2
\$\begingroup\$

I have a pressure transducer outputting 0-5V which is sampled by an ADC with a 0.5V minimum input voltage. I need to offset the sensor output voltage by a fixed voltage of 0.5V. I've looked at an op amp circuit for doing this. But I struggle with finding a practical way of generating a reference offset voltage that is temperature stable (I'm not sure about using a resistor divider down from Vdd).

Any ideas would be appreciated.

Incidentally, the upper limit of the ADC is irrelevant, as we will never get an output signal in that range. Therefore, I want unity gain.

LTspice circuit Simulation result

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

My first reaction is to get a better A/D converter. There are many many out there that can directly convert from the 0 to 5 V range.

You didn't say what the upper end of the A/D range is, but if it's 5 V, then you don't want to add a offset, but rather scale the signal towards 5 V. Put another way, you want to map 0 to 5 V --> 500 mV to 5 V. That's not adding a offset.

Fortunately, scaling relative to 5 V is easier than adding a offset. All you need is two resistors:

\$\endgroup\$
  • \$\begingroup\$ The upper limit's not really relevant, as we're never experiencing pressures that will output those voltages. But in any event, the method you're proposing is much faster to implement and fairly easy to compensate for in a GUI. But how will such a method load the sensor? We're using a Keller 23 SY Ei pressure sensor with 0-5V output, which has a max. 5mA output current. \$\endgroup\$ – Oystein Dec 16 '16 at 13:44
  • \$\begingroup\$ @Oys: At most this method requires the sensor to sink 450 uA. That comes from the full 5 V across R1+R2. \$\endgroup\$ – Olin Lathrop Dec 16 '16 at 14:26
0
\$\begingroup\$

I struggle with finding a practical way of generating a reference offset voltage that is temperature stable (I'm not sure about using a resistor divider down from Vdd).

If the ADCs internal reference is available then scale down from that - if that reference drifts, then errors tend to cancel out. If you are using an external reference for the ADC then it's the same story - scale down using precision resistors and any drift in the reference is ratiometric and produces no error.

If your external reference is Vdd then you might want to improve on that but, the scaled down offset will still be ratiometic and not produce a drift error providing good resistors are used.

\$\endgroup\$
  • \$\begingroup\$ How would you scale down from e.g. a 5V reference (not Vdd), without having the down-scaling (resistor divider?) affect the resistor network around the op amp? I tried to use a larger voltage source as offset in LTspice, and scale down with resistors, but ended up affecting the offset and unity gain. \$\endgroup\$ – Oystein Dec 16 '16 at 13:38
  • \$\begingroup\$ R1 and R2 initially reduce the gain but add the offset. With them both at 10k the gain halves so, if you wish to add back that gain (as you appear to have done), then R3 and R4 at 10k do this. If you use a 5 volt reference then R1 gets a lot smaller and that initial gain compromise isn't as bad so, you won't be needing R4 to be 10k but also smaller like R1. \$\endgroup\$ – Andy aka Dec 16 '16 at 13:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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