I've recently posted this question regarding input impedance when working with DC signal on HCPL-7840 Isolation Amplifier. The answers were very clarifying and helpful, but I've come to realize I need more specific help.

Just a disclaimer: I decided to use HCPL-7840 because it was the best option available on my chosen supplier that seems to fufil my needs.

My goal is to design a simple isolated amplifier for an NTC

  • NTC characteristics considered:

    Resistance [kΩ] Temperature [ºC]
    10 30
    15 20
    20 10
    25 0
  • HCPL-7840 characteristics considered:

    • The information on the "Recommended Operating Conditions" table of datasheet
    • Equivalent Input Impedance RIN — 500 — kΩ
    • Figures 12 and 14 of datasheet

  • Based on these informations, I designed the following simple circuit:
    • It's composed of a simple voltage divider, where R1 is fixed precise resistor, and R2 is the NTC.

enter image description here

  • As soon as I assembled the board and started making measurements, I noticed that the voltages on TP1 were not exactly what I expected to see

    • I was expecting a simple voltage divider between R1, R2
    • I figured the difference should be because of internal resistence of HCPL (Rin), between pins 2 and 3
    • This Rin is suposelly connected in parallel with R2 (between TP1 and GND iso)
  • I than started to conduct some experiments to empirically try and figure out the value of Rin. This way I could calculate adjusted values for R1 and R2 and keep the voltage on TP1 within the input limits of HCPL linear range

    • The experiment consists of trying different combinations of R1 and R2 and measuring the voltage at TP1.

    • The results of this experiments drive me crazy because different combinations of R1, R2 make me infer different values for Rin each time.

    • The following table summarizes the experiment results:

      Index Vdd [V] R1 [Ω] R2 [Ω] V measured @ TP1 [V]
      1 5.40 2150 100 0.213
      2 5.45 2150 68 0.158
      3 5.56 2150 40 0.101
      4 5.55 246000 11600 0.186
      5 5.70 246000 5800 0.119
      6 5.77 717500 11600 0.082
      7 5.79 717500 23500 0.113
      8 5.84 717500 51300 0.142
    • Vdd and V measured @ TP1 were measured with osciloscope.

    • R1 and R2 were measured with Ohmmeter before the experiment.

    • Outcomes:

      • Given Vdd, R1 and V measured @ TP1, I was able to calculate what would be the "real R2" that the circuit is submited to.
      • This "real R2" would be the parallel between R2 (from the table) and Rin
      • Because I have R2, I'm able to calculate Rin
    • After this procedure I ended up with a lot of different values of Rin for each experiment case, and I don't know how to proceed!!!

Any help is appreciated.


  • \$\begingroup\$ transferring an analog signal over an opto is challenging. You'l get much better linearity using the opto to transfer a frequency or pulse width. You could probably use a NE555 in both of those cases. Assuming you go with a pulse width modulated signal, analog filtering on the output of the opto will create an analog signal. Of course, you could use a single chip micro and just transmit uart data across the opto. Or go Bluetooth or WiFi and negate the need for an opto. \$\endgroup\$
    – Kartman
    Apr 29, 2022 at 10:44

2 Answers 2


Figure 14 of the datasheet suggests that the input current into the HCPL-7840 should be fairly linear (resistor like) for input voltages in the range of -200mV to +200mV. But acconding to note 12 on page 18 of the datasheet the input stage is actually a "switched capacitor", not a real resistor, so there is going to be current spikes. I would suggest putting a 0.1uF ceramic capacitor across R2.

I would also suggest making changes that make your circuit less sensitive to input current on the HCPL-7840 input pins.

  • One option is to buffer the + input of the HCPL-7840 with another rail-to-rail op-amp configured as a voltage follower.
  • Alternatively, you could lower the values of R1 and R2 enough that the input current into the HCPL-7840 doesn't affect the measurement as much.

from datasheet ...

The Broadcom® HCPL-7840 isolation amplifier family is designed for current sensing in electronic motor drives. In a typical implementation, motor currents flow through an external resistor and the resulting analog voltage drop is sensed by the HCPL-7840. A differential output voltage is created on the other side of the HCPL-7840 optical isolation barrier. This differential output voltage is proportional to the motor current and can be converted to a single-ended signal by using an op-amp as shown in the recommended application circuit.

You need using it as a "differential" device. Not single input(ok)/output(not ok).
So you need using something as figure 4 in the datasheet.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.