PDA

View Full Version : Home built encoder? Out of AOL Cd's?



ibewgypsie
09-03-2005, 11:58 PM
I need a way to cut the tiny slots into a cdrom. How do you suggest? I got about 1000 emitter detector pairs of leds.

Imagine this? a cable driven quadrature encoder mounted on the wall next to my lathe, a rs232 readout with a lil blinking led on it. When my computer tells me I am within 10 thou of my mark it blinks the led to warn me, no more overcuts.

Simple to do with the subroutines I have.

Except cut them stinking lil slots in.. a edge mill? like they cut slots with? How do you suggest? Since it is a plastic etching is out isn't it?

I learned something neat, the Armstrong tool holder is "flat" on the bottom.. you just lay the flat bottom on the HF tool grinder angle inclined platform and sharpen it to the same angle as a drill bit. CUTS CURLS like nobodies business.

------------------
David Cofer, Of:
Tunnel Hill, North Georgia

Bruce Griffing
09-04-2005, 12:26 AM
Look at this site. He printed it on plastic and attached it to the CD.

http://www.buildyouridea.com/hardware/pic/pic.html

hammerhead74000
09-04-2005, 12:40 AM
Are sure you want to? Do you have a laser printer? Why not just print an encoder pattern on a laser transparency?

I suspect that an ink-jet would work also, but I would be very careful about the ink smearing, and I would want to make sure that it had good ink coverage...

Another idea - the top of a bulk-pack of CD-Rs has a "CD" with nothing on it - no metalised layer, nothing, it's clear. If you have a printer that can print onto CDs, you could just print the encoder pattern onto the clear "CD" (just watch out that the registration is correct!)

hammerhead74000
09-04-2005, 12:42 AM
Arg! Beat me to it!

http://bbs.homeshopmachinist.net//wink.gif

ibewgypsie
09-04-2005, 10:22 AM
Someone suggested printing laser overtones on the cd and using a reflective led. No clue.

Too bad I don't understand how the actual cdrom reads the disc better. Imagine the resolution. A binary encoder, well they throw 8/16 bits out to denote exact position instead of a encoder/count up/dn logic. It takes more inputs but is exact. We used to use them on robots.

David

JRouche
09-04-2005, 12:26 PM
Ahhh, yer just like my Dad. AOL and everyone else send these things out. Dont wanna throw them out cause man, they have value. Lets see, oh yeah, hung a bunch on string at different levels and it makes a "sun sparklin thingy".

Dont forget they are polycarb so it may cut oddly.

How tiny are the led pairs? Bet they are small huh? What is the range on them (max dist. between)? JRouche

ibewgypsie
09-04-2005, 12:35 PM
J:

I got the datasheet coming.. I lost it. I have been reorganizing around here and building books on each project.
About a year ago I started developing a inhouse drive and bundled all the data sheets together. Yeah. I can't find that book.

Seems they were 3mm or less, They have a focus lens looks like a bead on each part. I can't even remember which is which right now, one is red, other clear. Polarity is not marked by pin length either.

Thinking about building a heat-punch to punch square holes in it on my indexer. I got a old weller solder iron w/plastic welding tip. Let the cnc do the grunt work.

[This message has been edited by ibewgypsie (edited 09-04-2005).]

BillH
09-04-2005, 12:51 PM
I took apart a cdrom drive to get the brushless motor out. The reading head with lense had the ability to move left/right, forward/back with some coils.

madman
09-04-2005, 03:16 PM
They are kinda tough but shatter when bent. I would suggest a small diamond abrasive and grind through while flooding with a coolant to prevent the disc from melting. Will work well.

jburstein
09-04-2005, 03:22 PM
Yeah, just print the encoder pattern you want on adhesive backed paper, and "adhese" it to the CD. Best way to do it is to make a little fixture that'll guide the cd down onto the printed pattern using a rod through the center hole. The sticker (the thing you printed) should sit on some foam, and preferable the center of it should be slightly higher than the outside edges (sort of cone shape), so that you avoid bubbles as you apply it.

As it turns out it doesn't matter too much if the encoder pattern isn't dead nuts concentric with the axis it turns on (as long as you use circle sectors as your windows--if your windows are just bars I don't think that applies). A small amount of error in concentricity will lead to a very slight position error. I have numbers someplace, but I don't know them off the top of my head.

-Justin

CCWKen
09-05-2005, 07:59 PM
Why do you have to have slots? Couldn't you just drill small holes? It doesn't take much to turn on a detector. They're pretty fast too.

