0
\$\begingroup\$

I recently bought one of those cheap digital clocks from my favorite Chinese superstore. Most of these clocks have one achilles heel: the power supply. They get their power either from a USB source OR from 3-4 AAA batteries. Unfortunately it's wired so that when you power it from USB, it passes some of that voltage on to the battery contacts. This is not regulated any way because it's not supposed to be a "charger". So your choice is to either run the clock from battery (and have to fiddle with the display every time you want to see it) OR run from USB (and have to reset the settings every time the power blinks). If you run it from USB with the batteries installed you run the risk of exploding the batteries.

I thought I could get around this by putting a diode in series with the battery contacts. To verify this, I did some testing. I used some jumpers to clip the 1N4001 diode in series with the positive battery terminal and then to a multimeter. To my dismay, the voltage was fluctuating around 100mV. I reversed the diode just to make sure I didn't make a basic mistake- when I did that I got 3.5 volts. So I put it back and hooked it up to a scope. I found that there was an odd-looking waveform and it was actually AC coming out the wrong end of the diode- between +100mV and -100mV to be exact. It was about 55Hz or so.

Why am I getting an AC voltage out of the wrong side of the diode? Furthermore, will 3 AAA batteries (@4.5v) be damaged if they are exposed to this small current for many months (years?) at a time?

Odd Voltage Waveform

\$\endgroup\$
3
  • 1
    \$\begingroup\$ Attach a load to the diode and report back, also, diodes do have leakage current, it is the famous Is in the diode equation, in a 4001 it is usually a couple of nA. \$\endgroup\$
    – S.s.
    Nov 8, 2020 at 3:12
  • \$\begingroup\$ You have a favourite chinese superstore eh? I dread to think which one is at the bottom of the list and how many times you used them before they hit bottom place. \$\endgroup\$
    – Andy aka
    Nov 8, 2020 at 11:14
  • 1
    \$\begingroup\$ How did you connect the scope? Thats not 55 Hz, thats too close to 16.6ms to be anything else than 60Hz mains voltage. Most likely the power supply is ungrounded 2 prong device and just has standard EMI capacitor which causes some AC coupling from mains input to low voltage output. \$\endgroup\$
    – Justme
    Nov 8, 2020 at 11:51

2 Answers 2

2
\$\begingroup\$

I found that there was an odd-looking waveform and it was actually AC coming out the wrong end of the diode- between +100mV and -100mV to be exact. It was about 55Hz or so.

Why am I getting an AC voltage out of the wrong side of the diode?

Your scope is picking up EMI.

When the diode is reversed biased it has a very high impedance, so electric fields can easily induce a voltage in the wiring at the 'wrong side'. Your scope probe also has a very high impedance, so it will show EMI that normally would be too weak to see.

The parasitic capacitance coupling the interference into the diode wiring is probably very small, acting as a high pass filter that accentuates rapid changes in the waveform. The peaks in your waveform suggest the original interfering waveform is a square wave, possibly coming from the clock display. Probe the clock display signals. If you see a frequency of ~55 Hz there too then it may be the source.

The current created by this EMI is very small, and since it is AC it will not charge or discharge the battery, so it is nothing to worry about.

\$\endgroup\$
1
\$\begingroup\$

They aren’t “supposed” to leak but diodes invariably do to some extent. Typically though the leakage current is very small. If you consider that a AA call has a shelf life of about ( years and perhaps 500mAh capacity then the self-discharge equates to about 10uA. If the leakage current is less than this (and with a 1N4091 it certainly will be) then I’d say that you have little to worry about. As an aside, alkaline cells are in face rechargeable, although they are only good for a few cycles.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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