Power supply protection networks

Dear all:

I'm designing the power supply end of a new gizmo (my first power supply in ages), and I'd like some advice on the protection network on the DC power input. (I've just had a nice iRiver MP3 player blown up by a failed no-name Chinese wall wart, so power-brick fault protection is on my mind.)

Back in the day, I'd just have used a crowbar SCR and a PTC thermistor, but I'm worried that doing that might make a failed wall wart catch fire. Sooo, I'm looking instead at PTCs and transient-protection zeners.

If the PS goes really bananas, the PTC will switch fast enough that the zener won't desolder itself from the board, but if the PS just goes out of spec (e.g. 24V for a 12V brick), the zener winds up dissipating a lot of power for a long time (seconds) before the PTC switches.

Soooo....what does the assembled multitude use for DC power input protection?

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal
 Click to see the full signature
Reply to
Phil Hobbs
Loading thread data ...

Was the blown wall wart a switcher? The transformer versions are pretty robust, I've never seen one fail. If you use an SCR crow-bar have it blow a series fuse to protect the wall wart and other components from fire.

Reply to
Bob Eld

PTC implies a repetitive over-voltage ?:-(

For expensive equipment I always opt for big-ass-cap/SCR/fuse.

But I'm sure penny-wise-pound-foolish Joerg will have other ideas, and whines ;-)

...Jim Thompson

--
| James E.Thompson, P.E.                           |    mens     |
| Analog Innovations, Inc.                         |     et      |
 Click to see the full signature
Reply to
Jim Thompson

Yes, it was a switcher, and so will the one for the gizmo be--it needs

10W or so, because it has a bunch of temperature controlled things in it. It's switchers I'm mostly worried about, because I can't tell in general what a failed one is liable to do. A fried diode or broken regulator in a 60 Hz transformer circuit is a lot more predictable.

I'm really hoping not to use a one-time fuse for circuit protection, because for this gizmo that's as bad as a circuit failure. I also want to make it reasonably idiot-resistant--there is no standard for coaxial wall wart plugs, so who knows what somebody is going to connect to it?

Built-in AC supplies don't have a lot of these problems.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal
 Click to see the full signature
Reply to
Phil Hobbs

Big nasty spikes will be swallowed by the tranzorb and the humbuster (a parallelled rectifier/1-ohm R in series with the cold end of the supply). It's the failed or mismatched PS producing some horribly out-of-spec supply voltage that I'm most concerned about.

A 50V input will switch the polyfuse before the tranzorb desolders itself, but on a 25V input the polyfuse might be too slow--I really don't want to have to spec a 50W zener!

It would be sort of nice to have three-terminal devices for this, where the built-in zener could heat the polyfuse. That would probably fix this particular vulnerability.

Alternatively, since I do need to have linear regulators inside the box (it's one of my usual high dynamic range gizmos), I could do something amusing such as using the overvoltage as trigger a PUT to drag the bases of the pass transistors to ground. PUTs are great for that kind of thing--not too accurate, but no surprises. The two-BJT SCR trick works OK too but is beta-dependent.

In general I'm pretty sure I can roll my own successfully, but it has to be a common problem, so I'm probably missing something ultra-obvious.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal
 Click to see the full signature
Reply to
Phil Hobbs
[snip]

"Bases"?? Rolling your own linear's? You can do the same thing with "317-style" regulators.

If you want accurate trip-off, use TL431 with positive feedback.

...Jim Thompson

--
| James E.Thompson, P.E.                           |    mens     |
| Analog Innovations, Inc.                         |     et      |
 Click to see the full signature
Reply to
Jim Thompson

I wouldn't worry about the wart catching fire; thry're not supposed to do that.

But we often use a polyfuse followed by a unidirectional transzorb and a big cap. Observation: all polyfuses are sloppy, and the surface mount ones are the worst. Use one of those flat leaded parts and you can put other parts below it if things are tight.

John

Reply to
John Larkin

AFAIK the wall warts don't get UL or equivalent approval unless they can handle a dead short without frying.

'course, you just don't know about what's coming out of China these days, or how well they refrain from bursting into flame when _they're_ broken.

--
www.wescottdesign.com
Reply to
Tim Wescott

Nope, just doing same on a current design I am wrapping up right now. But on the output side ... now the old debate begins :-)

No PTC fuses on this here design, I don't like things getting freaking hot on my boards.

--
Regards, Joerg

http://www.analogconsultants.com/
 Click to see the full signature
Reply to
Joerg

They are supposed to resist bursting into fire no matter what you do with the output load. We just bought a (small) boatload of them, and they not only have approved materials for the housing, but also an approved thermal cutoff set at something reasonable.

Reply to
Spehro Pefhany

Yep. You can make crowbars that trip at 5.7V +/-2% or so. Done it many times. TL431 -> transistor -> big fat SCR (but doesn't need heat sink).

--
Regards, Joerg

http://www.analogconsultants.com/
 Click to see the full signature
Reply to
Joerg

The one that gets sent to the lab, then production units, could be two different things.

--
Regards, Joerg

http://www.analogconsultants.com/
 Click to see the full signature
Reply to
Joerg

Not roll-my-own linears exactly--cap multipliers with very slow VR loops. I need something like 120 dB ripple rejection at 50-100 kHz. I'll look at the TL431 idea--not knowing just what's inside, I get a little concerned about turn-on behaviour. PUTs I understand.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal
 Click to see the full signature
Reply to
Phil Hobbs

If the budget is there: LTC and others make overvoltage protectors for automotive that turn off a series pass element instead of brute-forcing some sort of fuse down to its knees. But it'll be a buck or more.

--
Regards, Joerg

http://www.analogconsultants.com/
 Click to see the full signature
Reply to
Joerg

I've used the LT4356DE-1 overvoltage protecto device. You need an external pass transistor. Although it looks like you can use this as a voltage regulator at first glance, don't be tempted to. The part clamps at a set point and shuts down if the input voltage is over the set point after some milliseconds. These parts are rated for 80V spikes.

--
Mark
Reply to
qrk

Turn-on behavior is outline in the graph section of the datasheet.

--
Regards, Joerg

http://www.analogconsultants.com/
 Click to see the full signature
Reply to
Joerg

And sometimes, sound effects! :-) Ed

Reply to
ehsjr

TVS+polyswitch (plus some series Z if the load will allow, maybe a series MOSFET or diode, depending), and-- the big one-- I find that galvanic isolation is often worth the hassle/cost/noise.

Reply to
Spehro Pefhany

Yep. I think Joerg has watched too many Batman episodes ;-)

...Jim Thompson

--
| James E.Thompson, P.E.                           |    mens     |
| Analog Innovations, Inc.                         |     et      |
 Click to see the full signature
Reply to
Jim Thompson

Nope, never was into SciFi. Too much stuff that didn't jibe with the laws of physics. MacGyver was more up my alley, and now we occasionally have a guide dog named MacGyver.

But I did have my dose of phsssooosh ... *WHADDABAM* experience in my younger years :-)

In hindsight some of those were scary. It's a miracle that I didn't lose my right eye as a teenager when a big can electrolytic performed a lift-off and subsequently made a crater in the ceiling. That's when I learned that there is something like an ESR and that one shall not push one's luck with ripple current.

--
Regards, Joerg

http://www.analogconsultants.com/
 Click to see the full signature
Reply to
Joerg

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.