Open Source Scanfail Unit (phase 1)

[dave]

Ok,

Been doing a little bit of work this morning..

In the pic I have 2 differentiators summed into a absolute value ckt built up on baord. Ill do a bit of testing before going any further... What im expecting to see on the scope is positive going spikes that correspond to each new set of points.

What can to mind last night was to use a peak detector instead of the absolute value circuit. Fast attack, and variable decay, to allow some delay.. Turning the absolute value into a peak detector would be a matter of adding a diode, resistor, and capacitor...

Ill post a few cro screen shots when I get a chance to power this all up

[dave]

and another one

input and output of differentiator..

[dave]

and another one

top trace - position signal

bottom trace - after summing and precision rectifying..

[big error in this as positive and negative velocities are being summed and hence canceling out]

[aijii]

Is there supposed to be a delay between the scanner stopping/changing directions in top trace, and the bottom trace going to zero ?

Also - something I was thinking of a while ago. Rather than just having the safety circuit shut off the output once velocity drops below a certain amount, why not have it dim the outputs as the velocity approaches the cut-off threshold ? I wonder how hard this would be to do and if it would overly complicate things...

[dave]

Is there supposed to be a delay between the scanner stopping/changing directions in top trace, and the bottom trace going to zero ?

There isn't... I Just realised a big flaw with what im doing tho.. I think Ill have something working by the end of the week if I get a couple of nights t work on this

[drlava]

Glad to see some scope plots of this. You do need a filter on the diffenentiator, or as you can see it will produce spikes at the sample rate at every position change increment. Don't think you need a peak detector, though. Just a simple filter after the ABS circuit as in the schematic on the previous page.

[dave]

another pic.. Now we are getting somewhere

The top trace is the output of summing on Dr. Lava's circuit.

The bottom trace is after a little bit of filtering... This was pretty rushed, late last night, but I think there is potential here

I think I'm nearly ready to drop the signal into a comparator and start scanning some frames...

BTW, the traces are of a circle being projected.

[Miss A.]

Hi I'm back again because this topic is important.

It can be done with a few IC's !
See http://www.mylaserpage.de/html/scanner_sfp.html
For less than 80 Euro's you have your SafeScan PCB.
It swiches off <100mSec, it has 4 opto outputs and is smaller than a matchbox.

[Pitts]

Do you know if there is an English language version on this site somewhere.

Thanks

Ray

[drlava]

here is a babelfish translation of the operation in the manual:

"After switching on of the current supply on those flash Status LED on the Safety briefly alternating up. From this moment on a/D transducer examines the position of the scanners, waits one moment, measures again and compares the values. A measurement lasts clearly below 1ms. The path difference is smaller than the value stopped by potentiometers the lasers is switched off and the red LED shines. That is also in the condition By of the scanners the case. I start the show am thus after at the latest 5ms a movement detected and the lasers am again capable of being activated. Only one fails the scanner and second moves still becomes it Safety not actively."

so, it basically compares the current sample with the last sample and turns the laser off if they are the same or some approximation thereof. It doesn't sound like it does any higher calculations, although the processor is capable of much more.