Announcement

Collapse
No announcement yet.

Analog Input Loading - Redux

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Analog Input Loading - Redux

    Forwarned by the work of others, I decided to see if there was an alternative to removing the 560 ohm resisitor pack in the ADIO-100 as discussed elsewhere. It turns out the choice of the PIC 18F8455 as the processor in the ADIO-100 drives the requirement for a low impenance analog source. The 18F2455 specification can be found here http://ww1.microchip.com/downloads/e...doc/39632e.pdf; Section 21.1 states "The maximum recommended impedance for analog sources is 2.5 kOhm". The rationale appears to be related to the design of the on-chip ADC circuitry.

    The ADIO-100 560 Ohm voltage divider does two things, first, it drops the 10VDC maximum ADIO-100 input voltage to 5VDC which is the nominal maximum input voltage for the PIC 18F8455. Second, it guarantees the analog input impedance will be low, but, as noted by others, it loads down the analog sources such that the ADIO-100 measured voltage is not what the analog output of the sensor actually is.

    An alternative to removing the resistor pack (and therefore the voltage divider) is to interpose a buffer amplifier between the analog sensors and the each ADIO-100 analog input. The buffer amplifier presents a high impedance input to the analog sensor and low impedance output to the 560 ohm voltage divider. The buffer amplifier is simply an Op-Amp configured as a voltage follower (unity gain non-inverting); in this configuration, the powered Op-Amp will pull whatever power is required so that its output voltage matches the it's input voltage from the analog sensor.

    The downside of course is that you have to build this. The unregulated output of the ADIO-100 is a convienent power source for a 7810 voltage regulator to provide 10VDC to power the buffer amplifiers and anything else required; in my application I have 4 copper clad thermistors which need a power source so the 10VDC is used for them as well. The use of 10VDC also makes working on the circuitry somewhat safer with respect to the ADIO-100 (shouldn't be able to kill the analog inputs with over voltage), but makes the selection of the buffer amplifier somewhat more complicated.

    Since the range of input to the buffer amplifier is 0 - 10 VDC, the desired output of the buffer amplifier is also 0 - 10VDC. However, using 10VDC as the power supply to the buffer amplifier requires the uses of a class of amplifier known as 'rail to rail' (e.g. AD820 or LMC8482). This type of amplifier can swing its output to within 20 millivolts or so of its 'rail voltages' (0 VDC and 10VDC in this case); a typical off the shelf op-amp will only be able to get within 1-2 volts of its rail voltages.

    I did test the thermistor without and with the buffer amplifier; without the amplifier, the ADIO-100 measured voltage was significantly in error compare to a DVM as predicted. With the buffer amplifier the ADIO-100 measurement was on the money.

  • #2
    Funny you should discuss this as I was looking at ways of buffering the analogue inputs (I have posted this question on this forum, but no response) and also trying to eliminate the noise and sensitivity of my inputs. I'm using LM335's to measure temperature, but once over few meters of cabling they are getting to sensitive to noise.
    I did some googling and read about voltage followers - so one of my next experiments was to make one using a general purpose op-amp.

    Comment


    • #3
      <<but once over few meters of cabling they are getting to sensitive to noise.>>

      I had a look at a spec sheet for the LM335 family (revised March 2013 http://www.ti.com/lit/ds/symlink/lm335.pdf). I looks like the sensor is 10mv/C so I would expect your temperature values will be moving in 1 Degree C increments since the ADIO-100 minimum increment is 0.01 volts.

      The spec sheet also has some wire guage data, it looks like 400'-800' should be doable with 24 AWG.

      That said, sensor wire routing and shielding can be important. Are the sensor wires running parallel to any AC wiring? If you are using shielded pairs, is the shield grounded at only 1 end?

      Comment


      • #4
        How about a 4 channel isolator and voltage divider board, hobbyist grade board for 19.95 at MP3.com You can add resistors to change the measured voltage and scale up or down. It is shipped for 15 volt, since it's used for car stuff, but this says it provide isolation and gives scale to a 0-5 volt output.
        store.mp3car.com/4_Channel_Input_Isolator_and_Voltage_Divider_p/com-118.htm (add http:// to the front of the above link since I'm not worthy of posting links yet.) I've got 2 for this purpose, and have used them for such an occasion. Now would it then be wise to modify the ADIO100 to allow full scale representation and proper loading?
        Please advise if this is considered best practice for analog input protection & isolation.

        Comment


        • #5
          That 4-channel input isolator and voltage divider will be my solution too. Thanks for the link.

          It's isolator appears to be optical which is more than I hoped for. It also is not linear per the charts at //www.mp3car.com/fb-sensors/134060-4-channel-input-isolator-and-voltage-divider.html so you will have to do some math to get an accurate sensor response. Not a problem for me as I'm only interested in five voltages for perimeter and interior security purposes: Open Circuit/Tamper, Short Circuit/Power Loss, Detectors Nominal, Detector 1 Triggered, Detector 2 Triggered.

          Comment

          Working...
          X