Author Topic: Overvoltage clamp/limiter  (Read 1868 times)

0 Members and 1 Guest are viewing this topic.

Offline ubbutTopic starter

  • Regular Contributor
  • *
  • Posts: 100
  • Country: de
Overvoltage clamp/limiter
« on: June 02, 2020, 10:31:25 am »
Hi,
I have a circuit that needs to run from a 3.3V to 5V power supply and I want to protect it against over voltage. (Should withstand 15V). Current is below 100mA

I am therefore looking for a simple circuit that passes voltage below 5V with a maximum drop of 100mV and Clamps the supply voltage to 5V.

My first thought is to use an LDO, unfortunately most datasheets do not specify operation below the normal output voltage. And when they do (for example MIC5504 or similar) it is only for the 3.3V part, not the 5V part.

I also found other solutions, like the LTC4360, but it turns off the output completely rather then clamping it.

Also found the TPS25200 which looks very good, but seems overkill given its 2.5A max operation. Wondering if there is a simpler and cheaper solution ?
The datasheet also states:
Quote
TPS25200is designed for 2.7 V < VIN< 5 V (typical) voltage rails.While there is a VOUT clamp,it is not intended to be used to regulate VOUT at ~5.4V with 6 V < VIN< 7 V. This is a protection feature only.
Might be fine because of the low current I need, but certainly not optimal..

thanks
« Last Edit: June 02, 2020, 10:35:23 am by ubbut »
 

Offline Jeroen3

  • Super Contributor
  • ***
  • Posts: 4209
  • Country: nl
  • Embedded Engineer
    • jeroen3.nl
Re: Overvoltage clamp/limiter
« Reply #1 on: June 02, 2020, 11:21:27 am »
The TPS25200 will clamp until thermal limit. Which may be hit very quickly depending on circuit load. Just like a regular linear regulator.

But I think you are contradicting yourself here a bit.

You want overvoltage protection for a 3-5 Vdc circuit, which suggests there can be a failure condition that you want to harden your circuit against (eg: no damage).
Then you also write that you'd like to keep operating at 15V.

To me it then sounds like you're looking at a circuit which has a 3-15Vdc input that needs a voltage regulator.

There are parts for protection, like the NCP361. (UVLO might not fit here)
But I guess you'd be better of choosing the right power supply strategy with a buck-boost converter for example.

If all you want is to protect during a short non-repetitive temporary overvoltage and keep operating a power resistor and zener diode may be used.
I think you may also abuse an LM317 set at 6V for this, but I never used much LM317, you could also make one yourself with a depletion fet and zener diode.
 

Offline ubbutTopic starter

  • Regular Contributor
  • *
  • Posts: 100
  • Country: de
Re: Overvoltage clamp/limiter
« Reply #2 on: June 02, 2020, 11:26:52 am »
Quote
To me it then sounds like you're looking at a circuit which has a 3-15Vdc input that needs a voltage regulator.

The circuit needs to run at a voltage close to the supply voltage (which can be between 3..5V), but needs to be protected against a failure mode (user connecting it incorrectly).

So a boost, or sepic with a fixed output voltage would not work (because the voltage needs to be matched), neither would any circuit that only protects against short time over voltage.



 

Offline Jeroen3

  • Super Contributor
  • ***
  • Posts: 4209
  • Country: nl
  • Embedded Engineer
    • jeroen3.nl
Re: Overvoltage clamp/limiter
« Reply #3 on: June 02, 2020, 11:27:53 am »
Does it need to remain operational in the failure condition?
 

Offline ubbutTopic starter

  • Regular Contributor
  • *
  • Posts: 100
  • Country: de
Re: Overvoltage clamp/limiter
« Reply #4 on: June 02, 2020, 11:29:06 am »
Does it need to remain operational in the failure condition?

Would be great, but not necessarily
 

Offline ubbutTopic starter

  • Regular Contributor
  • *
  • Posts: 100
  • Country: de
Re: Overvoltage clamp/limiter
« Reply #5 on: June 05, 2020, 09:55:32 am »
For those interested:

I tested this LDO: NCP551 and am quite happy with it.

Dropout voltages are a bit worse than I speced above, but I think I can live with this:

Input voltage 2..5V, Drop is:
140mV @ 50mA
280mV @ 100mA

It also goes into thermal shutdown quite nicely and recovers without issues. I also got a lot of respect for the margin of the specs. It claims a maximum input voltage of 12V, but happily works up to 23V !
Between 23..25V, Vout = Vin-23V, but the part is still working :) Remarkable
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22436
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Overvoltage clamp/limiter
« Reply #6 on: June 05, 2020, 03:57:34 pm »
Yeah, simply regulating it down with an LDO will do.  You need enough voltage rating of course.  This comes to mind:
https://www.digikey.com/product-detail/en/on-semiconductor/NCV4274AST33T3G/NCV4274AST33T3GOSCT-ND/1967234
Though maybe its low voltage startup isn't enough, I haven't checked.  The 40V rating is typical of automotive parts, also see offerings by Infineon and others.  Many of which tolerate negative input and output, too.

15V dropping to 3.3V is 12V, at 100mA is 1.2W -- a small SMT part won't be able to sustain that for long.  If this is a momentary overload situation, that's fine; if the reg has a thermal limit built in it will cut off the load to reduce power, then cycle.  The SOT-223 above, will handle it continuous, under typical conditions and given some thermal pours.  SOT-89 similarly, SOIC-8 maybe maybe not, and anything through-hole with a heatsink tab definitely.

For more demanding applications, you might construct your own out of bigger parts -- a depletion mode MOSFET for example, can be used as a normally-on switch, biasing its gate to drain with a resistor.  (It's wired as a source follower.)  With a zener from GND to gate, its gate voltage is clamped at some voltage, and therefore its source will never be many volts above that.  With a precision reference (TL431 or the like) this can be improved to a proper accurate threshold (given some limitations in response time, which might be addressable with filtering).  I'm not sure how big DMOS are available (10s of W, maybe not 100s?), but they are available in quite high voltages, and can be stacked (cascoded) for arbitrarily high voltage standoff!

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf