Poor man's Android oscilloscope with photoelectric diode.

[Daniel López]

Hi,

It was nice to meet some of you guys yesterday.

One if the ideas I have in head is to find a Laptop with a flicker free screen (one that doesn't use PWM for screen dimming). 

That's quite complicated because manufactures don't use to publish the details of the screen of the laptop, so I guess my only option is to go to a shop and test by myself.

My idea was to build something called "poor man's" oscilloscope. The idea is to build a device that feeds a voltage into the cell phone's microphone jack and use an application [1] to analyze the current.

I found this article [2] that I think explains almost everything. I guess that the only thing I would have to do in order to detect light flickers is to connect a Photodiode to the cocodriles, isn't it? Would the preamp [3] be necessary?

Anyway, the point is that my knowledge in electronics right now is close to 0, so I would very much appreciate if someone could give me a hand with that project. I am quite sure I will burn my cell phone if I try to do that by myself! :P

Of course, I would be happy to become a member of the HackerSpace!

[1]: https://play.google.com/store/apps/details?id=ch.serverbox.android.osciprime

[2]: http://hackaday.com/2012/07/14/android-oscilloscope-built-from-parts-just-laying-around/

[3]: http://www.instructables.com/id/A-Preamplifier-for-Smartphone-Oscilloscopes/

[MRE]

We could simulate your test easily enough with one or both of the Oscopes we have in the space, and the multitudes of laptop and other screens that come into the space via members.

I suspect you will run into one or more of the following problems (which is why you should try it out with a REAL scope first as proof of concept):

1: Not sure what the pulse frequency of the lamps or LEDs behind the screen are likely to be. In fact, there could be several different frequencies, but I suspect they all are around the same range. It has to AT LEAST be 120hz or better. Likely MUCH higher. Depending on what this frequency is, it might exceed the capabilities of the audio range input jack. Audio tops out at 20Khz, but I suspect the bandwidth on a mic jack is a bit less than that.

2: You will need to find a photo transistor that does not get 'swamped' by the afterglow in between flashes of the backlight. In other words, the photo transistor needs to react at a rate significantly higher than the PWM frequency.


UPDATE;
A good monitor should have pwm rate of 200hz or higher. 100hz seems to be around the bottom end.
This video might demonstrate a much simpler method for you to use:
https://www.youtube.com/watch?v=oSiVii2qaJY

[Daniel López]

Hi!

Thanks for the answer!

It makes sense using a videocamera to perform the test but, in that case, I have two problems:

- I'd need to have access to the computer, open MsPaint, create the image and do the test.

- As far as I know, only high end videocameras have manual shutter spreed control. And I don't have one.

My idea is to build a device so I can go to a computer shop and test the screens of all the exposed computers. A little bit like that: http://youtu.be/zmRypg9UR0s but using my cell phone instead of a laptop as oscilloscope.

I am not saying that this is the most practical way to do that, but it will be a perfect excuse for me to learn how to build a simple electronic circuit. And well, it's cool! :P

The bad screens (like the ones the MacBooks are mounting) flicker at 100Hz. Some other brands have higher frequencies, like 400Hz. Also, I would like to be able to see the shape of the wave, so I guess I'd need ten times more resolution.

LED screens flicker like that:

   ---     ---

  |   |   |   |

__|   |___|   |___

Anyway, I guess that, first thing, I should go to Akihabara to see if I can find that photo-transistor.

Do you have any recommendation?

El divendres 19 de setembre de 2014 0:56:09 UTC+9, MRE va escriure:

[MRE]

OK cool so using the audio input as a scope will work fine.

As for photodiode.. nothing particular comes to mind. At those frequencies many/most ought to work. Buy a couple and experiment ;)

If you want to see waveform and not just digital switching, you will need to bias the transistor so it runs as a signal amp rather than in switching mode. A couple of resistors ought to do it.