Hello Everyone,
Just became a member like 2 seconds ago because this forum is awesome and I have been using it a lot, but just never joined until now.

I have looked around through other threads on clipping, where people have issues with it, or are dealing with audio, but I am dealing with a digital system (receiving data through white, room light via LED to photodiode) and I was wondering if I could receive some input on something simple.

Because of the medium, the signal becomes drastically noisy so I was looking to op amps to actually amplify and clip an incoming data transmission (19.2k baud) which would be similar to UART, using D-flip flops to clip, and I was wondering if there were specific op amps (not necessarily circuit implementations) I could use, or if it were better to not use them, and use something like an inverter (which would be very power consuming, but if it works, then great).

Mind you I haven't taken a circuits class, so I am learning theory on my own, I apologize in advance for ignorance.

Thank you,

Max

 

Could you show some pictures and/or drawings of your setup ?

For clipping a noisy signal, a comparator circuit may also be a good choice.

 

This isn't my line of expertise, but...op-amps are optimized for analog work. I think you need a digital type, like a comparator or a Schmidt Trigger.

 

A Schmitt trigger - a comparator with hysteresis - sounds like the way to go but we're only guessing until we see more detail.

You might get better overall results if you modulate the carrier at, say, 100kHz and then filter out anything that does not come in on that carrier. Retrieve your digital signal afterwards.

 

You might get better overall results if you modulate the carrier at, say, 100kHz and then filter out anything that does not come in on that carrier. Retrieve your digital signal afterwards.

I second that idea ! (Similar to IR)

 

Hey guys, thanks for the response, sorry it took so long. Here is what I was hoping would work.

I will definitely try both the Schmidt and put signal on a carrier. The only thing I am worried about the carrier is that this system is suppose to last for rough 3 months on a powerful LiPoly battery, but continuously.

Interestingly enough, I was going to use an inverter, but I accidently placed a Tri-State buffer in its place by complete accident (stupid me) but then it somewhat worked a little bit. I am going to investigate it soon.

http://tinypic.com/r/2iqer04/5

Under this condition:

http://tinypic.com/r/rwigis/5

Thanks for the help, sorry for late response, busy at school.
Also if you know of a better way to upload pictures that would be helpful, but I will read forum help stuff too.

Max

 

http://tinypic.com/r/288xffo/5

 

Hello,

You might be interested in the following PDF:
Photodiode monitoring with op amps

This is one of the links from this page of the EDUCYPEDIA:
PHOTO diodes

Bertus

 

You don't necessarily want to use an amplifier.
An amplifier will amplify the noise along with the signal.
What you need to do is reexamine why you are transmitting the signal the way you are doing now. You need to change or modify your scheme in order to improve the signal-to-noise ratio (SNR).

 

I will definitely check this out.

However, today I got the results out of the circuit that I was looking for. What I needed to do was reverse bias the Photodiode using Vcc across a resistor then put the signal in parallel and put that into the inverting side of the op amp.


Thanks again everyone,

Max

 

A schematic is worth a thousand words. Glad you posted it, even if it wasn't on this site.

 

If noise is an issue you can make a very narrow band-pass filter before amplification and an accurate carrier to transmit the data.
I was also trying out digital data transmission using led/photodiode and ended up using such a module
http://www.engineersgarage.com/electronic-components/tsop1738-datasheet
which does alot of the work for you .
All you need to do is generate a carrier at 38khz and modulate that (on/off) with your data signal .
disadvantage is that you have to use a 980nm Ir led . Not sure of the wavelength tolerance there.

 

Ah yes, I appreciate the help, I did in fact make a filter except because of specific reasons I used a wide-band filter, which I will explain for general knowledge. I am unable to use a carrier wave because my project has a very high concern over power consumption, and I am using a wide band active filter because of the scheme that I am using to send the data. Without the carrier, things clearly get more difficult and more noisy, but I am using very archaic PAM and I am also thinking about using IrDA protocol, which is meant for the LED above. I have to use 470nm PD because data is being transmitted underwater, and blue light has a low saturation level under deep sea water. So with the use of UART 1.0 (yes very archaic), PAM, and IrDA the frequency ranges from 9600 divided by the maximum bit package of one's or zeros (which is ten) and goes up to 3/16 of 9600, roughly 40k, so there is quite a bit of room for noise to enter, however my data is also in that range. I would post the working schematic which was finished today, but I just realized that it is IP and therefore I can give it up just yet (it doesn't solely belong to me).

I do appreciate the help though,

Max

 

3/16 of 9600, roughly 40k

Sorry meant 16/3 times 9600

 

I have to use 470nm PD because data is being transmitted underwater

interesting ...

What are you making , if you don't mind me asking

If you don't want to use a carrier then the only thing that comes to mind is at least using an optical blue filter to filter out everything from the red end of the spectrum to reduce the noise abit , and set a trigger at half the luminous intensity of the diode with a fair bit of hysterisis.

I am curious .... It is possible to have one photodiode pointed at the emitting diode for the transfer , and another diode which sits shielded away from the data emmitter and only records ambient light and then the two are fed into a differentiator op amp which filters out the noise .
I'd love to try this myself but unfortunately I have other projects planned for today :/