Hello everyone.
I am building a new machine and need lots of help. I have a decent CAD background, and know my way around a shop.
The machine will have 120" X 60" X 22" capacity, and will be used primarily to cut wood and softer materials. I like the gantry design, and contrary to the posted jpg's, this will have the A and B-axis mounted to the Z-axiscarriage. Every axis will be controlled by ballscrews, except for the A and B-axis. I believe those will be geared for power and precision.
My questions are:
How much servo will it take to drive this monster?
How much can I expect to spend on these servos/controllers/everything except my laptop cpu?
If I decide to use two screws for the x-axis, can two servos be sync'd up worth a darn? (Texas dialect, please ignore)
With all of the above mentioned questions, the budget for this project is $12000 us dollars, 2000 of which is for my machinist' friend.


I have already talked to Chris about the long ballscrew (10 feet) and I am opting for the rotating nut option.
Thank you all who have taken the time to read this.

Rob

Rob,

Mach2 and the newer version 3 can do the sync'd (I'm from Arkansas and understand your language just fine! :banana: ) axis' thing for you and still have 5 axis control. As to your decision on the ball screw...just curious if you have considered rack and pinion? Other than the z axis clearance, those dimensions are about the size of my machine which has used rack and pinion for 13 years now. I still get quality parts off of it. N/S/E/W Texas?

Mike

edited www.artofcnc.ca

Hey Rob, rather than edit agian I though I would repost. Your servo needs will be dictated by several things.

1. Linear rail design (friction)
2. Weight
3. Desired feed and rapid speeds

I can tell you that you want that sucker moving when it comes time for a tool change. 120" is a long travel even at 354 ipm which is my machine's max rapid. I will say that I am using steppers rather than servo's and could up that to around 500 imp with good servos.

Are you going to build your a/b axis or purchase? If you purchase, the $12,000 will get you about half way to the price of a new Colombo a/b head.
Spend plenty of time designing before you spend too much money and make sure you have good quotes on the machining. Don't ask me how I know.

Mike

You gentlemen sure are fast.
My intentions are to machine out the A B head myself, and power it with geared down servos. I will be designing everything, and my friend is a machinist.I left plenty of room in the Z axis to accomodate. I was thinking of using a small rotating table/indexer for part of it.
Correct me if I am wrong guys (and ladies):
I can pretty much build this machine and position servos anywhere, and the software will adjust to the machine. True/False?

As far as speed, is 40 fpm jog speed rediculous? I do not want to put coarser threads on this thing at the expense of accuracy. Say I want +- 0.005 inch accuracy. I am so new at this that I know I sound rediculous. So, without further ado, back to my rediculous demands...
I hope I can do all of this for under 12 grand...

Born Dallas/FT worth
Reside in Tenessee
Sounds like a country song...

Yet another set ot-so-well-thought-out questions:

My parallel port will operate the controller,
The controller will operate the drivers,
The drivers will operate the servos.

Is this correct? Nothing else to the electro-mechanical side of the house?

Hi Rob,
Looking at your Z-axis movement, I will be having a similar glide. Are you planning to move your z-axis drive motor and screw up and down with the z position while having the nut stationary ? On mine, this would require the motor to project upward increasing the height by the length of the motor, and moreso when the Z-axis is fully retracted....unless you can mount the motor out behind the gantry (y-axis rails) in a downward position.
I've seen other designs that have a stationary motor and screw while the nut moves the axis up and down, but this design won't lend to that type of function.

Bloy

I want tospin the ballscrew to move the z axis, with the motor mounted on top. Might as well fill up the whole garage with this thing, right?

Still trying to figure out how to interface my laptop to the servos... I guess this is a newbie question

I don't want to be an alarmist, but if you plan on applying large forces to the structure, I would make your z-axis rails a little wider to support the axis when extended. Also, you will probably have to beef up the gantry arms with gussets or some form of support as it will flex there too. the combination of just these two areas will prevent you from achieving good accuracy under loading.

... Just what I see from the drawings...

I'm keeping my X-axis rails high and close to the gantry cross members to eliminate the "arms"...only thing, you lose ability to access the table from the side easily, unless an "open" design is used allowing "reaching in".

The...

