Guy who was looking at the epilog rotary!

[matthew knott]

I typed a response but the thread vanished guess it got pulled but FYI this was my response, tips you may need

Sounds good in principle but a few things to think about,
First, you can buy for less than $250 a chinese rotary that will give you all the mechanics you need, (3 jaw chuck, tail stock, ect) We have one and they are not bad, they use a stepper motor (often 3 phase) but have all the belts and beraings, something like this http://i01.i.aliimg.com/wsphoto/v0/693070485/4th-font-b-axis-b-font-font-b-rotary-b-font-font-b-axis-b-font.jpg

Secondly and this is the real problem, the method epilog use relies on rollers that are set so as the part diameter changes this is compensated for, ie roller moves 1 inch, part on roller moves 1 inch, when you have a 3 jaw chuck, as the diameter of the part increases so does the distance travelled. The software does not support angles for example, this means you have to do a work around everytime you use the rotary, bit of maths etc, and its never really going to work well, as if you had say a 20' round part, every split move would be large, to compensate the graphic would have to be very small, therefore the density would be very bad. Epilog have changed all this for the YAG laser and now do a nice little rotary unit, but without firmware-hardware changes its never going to be right.

Our rotary we use on our YAGs allow you to enter the part diameter, and also allow us to enter degrees and are extremely accurate, we do lots of micrometer barrels and the like.

I like your thinking, but before you start on the mechanicals you need to work out the electronics-software side, it can definitely be done but i suspect it will be much harder than you may think. But if you like a challenge........

[Rodne Gold]

I deleted the post as it was advertising and commercial in intent , using the forum to sell , against the TOS on the site...

[matthew knott]

I deleted the post as it was advertising and commercial in intent , using the forum to sell , against the TOS on the site...
Pretty obvious why it got pulled Rodney, wondered how long it would last, not long enough for me to reply :)
See hes edited it a bit now !

[Joseph Gamble]

I deleted the post as it was advertising and commercial in intent , using the forum to sell , against the TOS on the site...


Alright, commercial bits put to rest, looks like ive got my work cut out for me. I'm continuing with the build. Hopefully it will be a cool rotary setup. Sorry about breaking the rules guys.

[matthew knott]

Having a think about this whilst eating my curry, and I have thought of a way you can do it, but you will need some electronics/programing -experience or help!
Heres what you need to do:

First you need to decode the signals going to the old motor, this should be quite easy to do with a lookup table, from this you can work out the step/direction pulses,
Next you need to work out what distance is covered by each step, for ease of the maths lets say 10 pulses = 1mm (not sure what it really is but easy to find out)
Now your new rotary has to have a way of being programed with the diameter of the part,
Lets say an LCD, Pic micro & keypad

Now lets say you use a 10:1 reduction gear box or belt drive on your rotary and you use a x10 micro stepper you new rotary would need 20'000 steps to do one full turn ! Fact

Now for this example lets say we have a 200mm part, 200mm dia = 628.4mm diameter . Now if you divide 628.4mm by 20'000 you get 0.03142mm
Now each new step moves the rotary 0.03142mm
The old steps coming from the epilog we said (*) 1 step = 0.1mm so if you divide 0.03142 you get 3.182
This means that for every old pulse you need to do 3.182 new pulses, you cant give part pulses so you need to round to the nearest whole number (3)
each old step must be converted into 3 steps but you have to keep adding the remainder, when this reaches 1 you give an extra step, so 4, then keep whats left in the remainder pot and keep adding to it. If you dont count the remainder you will get a big error
For example if you draw a 300mm in corel and engrave it, the old epilog signal will give (*) 3000 pulses (300x.1mm) the new rotary will output 3x this so 3x0.03142 = 282.78mm, this is less of a problem with smaller radius parts or smaller engraving area.
So if we had a 50mm part 50 x 3.142 = 157.07mm 157.07/20000=0.00785mm .1mm/.00785 = 12.73 round to 13, now draw a a 70mm box, this would give 700 old pulses, 700 x 13 x 0.00785 = 71.435mm.
Haven't got a clue if i have explained that right but i think it works out

The code to do all this is not that complex, i might get my pic development kit out and have a play

[AL Ursich]

Being a Retired Navy Fire Control Chief... I grew up with Tube Amp Analog computers and finished 20 years later with the latest Digital.. This post got me thinking....

Lots of Military Surplus stuff on the market...

Lets say you take a Epilog Stepper and have it drive the shaft of a Encoder or Servo. Even 2 Servos a Transmitter and a Receiver I believe would have the power to move a 3 jaw chuck...

There are even Servo Encoder and Decoder Cards...

Good Luck,

AL

[matthew knott]

Still have the same problem Al, the 'new' rotary has to know the diameter of the part, driving a 3 jaw chuck is the easy bit, knowing how the rotary angle change, relates to distance round the part is the tricky bit.
The epilog system is quite clever in that it does no need to know this, it takes the place of the Y Axis, if the Y axis is told to move 1mm, instead of moving 1mm the roller travels 1mm, anything sitting on roller goes 1mm. You cant do this if the motor is directly driving the part, unless by chance the part is exactly the same diameter as the epilog roller.
I think if you look at the new epilog yag rotary it is programed with the part diameter, i might be wrong !!!
Thanks
Matt

[Joseph Gamble]

There is a nagging issue on this, and that is, what happens if I print a 25 mm design, then scale it to 50 mm and print it again? According to what you're saying here, if the first design at 25 mm did 1 full rotation, then the 50 mm design would then do 2 full rotations.

That means that there is a 'standard' diameter. I wonder what this diameter is.

It is likely the diameter of the drive wheel that they selected in their design.

Their drive wheel is a reasonable 'average' diameter, and this would set the overall stepping scale range (so you dont ever have to make non discrete steps). remember too that rounding a number like 20000 steps to 20001, simply won't be a visible error. Hopefully there won't be a situation where the errors stack up though.

[Joseph Gamble]

On the topic of the magnitude of this challenge (building the rotary system), I and my group of colleagues have what is in my experience a unique combination of capabilities: I do mechanical design, cnc machining of all types, and intro level electronics. My colleagues have some of the best educations available in electrical engineering and control systems. I should be able to get this thing up and running within a couple of weeks pending any long shipping times for unexpected purchases. I think that I and many others can benefit from such a setup, and it should be right on the edge of becoming a reality here..

I also have a maybe not so unique situation, and that is that I am not rolling in cash and need to solve problems, complicated or not, on a budget. So, that should be an adequately descriptive statement about the driving forces behind the project that I am proposing.

[AL Ursich]

So all the unknown issues with size would not be solved by placing a machined metal shaft on a Epilog Rotary and having a 3 Chuck Jaw hanging over the end.... The part you install in the chuck really drives the size... not the pipe on the Epilog rollers...

AL

[matthew knott]

NO! Only if the part in the 3 jaw chuck is the same diameter as the shaft holding the 3 jaw chuck!!!
if you had a 3mm circumference part 360 degrees moves it 3mm ! A 300mm part moves 300mm for 360 degrees!!
as you say you can cheat by scaling the drawing in Corel so it comes out a size but it is a cheat and not a great one. It's because of using Corel and a print driver, the software and laser has no clue what your part diameter is, and for your 3 jaw chuck method it needs it!