DFR0026 ambient light sensor temperature compensation

Help & Support Robot Parts
1

Hi all,

Having got hold of a few of these, I see from the Sharp compenents datasheet for the PT550 that there is significant temperature dependency.
This is verified empirically by mounting a sensor securely in a reasonably well lit place to give a reading of say 1V, and hooking up the sensor to either a voltmeter or ADC channel and taking readings and gently warming the PCB from underneath. I went from 1000mV to 1185mV when warmed from a fairly cool PCB at 17C to 27C by applying a hot hand!

The Chinese datasheet shows a very unclear thick red line for temperature dependency, but nonetheless it is shown.

I am mounting the sensor behind a slug of pure PTFE to act as a diffuser and intensity reducer, but it will be in an environment which ranges in temperature from 20 to 38C, so this is not insignificant.

I am considering flattening a DS18B20 (down to 4.5mm so it is the same height as the PT550), and mounting it next to the PT550 and applying a compensation factor to my ADC readings accordingly by applying a linear compensation factor such as taking off ADC*(0.012*degC) for temperatures over 20 for example.

As I’m surely not the only to have noticed this temperature dependent behaviour, so I was wondering if anyone else has come up with a suitable/neat workaround?

Cheers,
Eclectica

2

Hi,

Your solution of adding a DS18B20 temperature sensor to apply a compensation factor to the ADC might work. The temperature sensor will need to be very close to the PT550 though.
You can either determine one compensation factor which will be used for, let’s say, temperatures that range from 20°C to 38°C by calculating an average between compensation factors calculated for various temperature readings from 20°C to 38°C.
Or, you can also use different compensation factors for each temperature range. For example, you use a compensation factor for temperatures from 20° to 22°C, a different one for temperature from 22°C to 24°C and so on… up to 38°C.
This being said, we also offer this Gravity Analog Ambient Light Sensor which uses a different phototransistor TEMT6000

Cheers

3

Hi Brahim,

Many thanks for your reply.

In the mean time, I’ve managed to thin down a DS18B20 so that it is the same thickness as the PT550 is high, and have thermally bonded it to on side of the PT550. The PTFE slug then sits flat across the top of both components, but mostly on top of the PT550.

Yes, the compensation factor - I’m just using a ‘rough and ready’ linear per degC factor of 1/75 of ADC taken off when temperature is above 25C. Although below 25, the same will be added on. i.e. ADCcomp = ADC - ADC * (degC - 25) / 75. This seems to give me a much more stable response to temperature change under constant light conditions.

Aha :bulb: - the TEMT6000 actually looks better all round, I’ve put in an order for one of those - in fact it is probably superior that the PT550 in its visible light response especially as we need it for plants; think P.A.R. 400-700nm.

So many thanks again for a great recommendation. :smiley:

Best regards,
E.