[Albigger]

Hello all

As this is my first post some info about me: college student/machinist wanting to build my own CNC. I've been registered for a few months, but only reading lately. So far I have learned a LOT and hope to continue.

What I want from my CNC:
1. Ability to cut Aluminum (most often) and steel (only IF Necessary).

2. Ability to use small bits ( around .045"), which means high rpm, but also hold up to 1/2" dia. tools

3. Repeatability to within .0005" actual, better if possible

4. Adjustable spindle speed (possibly need different spindles?)

5. Work envelope: like to have as much as possible (wouldn't we all), for starters i'll Say X 24", Y 15", Z 12" or there-abouts

6. Fast cutting (cut greater than 10 IPM and rapid at least 50 IPM)


I'll tell you what I was thinking for a basic setup (and hopefully someone can point out where I'm horribly wrong):

Moving table in X direction, overhead gantry for Y and Z directions. Aluminum or steel tube/plate construction, bolted together. THK or IKO rails for all axes. Servo Motors with Ballscrews. Spindle - not sure (reccomendations??) Will that get it done if built rigidly to meet my requirements?

Electronics end I will need more guidance in. I have computers and MasterCAM, which I hope to use to genrate toolpaths (have slight experience here) and output G-code. Then some free/cheap software to transfer it. Here's where I'm not quite sure anymore - I need a driver board (will this handle all 3 servos, or one per axis?) and a power supply to power both the servos and the driver board?

What I have at my disposal: Typical machine shop, Ebay, and hopefully some of the knowledge of the members here.

SORRY SORRY for the long post, just trying to provide all information at this point.

any help/direction/comments/criticisms/advice is much appreciated. Thanks.

--Jay

[sendkeys]

[wjbzone]

If you want to cut aluminum and steel, you should consider adding coolent into your design. (drains on the table and recirculation pump)

More then one spindle might be necessary for the range of bit sizes you want. A variable speed router for the smaller bits, and something like a Sherline or Mini Mill Machine Head for the larger bits. Some people have made their own spindles.


Cutting speeds of 10 ipm and rapids near 50 can be done with stepper motors. Servos cost a little more.

I started out by doing my electronics first, because that's where I needed more help. If you have a machine shop, the hardware will be a lot easier.

Bill

[Ferenczyg]

Even the more ambitious DIY CNC I've seen here do not attempt to mill steel. Your expectations of frequent alu milling with 0.0005 asks for a very rigid machine, sturdy but precise building and powerful motors. The other requisites are very restrictive for a DIY machine too except for the IPM. If I understand you well, you want to DIY almost a production machine.

And at the end but not less important.. How big is your budget?


Fer

[Albigger]

Originally posted by sendkeys
should be able to do everything but that "Repeatability to .0005" is going to be hard
But you can do it if you have the cash hehe. good ballscrews good rails good spindle :P

--I may be able to relax that requirement to .001, if necessary. But I thought servos and ballscrews were the best, is there something else to consider?

Originally posted by sendkeys

as for software i would use mach2 that gives you 44,000hz if you have a fast computer(1.5ghz or so)its around 150 bucks. seeing you want to cut steel and aluminum at the speeds wanted. you will need good size servo's this will need bigger controllers.

What does the 44 kHz refer to? How fast the servo updates its position? I'm not quite sure on this - hafta go read some more threads I guess.

Originally posted by sendkeys

So i would go with gecko or rutex(rutex looks best to me as of now) controlers there 150+-(20 bucks) per axis. all the 3axis boards i have seen are very small.

Ok - getting expensive, but ok.

Originally posted by sendkeys

as for power supply you should be able to make one just a transformer, 4-diodes and large caps.

And this will be able to handle the larger motors, or will that require several power supplies? (Or won't I really know till I get them?)

Originally posted by sendkeys

Main thing you didnt say is what type of budget you have :P You can do anything with money but if you only have 1000 bucks its going to be very hard hehe

Somewhere around 1000 (expecting the higher side). Up to about 1500 is probably about the max I would really want to spend on the machine itself (may reconsider)

Originally posted by wjbzone
If you want to cut aluminum and steel, you should consider adding coolent into your design. (drains on the table and recirculation pump)

Any special reccomendations for this job, if not I'll look around.

