off the wall nooobie design ideas

[project5k]

First I'd like to say howdy to everyone, and thanks for making all this information available.. I have learned so much and haven’t even touched the welder or soldering iron yet...

Oh, and please let me apologize for the monster post... but it's my first, so I have lots of stuff to cover...

Now, in all my reading I remember seeing something said that I should start off smaller and then work up. This just isn’t an option for me as 1) I'm impatient, 2) I just don’t have the resources to do something twice, and 3) I really don’t see that big of a difference in a small machine vs. a large one, everything is just bigger... ok, so at this point I think I have the perfect recipe for disaster... perfect, time to push forward...

So as the title says, I'm a total nooooobie to cnc. I work around cnc machines at work all day, but never have the chance to "crawl inside" to see what makes one tick. I have been doing lots of research, and for what I want (a 4' x 8' usable cutting area) I think I have decided to go with bipolar steppers and drives... I think that this will fall in my budget, and I feel confident that I can make it work for the accuracy I need.

Goals of my cnc machine:
1)Be able to cut and contour in 3d a full 4x8 sheet of mdf/ply for signs, speaker boxes, cabinets, and anything else I can dream up
2)Convert the machine from a router to a plasma table and have the ability to cut at most 1/2" mild plate steel for some prefab kits in designing
3)Have the ability in the future to add a 4th axis for doing all that nifty stuff, chess pieces, baseball bats, table legs, and so on...
4)Do it on the cheap, and I mean on the super cheap, I’m talking maybe $500 to $750 give or take to get the 3 axis running...

What I see that I need
1)A legal full copy of mach 3
2)The parts to build the driver circuits
3)Stepper motors
4)Steel for the table
5)Linear drive components
6)Steel
7)Bearings, linear slides, bolts wire, and a bunch of other little stuff….

What I don’t have and wish I did
1)Milling machine for making the parts to get my cnc going
2)More money

What I have at this point:
1)A large shop (2400sq-ft) to build, test, and run the machine
2)Computers, that will run mach 3 with no problems
3)Electronics background in design, and prototyping
4)Mig welders
5)A plasma cutter than can more than handle the 1/2" plate
6)A steel yard about 4 miles away from my shop
7)The ability (knowledge) in structural design to build pretty much anything that I can dream up given enough time
8)2 metal working lathes, one a mini 7x12 with 3 jaw chuck, an the other a 100 year old 10x24 lathe with 3 jaw chuck and independent 4 jaw
9)The demo version of mach 3
10)Rhino version 2
11)Some crazy ideas
12)Someone that just might pay for the mach3 for me
13)A boss (at my regular job) that is into this and wants to help (and is cool too!!) yea I know that’s a first
14)Someone that can mechanically or photo etch the driver boards for me for nothing more than the cost of materials
15)A design for the driver boards and sources for the parts that I don’t already have in house (I’m planning to use the l297/l298 design)
16)A 3/4 hp router and some bits
17)Lots of empty spaces in my breaker box, and the knowledge and ability to add more circuits as needed, and a “proper” amount of fear!
18)An austron 35amp 13.8v supply in a 24 space cabinet on wheels with 2 17Ah gell cell batteris for backup and momentary current draws over the 35amps..(im planning to put the drivers, computer and any other electronics in this cabinet as well)

My plan
1)Learn a bunch more
2)Finalize the design I want to use
3)Get the motors that I’m going to use
4)Build the driver circuits
5)Test the drivers and motors on the desk with mach3
6)Begin building the table
7)Assemble
8)Troubleshoot, troubleshoot, repair, redesign, build, troubleshoot, repair
9)Use the machine and giggle like a schoolgirl when it finally works
10)Upgrade the stereo in the shop so that it’s loud enough to be heard over the router. (and really annoy the neighbors at 3am) TURNABOUT IS FAIR PLAY!!!

