0
\$\begingroup\$

We have a sensor giving out a signal between -10 V and +10 V, and a signal acqusition device that only accepts voltages between -2 V and +2 V. So we need to scale down the signal into the smaller range. As possible solutions I have identified:

A. Use a voltage divider

Following the information in this page: https://learn.sparkfun.com/tutorials/voltage-dividers, I could use two resistance, R1 = 5000 Ohm, and R2 = 1200 Ohm.

voltage divider

  • I am not sure what am I allowed to use a Ground. Both devices (sensor and acqusition) have a ground connection, which are connected to each other. Can I use this common ground as for the ground of the voltage divider ?

  • The impedance of the signal acquisition device is specified as 1 MegaOhm. Why is this value given at all? Should I include it somehow in my voltage divider calculations ?

B. Buy some off the shell device

I have seen some voltage down regulators, but they appear to only address power applications. I assume they are too noisy to be used for sensor read-out. Furthermore, there appears to be no device offering downscaling negative voltages.

  • Are there commercial voltage downscaling devices made for sensor reading ? I am not familiar in this area and I might be just ignorant of a family of devices that already does this in a plug an play manner.
\$\endgroup\$
3
  • \$\begingroup\$ Datasheets or at least brand/model of both the sensor and the acquisition device, please. \$\endgroup\$ Commented Aug 11, 2020 at 11:59
  • \$\begingroup\$ Both are very uncommon devices for which no other information is given. Far from ideal, I know... \$\endgroup\$
    – Keine
    Commented Aug 11, 2020 at 12:27
  • \$\begingroup\$ This voltage divider makes Vout 1200/6200 of the way between the ground voltage and Vin. Which seems to be what you want. \$\endgroup\$ Commented Aug 11, 2020 at 13:34

2 Answers 2

1
\$\begingroup\$

I am not sure what am I allowed to use a Ground. Both devices (sensor and acqusition) have a ground connection, which are connected to each other. Can I use this common ground as for the ground of the voltage divider ?

If your circuit has a ground, then that's the one you need to use here.

The impedance of the signal acquisition device is specified as 1 MegaOhm. Why is this value given at all? Should I include it somehow in my voltage divider calculations ?

The impedance is in parallel with R2. If it is much greater than R2, you can ignore it. If it isn't, then you need to ensure that your voltage divider gives the correct scaling using R2 in parallel with the input impedance, rather than just R2.

Are there commercial voltage downscaling devices made for sensor reading ? I am not familiar in this area and I might be just ignorant of a family of devices that already does this in a plug an play manner.

I'm not aware of any, but that doesn't mean they don't exist. A high-impedance op-amp set up to give the desired gain would make an alternative to a simple resistive divider.

\$\endgroup\$
1
\$\begingroup\$

You can use a voltage divider. The grounds must be common, and any difference in potential will lead to errors. With your resistor values, the source must be able to drive a 6200 ohm load.

The ratio of the resistors is 1200/6200 so a +/-10V input will give you (nominally) about +/-1.935V. The resistor tolerances will affect what you actually get. 1% tolerance values (E96 series) that are close are 4.99K and 1.21K and even if the resistors are worst-case off it will still be less than 2V for 10V in.

If you were looking for an exact ratio you would take the 1M\$\Omega\$ input resistance into account, but it's around 0.1% effect.

For example, if you used a perfect 4.99K\$\Omega\$ resistor, the exact value of the other resistor for 10V->2V would be 1247.5\$\Omega\$ not considering the 1M\$\Omega\$ and 1249.06\$\Omega\$ taking the input loading into effect, or about -0.12% less.

To adjust the resistor for the loading the equation is:

Rnew = 1/(1/Rold - 1/Rload)

\$\endgroup\$
8
  • \$\begingroup\$ >the source must be able to drive a 6200 ohm load. So the sensor must be able to deliver 1.61 mA (Ohm's law with 10 V and 6200 Ohm resistance), right? Are there any dangers in choosing resistances with lower values, i.e. 500 and 120 Ohm? \$\endgroup\$
    – Keine
    Commented Aug 11, 2020 at 12:30
  • 1
    \$\begingroup\$ The sensor may not be able to drive the load. It might also make the circuit a bit more sensitive to where the divider is connected relative to the ground connection, because the wires will have a bit of resistance. Depends on how much you care about accuracy. 0.01% is quite a bit different from 1%-2%-ish. \$\endgroup\$ Commented Aug 11, 2020 at 13:00
  • \$\begingroup\$ We'd like as much accuracy as possible. I assume that choosing resistors with higher values brings the problem of R2 getting near the 1MOhm value of the parallel load resistance. \$\endgroup\$
    – Keine
    Commented Aug 11, 2020 at 13:07
  • \$\begingroup\$ "As much accuracy as possible" is not a specification. That will cost far more money than you have, and probably take longer. \$\endgroup\$ Commented Aug 11, 2020 at 13:09
  • \$\begingroup\$ As much accuracy as possible using simple off the shelf (E96 resistors) components \$\endgroup\$
    – Keine
    Commented Aug 11, 2020 at 13:10

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