ADC input protection (again...)

Discussion in 'The Projects Forum' started by daman, Jun 8, 2016.

  1. daman

    Thread Starter New Member

    Jun 14, 2012
    15
    0
    Hi,

    I have a small project based on MSP432 where I need to read analog signals on ADC inputs. The thing with the inputs is that the sensors (operating in 0-12V range) could be 200-300 feet away (either in the building or in the ground) so I'd like to add a voltage divider (as I can only read 0-2.5V) and some protection... Here's my initial idea... The thing is that I'm not even sure if it makes any sense and I'm pretty sure this will not protect from big voltage spikes. Would it make any sense to put clamping diodes somewhere in this circuit (eg. a pair of 1N4148 just after R1)? Any other ideas (or completely different ideas)?
    Any help would be appreciated...

    circuit.jpg
     
  2. ErnieM

    AAC Fanatic!

    Apr 24, 2011
    7,388
    1,605
    Clamp diodes are a good idea. How big a spike will your small signal diodes survive? You may want to use something heftier.

    Is there some compelling reason you chose an op amp designed over 40 years ago?
     
    daman likes this.
  3. daman

    Thread Starter New Member

    Jun 14, 2012
    15
    0
    Well... I'd prefer to power the op amp with 12v (there will be several devices powered by 3.3v, a lot of op amps and I simply don't want to put much stress on my 3.3v voltage regulator). Another reason is that... I have a lot of them in my drawer ;)
    Could you recommend better op amp (that I can power with 12v)..? How about OPA2743 (or LM2902) - that should be OK I guess...?
     
    Last edited: Jun 8, 2016
  4. crutschow

    Expert

    Mar 14, 2008
    13,014
    3,233
    For that distance it would be much better to go to a differential signal scheme to minimize the problem of ground noise and loops as well as other types of noise and spikes. Sending a single-ended analog signal over that distance is very problematic.
    You can use a differential analog output amp (such as one of these) at the sensor end and a differential instrumentation amplifier (examples) at the receiving end.
    If you want to go as cheap as possible, you can use two standard op amps to generate the differential output and a standard op amp configured for differential input using 4 resistors, but the performance won't be as good.

    You would send the signal over a twisted shielded wire pair with the shield connected at both ends to the electronics ground.

    For spike protection you can use two diodes, one between the signal line (cathode) and ground (or the negative supply if used) and one between the signal line (anode) and the positive supply.
     
    daman likes this.
  5. daman

    Thread Starter New Member

    Jun 14, 2012
    15
    0
    @crutschow - the thing is that single-ended signal is a requirement here. Another thing is my resolution is rather small (I'm only going to need 4 bit resolution ie. 1v) and I'll be doing (I think) moving average from the readings... Not sure if it makes any sense (but circuits is rather far away from my profession... ;))
     
  6. joeyd999

    AAC Fanatic!

    Jun 6, 2011
    2,681
    2,742
    +1.

    In fact, a 4-20mA loop would be superior, IMHO, at these distances. This also solves the problem of getting power to the sensor(s).
     
    AlbertHall likes this.
  7. ronv

    AAC Fanatic!

    Nov 12, 2008
    3,290
    1,255
    A link to the sensor spec would help.
     
  8. andre_teprom

    Member

    Jan 17, 2016
    31
    9
    For that distance depending on the region where you live, if the occurrence of electric storms are not too rare, I would also consider additional protection against common mode interference, such as circuits with sparking elements. This would not indeed protect the equipment against real strong atmospheric lightning discharges neither turn it completely safe, but at least tend to reduce a few the extent of damage, restricting to the equipment itself.
     
  9. jsallas

    New Member

    May 17, 2016
    9
    1
    Hello!

    From my point of view 4-20mA would be the best as joeyd999 said.
    Differential amps would also be a good choice.
    If you want to protect from voltage spikes you should try TVS Diodes instead of regular diodes, this will ensure better response to transient events.
     
  10. crutschow

    Expert

    Mar 14, 2008
    13,014
    3,233
    Why is single-ended a requirement?
    Who made that requirement?

    Considering the low required resolution I also think a 4-20mA current loop is likely the best way to transmit the data.

    What is the highest frequency response you need, i.e. how fast does the signal change?
     
  11. daman

    Thread Starter New Member

    Jun 14, 2012
    15
    0
    Well - I did... I'm reusing existing cabling (btw - I don't have sensor specs - these are "custom" temperature sensors)...

    I'll try what I can do with existing stuff.. This will be next option if I fail miserably

    The thing is that I only need to read a few (eight to be precise) values (ie. I need to know if the temperature went outside of particular "zone" for a period of time. A "period of time" could be as long as 10 seconds). So I'm only interested in "average" over 10 seconds or so...
     
  12. daman

    Thread Starter New Member

    Jun 14, 2012
    15
    0
    Going back to op amp selection - can somebody recommend "modern" op amp that can be used in my "original" circuit above (buffer/voltage follower, 12v supply, low power and - if possible - pdip package)?
     
  13. ronv

    AAC Fanatic!

    Nov 12, 2008
    3,290
    1,255
    I think you could just add the schottky diodes after the 10K resistor. You might also add a 4.7k resistor from the output of the 224 op amp to ground. They do suffer from crossover distortion and this will take care of it.
     
Loading...