Is a LED fast enough to strobe and is any color faster

[paulie]

I have no issues with wavelength or power output. All that is solved and faster than the steppers. Like with the LED discussion the main thing with a PCB exposer is on/off time. There seems to be a disparity between specs and my observed results. Keeping it close to the "knee" does not seem to work because of pattern spread. Practical experience with all on or off times are my main concern.

[LaserSteve]

If on/off times are the issue, check the diode driver for speed response. The less then 300 mW Laser Diode in most packages is good to 10 to 100 Mhz or more. It is the diode driver that affects device timing more then anything else. You may be hitting the rise/fall and bandwidth limitations of your driver.  Most of them become a nasty integrator if you start to push the speed.

This requires a good oscilloscope and a fast photodiode with proper termination at the scope input.

We usually try to find photodiodes with a very small surface area, ie about a square millimeter or less. We reverse bias them to increase response speed, and we load them properly at the scope input. If the scope does not have a 50 Ohm input, we add a BNC TEE and a 50 Ohm terminator on the O'scope input.

Take a look at Thorlabs DET10A and the schematic on its data sheet for an example of a moderately fast detector.

Steve

[eliocor]

I have the tools to capture an hi res spectrogram of a white led, but unfortunately I will not be able to do it until March 

[Seekonk]

"PWM will not significantly  affect the performance of the laser diode device, unless you are near the Knee in the LD Light Output Vs Current curve, where output starts to decrease with increasing current."

So how much current can you safely expect to to put into a LED.  The high speed kickstarter project seems to say they are using commercial LED at very high currents.  The LED I am using is 4W 35ma at about 150V.  If I stay under a 0.2% duty cycle, how high a current could I pump into this. Would 5-10 times be reasonable. They were claiming 2,000% overdrive.  Those were really short pulses.

[paulie]

From what I can tell the cheap pointers are using Sony SLD3236VF or knockoffs (knocked off the assy line) so driving them directly from an AVR pin at 40 or so ma should be about 30-35mw. Well below the 80mw pointer operation and 200mv CW spec. Low res tests show just right for my tiny motors that can't go any faster. Depending on the PCB file it may stay on for several ms so probably considered continuous. The problem is with 6-8mil traces where im getting smearing. It is turned completely on and off in a very short time.

I will try to characterize with a high speed diode as suggested to see if that is the problem or if it's the queerly oblong TEM pattern just not focusable. Thanks for the input.

[paulie]

(the grumpy zero was for paulie's catty post lol)

I prefer the term "snarky". Sounds more like a 1950s cruise missile. Of course a little pussy never hurts anyone.

[joeqsmith]

I'm helping a friend rebuild a couple SUN Distributor machines.  They have a spinning neop flash tube that spins in an assembly that lights up the dial 0-360 when the plug would fire.  I just got the first machine to work acceptably, but it is likely that the other machines will need new flash tubes.  These are custom made and cost $250. A 10W LED would fit in the space just fine. The dial would be about 2 1/2 inches from the LED. Most any color would be acceptable.  Is a LED fast enough, it would have to be better than 3 uS. 

Interesting project.  I am curious about how you came up with a 3 uS max requirement.   

If the old distributor type engine had a top end of 6,000 RPM.    6,000/60  = 100 rotations / second.   So one turn of the crankshaft every 10 mS.    The distributor runs from the camshaft.  Assume it's a 4-stroke so 2 turns of the crank to run a cycle.   10 mS  * 2 = 20 mS per rotation of the distributor.    Let's say it can resolve to 1/10 of 1 degree.   20 ms/3600 = 5.556uS.   

Seems very high performance for that old system.   Must be looking to do more than just scare the neighborhood kids.

[Seekonk]

You are correct. The one I'm working on only had a tach that went up to 6,000 rpm.  Another on of their lot only went up to 4,000 rpm.  They race at Bonneville each year and want the machine to do 10,000 and the distributor only does half that.  That will take a VFD and a flat belt in order to get that.  All the bearings have been replaced.  I am adding a micro and a once per revolution to the machine in order to get out hard numbers if it can be shown that the speed is consistent throughout the revolution.  A Hantek 6022BE was also purchased.  Looking at the ignition pulses and once per rev it should be able to perform the same calculations as the micro.  With open software available, it could be a nice ready made package.  There will also be a hand held strobe to stop motion in the distributor.  The eye can be very perceptive in seeing issues.  I have some other interests besides this machine. The number quoted is about twice as fast as needed in order to bring up any potential issues.  It has been an interesting discussion.  I am concerned about balance issues at those speeds.