SAR ADC Input Types Pseudo Differential ADC

[newbie_hs]

I was going through this document, which describes different SAR ADC input types.The below

image is taken from the same document.

I have couple of questions regarding this configuration.

1. Can I ground IN+ and connect the input to IN-.Any special care is required for this
2. Why this configuration is required, it is not giving the full benefit of differential signaling
3. Can you please tell some real life examples where this type of configuration is used

 

[danadakk]

1) -

2) Basically to do single ended measurements with fixed ground as reference.

3) Consider a DVM. Most of the time you ground the ground probe and measure with just the
other probe, thats a single ended measurement. But if you have a signal with a lot of common
mode using the differential configuration can remove the CM V from the measurement when using
the differential configuration. For example measuring a shunt in high side of a load to get at current.
Using differential will give you just the Vdrop of the shunt. If you did a single ended mneasurment
the reading is Vdropshunt + common mode V.


Regards, Dana.

 

[KlausST]

Hi,

Can I ground IN+ and connect the input to IN-.Any special care is required for this

Nothing special, just the usual considerations like input voltage ranges.

--> Indeed the given document is very clear in this by stating: "The IN+ input has the actual analog input signal, while the IN– input has a restricted range.

Why this configuration is required, it is not giving the full benefit of differential signaling

It is not required. It is provided.
It is the more economic solution to a true differential signal input.

--> if you need true differential input, then use true differential input
--> if your IN- is close to GND (often used to rectify GND_bounce_errors), then use the "pseudo-differential" solution.

Can you please tell some real life examples where this type of configuration is used

4 wire measurement on a GND-shunt. (Kelvin wiring)

Klaus

 

[newbie_hs]

(often used to rectify GND_bounce_errors)

May I know will GND bounce errors happen in Analog measurements

 

[KlausST]

Hi,

May I know will GND bounce errors happen in Analog measurements

If I´d say NO: Wouldn´t I contradict myself? .. since I already gave a "real life example".
***

And because you mentioned "analog" measurements: GND bounce is way less critical with digital measurements.

Klaus

 

[danadakk]

May I know will GND bounce errors happen in Analog measurements

Yes generally. A single ended operation is generally not disabling the differential input,
and if the bounce is substantial both in magnitude and width to the input sampler
you can get errors.

If you are using a micro with a SAR on it one consideration is to turn off all GPIO activity
to help mitigate the bounce. Or use averaging to help remove noise, but keep in mind
averaging degrades if noise has correlated components in it. Which micros have as clock
is central to most peripherals.

Also when doing bypassing work Polymer Tants best for bulk bypassing, and ceramics or
MLC for the high frequency stuff.

A good way of looking at your noise situation is put your scope on infinite persistence,
normal edge trigger at a low V, say 10 mV or so, and watch the pk-pk noise build up on
display.


Regards, Dana.