After establishing that we're not going to use our load bank in PWM mode (see here), I'm now trying to draft up a circuit that would be controlled by an µC (Arduino or else) and a multiplexer (this).

I'm also trying to protect the input from reverse voltage and overvoltage / making it able to cater with 1,8-24V digital input.

OK so here's my design thinking

  1. 1N4148 for reverse voltage protection
  2. 1,8V Zener diode for overvoltage protection (able to take 24V as a digital signal)
  3. Resistor for current protection (500R ==> 10mA@5V)
  4. This one I'm really not sure about : a 0,1µF in order for the whole contraption not to turn-off between two "cycles" of the multiplexer. I know that'd definitely be a no-go in a fast-switching application but that's not the case there... I'm also guessing that the MUX chip will not change an output state between two update cycles but...

enter image description here So anyway, I hope I'm not completely off on this and wait for your feedback.

So far what I take away :

  • My // diode is useless
  • I could use a P-channel transistor instead of a diode for reverse protection (avoids the diode Vf). Say something like this

But it also brings one question to my mind : would such an input protection be compatible with some high-speed input? Wouldn't the rise time of this p-channel transistor also add to the whole chain?

  • I cannot use a resistor to limit the current since my input voltage has a wide range

Where I'm still lost :

  • What should I use to limit current ? :-)
  • Is a zener diode used this way (through it's breakdown voltage) will really work at 10-40mA? If I refer to the datasheet Zener diode datasheet the Vz is way above the nominal 1,7V when above it's nominal 50µA. It's way above 3V at only 10mA
  • Would maybe just using a LDO such as this one solve all of that?

Thanks for the help!

Hi everyone,

OK so I'm still not confident in my zener overvoltage protection...

After some reading, I found that I have to make sure the zener is always going to be conducting some current or it might not be very effective at all... So I'm going to add another resistor in // with the zener to ensure this minimum current. Not just completely sure that's the right way to do it.

  • 2
    \$\begingroup\$ Last I checked, Zeners work just fine in a forward direction. You could easily cut the 1N4148. \$\endgroup\$
    – Dave
    Dec 28, 2017 at 23:29
  • \$\begingroup\$ Hi @Dave. You're right, I completely missed out on this... Thanks for the comment! \$\endgroup\$ Dec 29, 2017 at 7:54
  • 1
    \$\begingroup\$ Interest only: Zeners tend to have a noticeably higher forward voltage than standard silicon diodes. From memory 1V fwd drop is not unexpected. With some devices the higher voltage may matter (but almost certainly not here). \$\endgroup\$
    – Russell McMahon
    Dec 29, 2017 at 9:34
  • \$\begingroup\$ Very minor detail: if you make the schematic horizontally more compact (shorter traces and shorter space between components, U1/Q1 for example), the image will probably look bigger and clearer. (Or you can use the embedded editor) \$\endgroup\$
    – Wesley Lee
    Dec 29, 2017 at 23:07

1 Answer 1


Your Specs: (my summary)


  1. Input from 3.3V or 5V uC output.
  2. input reverse voltage protection.
  3. 1,8~24V digital input ( if possible)


  1. External 12V supply to series low-side switch , galvanic isolated from input.
  2. Output switch RdsOn ~0.0012ohm

IR opto LED characteristics

enter image description here

All LEDs can be simplified by \$V_f = V_{th}+I_f*ESR \$ From the (typ) graph, \$V_{th}=1.36V~ @T_c=25'C ~~ ESR = 2.86Ω~ \$(= 0.1V/0.035A)
However there is a large tolerance on this as indicated by VOM1271 datasheet table LED forward voltage \$V_f= 1.2V_{min}, ~1.4V_{typ}, ~1.6V_{max} @I_f = 10 mA ~~\$

Comments on proposed design in question

  • It is the wide tolerance on ESR that affects Vf vs If as well as Temperature shift in threshold voltage Vth, that makes using a fixed series R not suitable for a wide input voltage range when a limited current range is desired.
  • An input 3.3V to 5V is possible but one must analyze the ESR of the logic voltage and include this in the voltage drop using Vol/Iol.
  • for reverse voltage protection rather than shunt diodes, you use a series Shottky diode or seriously prevent reverse voltage by design.


Review example specs for Automotive Smart low Side switches, there are lots of these with better performance.
Then revise your design specs as req'd and start over. eg.



  • \$\begingroup\$ Hi Tony. Thanks for the answer. A couple of precisions though : I'm trying to make the input compatible to the widest voltage possibles, from 1,8V to 24V. But as I see it with ESR tolerance problem, that's gonna be tough... Also, the 10mA was pretty much out of the blue, I'm seeing that it can take up to 40mA and in that case, I'll up the current as long as I don't risk frying it, it should also speed-up turn-on... Regarding the Output, I'm controlling a 36V / 35A load which is too high for these automotive low-side switches... Too bad because they're pretty great ! \$\endgroup\$ Dec 29, 2017 at 7:48
  • \$\begingroup\$ Regarding reverse voltage protection, could you explain in what way is a series schottky superior to a // shunt diode? I guess in the shunt diode, you risk frying the supply, right? But in series, you have to swallow the diode Vf... :/ I was also looking at this way of doing with a small P-channel transistor, what's your take? \$\endgroup\$ Dec 29, 2017 at 7:50
  • 1
    \$\begingroup\$ Correct. better with low RdsOn losses \$\endgroup\$ Dec 29, 2017 at 15:26

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.