My current design and thoughts:
Build the table pretty much like the one that I saw on http://www.inshorepowerboats.com/cnc.../CNCROUTER.htm
With a few changes… I like the y and z-axis setups, so I will probably replicate this. But the x-axis is using rack and pinions and that’s a little outside my budget, so here’s my thought. I have some and can easily get more #40 drive chain. My thought was to either lay the chain along the top rail of the table frame and have the gears ride in that (still need to investigate this and how much slop there would be) and use that like a rack and pinion. The other idea I had was to suspend the chain along the insides of the rails, and have the x-axis drive gears interface the chain there. The chain would be pulled taught and would make an upside down “U” where the drive gear would be at the top, and then 2 idlers would make the other bends, this way the chain is around 180deg of the pinion gear. Either system would be replicated on both sides with a connecting shaft between, so there would be no chance of the gantry racking diagonally… then use one of the 500oz*in motors I found on the net geared 10:1… figuring 300rpm max (I have no idea what this should be, I just picked a # outta the air) on the motor and a 1” diameter gear for the pinion, my math says that I should get a max speed of about 90ipm. So, if ya’ll will, please rip this design to shreds and show me what I’m missing…

Project5k

Similar Threads:

• Build Thread- First DIY CNC, opinions on our design ideas?
• New Machine Build- Some design ideas!
• some design ideas
• need ideas on Y axis design
• My CNC links for design and ideas

[project5k]

oh, while im thinking about it, i was considering using something like 2"x2"x1/4" angle top and bottom on the x axis for the long slides.. i could turn some v groove wheels on the lathe... my worrie is how much variance will there be in the angle iron that will cause the slides to not be absolutely true straight and flat, and how much will i cause when i weld it all down...

[ger21]

Well,, at least you know you have a recipe for disaster. Most people building an 18"x24" machine will spend $750. Building a 4x8 machine for that is totally unrealistic, unless you get virtually everything for free.

It's likely that chain drive will have far too much slop. You'll quickly find out that 90ipm is painfully slow for a 4x8 machine. A 3/4HP router is really too small. If you plan on running your stepers on 13V, you won't get anywhere near 90ipm. While a lot of people want to use their machines for both routers and plasma, I'm not sure if I've really seen anyone that does. They have different requirements.

Good Luck. (you're going to need it)

[project5k]

Ok fair enough. This is exactly why I'm on here, to pull from the vast experience that every one has. If you noticed, the #1 item in my plan is to learn a bunch more before going any further...
One thing that I had considered, but was leaving as a fall back plan was to build the table 48x48 with support wings to either end, so that I could index my stock, cut half, move it, and then continue. This is of course not my preferred method, but I think with enough careful planning, I might be able to make it work good enough for what I’m going to use it for...maby not.. anyone got experience doing it this way?

Regardless of what the final table size, my main interest at this point is to see if some of the principals of my design are flawed, and I’m seeing that they are. If the chain has too much slack in it, then I will have to figure another method... that’s fine, that’s what I’m trying to do, is to learn.

Ok, so let’s go through this, one step at a time....
Chain, bad idea....

How about the angle iron "tracks" and v groove wheels for the x-axis? I can see the potential for there to be some slack in this as well, but I was wondering if anyone had ever done this, considered it, and so on…

12v isn’t enough... ok, here’s my thinking on this... I figure to have any kind of margin, I need at least a 30amp supply... (I like to design my electronics with at least 25% extra capacity) so what should I be shooting for minimally? 24? 48? I’ve seen lots of different #'s on this...

90ipm did seem a bit slow, but I was using this figure more to see if people agreed with my math, and how I got to that #... having it take over a min to travel the length of a 4x8 sheet would be aggravating, but would work, yes?
What kind of ipm should I be shooting for? In my example I have the motor geared 10:1. Figuring that I’m guessing my gantry will be around 50# and working the math backwards, I would need 1600oz*in on the x-axis shaft with a 1" pinion. 1600/10=160. Now, this should accelerate the gantry at 1g. Not accounting for frictional losses and so forth. So lets call it 200 to make the math easier. If I use one of the 500oz*in motors, that would mean that I could cut my 10:1 gearing down to 5:1. This would make the motor minimum 320oz*in plus friction. And would make my feed rate 180ipm...now I understand that the faster you run a stepper, the less torque that it can produce at high speeds. Increasing the feed V will help with this some, but not totally overcome the torque curve...

