OPT4060: Fast optical sensors

Hello Santi,

At this time, we do not have a device in production with both RGB and IR. The OPT4048 and OPT4060 are color sensors, not color + IR. The conversion times of these devices are measured per channel, so the sampling rate would be 4 * 600 us, much slower than the 0.1 ms you are needing. 

As for IR, the OPT4003-Q1 has an IR channel, but does not have color.

These are currently the only devices that could fit your application. May I ask for more details on what you're trying to do?

Thank you,

Daniel Balmaceda

Hi Daniel,

Thank you for your answer. The application is a counting device for biological particles in a liquid flowing through Ø3mm transparent pipe. Particles can be 0.5mm to 3mm long and have different grades of transparency and colours (not bright colours).  So at one point of the pipe there is a light sensor and a light source on both sides, and a microcontroller is sampling as the liquid flows so it can analyze and count the particles.

A fast sampling rate is important so we can set a higher flow speed for the liquid because there are a lot of litres to analyze, but we have to "squeeze" them into that Ø3mm tube. Sampling rate shall provide about 5 samples per particle, so there is enough information per particle in order to decide what kind of particle it is, based on its "shadow" pattern (five transparency samples) crossing the exploration area as it flows.

The light intensity is our decision, so we can have the sensor working at its best and fastest work-point.  And the light quality would be, lets say "broad white",  according to the sensor's channels.  We don't know beforehand about the colour grades and transparency of the target particles, that's something to determine experimentally, and it may be different from some environments to others, and software selectable, and that's why we need different channels and even IR and UV because they may be key to differentiating biological types. That's why IR, NIR, or even UV may be interesting.

After some maths a sampling rate of 0.1ms would be ok, but faster would be better. I've seen your answer about the 4 times 600us for  4 channels. I guess I get it, it's not only related to how we can read the values in the registers one after the other, it's mainly that the channel sensing is "not parallel", so each channel needs its own time, one after the other. In this sense what's the difference (or how do they relate) between the conversion time and the light integration time (stated in the OPT4060 as 25ms minimum).

So Daniel, we'd really appreciate your advice on TIs optical devices towards a solution for this. Thanks in advance!