Laptops contected to the driver ...
The drivers connected to the motors ....
The motors connected to the ballscrew
The encoder feedsback to the driver
And here's the word of the lord

So yes you are right. The easiest way to run your machine is from the parallel port, two pins are assigned to each axis, one for direction the other to tell the drive to step. You might think that servos don't step but the encoders are digital and using a driver such as a Geckodrive (www.geckodrive.com) you can make them step. The result is really smooth motion, you don't notice the stepping.

Your A and B axis will have to be geared down heavily using either wormdrive, harmonic drive or a massive great gear or a combination, not easy at all, this you need to sort out before making the rest if you can as it will effect the design. You also need to consider the backlash etc.

Do you know what software you will use? Driving 5-axis is easy, turboCNC (runs in dos free to try and excellent) does upto 6axis and Mach2 (runs in windows also good) I think will run 5. However software for creating tool paths and g-code is less easy.

Check out this site:

http://www.rainnea.com/cnc.htm

He has written a free plugin for creating toolpaths, it is good but a little complicated. The software is for 3Dstudio max but there is a program called gmax which is a free limited version but you can import into that from Rhino and similar 3D programs.

Home this is some help.

Graham

oh yeah, pairing up servos is a bad idea, use one motor and couple the two ballscrews with timing belts, or use timing belts instead of ballscrews and couple two of these. Don't be scared by talk of timing belts or even rack and pinion, they work well and often better than people actually need.

p.s. Geckodrives are on special offer now I think

However software for creating tool paths and g-code is less easy.

Check out this site:

http://www.rainnea.com/cnc.htm

He has written a free plugin for creating toolpaths, it is good but a little complicated. The software is for 3Dstudio max but there is a program called gmax which is a free limited version but you can import into that from Rhino and similar 3D programs.


If you don't use the above mentioned software to create your 5-axis toolpaths, expect to pay well over you're entire budget for the CAM software alone. :D

Greetings, and happy turkey day.

I have put away the crack pipe, and decided to use stepper motors. I am going to use 750 oz or better, at least for my rediculous X axis.
Anyone know how I can make 5 drivers connect to my CPU? I have seen packaged controllers of 3 or 4 axis, but not so with the 5 axis.


Rob

TurboCNC can control up to 8 axis with 3 parallel ports. Mach2 can do 6 axis. You may need to make your own breakout boards, but it can be done.

You might look at the two controller kits from EAS. I am new to the world of CNC and have been searching the net for my own project. I am thinking about trying 2 of the MicrostepPLD drivers. These kits will handle up to 3amp at 55volts.

http://www.embeddedtronics.com/

I plan to use the opto-isolated interface card from CNC4PC.

http://cnc4pc.com/Parallel_Port_Interface_Card.htm

Yet another set ot-so-well-thought-out questions:

My parallel port will operate the controller,
The controller will operate the drivers,
The drivers will operate the servos.

Is this correct? Nothing else to the electro-mechanical side of the house?

If you run Mach2 or any of the other "parallel port" programs, then your computer will be the controller, which will operate the drives through the parallel port. No external controller hardware needed.

There are also dedicated HW controllers that communicate with the computer via RS232 (serial). Here the "controlling" is split between the PC and the hardware. And of course there are more expensive solutions as well...

You might get a problem with step rates, depending on the resolution you want: I think the parallel port is limited to about 45000 steps/s (?), and if you want a high jog speed and a high resolution this may be a problem. 480 IPM = 8 in/s, 45000/8 = 5625 steps/in - should be enough, I guess, so forget what I said!

Arvid

Two quick concept sketches.
I am considering either gears, or chain. I believe with decent chain and an adjustable tensioner this would work.
Also on the board is finding two small, lightweight rotating tables (like you would see used with a mill) and adapting those clumsy handwheels into motorized happiness.

I smell turkey...

"The easiest way to run your machine is from the parallel port, two pins are assigned to each axis, one for direction the other to tell the drive to step."

So you can use say a 4 axis driver and a single axis or even six single axis boards. This can be done with just wires from a d-type connector or proper boards, the choice is yours.

There is nothing mad about servos, did you look at Geckodrive?