My question is this. Since I don’t have that much experience with this, what would be a decent ball park RPM# to plan for on a nema 23 motor at say 24V or 48V? This is one thing that I haven’t found enough information on... My original guess was something like 300RPM. Is this too fast? On a 200 step/rev motor this would be 1000 steps per second. My electronics should be able to handle this. But will the motor do this with any torque?

If I figure 150RPM motor speed, the 5:1 gearing, 1” pinion then I’m back to roughly 90ipm rapids… I could go to a larger motor and less gearing, but I’m trying to get a baseline so start from…. If you all agree with the math and how I’m getting my #’s then I can make adjustments to the design once I get some kind of ballpark figures… While speed is an important factor, its not nearly as important to me as it might be to some…

One thing that I think I understand is that a lead screw would take less torque to move the same load than the rack and pinion, but will require a higher drive speed… a 50# gantry accelerated at 1g on a 5tpi screw would require approx 25oz*in not allowing for losses. But to get 90 ipm I would have to turn the screw(or nut) at 450rpm…. Ok this seems a little fast to me… but at 25oz*in, couldn’t I get away with say 1:2 gearing on the 500oz motor…that would then put the motor speed at 225rpm and the input torque at 50oz*in…but would also reduce my move per step accuracy from .001” to .002”

As for the cost, the $750 is a target, but is not the absolute final #. The software will most likely be paid for, and the steel is one thing that I have quite a bit of.. and can get more at a somewhat decent price. 2x2 box tubing comes in 24’ sticks and is $28 for 16ga and $47 for 11ga. I figure I’ll need something like 6 sticks to make the table frame, and then I can use something else for the actual bed… so lets just call that $300, the motors im figuring will cost something like $200, so that’s $500, the drivers are going to cost me $20 each to build (I have most of the parts already) so that’s $60, so $560 total, so that leaves 190 for everything else…so $750 does come up short… and that’s fine, it was just a ballpark # to aim for… Even if I end up putting $2k into it to get it going.. that’s ok, it will just take more weeks to do it…I’m really aiming for a up and running date of the first of next year.. so I have a little time to get my plan together, save up some $$$ and build it…

[ger21]

I have seen a router driven by chain, and the builder claimed it worked well for him. But it's the only one I've seen. (It's not on this site, I don't think)

Shopbot's use V-Rollers on angle iron, or at least the older ones did. www.shopbottools.com

Also, see this: http://www.cnczone.com/forums/showthread.php?t=3673

Motor supply voltage. At least 10 times the motors rated voltage, up to maybe 20 times. More voltage = more speed, up to a point. To much voltage = too much heat.

What speed (ipm) shold you shoot for? As much as possible. The router motor you use as a spindle will probably be the limiting factor in how fast you can cut, and 90ipm might be adequate, but I'd try to shoot for rapid postioning moves of 200ipm or more, the faster the better.


Your gantry will probably end up weighing close to 150lbs., if not a lot more. Mine is a wood torsion box with about 32" of travel (40" overall length) and with the Z-axis, probably weighs about 70lbs.

How fast can you spin a Nema 23motor? Depends on the torque curve for that particular motor, drives, and supplied voltage. Maybe 200rpm, maybe 1000 or more. Can vary greatly. You need to consider the entire drivetrain package, and optimise everything to work together.

$2K is a lot more than $500 to $750. To build a decent 4x8 router (not good, just decent), figure $2K minimum. Jmo

[project5k]

Thanks for the information... i'm still doing more research.. and i started drawing this whole thing in rhino, so i'll try and post some renderings of it for yall to take a look at...
also while im thinking about it.. i found this site that has some guide rails for sale, and wanted to see what ya'll thought....
http://www.cadcamcadcam.com/index.as...ROD&ProdID=133

i dont know why, but 1/4" wide seems a bit narrow to me, but what do i know....

[ger21]

That #2 track will work fine. Keep in mind you'll need 2 tracks per side, at about $100 per track. $400 + 8 wheels = $500 for just the X axis.

