View Full Version : Laser tool path display....( IDEA )

08-30-2004, 04:08 AM
Maybe this has been done b4 but I have never heard or seen it.

How many time have you wondered what the actual cutting path on your work piece was going to look like before it was cut?

Have you ever had a sheet with a blemish on it in a small place and wanted to make sure you avoided the blemish?

Have you ever wanted to know where you could safely screw down you work without driving the router through your screw?

Anouncing the "Ynnebulator laser tool path displayer"

Well its more a concept than a finished project. If you had a $4 laser diode and a small mirror on an X,Y pivit that was controlled by rapidly moving solinoids ( Similar to DLP video display technology) With the correct driving electronics you could project your tool path directly onto your work.

The idea is free to exploit by any talented electronics person. I guess it may need to be worked into a program like Mach2 in order to produce an output to show the tool path.

I bet the next post made will be someone telling me its already been done. :)

Graham S
08-30-2004, 06:40 AM
Essentially a difficult project in its own right. A very cool idea but is it needed if you do your drawings within a box that shows the working area of your machine and then zero the machine to that box.

If I could make such a thing I would build a rapid prototyper with it rather than a collision detecter. Assuming I could find a UV laser :)

08-30-2004, 07:20 AM
Why don't you just buy one. :)


08-30-2004, 07:24 AM
Gerry there is a word that rhymes with bird and starts with T.

He he......i had a feeling it would have already been done. Is there anything in this world that hasnt already been invented. Arrrhhhh

Graham S
08-30-2004, 07:38 AM
I can think of a reason you wouldn't buy one and it begins with h.

HOW MUCH!!!!!! :O

08-30-2004, 08:07 AM
I have experience with laser show equipment.

If you were to make a laser tool path display on your own, there are two issues:

1) Unless you are prepared to have a very dark room whenever you use it, you would need more than the <5 mW a standard red laser pointer puts out. And this means expensive.

2) The scanners are very expensive if you want any accuracy at all.

I actually tried to build a laser scanner using a 5 mW 635 nm (i.e. high visibility red) laser pointer, with first surface mirrors glued to metal axles that turned in ball bearings, driven by (slaughtered) speakers. It worked for simple "moving to the music" displays but were very weak even in a very dark room.

Also, no accuracy. You need a closed loop scanner (this is a servo system just like a servo motor/driver) to get positional accuracy.


08-30-2004, 09:42 AM
Unless you are prepared to have a very dark room whenever you use it, you would need more than the <5 mW a standard red laser pointer puts out. And this means expensive.

These days you can get very bright and cheap lasers that can be clearly seen in daylight.

2) The scanners are very expensive if you want any accuracy at all.
Accuracy is not that important for just knowing where to put a hold down screw, or where to avoid a bad patch in the material.

08-30-2004, 10:42 AM
Benny, it was not my meaning to "cast down" your idea, I'm just sharing my (quite substantial) experience with lasers in display applications.

You will need a lot more laser power when you scan a connect-the-dots picture than when just holding the dot stationary. A normal laser pointer is clearly visible indoors in the middle of the day, but even scanning a simple circle on the wall it is difficult to see under the same conditions.

When I talk about accuracy here I'm not talking 1/10 of a millimeter. It would be difficult to build scanners that can hold even say +/-5 mm (+/-0.25 inch) at 0.5 meter from the scanner or so. I'm just saying it's a difficult project if you build it yourself, or expensive if you buy it.

Feel free to use my knowledge in this area, I'd be glad to help! It is a cool project :).


Graham S
08-30-2004, 10:42 AM
Benny, if you are scanning the laser to make the image then it is effectively only at a given position for a very short time that means that you need a bright laser to be able to see the image.

If you want the beam to follow the tool path slowly (slower) then that is different.

If you just want to have a go then get a pair of steppers and a pair of mirrors. You could use belt drives to increase resolution, some quick calcs should tell you how much you need. You then need to just run the g-code, the problem is that the x,y coordinates in the g-code need to be modified because the angle will be arctan(a/b) where a is the distance in x from the point beneath the laser and b is the height of laser above the bed.

Words to that effect anyway.

You might be able to write a gcode converter program to do this.

It would be fun to make some nice patterns anyway using this technique but without the conversion you will get compression of the tool path in certain areas and stretching in others.


Graham S
08-30-2004, 10:44 AM
Ooops some of my post repeated Arvid's.

The solution is slow scanning then. Resolution may still be a problem even if you measure it in clamp widths rather than microns.


Graham S
08-30-2004, 10:45 AM
I'm gonna make some pretty patterns anyway :)

08-30-2004, 10:47 AM
Posting within the same minute! :)

I didn't even think about the possibility to go slow. Sometimes working for some time with a technique can make you very blind to alternative solutions!


