Home > DIY, Engineering, Hacking, Photography > AF Assist V2 (Update 17 May 2011)

AF Assist V2 (Update 17 May 2011)

Continuation from this post. Build and experimentation log continues.

I built the unit and tested it last week. The laser seemed lower power then the first one I built (which was just the laser pointer hooked up to a set of 3 AAA batteries along with a pushbutton).

Last night I had time to take a multimeter to it. The laser LED is seeing a total of 3.4V from the output of the LM358 OpAmp. The rail voltage was 4.6V (3 AAA bateries) yet the maximum I could coax the output to put out was 3.3V – 3.4V. This explains why the laser beam seemed weaker at first try.

I’m really going blind here using an OpAmp. Did lots of googling, but found nothing to explain this behavior. I tried several LM358, and they all worked the same way. It must be how they work.

So after an hour of dead end attempts and research, I decided to add a 4th AAA battery (for a total of 4 AAA) to give the op-amp 6V, and that did the trick. Now the laser LED sees 4.3-4.4V, and the brightness is back to what I remembered.

One other thing to note about the laser LEDs. When soldering to the + side of the led (taken from the dollar store laser pointers) care should be taken not to overhear the LED. Minimal amount of solder and heat should be used. Excess heat will destroy the LED/the brightness will degrade VERY quickly. If the soldering is done quickly (under 1 sec of applied heat) then there is no effect on the LED brightness. I came across this problem when I ended up soldering and de-soldering wires to the same LED several times.

Will source better laser pointers in the future, where soldering is not required.

One other thing I need to add is a second LED. I want this tool to work at relatively long ranges (so I can reliably focus on band members on stage) as well as close up for people dancing.
I will be using two laser LEDs (hooked up in parallel) and de-focused by different amounts and aimed at different locations. One bean will be really defocused to give good coverage at close distance, and another laser led will be defocused less (less then the dot it normally produces) and aimed so that the center focusing point lies on it at longer ranges.

UPDATE 18 May 2011

I played around with the AF tool last night. I am happy with the light it gives. I adjusted the spot size and aimed the laser LED a little better, and I now have a good range covered. I have covered anywhere from 10ft to the length of my house (maybe 80ft or more) and it can reach farther. In absolute darkness, the beam is still quite strong at 80-ish feet.
I also tried adding a second laser LED, but I the only other laser pointer I had around was not very bright (must have damaged it previously while soldering to it).

The other observation is that even though the LED visually shuts down when the shutter is tripped, it still gets recorded on the image. I guess it must take the op-amp a fraction of a second longer to drop the voltage of the LED. I will live with the consequence at the moment as I am out of time to fix it. I can always de-focus or turn down the intensity so it’s not so bright (if this becomes a problem), but will certainly need to fix this in the next version.

UPDATE 18 May 2011 (2nd  update)

Based on the response time of LM358 found here in figure 5, I estimate that the pulse response is 400mV per 1µs, which translates to 11.5µs (or 0.0115seconds) for a voltage adjustment of 4.6V. The figure of 11.5µs is the same as 1/80 of a second.
If my assumptions are correct about the pulse response of LM358, I won’t be able to use it to shut down the LED at the time of exposure. I will think about just shorting through the camera’s hot shoe or another method to manage this. –> Continued:

Just discovered that the parameter I am looking at is called “slew rate”, and the LM358 has a slew rate of 0.6V/µs
The NE5532 I also bought (at the same time as the LM358) has a slew rate of 9V/µs. So with a simple swap I should have my problem fixed. The NE5532 should make the 3.5V switch in 0.5µs which is 1/2000 of a second.

UPDATE 19 May 2011

See here for upate

Categories: DIY, Engineering, Hacking, Photography Tags:


five + 1 =