I've got a couple of HD platters, if you want to play with those. I can't tell what they're made of--Not plastic, Not magnetic and hard. Email your addr, if you want them. Libre, brother.

Your Old Dog
09-05-2005, 08:19 PM
This might be a stupid idea but I'll throw it out there.

If you mated a cd to a piece of masonite and then afixed it to a rotary table, you could cut slots in it using a dremel with the Quarter sized saw blades they have. It would be tedious to be sure but it should work if all you want are thin slots equally spaced around the CD. Only effort would be mounting the Dremel tool so that you could lower it into the cut.

ibewgypsie
09-05-2005, 10:28 PM
CCW:

I got a couple of them too, "Birth of a nation" and a "old submarine movie U88?" I don remember.. I got a player too somewhere.

This lil encoder into the parallel port works in dos, don't work in windows visual basic.

I been at it 14 hours. It is kicking my butt, The DMA board can read inputs every 2ms-3ms, but I keep losing counts and I can't figure it out. Using a 128 ppr encoder so the pulse width is not that narrow. Its funny a stamp can read it without losing counts.

I mounted a router into a homemade tool-holder for my mill.. It'll turn a dremel bit quite well.

David

[This message has been edited by ibewgypsie (edited 09-05-2005).]

fixxit
09-06-2005, 01:47 AM
At the last Cabin Fever expo in January 2005 I had a chance to talk to the fellow who designed the "Shooting Star Technology" system. He was a techie who was VERY willing to talk. The company regularly advertises in HSM.

I basically outlined the same system to him, and he shot a bunch of holes in that method of measuring. The problems were associated with using a cable drive to the encoder, not the encoder disk.

Shooting Star eventually settled on a rack and pinion system. I can't remember the critical flaws, but there were several, and they were enough to stop me in my tracks.

It might be worth a deeper look for the flaws, before investing a whole lot of time and money.

------------------
457863656C73696F72212020466C69636B206C6976657300

[This message has been edited by fixxit (edited 09-06-2005).]

JPR
09-06-2005, 02:10 AM
<font face="Verdana, Arial" size="2">Originally posted by JRouche:
Ahhh, yer just like my Dad. AOL and everyone else send these things out. Dont wanna throw them out cause man, they have value. Lets see, oh yeah, hung a bunch on string at different levels and it makes a "sun sparklin thingy".
</font>

some other ideas http://stupidco.com/aol_throne_intro.html




[This message has been edited by JPR (edited 09-06-2005).]

Stanko
09-06-2005, 05:33 AM
Cant you just draw it save as a dxf then take it to a water cutter place or a laser place.how many slots? what about a punch ?

Weston Bye
09-06-2005, 11:16 AM
I did a quick layout. Drill 100 holes, #50 drill, 3.6° spacing, on a 2.250 radius circle. This gives nearly equal "light" and "dark" areas for interrupting the light beam. Add 2 optointerrupters approximately 10° apart to sense the holes and you get quadrature output, so you can sense direction.

Because the plastic may be reflective, even in the holes, I would paint the area with a thin coat of flat black paint. That will improve edge repeatability.

If your leadscrew is 20 turns per inch, the encoder will give you .0005" resolution.

1"/20=.05" per turn, .05"/100=.0005" per hole.
If you get fancy in software and count positive-going and negative-going edges the resolution jumps to .000125"

Still fancier software and additional properly spaced optointerrupters to provide more edges to detect can increase resolution further. The limiting factor in accuracy then becomes the accuracy of the spacing of the holes.

Wes Bye
Grand Blanc, MI

ibewgypsie
09-06-2005, 12:47 PM
Software is working, must be dedicated while reading the encoder.

Not missed a count since.

I bought 3 clarostat encoders 128ppr quadrature optical... At $50 each.. ouch.

I plan on making some homebuilts out of the cdroms.. I think about using the flood coolant and the router attachement I built.

Right now? I am getting over the flu.

I can post the exe code using the parallel port when I get done with it. ANYONE can use the parallel port. It supplies voltage and reads high speed data.

darryl
09-07-2005, 03:49 AM
I wonder how well the cdrom would work if you indexed a bunch of radial scratches through the metalized side. You might not need to make slots right through the disc. If the detector was put on the scratch side, the emitter would flood the clear side of the disc, and light would come through the sctatch lines.
By the way, I think the detectors on those led sensors are the dark ones. The clear ones are the emitters.