how bad are LDO regulators ?

[Simon]

in order to get down past -46 the only regulator I can find is a low drop out type. Being safety critical I'd rather not use one, or are they ok these days ?

[Kjelt]

I am also interested in that question, the few times I used a LDO reg. I really made sure that the output caps ESR complied 100% to the datasheet in order to prevent oscillations but I never felt comfortable using it in a critical application, also not knowing how the capacitors specs will change over time.

[luky315]

LDOs are ok as long as they have the right output AND Input capacitor. Modern LDOs work with ceramic capacitors and if you use high quality (X7R or better) material they are stable over time / temperature. To be on the safe side use slightly more capacity then the minimum specified values.
Look for AEC-Q200 (passive) and AIC-Q101 (active) qualified Parts.

[Simon]

Thanks, seems I'm stuck with one, I found this one: http://www.ti.com/lit/ds/symlink/tps7201.pdf that is supposed to work down to -55 but all of the specs stop at -40.....

[Kjelt]

Time to ask your boss to buy that fridge that goes down to -50 for testing and fast cold drinks in the summer

[Simon]

looking at the list at the bottom of the datasheet it shows that while some of the variants go down to -55 the 7250 or 5V series does not go below -40. With safety critical applications I can'r rely on a couple of chips from the same batch working for a few hours as being representative of the parts on the whole.

[Precipice]

Thanks, seems I'm stuck with one, I found this one: http://www.ti.com/lit/ds/symlink/tps7201.pdf that is supposed to work down to -55 but all of the specs stop at -40.....

You've seen page 25, where you have to provide a somewhat specific output ESR to keep the chip stable? Unless you're driving a nice simple load, that's tricky to do. Are you _sure_ you can't find a less unstable device? Doubly so if you think your capacitors are going to change wit temperature...

[Simon]

I'm not using this part, I'm looking at something like: http://www.ti.com/lit/ds/symlink/tps73115-ep.pdf now, the main problem is making sure I can dissipate enough heat, not sure if my scotchcast N8 potting is a blessing or a headache. I'm going to have to put a SMPS regulator in front of the linear reg. This new one uses an N channel device, I think to get the LDO they boost the voltage internally to drive the gate of the NMOS

[Simon]

What on earth does: Maximum power dissipation for the SOT23 package, is calculated assuming that the device is mounted on a ceramic substrate measuring
15x15x0.6mm.

supposed to mean ? ceramic ? usually they ask for copper.

[jpb]

What on earth does: Maximum power dissipation for the SOT23 package, is calculated assuming that the device is mounted on a ceramic substrate measuring
15x15x0.6mm.

supposed to mean ? ceramic ? usually they ask for copper.

You do get ceramic heat sinks:
http://uk.farnell.com/amec-thermasol/mpc151525t/heat-sink-ceramic-15-15-2-5-std/dp/1892471

But it sounds as if they are referring to the whole circuit board being ceramic (Alumina ??)
eg
http://www.bestpcbs.com/products/ceramic-substrate.htm

[Simon]

I'll have to go with as much copper I can afford, If I get it right my SMPS will be 5.5-6V output so very little left for the linear reg to drop.