I want to match the outputs of two light sensitive sensors when they see the same light level.
When I measure the outputs from the two sensors they don?t follow each other. So I have an idea that I should run one of the signals through an variable gain amplifier to piecewise match their gain.
The amplifier should look up the gain value from a table. The correction gain should be a piecewise linearization algorithm.
What is the best way to perform this piecewise linearization?
You need to provide a bit more information: What type of light sensors are you talking about, and over what range do you want them to match.
If you are using photodiodes and a proper current-to-voltage converter, the outputs should be linear over several orders of magnitude of input energy. You should not need to linearize anything---just provide a single gain matching circuit and offset adjustment.
You'll need to specify the detector and amplifier characteristics before I can provide any pertinent links.
What sensor and HOW are you ensuring they are seeing EXACTLY the same light? Too many setups have had their 'calibrations' setup by what was thought to be the same light, for such obscure reasons as -
one sensor was hotter than the other one light source to two sensors at same time was NOT the same level to multiple sensors internal reflections/shadows giving a difference different temp co-efficients different colour spectrum response between similar devices Tolerance differences between sensors
Check that they are actually seeing the same light, and better still first check each one INDIVIDUALLY against the SAME light source DIRECT with no other obstructions or reflections or optics.
How far out are they and is that within the sensor and its front end component tolerances?
You are of course checking the light levels with some form of light meter for reference?
First be sure you are looking at the correct problem.
Depends on being sure you are looking at the correct problem.
--
Paul Carpenter | paul@pcserviceselectronics.co.uk
PC Services
GNU H8 & mailing list info
For those web sites you hate
Although not for light sensors, I have done this exersize about 10 years ago to match the channell resistance of a pair of JFETS and to linearize the measurements - a gamma spectrum - to a 12 bit linearity, and to thermocompensate the whole mess. The technique is viable for relatively low volume products. I used to select pairs of FETs (no good enough duals on the market) for a start, then a table with temperature dependent gain correction coefficients was built and stored in the products MCU EEPROM (it was a HC11, but it only stored the values, the bigger host made use of them). For linearization, it turned out that the non-linearity of FETs was pretty well described by only 3 coefficients (second order), which were again measured on a per unit basis and stored in the EEPROM. The host used the values to create a correction curve, which was used to do the linearisation transformation of the spectrum. I hope this is more or less what you are looking for, although your bandwidth will probably be much much lower. Nowadays, we have unbeliavably fast and precise ADCs and we use DSPs etc., life is much easier....
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.