11
\$\begingroup\$

Followup to my previous question.

I'm designing a product powered from a 5V supply that feeds it into a TLV1117LV33 LDO. Typical current draw for the project is less than 500mA, but the regulator is cheap and stable with ceramic output caps so I like it so far. Anyway, I was advised in the previous question to put some ESD and transient protection on the Vin line, which makes sense to me, so the solution I'm thinking of is a PTC polyfuse in series, followed by a standard unidirectional TVS diode to ground.

However, all the diodes I've looked at (eg Littelfuse SP1003, Vishay GSOT05C) have quite a high clamping voltage. I understand that this is standard, ie that the maximum clamp voltage of a TVS diode will be a lot higher than its max reverse standoff. I'm also aware that lower clamping voltage and lower dynamic resistance is considered a figure of merit for TVS diodes.

However, my regulator's safe operating input range is up to 5.5V (6V absolute max). These diodes have 5V reverse standoff and typically ~7V clamping voltage at 1A (~12V at 30A). That's gonna fry my regulator easily! I've seen devices with lower clamping voltages, but when I say "lower", I mean 10V as opposed to 12V (at 30A). How do I reduce the clamping voltage so that, if an inductive transient appears on my 5V rail (or something similar), my regulator will survive?

Devices like MOVs seem to have even higher clamping voltages, so as far as I can tell, that's not a superior solution. I've also heard a bit about multi-stage ESD protection, but no advice on how to do it. Am I overthinking things?

\$\endgroup\$
3
\$\begingroup\$

You can look at Zener diodes instead of TVS diodes for lower clamp voltages (but worse power handling.)

How bad an input spike on the 5V rail would you want to worry about? Or do you have higher input voltage? You could put a non-LDO regulator with a high voltage rating (like LM350 or whatever) before the 5V, so that the 5V will already be well regulated. A normal TVS will clamp well below the 35V rated input voltage of the LM350.

However, if your power input, which may contain the spike, has to be 5V, then this won't work.

Another option is a resistor/Zener. A 2 Ohm resistor will drop 1V at 500 mA. Follow that by a 5.1V Zener and your regulator, and precede it with the TVS, and you'll have a somewhat more robust system. The TVS takes care of the really bad spikes; the resistor+Zener takes care of getting the voltage "in range" without blowing up, and the regulator can then regulate.

schematic

simulate this circuit – Schematic created using CircuitLab

Disclaimer: I'm a hobbyist. If someone who does this for a living tells me otherwise (and why/how,) I will gladly concede I'm wrong :-)

So, consider a 7V spike makes it past the TVS. This means the Zener needs to be rated for 1A temporary dissipation to survive.

\$\endgroup\$
  • \$\begingroup\$ My device is being powered from a phone charger style supply, so it's fixed at 5V@1A. Daisy chaining regulators will probably involve too much dropout, because there's not that much headroom between 5V and 3.3V. Series resistors are also out of the question - if it's big enough to attenuate a spike, it'll also incur big power losses and push the LDO into dropout. One other thing - if I could use a higher input voltage, I probably wouldn't bother with a linear reg at all. Past about 8V, the losses from heat would be high enough that I'd get a switcher, which could withstand a 12V input. \$\endgroup\$ – tummychow Aug 21 '13 at 4:51
  • \$\begingroup\$ The idea of the secondary zener is interesting though. Has anyone had experience with that technique? \$\endgroup\$ – tummychow Aug 21 '13 at 5:15
  • \$\begingroup\$ Your assumption that the series resistor will cause loss to drop the secondary regulator out goes against Ohm's law. There exists a resistor where the max draw of your device (500 mA) will drop the input voltage (5V) less than the safe margin for the LD117V33. 2 Ohm is safe, as you can verify yourself. Follow by a second Zener, and you're good! \$\endgroup\$ – Jon Watte Aug 21 '13 at 21:51
  • \$\begingroup\$ 2 ohms incurs a 1V drop on the line. Combined with variances in USB power specifications (as low as 4.75V is considered admissible, and I wouldn't be surprised if my fake USB wall supply was sagging to 4.5V), the regulator might now be pushed to marginal operation. At 500mA its dropout is approximately half of what it is at 1A, ie ~250mV, so 3.5V Vin is definitely cutting it too close for my tastes. \$\endgroup\$ – tummychow Aug 22 '13 at 0:33
  • \$\begingroup\$ So why are you worrying about large voltage spikes? Add a 220 uF capacitor and a 5.1V Zener on the input and you're fine for all normal uses in that described use case. \$\endgroup\$ – Jon Watte Aug 23 '13 at 0:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.