In the datasheet for the MAX15500, it says I need to use several diodes in series:

... to ensure a voltage difference of 2V to 3.5V from AVDD to AVDDO and from AVSS to AVSSO.

Diodes in series

It seems to be that a 2V- 3.5V Zener diode should be able to replace all 5 of these diodes. If so, why are Maxim suggesting such an overly complicated circuit? Am I missing something?

  • 1
    \$\begingroup\$ Indeed a zenerdiode could also be used. Maxim may have chosen to use diodes as diodes are more commonly used. There's a good chance there are already diodes used elsewhere in the design and maybe using a couple more is easier than finding a suitable zenerdiode. Also, most designers know that you could also use a zener instead so they would do that anyway if they preferred. \$\endgroup\$ – Bimpelrekkie Oct 27 '15 at 15:23
  • \$\begingroup\$ @FakeMoustache: One could turn that into the other direction. Use some simpler Zener in the design and say that most designers know that they can replace it by some diodes that already have at hand... \$\endgroup\$ – PlasmaHH Oct 27 '15 at 16:10
  • 1
    \$\begingroup\$ Low voltage (<6V or so) Zeners have rather more poorly controlled/softer curves. Those being supply pins, the current probably varies a lot... but then the datasheet doesn't seem to mind significant variation in this voltage drop... so that's probably not it. \$\endgroup\$ – Fizz Oct 27 '15 at 16:22
  • \$\begingroup\$ You have to be careful with the minimum current though. The datasheet gives only a typical and max current for those pins (5mA, 7mA) no word what the minimum is. You need to use a Zener that will be in regulation below that [unknown value]. There are very low current Zeners these days like SZL2.4A. \$\endgroup\$ – Fizz Oct 27 '15 at 16:40
  • 1
    \$\begingroup\$ @RespawnedFluff - The good thing is that I'm actually going to power 8 of these chips at once, so the current should be enough. \$\endgroup\$ – Rocketmagnet Oct 27 '15 at 17:02

Theoretically, you could use a Zener diode.

However, there are several practical reasons why you would NOT do so.

1) Low-voltage Zener diodes are very leaky. The threshold knee is soft and not well defined. You have to run a fair amount of current through them to reach the rated Zener voltage.

2) Low-voltage Zener diodes are hard to manufacture. The lowest-voltage Zener diode that comes to mind is either 3.0V or 2.8V - (I don't recall which).

For a low-power circuit, a string of series diodes performs much better.


As you know, a zener diode differs from "normal" diodes in that it will allow current in the reverse direction, but not until the voltage rises above a certain level. In your circuit, there might be some initial condition in which the diode is not conducting in reverse, even for a microsecond, and your circuit might have strange effects because of it.

With the "normal" diodes in the circuit, their behavior is invariant, so they will always perform the expected function.

I suspect that you could get away with using a zener, but there might be a 1% boundary case where the system would behave strangely.

  • 1
    \$\begingroup\$ The "normal" diodes used will not conduct in the forward direction until the voltage rises above the combined 'knee' voltage of the four diodes - so what's the difference, a zener diodes behaviour is just as "invariant". \$\endgroup\$ – Icy Nov 11 '15 at 9:20

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.