[SOLVED] photosensor selection - help needed

[prateek_k_chd]

Hello folks

I'm designing a photometer for visible spectrum light.No need for a huge bandwidth or high speed,just need it to be simple and accurate. (Calibration will be available to me)

I need to choose a sensor for this.

From what I know (i'm not an instrumentation engg),my choices lie between;

(a) photodiode
(b) phototransistor
(c) sensor+chip combos like ADPS9002
(d) ?????? (does anything else exist too?)


So I need suggestions on these ::


(1) i NEED linearity . Which of these (a/b/c/d) is best at it?
(2) Application simplicity (in terms of circuitry) would be a big relief.


So can anyone please guide me on this?

Also,can anyone point me to a guide for selecting PHOTOsensors ??

Thanks a ton !

 

[hiren.dave@bhuved.com]

you can use simple photodiode .

 

[prateek_k_chd]

So,is a photodiode more linear than others? Or do you suggest it for any other reasons ?

 

[Old Nick]

When you say you're wanting to measure the visible spectrum, what do you mean by that? Any photo-sensor is going to react to different wavelengths of light differently. If you shine 1 Watt of red and 1 watt of blue light onto the same photo sensor you'll get a different result. If you're measuring white light then this will not matter as long as the spectrum of the light doesn't change, you would need to calibrate for sunlight/fluorescent light/incandescent etc. when making measurements. Depending on how important linearity is to you, you'll maybe want to perform a calibration afterwards anyway so I wouldn't worry too much about that. If you give some more details on what it is you're trying to do then it'll be much easier to guide you in the right direction.

cheers

 

[prateek_k_chd]

Thanks Nick
I'll be measuring light through a monochromator...so at a time I'll have only one (or a set of very closely spaced) wavelengths)
By linearity,i imply that after calibration for a particular wavelength,the output should be proportional to the power of light falling on it for that wavelength.

also the measurements will be performed in a dark chamber so there will be no external interference.

Keeping this in mind,which type of sensor shall be best (and,hopefully,simple to implement) ?

thanks again

 

[keith1200rs]

Use a photodiode.

Phototransistors have an output which depends on a transistor current gain which is highly variable with temperature an will also introduce non-linearities (gain varies with signal level).

The ADPS-9002 has a terrible accuracy - varies all over the place with temperature.

Photodiodes tend to be pretty stable and linear. The main thing to watch out for is leakage (which is also highly temperature dependent). If you end up needing a very large sensor then you may want to keep the bias voltage low (i.e. zero) to minimise leakage currents. It depends on the exact photodiode. Hamamatsu would be someone to look at for a huge range of devices which are generally well characterised.

Keith.

 

[prateek_k_chd]

Thanks a ton,Keith !