Originally posted by wjbzone

More then one spindle might be necessary for the range of bit sizes you want. A variable speed router for the smaller bits, and something like a Sherline or Mini Mill Machine Head for the larger bits. Some people have made their own spindles.

Ok, figured I might need 2. Kind of leary about making my own, but that's a ways off so I'll look into it some.

Originally posted by wjbzone

Cutting speeds of 10 ipm and rapids near 50 can be done with stepper motors. Servos cost a little more.

I really wanted to use ballscrews figured why not do it right the first time with servos. I thought they were more precise anyway.

Originally posted by wjbzone

I started out by doing my electronics first, because that's where I needed more help. If you have a machine shop, the hardware will be a lot easier.
Bill

Yeah, the mechanical construction I'm not so worried about, so maybe I'll take your route and start the electronics (for at least one axis)

Originally posted by Ferenczyg
Even the more ambitious DIY CNC I've seen here do not attempt to mill steel. Your expectations of frequent alu milling with 0.0005 asks for a very rigid machine, sturdy but precise building and powerful motors. The other requisites are very restrictive for a DIY machine too except for the IPM. If I understand you well, you want to DIY almost a production machine.

And at the end but not less important.. How big is your budget?

Fer

Production machine capabilities would be nice, but I'm realizing it may be out of my budget. If I just spend a little at a time its not so bad though. Plus I'm hoping some of the stock for the machine body I can salvage from work scrap - but that seems to be the least expensive of the project.


A couple of you have stated it will take large motors, so what brand/ how many OZ-in. are we talking (at least so I can start looking to get a price range on them).

--Jay

[JFettig]

Hey man,

It looks like you might be looking into something bridgeport size? the only thing about a bridgeport is the spindle speed isnt adaquate for what you want to do.

From littlemachineshop you can get part of the mini mill head with a spindle. You can change the bearings out and hit 10k(I have 10k max bearings in mine)

Albigger: Cince the last time we have talked, I converted my mini to cnc. with higher voltage and my slow laptop I can get 96ipm out of my steppers, I have it set at 60ipm because I dont have enough high speed torque.

Jon

[sendkeys]

[Albigger]

Jon - How's it going man? Didn't know you converted your mill. Do any new projects on it (I haven't been following the 'other' forum quite as closely lately).

- I'll look into the spindle you suggested, 10K would be fine for me I think.

Thanks all for the good advice everyone. As expected (I'm sure) I have more questions:

Most notable repeatability vs. accuracy. In theory, If I build a "perfectly" rigid machine with no-backlash, good servos and rails, etc... the REPEATABILITY would be right on.

But ACCURACY would depend on - construction of the frame to parallel and perpendicular specs (think I can get that done), and mostly the ballscrews? From what I've seen, most "cheap" ballscrews can vary as much as .004" per foot of length, which would mean I could be out by .008" over a 2foot X-axis machine? I know things like glass scales are used in real machines but I'm not going that expensive, so my question is if I buy cheaper and less precise ballscrews, then measure the "inaccuracy" over distance, is there some way to compensate for this with software (mach II or other)?

Also, I've put together a preliminary price list of what I will need:

Mid-size machine (travel of axes - X: 20-30 inches, Y: 12-18 inches, Z: 6-10 inches)

(All in U.S. dollars)

Frame/Construction Materials ~$300
(steel/aluminum bar or tube - undecided as of now)

3 Servos with encoders ~$400
(greater than 300 oz-in.)

Drives for servos ~$450
(rutex or geckos or ???)

3 sets ballscrews/preloaded nuts/bearings ~$250
(precision ground if possible, rolled if can't afford better)

3 sets linear rails ~$300
(THK or similar - is there anything else?)

Power supply/switches/misc elec. ~$125
(home-made or cheap purchase)

Spindle/Motor ~$??? (still have to look into this)
(have dremel, can purchase router ~$100 for high-speed, for low-speed ??)

