using bus switch instead of voltage regulator?

Bad idea! The bus switch would be a very bad regulator, equivalent to a fet source follower with drain and gate both at +5.

Use a real voltage regulator.

John

Absolutely not. Use a nice little switching regulator module from TI, like the PT5100 series. Very efficient.

Why don't you quote the post you are replying to?

And why recommend a switching regulator for a 90 mA max load?

The LM1117 might be a good choice.

If the average current is expected to be steady, the OP might consider a shunt regulator like the TL431.

--Mac

The 1117 is a great little "MDO" regulator, cheap and very rugged. I tried to destroy one by applying illegal potentials to various leads and it didn't mind at all.

It does need a big (22uF or so) output cap to be stable.

We use them for FPGA supplies a lot, switchers rarely.

John

Do you get the idea that every time a new component becomes a candidate for use in a product that Larkin rushes down to that lab to make sure he's in on this?

I once saw a job ad for a "Non-Destructive Tester". Think they would have hired John? 8-)

I did an 8-channel load cell/strain gauge signal conditioner...

that needs a separate remote-sense power supply for each load cell. I used an LM1117 with its adjust node driven by an opamp, basically using it as a power amplifier. Under sundry fault conditions, the various pins could be forced into all sorts of unnatural polarities and levels, so I tried abusing an 1117 to see if any of the feasible states might break it. None did.

But right, I do like to torture tiny, defensless parts. *Electronic* parts, I mean.

John

Heh. "MDO." I like that. It IS a bit of a stretch to call it an LDO, but it is better in that regard than an LM117.

Yeah.

I used a switcher for an FPGA, but only for the 3.3V i/o banks (The external supply is 5V). For the core voltage I used an LDO, with the 3.3V as an input.

The total power consumption at 5V is almost an amp, so I think the switcher was a good idea.

--Mac