[edit: spelling]

Graham S
08-30-2004, 10:48 AM
What about using a tool path verification program to print out the tool path and then make a projector to project it.

Now I really must do some work

08-30-2004, 11:01 AM
Regarding resolution:

With the scanner 0.5 m above the table, the movement of the spot for one full step of a 200 steps/rev stepper is about 16 mm right below the scanner (it gets better the further to the sides you get).

With 3:1 reduction and microstepping it might work? Let's see, one full step at 600 steps/rev gives about 5 mm/step, so even using only half-stepping the resolution is down to 2.6 mm. Should work! :)


No, I realised this is wrong. The beam angle is twice the angle of the mirror, so you would get 31 mm/step for the 200 steps/rev stepper. And 5.2 mm/step with 3:1 reduction and half-stepping.


08-30-2004, 11:59 AM
I just came back from the IWF show in Atlanta. These things are available for about $16,000 to $18,000 U.S. dollars.

This model along with the Carter system seemed very popular.

08-30-2004, 12:10 PM

Interesting idea!

Your concept, I believe, is being studied but, applied as a solution to a different type of problem: “embedded servo” for high-precision CNC – this is a technique analogous to what is used in a disk drive – get the position information as close to where the data is for the R/W heads to follow or, in the case of CNC, get the position information on the material for the cutter to follow.

Unlike today’s techniques where position information is relatively far removed from where the cutter is, thus flexures between where encoders are and the point on the material at the cutter are usually not accounted for, an embedded technique would essentially use a scheme of having the cutter close the loop and follow a point-source of light that is projected and traced out on the material.

Still, many problems to solve with this method. . . You might have a chance to be a pioneer.


08-30-2004, 01:19 PM
Just a thought...how about using several cheap laser diodes and use them together to give out a brighter beam? I'm sure there is some way to "bend" the beam and illuminate all in the same path.

Graham S
08-30-2004, 01:37 PM
For you reference a just waved my 1mW HeNe about as fast as I could by hand and could see the beam fine. I should think that is faster than steppers can manage. Galvo's can of course go MUCH faster!


08-30-2004, 03:22 PM
Why not use mass production items?
LED projectors are not very expensive now. Especially if you can get an old low-res projector on the used market. Then use on of the available backplot programs to generate an image of the toolpath. Hang it above the machine at a height that give the correct image size and off you go. It will not have resolution to make fine lines, but for placing the cut on a sheet, avoiding clamps, holes already cut and such it will be fine.

08-30-2004, 06:27 PM
All I know is that the laser I have, shines onto two offset spinning angled mirrors in series will sufficiently make large paterns on my wall in daylight. I know the lasers of old and how dim they are, but these lasers will work for what I am thinking of.

As for for accuracy, I guess that gets down to the driving electronics for the mirror solinoids. It is my guess that while the circuit may be difficult to design, the actual circuit would only cost $50 - 100 bucks max.

Graham S
08-30-2004, 06:38 PM
But what do you even feed that circuit with? Are you going to try and make step and direction galvanometers?

I am going to play with steppers, even if it is just for pretty patterns :)

edit: for speeling

08-31-2004, 11:09 AM
I don't think the electronics for a scanner would be too difficult; step and direction could easily be implemented with a microcontroller that counts steps and outputs the result to a DAC as a +/-10V absolute position reference. This is how most scanner galvo's are controlled.

(For those of you that don't have experience with laser scanning equipment, galvo's - or galvanometers - are a special kind of "motor", something like what you can find in ordinary analogue panel meters :idea: (!)

- hey! there we have it! Maybe it would be possible to rebuild a couple of panel meters for slow scanning? Remove the arms and attach small first surface mirrors instead; drive them directly with the voltage from the DAC above!

What do you think?


08-31-2004, 05:10 PM
Meters have built in damping, not good for rapid movements.
The head movement of all 3,5" and smaller hard drives now have a voice coil movement. It is built to make rapid movements (just a few milliseconds to stabilize in a new position) and also to do this millions of times over (high accuracy ball bearings).

So there you have a near perfect movement. The mirror will prolly weigh less than the head and arms, so it may even be faster. What about the servo loop? How to sense where it is at any moment? Array of photo sensors? Hard drives have servo info embedded in the platter, which will not be there any more.

BTW: The head coil out of an old 5,25" HD is strong enough to drive a filing machine with swiss files. Just cix center pos with springs and apply AC and it will vibrate violently.

Graham S
08-31-2004, 09:59 PM
If you try and use something like a moving coil meter then you rely on the spring for movement in one direction. Slow scan indeed.

