Naive question here. I've been toying around with the idea of making a simple light meter. From what I've read, photodiodes generally offer the best sensitivity and accuracy (as compared to photoresistors etc.) The simplest op amp circuits for photodiodes seem to require at least 12-15V. However, there are commercial lightmeters which use photodiodes and run on a single 3V battery (e.g. http://www.sekonic.com/products/l-758dr/specifications.aspx) Is it possible to achieve this using a "rail to rail" op amp, or is it more likely that these devices obtain a reading from the photodiode without using an op amp at all? (The latter is certainly possible but I wonder if the cost in sensitivity/accuracy would be too great.)
\$\begingroup\$
\$\endgroup\$
6
-
1\$\begingroup\$ Why not use a MCU with a differential and gain ADC? \$\endgroup\$– Ignacio Vazquez-AbramsCommented Mar 25, 2014 at 20:52
-
\$\begingroup\$ If you're using the diode in photovoltaic mode the voltage across it is very small, so although an op amp isn't strictly necessary you get a much better signal. (That is my understanding anyway.) \$\endgroup\$– foldlCommented Mar 26, 2014 at 4:57
-
\$\begingroup\$ Using 100x gain and a 1.1V reference should do it, no? The ATtinyX41 has both of those, can run on 3V, and only uses about 80uA at 200kHz. \$\endgroup\$– Ignacio Vazquez-AbramsCommented Mar 26, 2014 at 5:01
-
\$\begingroup\$ Oh I see what you mean, sorry. Is there any information available on the pros/cons of using the microcontroller to amplify the signal vs. using a separate op amp? \$\endgroup\$– foldlCommented Mar 26, 2014 at 16:55
-
\$\begingroup\$ It's basically "no separate op amp" vs. "dependent on the quality of the analog circuitry in the MCU". \$\endgroup\$– Ignacio Vazquez-AbramsCommented Mar 27, 2014 at 0:50
|
Show 1 more comment
2 Answers
\$\begingroup\$
\$\endgroup\$
2
Just because the battery voltage is 3V doesn't mean there isn't some sort of switching regulator to boost the voltage. However, even if there isn't it is possible to achieve using just a single 3V rail, 2 op-amps, a couple of resistors, and a photodiode.
You just need an op-amp which can operate on low voltages. You will either need to bias the non-inverting reference point, or put an capacitor switch pump to generate a negative rail.
Here's an example with a dual op-amp package, one which is used to provide a virtual ground reference.
You'll want an op-amp which has very low noise, low input bias current, and low input offset current. Something like this TL032.
simulate this circuit – Schematic created using CircuitLab
The output is \$V=I_D R_G\$. Make \$R_G\$ small enough to avoid saturating the op-amp, but large enough to give you the full resolution. This is designed to operate with a differential voltage sensor, you can in theory measure Vout+ single-endedly, but it won't necessarily work as well because you'll miss out on fluctuations on the 1.5V reference.
Oh, and it probably goes without saying that this circuit doesn't include any of the logic/adc stuff. That is all extra, but simple microcontroller with a differential ADC should be sufficient to get a reading (these can operate in the 3V range as well).
answered Mar 25, 2014 at 21:15
-
\$\begingroup\$ Why R3 and not a direct connection? \$\endgroup\$ Commented Mar 26, 2014 at 3:14
-
\$\begingroup\$ input bias current, though I suppose arguably it's not a huge issue. \$\endgroup\$ Commented Mar 26, 2014 at 3:21
\$\begingroup\$
\$\endgroup\$
Heck, I've seen photodiodes directly connected to a DIGITAL IO pin used to measure the ambient light(I don't think it was accurate but it told you if the lights were on) by using the photocurrent to discharge the paracitic diode capacitance and measure the time.
Tons of op amps run at 3v. In fact some run at 1.8v. Totally doable at 3v, or else those little IR receiver modules wouldn't work at 3v.
I don't see any reason to use bias the non inverting reference point or anything fancy. Just use a single supply op amp that doesn't expect a negative voltage.
