New CNC Machine - Building Stage!

[JavaDog]

Ok, so I am starting to aquire peices and doing planning on my first CNC 3-Axis machine! Here's to being broke for the next few months.

The design will be similar to the Data-Cut. I am planning on a working area of 24in (Y-Axis) x 30in or 32in (X-Axis). The Z-Axis will be using a 7.5in THK Rail.

1) THK Rails for the X, Y, Z Axis.
2) Ball-Screws for the Y, Z Axis.
3) Belt-Drive for the X Axis.

All of the above will be from Ebay. By my estimate, will run about $250-$350.

The parts I am not sure on:

1) Router. I was thinking of a 1/2hp Porter-Cable?
2) Servos. Not sure here.
3) Controller and PSU. I would like to go Gecko, but they are $$$.
4) Table/Frame. Thinking of Alum. Extrusion like 80/20.

I am mainly going to be machining Acrylic, Plexi, Poly-Carb, Delrin/UHMW. I would also like to be able to machine Aluminum stock up to 1in thick (obviously would be multiple passes).

I am trying to stay under $1000 (hopefully $500) total cost.
So, thoughts? Ideas? Things I should change?

Thanks in advance for the help. This is a great community!

Cost so far:

1) Two 26in Slide Assemblies: $130 (Ebay)
2) Porter Cable Router: $136 (Amazon.Com)
3) One IKO Linear-Rail (Z-Axis): $15 (Ebay)
4) Framing:
- - A) 80/20 Aluminum Structure - 47" x 2: $33 (Ebay)
- - B) 80/20 Aluminum Structure - 36" x 3: $36 (Ebay)

Total: $350
Budget Max: $1000
Left to Spend: $650

[BigDaddyG]

JavaDog,
Sounds good so far! A big yes on the Extrusion frame! A bit more expensive but well worth the money, and the fab is easier than most. Are you set on Servos, because I mentioned on a previous thread, Automation Direct just came out with a Stepper combo that is hard to pass up on! All new stuff, with up to 20' of wire (Driver, Motors, Power Supply all configured together and delivered in a few days to the house, all brand new with warrenty...can't beat it).

Gotta ask, where in W. New York. I am in Fla now, but grew up in Alden. Good luck, ask if you need a hand or have any questions.

Regards,
Glen

[JavaDog]

Originally Posted by BigDaddyG JavaDog, Sounds good so far! A big yes on the Extrusion frame! A bit more expensive but well worth the money, and the fab is easier than most. Are you set on Servos, because I mentioned on a previous thread, Automation Direct just came out with a Stepper combo that is hard to pass up on! All new stuff, with up to 20' of wire (Driver, Motors, Power Supply all configured together and delivered in a few days to the house, all brand new with warrenty...can't beat it). Gotta ask, where in W. New York. I am in Fla now, but grew up in Alden. Good luck, ask if you need a hand or have any questions. Regards, Glen

Yeah, I might have a source of near-free extrusion locally - so that makes it easier to go with!

Is there any drawback in accuracy (or speed) or more complication in going with steppers? If not, I am not married to using servos - and in fact - I can source stepper motors for older high-end line printers for free (what specs are important? Amps? Oz? Turns? Steps? What is ideal for Plastic & Alum?).

I am in Buffalo, near Alden. It's funny, every forum I am on I know like 3 people from WNY. I guess we are just to friendly to stay quiet!

[buscht]

JavaDog, you have champagne tastes on a beer budget

Servos, Geckos, THKs, Ball screws, etc. I'd guess that your budget figure should be more like $1500 to $2000 unless you are a great scrounger.

It seems to me that stepper motors are much easier to set up, but servos are capable of faster more reliable motion. If you don't overtax your system, then steppers probably are the way to go.

My recommendation is to set your budget.