If you think you'll need bigger, go to www.bwc.com

[project5k]

ok, so here are a couple of the renderings that i have done in rhino so you can more easily see what im planning, the cavities in the v groove wheels is where the bearings will go, right now the plan is that the outter flange of the wheels is 3" diameter(its what stock i have) and the bearings will be 2"OD... all of the frame is 2x2 box tubing, and the y axis slider is 1x2.. i still have to work up all the propultion stuff, but im still thinking that i might just try and go a different route with the chain drive, im starting to think that maby crossing the chains is the way to go... this would put the x axis drive motor on the gantry, which may be a weight issue, but its just a thought.. at least doing it that way would eliminate any chance of it racking...

ok, question, 2 tracks per side? im guessing thats so you have rollers on top and on bottom.. well thats a thought, but i was thinking more like 1 track per side, and then just have a flat "follower" wheel under the rail... maby on a spring loaded arm with some really stiff springs, and some form of adjustable mechanical limiting...
i'll draw up what im thinking later on....

ok, so this is the first time im going to try and attach images... so i put the web address where they are, and also i inserted the images... forgive me if it dosent work correctly.. and there are larger images on the website as well...

http://www.jpcustomcrafts.com/shopmachines.htm

[project5k]

oops... wrong button

[spalm]

Hey,

I somewhat successfully used those V-bearings on angle iron, so did Benny (Ynneb). It is on my third machine, first being a JGRO, second was a torsion box – skate bearing machine. It works quite well and is a magnitude cheaper than using the proper rails. My machine is a 24” X 48” and mostly wood. If you can figure out how to support the angle iron for the length you have in mind, you might want to give it a thought. You can see my build at:
http://www.cnczone.com/forums/showthread.php?t=15270

You can see Benny’s larger metal machine at:
http://www.cnczone.com/forums/showthread.php?t=6143

My recommendations: start a bit smaller until you get the hang of this. Large machines are a PITA to build and costs get large very quickly. You can always build a second machine and re-use many of the components. Chain drive seems scary to me (too loose).

Steve

[project5k]

yea i hear ya, actually after pricing some of the parts that im looking at using, i'm really thinking about cutting this thing in half.. it would be cheaper, take up less room in the shop, and would let me see if i'm really gonna use it as much as i think that i am.. im thinking about actually building it to work a 4x4 or half sheet... my buddy/boss, has a tiny unit that we've been trying to get to work, but theres no communication from his parallel port, so im gonna take the machine that im planning to use over to his place and see what happens....

one of the things that i first noticed about your machine is that you used the angle on its edge.. thats not something that had thought about... i had been planning to put it down open side against the 2x2 frame and spot welding it into place, the one thing that really scares me about that is that theres not adjustability and the other thing is that im afraid that the welding might cause the angle to pull and not be straight... bolting it down like you guys did would allow some adjustability, and i really like that....

the v grove bearings just dont appeal to me... im going to machine my own wheels and then use standard issue bearings, that way i can change out any one part of any one truck if it wears or gets damaged... plus its an excuse to be out in the shop and on my lathe... which in and of itself is a great joy for me...

[project5k]

ok, so it might be a mistake, but i did it anyway, i ordered my electronics for to build the drivers, l297's and l298's and some diodes... and then just now i ordered my motors.. i got 3 nema 23 500oz-in 4 wire bipolars for $175... so how'd i do??

the way i figure it, ill spend the weekend building the driver boards, then early next week the motors will get here from california and i can test both them and the drivers... pluss then i can see what the no load speed of the motors is, and that will give me some kind of an idea as to what i need as for the rest of the drive train...

I did have this thought... what if i mount one of the motors onto something solid, and then mount a large set of fan blades.. i know theres not really a way to measure the amount of load, but it would allow me to see how much the rpm's fall off and or if it starts cogging, and give me some idea as to what kind of ramp up to play with.... Ideas? suggestions??