Coolant system/pump ~$?? (still have to look into this)
(any good sources for this?)

Software ~$150
(CAD/CAM i have plenty of programs, probably buy Mach II or similar to go from G code to drives)

If I want to start cutting costs, I see my options as follows, in order of preference:

1. Cut down axis travel to save on ballscrews and linear rails
2. Use smaller servos but gear them 2 or 3:1, possibly allowing the use of a cheaper controller, and at the expense of cutting/rapid speed

Is there anything I'm overlooking to cut costs (besides changing from linear rails to rolleblade bearings or things like that...)?

!!! So many decisions!!! I guess I REALLY have to decide what this machine is particularly for so I can focus more on the requirements.

Thanks again for all your help - you guys have given me the idea and confidence that this can be done.

[Cold Fusion]

First of all, welcome to the forums Albigger. You will find some incredibly knowledgeable people on these forums to help you in your quest.

As far as your budget goes. You CAN do it for that much. Ebay is your best friend. I suggest staking it out for a few weeks in search of THK rails, ballscrews (seen a lot of those lately), and motors.

My machine is all aluminum, ground .0004 ballscrews, Parker Z axis, THK, ect. The whole system cost me $1200. My workable area is 26x24x6.

I went out and bought 4 Thk 30 inches with 8 blocks for $300, so yes, your right on the mark for those prices.

It's a lot easier and sometimes cheaper to buy a pre-made Z axis. Look for a Parker slide on ebay. Precision, all aluminum, the works!

Mach2 can be adjusted to compensate for backlash, but it's ALWAYS better to get rid of it at the source. This will give you cleaner cuts on top of better precision.

[Albigger]

Thanks Cold Fusion - I just read the entire thread about the design and building of your machine last night (when I SHOULD have been studying for my finals).

Anyway, at least I know I'm on the right track. You have an impressive work envelope and machine for a $1200 budget - this gives me hope.

I'm scouring ebay recently, but getting anxious to buy. Trying to resist and look for good deals.

At least it seems there is enough available at the moment.

It is a bit of a pain trying to look up the product specs for each ball screw I find on ebay - any tips on that or just practice (and bookmark the companies product catalogue pages)?

Also - I'll have to go re-reading your thread to find out what kind of tolerances your machine gets (unless you want to tell me here)

thanks.

--Jay

[Cold Fusion]

I couldn't tell you actual tolerences because the machine isn't cutting yet.

A rule of thumb is any ground ballscrew is atleast a c5 tolerence (.001). Ground ballscrews don't have the ridges in the center of the thread like roled do.

Ground:
http://cgi.ebay.com/ws/eBayISAPI.dll...806456822&rd=1

Rolled:
http://cgi.ebay.com/ws/eBayISAPI.dll...813611793&rd=1

Dang! I lost an auction for a perfect ground ballscrew while writing a post.

[sendkeys]

[avsfan733]

i've always been of the opinion that a .001 tolerance part requires a machine made to .0001 as a general rule...meaning you would need to be able to make your machine perfect to .00005 or f hundredthousandths of an inch. Call me a little skeptical that even the .001production tolerance is practical on your price range in steel. not meaning to discourage just offering a poitn of view good luck either way

[Albigger]

For the G-code software I was planning to use mastercam because:
1. I have access to it already at no cost
2. I have some experience with it

Oh and the dremel thing was more like something I could use till I get the cash for a real spindle, but on second thought I bet I won't end up using it at all.

avsfan - you bring up an interesting point. I will keep this in mind during construction, and as I said above I have a machine shop (and some very experienced machinists who would probably be willing to help), but even so there is no way I can see getting any closer than a few tenths, let alone hundreths.

[sendkeys]

Goodluck with everything Albigger sounds like you got everything under control

[Albigger]

Originally posted by sendkeys
Goodluck with everything Albigger sounds like you got everything under control

Ha-ha....

not quite, but enough to get started hopefully. Thanks for the encouragement.