Gimbled laser head for 5 axis 2.X design

[awesomenesser]

I didn't realize you had other reasons for using the arduino. (I just prefer to use AVRs because they don't have the arduino bootloader and overhead)

I still believe you could do this with interrupts (even if the arduino gets hung up on something the interrupt always triggers) but yes your method will work and is guaranteed.

I just looked on digikey and they had a variety of up/down binary counters unfortunately they are out of the 8 bit ones but you could use multiple 4 bit models.

Since the supermodified servos have 12 bit encoders for their 360 degrees of travel that means for our 90 degrees we need a 10 bit counter or 0 to 1024. So we will need to cascade three 4 bit counters. This also means that we will need a parallel to I2C expander for each counter (they are addressable so we can use the same i2c line in fact the servos can share the lines too).

We will also have to make sure our counter chip is capable of having a preset value. This way when you press "clear" it gets set to 512(if we count our max as 1024 or we could use our actual counter max which is 4096 and our preset as 2048) which is "zero".

All of this circuitry could very easily be fit onto a arduino shield board.
- 1 parallel port
- 6 4 bit counters
- 2 IO i2c replicators
- 1 output connector [5V, GND, SCL, SDA] (DB9?)

[r691175002]

IMO This isn't worth the trouble.

Mechanically it is a pain in the ass to keep accurate, but that's beside the point.

One problem is the tool paths. Five axis CAM software is generally in the 10,000-30,000$ range. Your only realistic option is to hack together your own half-baked system that lets you specify angles for individual lines. It would be a major project and given how specific it is you would probably be on your own.

You then get screwed on cut accuracy because when the laser is angled, every millimetre up or down the material moves will shift your cuts. Lets say you want to cut a 1"x1" square with a 45 degree bevel. If your z-axis is off by 0.01" then the square is going to be larger/smaller by the same amount. Measuring the exact distance from the material surface is going to be a pain.

Suddenly you are spending a pile of money on an accurate z-axis system, two extra motors and a bunch of electronics. You are getting a fairly inaccurate low power hobby laser and have no way of generating toolpaths.

It just sounds like too much work for a single person, especially when you consider that all you are getting out of it is angled cuts.

The above is just my opinion worth two cents etc.

[macona]

A head for doing 5 axis wouldnt be too hard to build. You are going way overkill using little servos and the like. All you need is a mechanical setup similar to whats used in medical laser delivery systems with rigid arms.

But, as said before, the machine is simple. The software is not. 5 axis cost a lot, as in the price of a small car lot. And even when you have the software its not that simple. Guys that can program 5+ axis machines earn their keep.

[wdyasq]

The idea of a variable angle is interesting. As the angle of the tilt of head and mirror angle should be exactly a 2:1 ratio it should be a simple matter to build a timing belt setup to coordinate the mirror/lens. Alignment might be interesting, to say the least. A large bushing mechanical plastic bushing could hold the rotating lens assembly

The geometry of length of focus and Z axis height should be fairly simple math. A spreadsheet could be built for the conversion - I think?

While we are at it, I am needing to cut materials long enough where an elevating bed is a problem. I'll make another post on a possible solution to that.

Ron

[dirktheeng]

The idea of a variable angle is interesting. As the angle of the tilt of head and mirror angle should be exactly a 2:1 ratio it should be a simple matter to build a timing belt setup to coordinate the mirror/lens. Alignment might be interesting, to say the least. A large bushing mechanical plastic bushing could hold the rotating lens assembly

The geometry of length of focus and Z axis height should be fairly simple math. A spreadsheet could be built for the conversion - I think?

While we are at it, I am needing to cut materials long enough where an elevating bed is a problem. I'll make another post on a possible solution to that.

Ron

This is exactly what I set up but with 2 timing belts each set of pulleys is a size ratio of sqrt(0.5).

[dirktheeng]

IMO This isn't worth the trouble.

Mechanically it is a pain in the ass to keep accurate, but that's beside the point.

One problem is the tool paths. Five axis CAM software is generally in the 10,000-30,000$ range. Your only realistic option is to hack together your own half-baked system that lets you specify angles for individual lines. It would be a major project and given how specific it is you would probably be on your own.