Let's say $1000.
That $232 per AXIS
rails $70
ball screw $65
Stepper Motor $65
Cables $10
Switches $2
Aluminum plate $20

That leaves $304 for general machine construction
controller Xylotex $150
Fan $5
DC power supply $50
1/2 hp router $100
frame free???

Used PC free??? or $50

Control software
Mach2 $150 or TurboCNC $60
Programming software
Sheetcam free in beta

Total costs not including software $1001

Anyway, you can see that if you write it out, you can keep track of costs better and it will help you on your Ebay bidding. You can also see if your original estimates are inline or way off base.

Good luck.

[JavaDog]

Well, I am a good scrounger, but $1000 is going to push this machine out until maybe Dec 2005.

I would really rather go with actual Servos.
What should I look for in a servo when I am hunting?

I have plenty of spare PCs and Monitors for the control system.
Is Autodesk Inventor capable of driving a CNC machine directly. I do all of my CAD work in Inventor. I can export in DXF or DWG format...as well as others.

As for the controllers...are there good Servo controllers that I can build myself? Any with free plans?

Thanks for the input Buscht, it does help quite a bit.

[buscht]

Check out this website,
http://www.homecnc.info/cnc-main.html

I don't know much about servos, other than the motors,encoders, controller, add up to about $1000 minimum without the mechanical part of the machine. That's why I stayed away from them.

Check the classifieds on this forum. Balsaman is selling his 2nd servo router for about $2500. Read his thread on building the machine. Its one of the best ones here.

To my knowledge Autodesk inventor will not drive a CNC machine directly. It just CAD. Once you export in DXF format (for example) you usually need another program to take this drawing and produce G code to run your machine. www.sheetcam.com would work. Then you need a program like MACH2 or TurboCNC to read the G code and control your machine tool.

I just pulled the $1000 out of the air. You can build a unit for $500 for sure.

Use stepper motors.
Don't expect high end speed or cutting power. If done properly, you should get good accuracy however.
You had a pretty reasonable machine, (costwise) until you mentioned cutting 1" aluminum. That requires heavier cutting forces than some of the other materials you want to cut.

Good luck.

[Graham S]

Servos don't have to be so expensive, you can get the motors for $17.50 each (search for "my cheap servos" thread), encoders for ~$25 (www.usdigital.com) and drivers for $114 per axis (www.geckodrive.com).

[JavaDog]

See, though, half of my projects are going to involve Alum....

I guess I could start with just Acrylic and use that to do a $2500 Servo based router later...it's not like the desire will go anywhere!

What do you think of the FET3/Cruiser system from StepperWorld?

Originally Posted by Graham S Servos don't have to be so expensive, you can get the motors for $17.50 each (search for "my cheap servos" thread), encoders for ~$25 (www.usdigital.com) and drivers for $114 per axis (www.geckodrive.com).

What type of power and accuracy can I expect from such low-cost servos though?

[buscht]

I don't know much about it other than after reading up on this forum, most people recommended the Xylotex instead.

There is a thread by Cold Fusion on this forum, he built a DIY router and cuts aluminum quite a bit. I think that he had a $500 budget and ended up spending $1400. Check it out.

By Graham S's figures, you would spend $469.50 on a servo system. You still need a DC power supply, cabling, control box, and most people use a breakout board. That's roughly $250 more.

I'm not sure what reads the encoders, is it the Gecko's or something else?

JavaDog, Balsaman is selling it for $2500. My guess is that it cost about $1000 to $1500 to build.

[ger21]

Originally Posted by JavaDog What do you think of the FET3/Cruiser system from StepperWorld? What type of power and accuracy can I expect from such low-cost servos though?

Everyone I've seen comment on their Stepperworld board wonders why it's so slow. A Xylotex should give you at least twice the speed, more efficiently for the same amount of money.

The low cost servos will give you the same accuracy as the more expensive ones. The encoders actually provide the accuracy with servos. I believe they will provide enough power for your application. You can always gear them down a little for more power (trading off speed).

