0
\$\begingroup\$

Background

I have now three smartphones (SGS3 [neo], max 77686 or so) lying around, all more or less dead due to a broken voltage regulator. Since I am not a fan of coincidences, I suspect a broken wall wart somewhere around in my apartment, but I not only have too many to just throw them all away, I also currently don't have a scope to measure things.

Required Idea

Since I am suspecting transient too-high voltages that destroyed the S3s regulator, I want to build a simple circuit that can detect those. For normal DC (or rather low frequency AC) I can come up with some ideas (monoflop, comparators, zener etc.) that can be the base of a circuit, but I have no idea about how fast the transient can be.

Since those are all smpss, I think their working frequency is at most 200kHz so the voltages to detect should probably be in that timing area. I don't have much experience in what circuits would be a good idea to detect the voltages that I am searching for fast enough.

So what would be a good design idea for a circuit that detects, lets say anything above 5.5V that lasts at least 5µs (or if you have experiences with mentioned voltage regulators smaller times when they are likely to damage the regulator circuit)? Although it is sylvester soon and everyone appreciates some blowing up tantalum capacitors due to overvoltage, I would prefer something reusable with not too much components, that is also nothing that requires a microcontroller or similar.

\$\endgroup\$
  • \$\begingroup\$ It could also be a charger that has a current limit too high for your phone, or one that doesn't play nicely. From a 5 V USB port, my phone charges at 0.5 A. Slow charge from a strong source, ok. But on a compatible charger, it connects the lithium battery directly to the charger output and charges at 1 A, only ~4 V, and the charger is now limiting the current which is smart, it's already a switch mode supply. A charger that thinks you have an iPad wanting 2 A might fry your phone. \$\endgroup\$ – tomnexus Dec 27 '14 at 14:54
  • \$\begingroup\$ @tomnexus: All have below 1A and the manufacturer one is rated 1A so it won't fry per se the phone. They have been serving phones here for many years now, and one of those that died was relatively new, so I think there is really one broken. How it is broken, I don't know (nor which one it is) so I start with the most likely failure mode that comes to my mind. \$\endgroup\$ – PlasmaHH Dec 27 '14 at 16:13
  • \$\begingroup\$ Have you checked that the outputs are nominally around 5 V? (Slightly loaded, that is.) \$\endgroup\$ – copper.hat Dec 27 '14 at 17:02
  • \$\begingroup\$ @copper.hat: Yes, but I only have a multimeter at hand. It won't detect if there is a 250V spike for 10µs \$\endgroup\$ – PlasmaHH Dec 27 '14 at 17:41
  • \$\begingroup\$ I presumed so, but I often check the obvious last :-). \$\endgroup\$ – copper.hat Dec 27 '14 at 17:42
1
\$\begingroup\$

If I were going to build a detector device such as you propose I would spend some time to make it really useful. As such I think you should consider the following list of features.

  1. Have a fast voltage spike detector that detects spikes in the > 6 volt range.
  2. Have a fast voltage spike detector that detects spikes that go negative more than -1V.
  3. Equip the circuit as a pass through device - charger input - USB cable output to device.
  4. Provide an output clamp circuit that clips any spikes on the output to voltages 5.6V.
  5. Provide a negative clamp circuit that clips any negative spikes on the output to -0.6V or less.
  6. Provide an LED that lights when there is either a positive or negative spike detected.

Additional nice features if you could see fit to equip the device with an MCU.

  1. Monitor the charging current through a current shunt and switch off for short or overload current conditions.
  2. Monitor voltage to make sure that it stays nicely within a range of 4.75 to 5.25V.
  3. Add a small LCD display to show unit status such as voltage, current, spike detect count and a charge time counter.
\$\endgroup\$
  • \$\begingroup\$ That really sounds like a product people would pled for on kickstarter... anyways the main problem I have is with thinking of the fast part here (100ns as per comment elsewhere). \$\endgroup\$ – PlasmaHH Dec 27 '14 at 18:44
0
\$\begingroup\$

Are you preferring digital or analog? If analog then you need a peak detector and a latch -one PNP and one NPN transistors. If digital - then two resistors, a transistor and an Arduino-like device.

\$\endgroup\$
  • \$\begingroup\$ updated the question \$\endgroup\$ – PlasmaHH Dec 27 '14 at 16:04
  • \$\begingroup\$ I read it again and sounds like a fast comparator, RC-chain to filter out too narrow peaks and a trigger should do fine. \$\endgroup\$ – Ilia Dec 27 '14 at 16:33
  • \$\begingroup\$ If the spike peaks on the charger are high enough voltage they could fry the target device even with a few hundred nanosecond width pulse. \$\endgroup\$ – Michael Karas Dec 27 '14 at 17:33
  • \$\begingroup\$ @MichaelKaras: So the "specification" of my "detector" should be for spikes of, say, 100ns to be safe? \$\endgroup\$ – PlasmaHH Dec 27 '14 at 17:45
  • \$\begingroup\$ I would recommend that you design for spikes that are as narrow as you can detect without spending huge amounts on high end components. 100nsec spikes are possible on the output of a faulty switch mode charger (or a cheaply built pos). Many modern comparators should be able to detect pulses in the couple hundred nanosecond range or less with careful circuit design. This may be a part that costs 10x the dirt cheap garden variety comparators but still well within reason when you consider the pretty price tags on smart devices these days. \$\endgroup\$ – Michael Karas Dec 27 '14 at 18:15
0
\$\begingroup\$

Quick and dirty way: add a 4700uF (or bigger, the bigger the better actually, as long as it is rated at higher than 5V and can hold off transient spikes) cap and 1K bleeder resistor between the power rails and the cap will smooth out those spikes.

And if that cap blows, you have too much a spike.

\$\endgroup\$

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.