PDA

View Full Version : Zener on input pin OK ?



luminas
- 27th September 2011, 14:17
I want to limit the voltage on analog input pin, can I use a zener diode ?
my setup is as follow : v in ---> opamp --> rc filter ---> PIC
does it have side effect if I put zener between RC and PIC ?
Thank you

Charles Linquis
- 27th September 2011, 14:20
A Zener diode will not have a significant effect. However, you must consider that most diodes have leakage, and depending on the "R" in your R/C network, that leakage may introduce an error.

luminas
- 27th September 2011, 14:36
I am wondering , what will happen when a millisecond peak voltage above the zener limit hit the analog pin ?

Charles Linquis
- 27th September 2011, 15:09
What are you saying? The Zener is fast enough to catch a millisecond peak. Also, an analog pin won't be damaged with ANY voltage above Vcc - as long as there is current limiting to less than 10mA and you have a few hundred pf of capacitance on that pin to GND.
The cap is needed in case a picosecond pulse (spark discharge) has a chance to raise the voltage at the input pin faster than the body diode (the input pin to Vcc diode inherent in the PICs design) can start conducting.
But you already said you have an R/C network. That will be enough.

luminas
- 27th September 2011, 15:28
Thank you Charles, great information :)

I will not use zener as it is not required now

sougata
- 17th October 2011, 20:55
Hi,

Darrel and myself once tried to profile the leakage and breakdown of the zener diodes on the analog pin. For faster acquisitions of the s/h cap a low impedance path is always preferred to the ADC which can kill the PIC due to shootups or downs (negative).

I had been using zener without any problem cause I almost always use a temperature compensated reference for Vref mostly 2.5V. So the 5.1V zener did not come into picture. But Darrel did an actual test with a modulating input and found that the zener mis-behaves (turns leaky) near its threshold. I found ambient temperature variations could create large error in the measurement if you are using VDD (+5) as your reference.

So the bottom line is if you could lower your Vref much lower than Vdd then zeners of Vdd ratings can be a safe bet.