Graham

I looked at geckodrive. I like the price of their drives for Steppers. I am still not sure how to connect them all.
If I use two 636 oz stepper, a 1080 oz stepper, and two 212 oz steppers (for the A-B head), the appropriate gecko drives for each, what would you recomend to connect this mess to my parallel port? I like the idea of using my CPU as a controller. That sounds quite nice.

I smell sweet potatoes...

I'm sorry but I don't quite get your question, what part of the connection don't you get? If you use TurboCNC then you can I think drive 5-axis from a single parallel port but you would have to do without home switches and perhaps just stick with simple limit switches. The parallel port has pins, two can be asigned to each axis within turboCNC, you then connect with wires from parallel port socket that will connect to the cable from the PC to the drives OR you make a proper little "break-out board" which does the same thing but looks neater. There are some commercial ones available (try google) or you could knock one up.

You really don't have any other option than using a PC as the controller for a hobby machine.

Graham

p.s. Read peoples replies more carefully, I think you may be missing good info.

Rob as Graham said, let's take a breath, and take this a little at a time. This machine you are proposing is quite a bit to bite off and if you are serious about doing it take it slow, careful and pay close attention to those giving advice.

First things first. What type of work do you plan to do on this machine? I know you mentioned wood and lighter materials, but is it going to be used in heavy production? Are you carving wood or simply routing wood. The lighter materials won't make much of a difference, but if you intend to do heavy routing or carving of wood the machine is goin to have to be stiff and the size you have laid out will probably have to be made from steel to be made stiff. Even using steel the design will have to be right or you will still have problems.

Your 480 ipm (40fpm) is achievable but not with 750oz steppers. Steppers are limited in a couple of different ways one of which is speed. The other is torque at high speed and that is what you are going to need to be able to move a machine this size at that speed especially since you don't want to use a coarse thread ball screw. I won't even get into the accuracy thing because these other items need to be addressed first.

I told you in my first post my machine is nearly as large as the one you want to build....but it is made out of aluminum tubing and it is neither stiff nor heavy and that is why I can achieve 354 ipm with 450 oz steppers. If my machine was made from steel I bet my rapids would not be half that.

Oh...you can breathe now! :p

Mike

Thank you all for your patience, I am breathing again.
About the gecko drives: If I understand now, I could solder wires directly from a 1 axis driver to my parallel port. Is this correct?
Its hard being a newbie.

I will have to lose some speed to make this monster. That is fine.
I want to use a 1080 oz stepper for the x axis (spinning the nut), and then 600 oz or so for the y and z. I think I am leaning toward the flashcut steppers, and gecko drives, as long as there are no compatibility issues.
As far as accuracy goes, I am not really building space shuttle parts. I just like the idea of producing what professional machine shops produce.
So, like everyone else, I want it cheap, light, strong, accurate and the ability to machine titanium and solid carbide. Is this really asking too much? I mean really?

Thank you all for your knowledge and time.
Rob

I want it cheap, light, strong, accurate and the ability to machine titanium and solid carbide. Is this really asking too much? I mean really?


Rob

Yeah, based on what I see in your pics, you are asking too much. I think that is going to have way to much flex and/or vibration for metal work.

As for the different size steppers. If you can get the different size steppers you need with similar voltage ratings, you might be able to run them all from one power supply. If the voltages vary by a bit, you may need multiple power supplies. As for connecting 5 Gecko's to the parallel port, it can be as simple as a parallel port cable with the appropriate lines running to each Gecko, to multiple breakout boards such as these: http://www.campbelldesigns.com/breakout_board_plus.htm

Hello everyone.
I am building a new machine and need lots of help. I have a decent CAD background, and know my way around a shop.
The machine will have 120" X 60" X 22" capacity, and will be used primarily to cut wood and softer materials. I like the gantry design, and contrary to the posted jpg's, this will have the A and B-axis mounted to the Z-axiscarriage. Every axis will be controlled by ballscrews, except for the A and B-axis. I believe those will be geared for power and precision.
My questions are:
How much servo will it take to drive this monster?
How much can I expect to spend on these servos/controllers/everything except my laptop cpu?
If I decide to use two screws for the x-axis, can two servos be sync'd up worth a darn? (Texas dialect, please ignore)
With all of the above mentioned questions, the budget for this project is $12000 us dollars, 2000 of which is for my machinist' friend.