If you want real speed you need pairs of coils or current reversing, servo type control is not a given on galvanometers, some do not even have a DC response. Remember the moving meter gives the effect of control because the force produced by the coil is balanced by that of the spring at some position. To get servo type control I think you need coil pairs

HD heads might be just the thing but I am not sure. It all depends what you call difficult and how much you can be bothered to do it.


08-31-2004, 10:19 PM
I think an adaptation of this could work.

This flash video may take a while to display so please be patient. It may not work in all browsers either.

In order to rewind and see the video again, just right click on it and select the rewind option.

If it doesnt work for you, you can try and visit here http://www.dlp.com/dlp_technology/includes/demo_flash.asp?bhcp=1
<meta http-equiv="Content-Type" content="text/html; charset=windows-1252">
<title>New Page 1</title>
<meta name="Microsoft Theme" content="cascade 001, default">


<iframe name="I1" src="http://www.dlp.com/dlp_technology/includes/demo_flash.asp?bhcp=1" width="588" height="416" scrolling="no" border="0" frameborder="0">
Your browser does not support inline frames or is currently configured not to display inline frames.



Graham S
09-01-2004, 07:57 PM
In what way would it be adapted?

I have seen something quite similar with a single mirror and flexure pivots and 4 coils. That was for a laser scanner as part of a microscope however it was not servo controlled just fed with sinewaves I think to get a sort of interlacing effect.

It's the servo control bit that makes it hard to scratch build I think


09-02-2004, 02:53 AM
Graham, I never suggested servos or steppers in the first place. Thats what i meant a mirror on a pivot with solinoids doing the piviting work. It would only need one small mirror and 3 or 4 solinoids that was controlled by an enlectronic interpreter. The elecronics is out of my technical scope but I am sure it can be done quite cheaply.

09-02-2004, 05:19 AM
The term "servo" does not imply a "servo motor". A servo is anything that uses feedback to go where you tell it to.

I think you are right in that it would be quite easy to build a driver for a few solenoids, the problem is that if you want to use this simple (i.e. open loop) driver you will need a very sophisticated mechanical construction of the mirror and pivots, to get any kind of accuracy (+/- 5 mm corresponds to about 1/1250 of a turn of the mirror if the mirror is half a meter from the table!).

The other way is to use a less sophisticated mechanical construction and use feedback. The biggest problem here is how do you build a feedback system sensitive and accurate enough to sense such a small angle?

Perhaps one could put tiny magnets on the mirror and use hall effect sensors?


Graham S
09-02-2004, 06:43 AM
Indeed I did not mean servo motors just servo control. You need calibrated control over the angle and if you want speed this can't just be against a spring. Voice coils can be though of as being torque proportional with current, if you work them against a linear load you can get linear control but the speed is limited by the spring. By using them in pairs you can get the speed back but control becomes more difficult unless it is simple AC oscillation. If you can pick up a galvo on ebay that gives a linear angle change with voltage then you would be sorted perhaps but you still need electronics to intergrate (count) the step and direction pulses and produce the correct voltage with linearizing factor.

One feedback mechanism might be a linear photodiode, they can act a bit like a potentiometer but with light as the slider. Would need another light source and possibly a second mirror too. Good idea on the hall sensors, they are now being used on some micro RC servos and work pretty well however noise will become an issue of course.

Why not give steppers a go, I would, still going to be much faster than the router in checking the tool path. In fact I intend to try this come October when I am free to play not just talk about it.


09-02-2004, 05:31 PM
Or, you could do it this way:

http://elm-chan.org/t/vlp_pre/ (the page is in japanese; use babelfish.altavista.com or similar to get an almost totally unintelligible english translation :D)



Graham S
09-03-2004, 10:08 AM
cool, homebrew capacitive feedback. Certainly cheap but I doubt we shall see a rush of people making them.


09-03-2004, 11:40 AM
Hi all,
This sounds perhaps to simpel, but can we not put the lasediode in the toolholder of the router, then run the G-code, of course without powering the router? this is of coarse slow, but it surely will work.

Graham S
09-03-2004, 03:40 PM
I still thought it would be nice to make a laser fly around the room with stepper motors so I started knocking something up. First problem was the mirrors. Most household mirrors are glass with a metal coat on the back which then has protective pain layers. This means if you start bouncing lasers off it that you get reflections off both surfaces. What you really need are single surface metal mirrors. So I made some:

1. Take a self adhesive vanity mirror and cut it to size using you prefered method.

2. Remove the sticky back.

3. Dunk it in paint stripper.

4. Leave as long as you can and gently rub off the paint layers with a cotton bud (Q-tip). Do it slowly and reapply paint stripper if required.

5. Rinse and you have a mirror that is perfectly good enough for bouncing a laser around for fun. You will find they are slightly transmissive but not very much and good enough.

Next to glue them to a stepper shaft for a trial run.



