5V 200mA LDO with 0.5% accuracy

Why would they put a precision voltage reference in the same package as an LDO?

The application notes are full of examples where an LDO is used as the protected output device on a precision regulator.

Semiconductor manufacturers make integrated circuits that they can sell in huge volumes. Collaborators with physicists have the task of using them to do the jobs that a very few physicists want done.

The Venn diagram gets a little slivery when you combine power + tight tolerance + high voltage but here are a couple that are pretty good, and a few more that are close. I didn't do an exhaustive search.

LDFM50PT-TR 16V 1% 500mA 0.125V dropout at 0.5A $1.36

Microchip MCP1702 flavour is my favourite. 13.6 Vin though...

Seiko S-1142D50I-E6T1U

Regards,

Boris Mohar

Got Knock? - see: Viatrack Printed Circuit Designs (among other things)

void _-void-_ in the obvious place

Hey, Spef, for a non-exhaustive search, you found a lot! More than I found in AoE-III Tables 9.3 amd 9.8, plus another 30 minutes of poking around! Thanks!!

You also included SOT-223 stuff, which I was going to add to my list of qualifications, since it's a nice small power package.

I liked ST's LDFM50, good for 500mA in DPak-5. The MIC5201 is nice, but spec'd for 200mA, I then noticed their MIC5209, 1%, 500mA and SOT-223

Linear Technology LT6658 can get 200 mA at 0.05% accuracy, with up to

I haven't used it myself, I just happened to see the advert a couple of minutes ago.

Thanks, Boris! S-1142, HSOP-6 package, 5x6mm, very interesting!

Do you have a voltage reference handy? An opamp, an LM317/1117 type, and a few discretes will make a good precision regulator.

Sort of like this.

Regulator self-heating doesn't affect voltage accuracy... which the fan doesn't need. The 317 isn't inside the feedback loop here, but could be.

A depletion fet would be interesting, too.

We are discouraged from adding a lot of new parts to the stockroom, so we often make circuits like this.

Huh... Can you get any sort of speed out of that if you wiggle the input around? (say a few kHz.)

George H.

The big output cap slows things down. The fan doesn't care. We'll sense PCB temp and probably do a simple slew-limited bang-bang control loop, basically an up-down counter.

Reminds me, I've got to check that a 317 is stable with that polymer cap. I'll breadboard that, but I do think I have an LM317 Spice model around somewhere... of unknown accuracy.

LM317 should be OK. LM337 is not.

Jeroen Belleman

Ordinarily the old pre-regulator trick using two regulators would work but you threw a wrench into the works with that USB source stuff. Doesn't the U SB ideally supply 5.000V, so how you expect to keep a 5V regulator within 0 .5% (25mV ?!?) at 200mA loading?

Interesting part; it has a precision reference, and two four-terminal regulators (separate pins for IN, OUTforce,OUTsense, GND), so even though the output is limited to 2.5 to 6V, you can build a ~33V LDO by using a PNP pass transistor and driving the base from the IN pin. Just resistor-load the OUTforce pin rather than connecting it to the real output.

Can't see the datasheet, though: some kind of error in my browser.

My testbed board can be used in a variety of ways and with a variety of power sources. Of course in the event of a USB source, an accurate 5V goes out the window. But that's not too bad; I'm using a 4.096V reference IC for the onboard ADC.

I thought the data sheet said 0.4% ? Not that it's an issue for me...

No problem. In the past I've sometimes used two regulators or a regulator and a 5V series reference. They are powered from the same input.

There is no heating in the reference/reference regulator so the accuracy is improved and the other one does the heavy lifting current-wise. Since they're powered from the same source, sequencing issues are mostly absent.

--sp

Cheers

Phil Hobbs