Switchemode PSU failure mode

Go To Last Post
14 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I am looking for comments, opinions experiences with on possible failure modes of three terminal LM78xx replacement switch mode regulators and a need for output crowbar protection scheme.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I had some 30% failure on a project using about 100 of the Recom brand and I will never use that brand again.

After messing around with Recom for more than a year they blamed our production for blowing them up with static.

I know from some inside information that they were using dud capacitors in some batches and that was the reason why they were failing but they never acknoledged that and never got a refund.

The failure ranged from no output, very low output and short circuit input to output, fortunately I had a beefy TVS in parallel with the 5V supply. Only a few boards suffered damaged chip which I had to replace.

..hmmm maybe this is a bit of payback... :twisted:

John Samperi

Ampertronics Pty. Ltd.

www.ampertronics.com.au

* Electronic Design * Custom Products * Contract Assembly

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

John,

Thanks for the comments.

I need to operate a number of optical scale reader heads. Each head is a tad over a thousand dollars, sealed non repairable.

I want to ensure that the power supply cannot fail in the input to output short mode.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I am always a little nervous with switchers and expensive loads.
As John suggest, good output clamping is the way to go with some resistance on the input to the switcher. There goes some of the efficiency gains! It makes linear regulators attractive!

Charles Darwin, Lord Kelvin & Murphy are always lurking about!
Lee -.-
Riddle me this...How did the serpent move around before the fall?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

How about a switcher to some suitable voltage, followed by a linear regulator, maybe an LDO? Depends on load current though, I don't know how much current is needed in this application, so I don't know if you can use this idea.

Still, depending on the regulator, it may or may not survive if the switcher blows up shorting input and output. Actually, I think I have never experienced that. Depending on input voltage, some regulators might just overheat and shut down, some regulators might blow up if the input voltage exceeds maximum. But also the capacitors on regulator input should be selected to handle higher voltage.

So maybe just some crowbar would be nice to protect the sensitive components.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Comrade Stalin was right...

Trust is good, control is better

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

John, interesting that you had problems with Recom. My experiences are exactly the opposite. I had some strange voltages on the 5v rail and i blamed the regulator. Sent some back to Recom and they did extensive tests. Turned out the problem was some tpic6595 devices which i suspect were counterfeit - these were leaking 24vdc from the outputs back into the 5v rail as well as the outputs going leaky to the turn of 1kohm. I've used a few thousand of the Recom 785xx items and not had a problem attributable to them- my apps are 24/7 in 70c ambient.

One thing though - exceed the input voltage on the Recom devices and they will fail. Normally DOA

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Kartman wrote:
- exceed the input voltage on the Recom devices and they will fail. Normally DOA
Meaning? ... zero output? short input to output? I think that is what ignoramus wants to know.

Cheers,

Ross

Ross McKenzie ValuSoft Melbourne Australia

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

If there is room for a decent redesign of the powersection of these $1k-up devices, I would skip the switched version of the three terminal regulators all together. The approach as described by Jepael is both efficient and safe. I am using that approach in a current design.

More general: A good powersupply is like loudspeakers in an audio system. :)

A GIF is worth a thousend words   They are called Rosa, Sylvia and Tessa, You can find them https://www.linuxmint.com/

Dragon broken ? http://aplomb.nl/TechStuff/Dragon/Dragon.html for how-to-fix tips

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Spot on Ross

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Can use a pass transistor to clamp, or circuit breaker, instead of a TVS though a TVS is simple and can be rugged.
One way: http://www.linear.com/product/LTC4366
A number of functions are available (circuit breaker, surge limiting, over-voltage protection, etc.).
Some of the low cost Cortex-A motherboards (BeagleBoard) have a low cost over-voltage protection (5V in, opens at like 5.8V); function is described in the board's hardware manual.
BeagleBone Schematic (go to page 2, upper left).
NCP349: Positive Overvoltage Protection Circuit with Internal Low Ron NMOS FET

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
I've used a few thousand of the Recom 785xx items and not had a problem attributable to them
As I mentioned I had some "insider" information saying that they had some bad batches of a cap inside the regulator.

I never saw THOSE xray etc. but they sent me the ones they wanted me to see.

This was a couple of years ago when counterfeit parts were roaming free, maybe I just got "lucky".

John Samperi

Ampertronics Pty. Ltd.

www.ampertronics.com.au

* Electronic Design * Custom Products * Contract Assembly

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

DOA as in dead. No output. Not sure if there was any let-through, but the mega128 survived. The boards in question have varistor protection and it was powered by a transformer with two 120vac primaries. The job was to use 120vac so the transformer was wired as such but to do the factory test, the engineer powered the boards directly from 240vac rather than through the three phase stepdown transformer. I wasn't present at the time but there was a loud bang as 14 varistors went pop. I should've put fusing on that circuit that would've protected the boards. Lesson learnt. I'm surprised the primary fusing of the transformers allowed enough energy through to blast the varistors ( they were 20J), maybe they arced? The transformers survived.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

An OTP bit slice ( fuse and crowbar ) are the order of the day.