Input voltage protection

Hi Guys,

I need to read automotive voltage using a Teensy. I need good accuracy between 10 and 20V and have set up the voltage divider such that the full range of 0-3.3V corresponds with 0-25V to provide some buffer for the zener errors. I just want to verify that the attached circuit looks good; my simulations seem to indicate the limiting factors will be the power dissipation of the 100K resistor and the zener diode. Thoughts?

If the Teensy is similar to the Atmega328, the input resistor should be 10K.

What are you trying to accomplish?

I want to prevent voltage transients from doing any damage. Automotive systems can have very large spikes from high current motors on the system.

I think you’re overthinking it.
The 2.2Megohm resistor is already protecting the input to at least 2000volt DC, because of the internal input protection diodes. The resistor might protest at these voltages though.
Spikes won’t get past the 2.2meg/100n lowpass filter.
The 100k/zener might introduce a measuring fault. Zeners are not perfect.
Those high resistor values are ok if combined with a 100n cap.
Leo…

Read this:- http://www.digikey.com/us/en/techzone/microcontroller/resources/articles/protecting-inputs-in-digital-electronics.html

Or / and this http://www.thebox.myzen.co.uk/Tutorial/Protection.html

Remember 200V spikes have very little power behind them, to be technical they have a high output impedance or to be colloquial they don't have much umph.

The DigiKey link is not working from here. I believe this is the same article... http://www.digikey.com/en/articles/techzone/2012/apr/protecting-inputs-in-digital-electronics

Link works here. Maybe you're on a blacklist, being watched...:) IMHO input protection diodes are not needed with these values. With normal operation, voltage stays in the input range of the micro. With constant overvoltage, the input current stays well under 1mA. And spikes have to be very long to change the voltage on the cap. Leo..

Thanks for your help guys. I was recommended by Paul at PJRC to reduce the resistance values as the Teensy should be able to take up to 10mA pumped in and it should reduce errors from the very high impedances. I was originally going to use a 1K in place of the 100K but was worried about blowing the resistor out. But I assume that since the transients don't have much "oomph" a surge current resistor should be alright with the occasion high voltage spike?

Average automotive (charging) voltage should not go above 13.8volt. Some (boat) charging systems have a fast/equalising charging system. Then the voltage can rise to ~15.5volt. If you see more than that on a 12volt lead/acid battery, it's stuffed. I would choose a 1:4 resistor ratio. That's 1/5 for the micro, or 16.5volt full scale on a 3.3v micro. 56k/180k or 68k/270k or 82k/330k all come close. Those values, with a 100n cap to ground, give enough protection. Read the link I gave you about using high value resistors.... Leo..

The automobile is a very noisy environment you can easly get spikes in the 50 V region. I would not use the internal static diodes to protect anything if you want any decent life out of the system.

True, but remember that there is a 100n cap in the circuit. Static voltage with a connected good battery will never get high enough with the values I have given. And transients can't come through the RC filter. A 180k/100n filter has a corner frequency of about 9herz. Even without cap and 50volt on that 180k, the current into the A-in pin is still under 200uA. Any internal protection diode can handle that. Leo..

Awesome information guys I missed the link on high input impedance the first time. I will go with 180/56K range resistors I think with a high power 1K resistor for the diode to reduce leakage. With the analog pins being able to handle up to 10mA injected I should have lots of headroom for spikes.