PSN-L Email List Message

Subject: Re: Publication on the Web
From: ChrisAtUpw@.......
Date: Mon, 26 Nov 2001 21:49:44 EST


In a message dated 26/11/2001, ivey@.......... writes:

> Chris,
> Just curious, why an incandescent bulb and not an LED?  Also, that's a 
> pretty large area photodiode.  Trying to keep the noise down?
>  Thanks,
> Jack Ivey

Dear Jack Ivey,

       The movement signal is proportional to the photodiode current. The 
noise is proportional to the square root of the photo current, so you need as 
large a photo current as is practicable to optimise the resolution (100 micro 
Amp?). (Phototransistors amplify the noise as well as the signal and are NOT 
useful here).You need physically fairly large, high sensitivity photodiodes, 
well matched for response eg the VTD34. These allow an accurate linear 
response for movements up to +/- 1 mm. The high Infra Red sensitivity of 
Silicon photocells matches tungsten filament lamps very nicely.

       Tungsten filament lamps can provide the fairly high light levels with 
ease and the output can be very highly stabilised using a bridge circuit or 
another Si photocell. A thin straight filament is ideal for the purpose. The 
photon output is not effected by high frequency noise in the supply current. 
If you reduce the supply voltage to about 3/4 the rated value, you get 
effectively infinite filament life.

       LEDs have four major disadvantages. The output decreases exponentially 
with increasing temperature and the sensitivity is high - the chip 
temperature needs to be closely regulated / compensated. LEDs have higher 
inherent photon noise levels than tungsten filaments, with red LEDs being the 
worst. The IR LEDs are fairly quiet. Any noise on the supply current appears 
in the photon output. Most higher power LEDs have a cup like structure 
surrounding the chip and a circular metal central contact area, which blocks 
much radiation. It is usually not possible to get an effective 'point' source 
- you see a central spot surrounded by a bright ring. The LID chip LEDs do 
not have this problem, but are of lower power. High power regulated 
specialist devices similar to laser diodes are available - Hamamatsu make a 
nice one for ~$200. Laser diodes themselves are extremely noisy.

       Regards,

       Chris Chapman
In a message dated 26/11/2001, ivey@.......... writes:


Chris,
Just curious, why an incandescent bulb and no t an LED?  Also, that's a pretty large area photodiode.  Trying to keep the noise down?
Thanks,
Jack Ivey


Dear Jack Ivey,

      The movement signal is proportional to the photodiode current. The noise is proportion al to the square root of the photo current, so you need as large a photo current as is practicable to optimise the resolution ( 100 micro Amp?). (Phototransistors amplify the noise as well as the signal and are NOT useful here).You need physically fairly large, high sensitivity photodiodes, well matched for response eg the VTD34. These allow an accurate linear response for moveme nts up to +/- 1 mm. The high Infra Red sensitivity of Silicon photocells matches tungsten filament lamps very nicely.

      Tungsten filament lamps can provide the fairly high light levels with ease and the out put can be very highly stabilised using a bridge circuit or another Si photocell. A thin straight filament is ideal for the pur pose. The photon output is not effected by high frequency noise in the supply current. If you reduce the supply voltage to abou t 3/4 the rated value, you get effectively infinite filament life.

      LEDs have four major disadvantages. The output decreases exponentially with increasing temperature and the sensitivity is high - the chip temperature needs to be closely regulated / compensated. LEDs have higher i nherent photon noise levels than tungsten filaments, with red LEDs being the worst. The IR LEDs are fairly quiet. Any noise on the supply current appears in the photon output. Most higher power LEDs have a cup like structure surrounding the chip and a ci rcular metal central contact area, which blocks much radiation. It is usually not possible to get an effective 'point' source - you see a central spot surrounded by a bright ring. The LID chip LEDs do not have this problem, but are of lower power. High p ower regulated specialist devices similar to laser diodes are available - Hamamatsu make a nice one for ~$200. Laser diodes the mselves are extremely noisy.

      Regards,

      Chris Chapman

[ Top ] [ Back ] [ Home Page ]

Larry Cochrane <cochrane@..............>