[Mack]

Try to float the pinion on the rack with spring loaded cam system. Works quiet well (will draw rhino example if requested). Typical effective stepper speeds is about 500rpm so calc gearing firstly, pulleys etc can get in the way of construction. Rigidity is essential. Weight also helps on fame etc, gantry should be light and ridgit. Belt an "y" axis works well. For 3d application keep "z" axis accurate, rolled ball screw with good linea and coupling. Do 1 thing at a time. Build one drive test it on cheap motor. Clamp motor to avoid resonance. Apply a little load to that motor ie flywheel or a bit of resistance. Good thing to build 3d model in rhino solid egde etc, helps with spacing and avoids bolting and assembly issues. Build one example of linea bearing, test it on small profile, get it smooth but ridgit. Avoid welding long flexable pieces of steel, try precision dowel and bolts. Watch for swarf and debris on crucial mechanical parts. etc, etc, etc

[project5k]

yes, if you will please draw me an example of that one, im just not getting it in my head...

ok, 500rpm, i can definately work with that....

i was worried about using a long belt, having no experience with them, i was worried about it being kinda springy, and giving and stretching under load and with sudden accel's and stops....

clamp the motor??? you mean to the frame with a c clamp? or are you talking limiting the current?

my rails will be above the table, so im hoping that will reduce the amount of chips, and im planning to make some little brooms that will clear the rails as the axis move...

im considering lots of options for the x axis drive, possably dual x motors and screws, or maby rack and pinion.. or maby even a crossed chain design... im just not sure yet... the other idea that i had was to run the x screw under the table... mount the x motor on the gantry and spin the heck outta the nut.. i have a couple of ideas on how to do this... my thinking is that this way i wont have to use as large of a screw to reduce whipping.. i'll only be worried about the push and pull forces... plus by having a stationary x screw i eliminate the end play of the screw...

the other off the wall idea that i just came up with and its cheaper than all the rest of the ideas ive had, was to have a length of 1" wide .5" pitch drive belt along each of the side rails.. then do the upside down u drive thing with the belt. By pulling the belt tight at each end, im creading a flexable rack... The belt should have little to no slop in it, and thus is better than the chain, and is also going to be lots quieter than the chain.. then just have a shaft that connects the 2 sides together and have the x axis drive motor mounted on the gantry.. again i know this adds weight. the point of this is to not have to have closed loops of belts.. but just use the cheap easy to measure for open ended stuff... i'll try and draw some of this up tonight when i get home... but this is sounding like the way to go...

another idea that i've been toying with is for my power supply, im thinking that i really want something in the 36-48 v range, so i was just wondering, about atx power supplies, i saw somewhere else on here someone was thinking about this.. but im going to wire it just a little differently.. im only going to use the ground and +12 lines.. what i need to find out is if the ground is common to the chassis of the power supply, or if its more like the CT on a transformer.. if its an "open" or floating ground, then im in good shape, i can series 3 or 4 ATX's together to get the 36-48v.. the question is going to be the current.. well i have someone willing to donate 4 computers for parts to this project, so maby i'll pull one of the PS's out and play with it, its the right price so that if i smoke it im not out... and yes i know i need a "load" resistor on each PS so that the switcher is happy... i just need to figure out how much load, and if i need one for each V out, or if one on any one of the v outs will work...

[ger21]

http://www.cnczone.com/forums/showthread.php?t=17548

[project5k]

o i c so i do have to find the chassis ground and dissconnect it.. ok, good information.. i gotta go get thoes machines and rip out the psu's and set it up.. if nothing else for the experience.. and maby some fireworks, and then just maby maby something useable.. maby for the cnc, maby for just a test bench power supply... spiffy...

[Mack]

Hello p5k.

lets take it one step at a time. There is a lot of litrature on stepper motors, some of the most accurate is from Marais at geko drives. He has a short sysnopsis or document explaining the relationship between power/current/voltage etc of these motors, it is definately worth a read.

Without getting very tech about this here are a few generalisations:

Steppers act almost inversly to a dc brush motor in that they have unique power to voltage curves etc. for example more volts can produce more torque but not neccisarly more speed.