You then get screwed on cut accuracy because when the laser is angled, every millimetre up or down the material moves will shift your cuts. Lets say you want to cut a 1"x1" square with a 45 degree bevel. If your z-axis is off by 0.01" then the square is going to be larger/smaller by the same amount. Measuring the exact distance from the material surface is going to be a pain.

Suddenly you are spending a pile of money on an accurate z-axis system, two extra motors and a bunch of electronics. You are getting a fairly inaccurate low power hobby laser and have no way of generating toolpaths.

It just sounds like too much work for a single person, especially when you consider that all you are getting out of it is angled cuts.

The above is just my opinion worth two cents etc.

if everybody said no right off the bat, we wouldn't get anywhere. And you are not right about the software, you can get 5 axis cnc software for 500-1000. Most standard cam software does this.

[dirktheeng]

A head for doing 5 axis wouldnt be too hard to build. You are going way overkill using little servos and the like. All you need is a mechanical setup similar to whats used in medical laser delivery systems with rigid arms.

But, as said before, the machine is simple. The software is not. 5 axis cost a lot, as in the price of a small car lot. And even when you have the software its not that simple. Guys that can program 5+ axis machines earn their keep.

you still need a way to actuate it... any suggestions beside this are welcome.

[dirktheeng]

I was wrong, the cost was a little higher than I thought. What I was remembering was options to 5 axis, but not realizing that they weren't included in the base software. I don't think it will be $10k to get a solution but it is going to be more than I want to spend. I'm going to have to think about that a little more. You may be right.

IMO This isn't worth the trouble.

Mechanically it is a pain in the ass to keep accurate, but that's beside the point.

One problem is the tool paths. Five axis CAM software is generally in the 10,000-30,000$ range. Your only realistic option is to hack together your own half-baked system that lets you specify angles for individual lines. It would be a major project and given how specific it is you would probably be on your own.

You then get screwed on cut accuracy because when the laser is angled, every millimetre up or down the material moves will shift your cuts. Lets say you want to cut a 1"x1" square with a 45 degree bevel. If your z-axis is off by 0.01" then the square is going to be larger/smaller by the same amount. Measuring the exact distance from the material surface is going to be a pain.

Suddenly you are spending a pile of money on an accurate z-axis system, two extra motors and a bunch of electronics. You are getting a fairly inaccurate low power hobby laser and have no way of generating toolpaths.

It just sounds like too much work for a single person, especially when you consider that all you are getting out of it is angled cuts.

The above is just my opinion worth two cents etc.

if everybody said no right off the bat, we wouldn't get anywhere. And you are not right about the software, you can get 5 axis cnc software for 500-1000. Most standard cam software does this.

[r691175002]

This project might interest you:
http://www.kickstarter.com/projects/109 ... -machine-p

You get tonnes of diy 3-axis routers but 5-axis versions are very rare mostly because of the software issue. This guy outlines a few options although I don't know how mature they are.

[dirktheeng]

Well,

Thinking about this a little bit, we can reduce the complexity of the problem a little bit by functionally calling this thing a 4.5 axis machine. That is, the Z axis is not really independant of the U and V axis. Once the U & V angles are set, one can calculate the Z axis height. This could be done in the Arduino becuase it will know the U and V positions and could send pulse signals to the Z axis to make it move as needed to keep the proper distance from the head.

This brings the CAM software price down significantly into more normal rhelms. Ez-Cam has a product Ez-Mill for $2520 with shipping. It is fourth axis capable.

After thinking about this, I started looking into Wire EDM CNC control. A wire EDM is very much like a laser and I think we can make use of that. There are several very affordable products out there. Of note is Dolfin CAD/CAM. They produce a fairly capable cnc software package which has 3 axis milling with some 4 axis options (I haven't looked into this very much). The hobby price for this software is $399. They also make a wire EDM which creates 4 axis tool paths, but don't offer a hobby price for that, it is $995. We will still have to look for G1 feeds and turn on the laser, but this is probably sufficient to do what we need to do.