# light detection circuit

Discussion in 'The Projects Forum' started by Hyperian, Aug 14, 2008.

1. ### Hyperian Thread Starter Member

Apr 25, 2008
13
0
I'm using a photodiode in parallel with a resistor in series with a LED to detect when light hits the photodiode.
The photodiode goes from 7kohms (with light) to 5Mohm (without light) and I'm using a resistor of 1.1Mohm.

I am planning on replacing the resistor (since it's big) with a transistor. If i remember my classes right, i should use a NMOS. any advice on the setup?

2. ### SgtWookie Expert

Jul 17, 2007
22,182
1,728
The bolded portion really doesn't make sense. A schematic would be invaluable here.
That sounds reasonable.
Why? You can get 1M resistors in 1/10 W, and even smaller if need be.

Let's start with a schematic on what you have now. If the physical size of your existing resistor is a problem, that is very easily solved. Adding a variable like a transistor would mean that you'd have to add more resistors to control the transistor - so your project would escalate in cost and complexity while plummeting in reliability.

3. ### Hyperian Thread Starter Member

Apr 25, 2008
13
0
I was trying to avoid having to draw a schematic cause i would have to do it with paint but here it is

Well i can get my hands on a NMOS easier than a specific type of resistor that my school might not provide.

I was thinking that the gate of the NMOS would be connected to a lower value resistor which connects to the power supply to get a specific voltage on the gate to allow for the NMOS to act as a resistor. (i hope i'm right it's been a while)

4. ### SgtWookie Expert

Jul 17, 2007
22,182
1,728
OK, I see the problem.

You need the photodiode and resistor in series to create a voltage divider network.
If the photodiode is on the + side, when the photodiode is in a low resistance state (light), the output of the divider will be high; when the photodiode is in a high resistance state (dark), the output of the divider will be low. However, the amount of current output available will be very small; in bright light it will be roughly 5V/7kOhms, or about 0.7mA. This would be enough to bias on an NPN-type transistor, which could then sink the current from your LED; which would need a current-limiting resistor.

See the attached.

File size:
8.1 KB
Views:
77
5. ### Hyperian Thread Starter Member

Apr 25, 2008
13
0
interesting, but i'm trying to replace the 1.1Mohm resistor with a transistor biased in the linear region so it will have a resistance of around 1Mohm.

6. ### SgtWookie Expert

Jul 17, 2007
22,182
1,728
It's easier to get your hands on resistors than it is to get your hands on transistors.

Do you have a particular transistor in mind? Without knowing what the actual specifications of the transistor are, it will be exceedingly difficult to recommend a way to bias it.

Meanwhile, it will require a pair of resistors to properly bias the transistor.

7. ### Hyperian Thread Starter Member

Apr 25, 2008
13
0
mm i thought with BJTs you need 2 but with MOS you need 1? anyway i was thinking of something like this

i think i saw something like this before but i cant simulate it in any of the circuit programs since i cant find the components i need, what circuit simulating program do you use?

8. ### blocco a spirale AAC Fanatic!

Jun 18, 2008
1,424
357
I have no idea what's going on here. Do you want the led to illuminate when light falls on the photodiode? What's the objective?

9. ### Hyperian Thread Starter Member

Apr 25, 2008
13
0
I want the LED to light up when there's light hitting the photodiode.

10. ### Audioguru New Member

Dec 20, 2007
9,411
895
The photo-diode should be reverse-biased. It leaks a very small current when it is in light. It needs a very strong light to develop enough current for the transistor to light the LED.

I corrected the circuit.

File size:
4.7 KB
Views:
75