Graham S
09-03-2004, 06:12 PM
In possibly the worst optical set up ever I just tested the steppers, well they certainly move without any load but as expected, ungeared the resolution is really poor, I played with single axis for a while and then managed to add the second axis after putting a biro shaft extension on the motor to get the mirrors as close as possible. The whole thing was held with superglue, bluetac and for suport a pair of socks. No kidding, the laser mount was a pile of jumpers and the optical bench my bed.

Depite this I managed to get a circle on the ceiling but only at low speeds, at higher speeds the motor was just vibrating too much and I got a square with a tiny circles at each step, looked nice but not much good for what we want. FYI I was just running a G02 code in turbocnc to get the circle.

Still I am quite excited about tracing a laser around with CNC, I have an idea about getting ultra high angular resolution over a small angle on a serious budget, actually it is not my idea, it comes from a book of mechanisms, I will post it later.


09-03-2004, 06:26 PM
Graham, I probably didnt mention that I once had this laser kit that had two mirrors.
both mirror spun at variable speeds according to music input. The laser hit one mirror and then bounced onto the other mirror. bevause the mirrors had a slight angular ofset on them it would produce a wobble. the two of them wobling would produce fantastic conatantly moveing projections onto a wall. This is how i know that the laser will be able to be bright enough even when spreadover a big distance.
Here is the kit I used. It is a pity they dont have any pictures to show the results. http://www.oatleyelectronics.com/kits/k115.html

Arvidb, That is the perfect link that shows the laser control device. It really is a pity i dont read Japanese. Thanks for posting the link, that does show exactly what I am thinking about.

EDIT: I didnt mention that the laser beam reflected in series off the two mirrors and not in paralell. The reason i feel to mention this is because if one mirror just did X and -x , and the other mirror did Y and -Y that would mean that the two mirrors only had to cope with 1 plain each. The end result is full coverage. I know I explained this cryptically.......ill do a drawing to show what i mean.

Graham S
09-03-2004, 07:23 PM
Those patterns are Lissajous figures and because they are repeated so fast you can see them as your eye intergrates over many repetitions. Same is true for an image although it depends on how fast you can refresh it. I think Arvid's speaker sytem which would probably produce a random pattern would be a worst case senario.

Anyway there is a lot of work between getting some gcode and making a servo scanning mirror system. Both software and hardware. This is why I am opting for the slow scan approach, not that slow but not lightning fast.

You have to put the mirrors in series, the only other option is a single mirror on a 2 axis stage. Mine were in series, one for each axis.


p.s. I have a video, will try to post somewhere

Graham S
09-03-2004, 08:48 PM
Here is the way of getting very high angular resolution on the cheap as it were.

The leadscrew is attached to a semicircular "platform member" with two metal bands. These sit one on top of the other. When the lead nut moves it pulls the platform member by the metal bands. causing it to rotate on it's bearings. If the platform member is very near to being in contact with the nut (like photo not diagram) then the movement of the nut corresponds to a specific part of the platform members circumference.

So here is the proposed system: a nema 17 1.8degree stepper, 6mm diameter 1mm lead threaded rod. Over length Delrin nut or simple antibacklash nut. Standard ball bearings as thrust bearings. This gives a linear resolution of 0.0025mm in half step mode. If say the platform member had a diameter of 2" then the circumference is 2X25.4XPI = 159mm. The resolution of the linear part is 0.0025mm which is 1.56e10-5 of the circumference and so also of 360degrees. So the angle moved with a step of 0.0025mm is 0.005 degrees. That is a angular gearing factor of 160:1 between motor and platform member.

Quite good considering it is just a few bits of shim.

Obviously with gearing comes slower speeds but even at 160:1 the movement of the beam will be pretty quick. There is no backlash in the flex mechanism only in the lead scew so it is not perfect but still pretty cool I think.


Ref: Mechanisms and Mechanical Devices sourcebook.

Graham S
09-04-2004, 05:43 PM
I did a little drawing of a unit for a Nema17 stepper and a 2" diameter wheel. With just 10mm travel on the nut I can get 60 degrees total beam deviation and with 0.005 degrees angle on the mirror (which equates to 0.01 deviation of the beam) I will get a resolution of 3mm.

At a range of 20M!! at least in the centre of the operating area :)


09-05-2004, 12:10 PM
Cool! I'm looking forward to see pictures if/when you get it done!


Graham S
09-06-2004, 12:35 PM
Hey guys, this is SO COOL:


Combine a laser scanner type system and a CO2 laser and have some fun! I intend to!!


09-06-2004, 03:16 PM
That is tremendously cool! I especially like the aluminium marking and the SS welding.

There are some interesting documents under the Documents menu.


04-10-2006, 11:19 PM
Any one looked is this idea anymore?

It sure would be nice to have a laser guide for setting up pre cut parts.