Voltage divider before ADC MKL17Z256

[samc]

Hello,

In a design with a MKL17Z256 ( Vcc = 2V8), I use the ADC16 to read the voltage of a battery -3V3 to 4V2). So I must to place a voltage divider before the ADC to reduce the voltage. At this moment I use a 27K/47K but thses resistors are so big and I have a voltage drop in my ADC result (and the datasheet p32 says that a little resistance source is needed to drive the ADC). For example, instead of 4V15, I read 4V07. It's a bit much....So I will decrease these resistors values.

My question is : must the 2 resistors be low ? or just one of the two ? (which one ?)

In fact I would like to use the more bigger resistors possible to avoid current consumption through this voltage divider, it's battery powered design. Si if I can use a big resistor on 2 it will be good.

Thank you,

Sam,

[samc]

No I use 1% resistor but it's not important to be very precise. As you said the problem is to use low value resistor, that's consumes current. So to keep big resistor to avoid consumtion, a follower opamp is needed after the resistor divider to buffer the ADC not before.

Thank you for your help.

Sam,

[jorge_a_vazquez]

Hi Samuel Conti

I didn't know you were using a capacitor to filter the input signal, yes, the value of the net input resistance is important. if you calculated resistance value is too low and you want to keep high the resistance value, as you mentioned, you going to need a opamp to buffer the divider. Please tell us if you could resolve you issue and any question that you have.

Have a great day,
Jorge Alcala

-----------------------------------------------------------------------------------------------------------------------
Note: If this post answers your question, please click the Correct Answer button. Thank you!
-----------------------------------------------------------------------------------------------------------------------

[samc]

It's OK, I know the problem so I can resolve it. For the moment I will test with 47pF capacitor instead of 1nF, it will better.

Thank you,

[samc]

You said : "So, answering your questions, if your resistances are for the voltage divider there shouldn't affect the result"

But I constate that these resistors have influence and the capacitors too. In more when I use formula to calculate Rmin I found about 380 ohms (with cap 33pF) then the value of the resistors is important ?

[egoodii]

Your concern is valid.  The net input resistance for the analog pins is 'not high'.  I might assume you are using 0.1% divider resistors, and a 0.1% reference, and to keep input-current errors below that same 0.1% still means resistive-divider values probably 'too low' for sustained battery operation.  You might look into a micropower opamp to buffer the divider.

[jorge_a_vazquez]

Hi Samuel Conti

The RAS resistance source that you read in the datasheet refers to an external resistance taht is used with CAS to reduce noise, and it should have low value to keep the time constant RAS/CAS very low.

The reference manual says

"Leakage on the I/O pins can cause conversion error if the external analog source resistance, RAS, is high. If this error cannot be tolerated by the application, keep RAS lower than VREFH / (4 × ILEAK × 2N) for less than 1/4 LSB leakage error"

So, answering your questions, if your resistances are for the voltage divider there shouldn't affect the result, but this can be a noise related problem. Please check the reference design manual file and the following link for more information of how to improve the ADC signal disquisition.

Improving my ADC design

Hope this could help
Jorge Alcala

-----------------------------------------------------------------------------------------------------------------------
Note: If this post answers your question, please click the Correct Answer button. Thank you!
-----------------------------------------------------------------------------------------------------------------------

[samc]

Hi,

I don't think it's a noise problem. I use 2 another ADC pins for another signals without this problem.

This said, yesterday I tried with 1K2 /2K instead of 27K/47K and it works well.Further to your reply, I didn't precise that I have a 1nF to filter the ADC input (this capacitor is then in parallel with the 47K resistor). Is possible that the problem ? (I will test deleting this capacitor).

About your document, in the equation to calculate Rin : what's alpha  ? What's the unity of the varaiables (s, F, V) or is there factors ?

[samc]

The problem is the 1nF capacitor which is too large. I delete it and I read the correct value with 27K/47K resistors.