Re: Ever smaller DC-DC converter modules by TI

Their little SOT23 switcher chips are fabulous too. The TPS562208 costs us 22 cents, is rated for 2 amps, and runs spread-spectrum.

It's nice and quiet. We used some LTC integrated-inductor switcher bricks that were EMI nightmares. I wish people would stop making their n-channel substrate diodes into SRD's, which happens too often.

I noticed in the datasheet to avoid disabling it under full load. To me that reads the shrapnel would reach the ceiling...

Arie

After better reading the datasheet of the boost converter, TPSM81033, most interesting for me right now. And not disabling but removing the input voltage. Like with the good old LM78xx regulators.

" Device Limitations (intended to be removed for final material):

1. Inductor current limit function is not optimized. TI don't recommend to use TPSM81033 at the condition that the input current is higher than 5A. Final material is expected to improve the current limit function.
2. TPSM81033 doesn't support to remove input voltage at heavy load. TI recommend to remove input voltage at no load, or connect EN to GND first with load and then remove input voltage. Final material is expected to support remove input voltage at heavy load. "

Arie

I was amazed at the TI ISOW1412DFMR on-chip isolated RS485 power supply, but then I'm no designer.

-- A host is a host from coast to snipped-for-privacy@panix.com & no one will talk to a host that's close.......................... Unless the host (that isn't close).........................pob 1433 is busy, hung or dead....................................20915-1433

I’ll certainly be reading the datasheet carefully before using the part, but I expect that it’s not a big worry in real life.

You’re going to have some sort of reservoir cap on the input, and with a 76 MHz switching frequency, its discharge will be nice and smooth.

It would take work to cut off the input supply fast enough to cause inductive problems.

Cheers

Phil Hobbs

I'm always suspious when I see a note like that on a data sheet. Oftimes, they are hiding a latchup or something.

If it does latch up, a brief brownout, or some spike, might fry the thing. I'd get an eval board and play with one before designing it into anything.

Sometimes the current ratings are optimistic too.

No fear. I don’t use any switcher chips that I haven’t tried out, good and hard.

My soul was seared by a TI part that the datasheet swore would keep on PWMing right down to zero load current. They made a big deal about it.

Aha! I says, says I, “Perfect for my fancy isolated triple-output flybuck!”

After I’d designed it into the guts of a biggish board, it turned out to be all lies—with a light load on the main output, it went into burp mode and the isolated outputs collapsed.

Barstids. Never again.

Cheers

Phil Hobbs