A rule of thumb would be to start out with 20 to 25 times rated voltage, but reduced current to the motor, heat is a good indicator as to wheather the motors are running near there capabilities. Hot to the touch but not excessive. If you can hold the motor for say 10 seconds it probally is running well. This sounds kinda strange and not very scientific however there is a lot to these little devices. Increasing the current will generate more heat, vica versa. Resonance is a small issue as well, at certain speeds and a certain current these motors can produce sporadic results also depending on the load they are trying to drive. It is always good to have some load on the motors do dissapate energy that can return to your driver. It may reduce resonance as well, free running motors are not a true indication.

A lot of steppers availabe produce good power/speed results at about 500 to 700 rpm. This brings me to gearing.

You would be surprised at how accurate a shop built machine can be for say cutting mdf, ply etc with a reduction ratio that translates to:

// I hope my math is correct, we deal in metric units over here. I know my spelling sucks. //

lets say 100 full steps to 1/2 inches (approx 6.35mm) of travel, yielding quit a good speed with resonable accurace, however you can micro step these motors ten fold producing very respectable speed and accurace, however the step output frequency from your parralell port will also increase.(there are limits to the parallel output normally/reliably around 20k.)

ie a 1.8 degree stepper has 200 FULL steps per rev. therefore 600rpm translates to 120000 steps per minute or 2000 steps in a sec.

or at 200 steps per inch producing speeds of up to 10ips (10 inches per sec or 250mm odd per sec) and microstepped will give ya accuracy of 2000 steps per inch. remembering your pc output or driver might only allow 20k on the limit in this case.

With this resolution in mind, i have had no trouble with with using a reinforeced belt driving a 2metre y axis or 6 foot axis with less play than the capable resolution of a stepper.(especially with the drift between its holding points)

as far as the x axis goes, i highly recomend rack and pinion. Firstly its cheap, strong easy to assemble (will send drawing of floating mount) yields great accuracy.

As far as power goes, motors of most sorts like simply regulated supplies , ie a torrodial with reg and caps, maybe a resistor to draing cap. Over here they are pretty cheap ie 400va 36v ac torroid is about 30usd. Switched mode supplies can produce odd results, but i am no expert in this fiels.

[project5k]

ok, thats some good guidance... thanks.. well the drivers that im building can handle up to 46V so im going to aim for about 40 for now.. and then i can allways build bigger and better drivers later on.. or buy some, whichever the case may be.. but for now thats a decision i made, and so im going to live with it.. cause the parts are here, and i actually have one of the drivers breadboarded up and working some led's with the demo of mach3...
i cant add the video here, so heres a link...forgive the yellow snail.. i was trying out some new free software...
http://www.jpcustomcrafts.com/shopmachines/Video_09.avi

oh and i did some more drawing, this time showing the long belt mech, x axis drive motor, and cross table timing shaft.. i noticed several things when drawing this, so theres gonna be some redesign... first thing that i dont like is the thought of 6 feet of unsupported spinning shaft, so i'll probably but some sort of bearing block in the middle of the x slider... secondly for the 5:1 gearing i noticed that the pully is going to be in the way of the y and z axis when its at one end of the travel.. so i'll have to come up with something, maby a jackshaft arangement, depending on how much slack there is in the belts and pullies...

i arranged the belt with the "teeth" downward so that they wouldnt collect swarf and or dust... and the 2 lower pullies are to guide the belt and to maintain a higher % around the drive pully in contact with the belt..... lemme know what ya'll think..

[Mack]

Two motors on the x axis seems far more appealing than that cross link shaft. You said you may want to router with this machine, but i do not understand how you plan to overcome the force applied by the router on the material against the lift of your cross travel slide. Those belts will have to be tensioned quiet correctly (not a big deal) but may lead to some drift. You could always re-calibrate mach. Those gearing units will have to be well secured to the gantry. Why not rack and pinion? (I dont really mind that belt idea, it will cut fairly accurately).

[Mack]

Another thing, I believe that cheaper rack is cut with better tolorance than belt tooth spacing. But i cant confirm this, mabe a good idea to look into it.