[ger21]

The design will be similar to the Data-Cut. I am planning on a working area of 24in (Y-Axis) x 30in or 32in (X-Axis). The Z-Axis will be using a 7.5in THK Rail. 1) THK Rails for the X, Y, Z Axis. 2) Ball-Screws for the Y, Z Axis. 3) Belt-Drive for the X Axis. All of the above will be from Ebay. By my estimate, will run about $250-$350.

I'm not sure how you can build a usable z- axis with 7.5 inch rails. For what you want to do, I'd recommend that your Z-axis have at least 3" of travel. If you can get 3" of travel with 7.5" rails, I doubt it will be very sturdy.

Also remember that you'll typically need about 8 or more inches longer rails than your intended working area, so you'd probably need at least 32" Y rails and 40"+ x rails. These longer rails start to get expensive (even on Ebay). One placve you can save some money is to use the smaller 15mm rails, instead of the more expensive 25 or 30mm. The 15's will still be plenty strong enough for your application.

1) Router. I was thinking of a 1/2hp Porter-Cable?

Do you mean the 1-1/2HP 690? I'd recommend the 892 (2 1/4HP), but it's almost $200. You can get the 690 for maybe under $150. Get one with variable speed, though. Plastic melts fast at high rpm.

I'd say it's pretty doubtful you'll be able to stay under $1500, unless you have a free source for the aluminum extrusions. Even with a Xylotex, you'll probably have to pay around $50 each for decent steppers, and that's $300 already. You might want to look at good acme instead of ballscrews. Multiple start screws can give performance close to inexpensive ballscrews, for about 1/2 the price. And if you want to cut 1" aluminum, you might want to think about a smaller machine. Easier to make more ridgid. Good luck.

[JavaDog]

Originally Posted by ger21 I'm not sure how you can build a usable z- axis with 7.5 inch rails. For what you want to do, I'd recommend that your Z-axis have at least 3" of travel. If you can get 3" of travel with 7.5" rails, I doubt it will be very sturdy.

So, 12" for the Z-Axis would be about right then? I can do that.

Originally Posted by ger21 Also remember that you'll typically need about 8 or more inches longer rails than your intended working area, so you'd probably need at least 32" Y rails and 40"+ x rails. These longer rails start to get expensive (even on Ebay). One placve you can save some money is to use the smaller 15mm rails, instead of the more expensive 25 or 30mm. The 15's will still be plenty strong enough for your application.

I've noticed. I think I may have to reevaluate my needs.

Originally Posted by ger21 Do you mean the 1-1/2HP 690? I'd recommend the 892 (2 1/4HP), but it's almost $200. You can get the 690 for maybe under $150. Get one with variable speed, though. Plastic melts fast at high rpm.

Yes, the 690 was the one I meant. Typo on my side. Without a doubt I need to go variable speed.

Again, this is the reason I am looking for good steppers or servos - with plastic and a high-speed router, I can push the table pretty fast and keep the cuts nice and clean.

Originally Posted by ger21 I'd say it's pretty doubtful you'll be able to stay under $1500, unless you have a free source for the aluminum extrusions. Even with a Xylotex, you'll probably have to pay around $50 each for decent steppers, and that's $300 already. You might want to look at good acme instead of ballscrews. Multiple start screws can give performance close to inexpensive ballscrews, for about 1/2 the price. And if you want to cut 1" aluminum, you might want to think about a smaller machine. Easier to make more ridgid. Good luck.

So, let's say I went with a 20" x 20" of usable cutting space. That would mean I would need 28" rails, correct? (Are these a good price for such a rail?)

Now, don't forget, I was planning to use belt drive (since I have a free source for gears). If belt drive is (i) accurate enough and (ii) strong enough for my stated uses, I will use it on all three axis. That should help cut the costs down since I don't have to buy ballscrews then.

Thanks again guys for all the great info. Really is a learning and fine-tuning process.