to wrap my head around is how current limiting actually works. I have this simple circuit using LM317 (or LM337 for negative rail):



Will that configuration work at all? If I understand that circuit, LM317 will start regulating(dropping voltage) when voltage drop across R1 is greater than it's reference voltage of 1.25V. If the load will draw less current, LM317 will not regulate and drop some minimal voltage. Right?


Now for the second part. I'd like to indicate that LM317 started limiting current. Is it enough to stick a comparator between input and output of LM317 and turn on the led whenever I(out) < I(in)?

Thanks a lot for your help!

[T3sl4co1l]

Not sure what you mean by I(out) and I(in), current is conserved through the regulator -- it's just a series element.  But Vin and Vout will differ, when the reg is doing its job.  You wouldn't get a sharp yes-or-no indication with an LED+resistor, but that would work, yes (lit, slightly or a lot, versus completely dark).

The main downside is the supply isn't a constant regulated voltage source, but has a Thevenin resistance equal to R1 when it's not in current limit mode.

It's not a big deal in the first place, because SMPS like that have internal current limiting.  They can't deliver DC fault current, not like a transformer and rectifier can.  But they do have a lot of capacitance on the output, which can destroy transistors in a hurry.

An active current limiter strictly prevents an excursion of current, whereas a current limit upstream of capacitors cannot do anything to limit the current drawn from the capacitors.  So that's the primary benefit you stand to gain, using a circuit like this.

HTH,
Tim

[newbie666]

Not sure what you mean by I(out) and I(in), current is conserved through the regulator -- it's just a series element.

I typed too fast. What I meant is I wanted to use a simple opamp comparator and use voltage ( ) before LM317 as reference voltage and then voltage on the Vout pin of LM317 as input voltage. Then whenever Vout of LM317 is lower than Vin of LM317 I would turn the LED on (on the output of comparator) to signal that current limit is on.

Of course i would have to drop reference voltage of comparator by whatever is the voltage drop of LM317 that is not regulating. From the block diagram of LM317 I see there's a pass transistor and some sense resistor that would cause some voltage drop.

[ogden]

LM317 will drop voltage even when it is not regulating. What's worse - this drop is current-dependent, will increase as current is approaching "current limit treshold". It means that you more or less don't have regulated voltage on load anymore. If you want predictable voltage out of such oversimplified "lab supply", current limiter shall be put *before* linear regulator.

Good idea is to Get LTspice and verify/simulate your circuits there.

[newbie666]

Thanks Ogden, that's a good point I didn't think about.

Looking at the block diagram of LM317 (attachment) there will be a constant voltage drop across pass transistor and variable drop across the resistor. Now I don't know the value of sense resistor, but if it was let's say 1R, my max current of 0.8A will cause a max voltage delta of 0.8V which I guess I can live with.

Anybody has an idea what could be a value of that sense resistor?

[ogden]

Looking at the block diagram of LM317 (attachment) there will be a constant voltage drop across pass transistor and variable drop across the resistor.

Don't dream to have constant 0.8V across regulator for different currents that does not approach regulation treshold.

You shall pick resistor value so voltage across it is 1.2V when target current reached. Ohms law, you know.

Attached: LT1117 as ~100mA current limiter with 12ohms resistor, 20V input

[ogden]

Attach: same circuit, just load up-to 100mA, w/o tripping current limit.