LM3478/3488

if those chips are anything like the recent discovery of the LM2940 series LDOs regulators from TI, I can understand it.

You're problem maybe due to QC issues and not yours.

Recently I had a box of encoders dropped on my desk for evaluation, these units had become damaged with in short time operating in systems that were powering them via a stand alone 24volt supply. The encoders are spec'd to operate from 5 to 26V DC, they were operated no higher than 24VDC

After examining them, it was the same failure for each, the LDO reg had shorted through, one IC on the board was not able to handle that 24v.

Looking at the PDF from TI, it clearly states that the input voltage for the one we were looking at "LM2940-5", which was a 5 volt type, could only handle 5 to 26VDC input. This to start with is obviously a design error on the part of the design that is employing it however, the PDF states that the ABS ratings are up around 36V or something like that, but there is a catch, anything over 26 volts falls with in this ABS range and it can only handle it for 1ms. To me, if it could only handle that for one ms, then I guess the 26VDC is the ABS max, so why are they boasting it any higher?

I order a tube of these, tested some and sure enough, 26.5, poof! and this is with no load.

We called the company that used these in their encoders, I guess they are aware of the problem. We are now looking at other manufacturers of encoders.

It's too easy to modify a PDF to cover one's tracks.

Kind of reminds me of a friend of mine that got a new ford truck years ago that had a large V8 in it. He noticed between oil changes the truck appeared to be burning oil, there was no leaks that he could see so he took it to the dealer and they checked it out and the rings were bad for some reason?

They replaced the rings, this is a brand new truck from ford we are talking about. So when he went to pick it up, they then gave him a 3k dollar repair bill. He kind of didn't go for that since the truck is new and still under warranty.

Now you'll like this one, the dealer showed him a quote from the owners manual for that truck and it stated that it may use oil between changes. That in it self indicated that ford knew of the problem and decided to reprint the manual to cover their butts and how did they come to this conclusion? They looked at an owners manual for a truck that didn't have that engine, that comment wasn't in there.

Well it didn't go over so well with him, the dealer or FORD, who ever, did end up paying for the rebuild and the engine never burned any oil between changes after that. He ended up giving that truck to his father. it's rusty as hell now (common trait for Ford) but it still runs good!

Jamie

"Jamie" wrote in message news:XtHKs.76999$ snipped-for-privacy@newsfe22.iad...

I've seen some switching regulators, for instance, which are rated in this way. There's an LT part specified for 36V maximum DC, 60V peak. The purpose is logical: it allows for significant overshoot in hot-plugging conditions. Which, LT being the informative people they are, gets explained in detail.

I don't see what they do on the silicon for that. If it avalanches at

60V, why wouldn't it simply operate at 60V, switching transients (if applicable) notwithstanding? For BJT type devices (like the LDOs, and maybe switchers coming out of saturation?), there's second breakdown, but that's thermal and current dependent. So it shouldn't happen statically at zero load as an absolute rating.

Tim

--
Deep Friar: a very philosophical monk. 
Website: http://seventransistorlabs.com

Is there a problem performimg partial differention or implicit differentiation so as to derive total devaitions and sensitivities of some sensible measure of margin from disaster as a function of this or that variable dependent on those resistors????