ADC measurement corrupted if some inputs are higher than 3.3

Go To Last Post
9 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Ok I have some opamp driven RTD to analog board I have designed....the board is powered by 5V. So the maximum output voltage when no RTD is connected, is about 4.2V (close to supply rail). The small voltage drop is due to the opamp output rail not going close enough to the supply rail.

Now this creates problems...I have 5 RTD inputs and hence all 5 outputs from the RTD board go direct to UC3A1256 ADC inputs which accept from 0 to 3.3V max (up to ADCref).

I noticed if I only connect one RTD sensor to my RTD board, and leave the other 4 open. I get 4 of the 5 ADC pins having about 4.2V each...ie 4 of the 5 ADC input pins coming into UC3A has higher voltage. This affects the firs input drastically. For example where I am expecting a raw value of about 200 at the first ADC input of the UC3A (due to the first RTD being connected), the UC3A is actually reporting a raw value of 550 ish...

I then disconnect the other 4 ADC inputs and my first ADC input shows the expected raw value of around 200.

So why does having some ADC inputs higher than 3.3V throws the first ADC input result off? I am thinking of having all ADC inputs go through a 3.3V zener diode and a resistor...

Whats your though guys?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I would use resistors divider instead of zeners as they are not very precise and will distort the readings.

When inputs go higher that VCC who knows what goes on inside the chip.

John Samperi

Ampertronics Pty. Ltd.

www.ampertronics.com.au

* Electronic Design * Custom Products * Contract Assembly

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I dont need it to be precise... I just need to it to start capping the volts to 3.3 if it goes higher...

Dont like the resistor solution as its hard to find accurate ones... I could I guess try the 0.1% ones...

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

1% zener?

Ok ...Try adjustable PNP clamp on the output of Your opamps

Try Zener clamp in the feedback path of Your opamps this might reduce the "softness" of the zener knee point

For more precision and repeatability arrange a comparator based clamp on output of Your opamp.

Use the 3.3V rail as reference for clamps

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Replace your opamp with a rail-to-rail output model and power it from the 3.3V.

Warning: Grumpy Old Chuff. Reading this post may severely damage your mental health.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I am using OPA2335....datasheet says full output swing should be just bellow 100mV from supply...so I might try the 3.3 supply option...let me try this tomorrow. Will get back to you guys.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

OK I made the supply 3.3V and amazingly the opamp holds the output to full range 3.3V!! wow...pretty good.

However, I have an opamp stage where by I am trying to feed constant voltage to my RTD supply. I get constant voltage using an LM336 2.5V reference and the OP2335. Surprisingly, there seems to be offset voltage at the voltage input :( it is in the order of 2-3.5mV even though the datasheet says 2uV offset! The 2-3.5mV is quite big considering I am only measuring signals between 100->150mV...

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Input saturation of analog muxes is a common problem. The usual 4051 family of devices suffer from this phenomenon as well. You can get specific muxes that avoid this problem.

As for your offsets, read the op amp datasheet carefully - there may be some gotchas regarding offset vs supply voltage. Also consider noise and voltage drops could also be contributing.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Is it possible to use the ADCref for the RTD supply? If the system is ratiometric errors may be less.

It all starts with a mental vision.