I have already talked to Chris about the long ballscrew (10 feet) and I am opting for the rotating nut option.
Thank you all who have taken the time to read this.

Rob
Rob,
just read this thread today, I am in the process of designing a new cnc machine to do 3 D carving my budget is up to $50K to get it right first time.
But the question I have is why 5 axis surely 3 can do most if not all of the required jobs a 5 axis can do or am i seriuosly missing something.
I believe that if I can interpolate all 3 axis similtaneously what advantage coud the other 2 axis add.
I realise this may sound dumb to most of you but please indulge me
chris :drowning:

Crazy story there...
My dad wants to invest in and have me design and build this machine. I argued with him until I was blue in the face that we should settle for 3 axis and a rotary table (4th axis) mounted like a lathe (with tailstock) on the table. I think the 5 axis setup makes it that much more versitile, such as carving a 3d object from the underside. As my dad put it, "I want to be able to stand a small statue on the table, and pick it's nose with the cutter" .
I sure am having fun looking searching for thrust bearings, helical gears, and the perfect design for the a-b head though.
Any plans out there?

That's refreshing to hear (dad wants to invest), it seems mine always wants freebies.

I think for the rotary axis you would be better off using very large timing belts. It would be very hard to get a geardrive to work with no backlash, and gears would need to be protected from the cuttings much more so than a belt.

Rob,
just read this thread today, I am in the process of designing a new cnc machine to do 3 D carving my budget is up to $50K to get it right first time.
But the question I have is why 5 axis surely 3 can do most if not all of the required jobs a 5 axis can do or am i seriuosly missing something.
I believe that if I can interpolate all 3 axis similtaneously what advantage coud the other 2 axis add.
I realise this may sound dumb to most of you but please indulge me
chris :drowning:

One simple thing is drilling a hole at an angle. With a 3 axis, you got one option, straight up and down. Another easy one is bevels. A five axis makes it quick and easy to tilt your tool at say 45 degreee, and cut a long bevel with a standard endmill, in one pass. You could 3D surface a bevel, but it would take many passes, and the finish would not be as smooth.

The real trick is generating 5-axis G-Code. Simple positioning of the A-B axis and then cutting is pretty easy, but simultaneous 5 axis movement is VERY difficult. I use EdgeCam at work every day, and even with $20k dollar software its diffucult.

Rob, have you thought of making the spindle tilt in one axis only, and adding a large rotary on the table face up. This would make designing the head much easier. Also make sure you have some locking bolts for the 4th/5th axis to get the more rigidity when doing 3-axis only milling.

Michael

I had no idea how much the software would cost to run these 5 axis. Building/machining them seems straight forward, I think helical gears would eliminate some backlash. Seems like I would do better to go with a 4 axis and leave room to upgrade later.

I am thinking once more about servos instead of steppers. Can anyone give me a reasonable estimate of what motors it would take, and the cost of these servos/controllers/and the power supplies? I think I can keep the gantry (with tooling) below 180 lbs. I want this thing to jog fast, and cut wood at whatever speed delivers an acceptable finish. Is 40 fpm a rediculous number? I saw a production machine do at least that.

Is 40 fpm a rediculous number? I saw a production machine do at least that.

Our router at work will cut 10 meters/minute, which is about 32 fpm. It won't move any faster. Some production routers can cut at 2000 ipm, but they are quite expensive (~$200,000). It is probably achievable, but it will be expensive. You'll need big servos, and high lead ballscrews (or rack and pinion).

I think for very high speeds a chain or belt drive would be the least expensive. Ballscrews that large would be super $$$ and a rack and pinion is pretty costly too.

P.S. if anyone knows of a cheap source for rack gears...let me know :)

As far as 5 axis software goes, you don't really need to spend the big bucks to buy it, there are places around who can for a fee process the cad drawing of a part and generate the code to run the machine. Dig around a bit and see if you can find someone who can do this for you and see how much they charge, they don't even have to be local to you as the could be half way around the world and transfer the files over the net. Some of small shops do this if they dont have a person or software who is able to produce complex code for things like molds. There may even be somone here in this group willing to help you out who has access to 5 axis cam software at their work. It's worth looking into before canning the 5 axis project.
Cheers
Splint