Hi all,
i googled over technical paper about protection for ADC when it power down but i didn't found any clear documents.

Usually datasheet says that we can't apply input voltage higher than Vref/Vcc + 0.5V. Then, if an hypothetic input is tied to 3V while ADC is power off, what happen and does we have to protect ADC?

Example, Arduino board and similar haven't any protection on analog input, is it right?

If protection is needed, how do you make this protection circuit?

Thanks

Regards

Marco

 

http://www.ti.com/lit/an/slaa593/slaa593.pdf

 

Thanks kermit, i already read it, but in my opinion it doesn't show right solution.
Suppose i've to measure battery value, i can't drain all power through clamp diode while ADC is power off, neither with AO that solution simply moves the problem into AO input pin.
Marco

 

Hi all,
i googled over technical paper about protection for ADC when it power down but i didn't found any clear documents.

Usually datasheet says that we can't apply input voltage higher than Vref/Vcc + 0.5V. Then, if an hypothetic input is tied to 3V while ADC is power off, what happen and does we have to protect ADC?

Example, Arduino board and similar haven't any protection on analog input, is it right?

If protection is needed, how do you make this protection circuit?

Thanks

Regards

Marco

That is a good question, I would say that you would have to first check how the power switch disconnects power from the circuit AND what the MCU does to the multi-function pin that accesses the ADC. It may be floating or protected.

I have never heard of anyone seeing damage to the ADC from simply powering down the MCU while still connected to the sensed voltage. I am sure there would be all types of warnings and more information on the web if it was a problem. I think you will be ok.

 

The part about using a 'buffer' op-amp is not useful?
the inputs would not draw any power when the circuit was powered off, if the sensed input was input into a buffer before the ADC. It is the logical thing to do IF your concern is about input signal causing problems in the ADC.

 

In my case i'm planning to use ADS7951 as ADC. I'm agree with you i never broken an ADC in this way too. But it sounds very weird, some productor suggest to use analog switch to protect input, but i think they are useful only for high voltage.
Maybe S/H switch is opened when it power off, saving other circuits. there are also channels selector mux before S/H, i don't know how it can behave.

 

The part about using a 'buffer' op-amp is not useful?
the inputs would not draw any power when the circuit was powered off, if the sensed input was input into a buffer before the ADC. It is the logical thing to do IF your concern is about input signal causing problems in the ADC.

I think so, but AO also have datasheet constraint for V+ and V- have to remain into VSS-0.5V and VDD+0.5V. In this configuration probably ADC is safe but we could break AO.

 

You haven't told us anything about the characteristics of the signal you are converting with the A/D, and those details are important in selecting the protection method. Signal amplitude peak-to-peak, AC or DC coupled, bandwidth, source impedance...they all matter.

ak

 

Ops, Here you are:
Input signal: LiIon Battery Min: 2.8V Max 4.2V R = 80mOhm DC Coupled Bandwith: no matter
Filter: simply RC filter R=1k C=100nF
ADC: ADS7951 VRef =5V SAR 12bit 1MHz


Suppose that ADC is power off and battery still connected to ADC.

 

Change the series resistor on the RC filter to say 100kΩ. Then there's only a small current when the ADC power is off that shouldn't significantly drain the battery or damage the internal protection diodes of the ADC.

A more complex solution would be to put an analog CMOS switch in series with the ADC input that is powered from the battery but controlled by the ADC voltage.

 

With big series resistor i need also big capacitor altough input impedence is too high for charghe S/H capacitor in few seconds. However,i think this is a good solution for Op.Amp. input protection.

For ADC best solution seems to be CMOS switch biased from ADC power supply.

 

With big series resistor i need also big capacitor altough input impedence is too high for charghe S/H capacitor in few seconds. However,i think this is a good solution for Op.Amp. input protection.
................

The series resistor doesn't affect the S/H charging time, only the time to charge the capacitor for a change in the input voltage..
The S/H capacitor is very small so the 100nF is plenty large, independent of the R value.