New Build -- Moving Gantry Milling machine

[007CNC]

Hello everyone. Newbie here.
I am starting a build log of a moving gantry milling machine.
I am not completely new to CNC as I have built one in the past. I have learnt so much from the CNC veterans on this site, and I hope to combine my learning here with my mistakes from an earlier build to come up with something better.
Your inputs are highly appreciated.
Here are some stats:
Machine type: I am aiming for a 36"X24" moving gantry mahine built from hollow rectangular steel tubes
Usage: Mostly used for milling aluminum and a few plastics. Occasional steel drilling.
Accuracy: I am choosing repeatability over accuracy. I am happy with a very rigid and repeatable machine at 0.01" accuracy.
Mechanical components:
42" hiwin linear guide rails on the Y axiz
32" THK linear guide rails on the X
some other (unknown) 12.5" linear guide rail for Z
Acme screw or ballscrew for X and Y, ballscrew for Z
2 nema 34 640 oz-in motors for Y, and 1 for X
1 Nema 23 motor for Z
Electronics:
4 microstepping driver (M542H). I believe its chinese.
Power supply:
48V 7.5A for X and Y, 24V 30A for Z and my spindle.

I have been working on a CAD model for a month now. I will post design pictures shortly.
Your critique/inputs are highly appreciated.
Thanks

[007CNC]

Here is what I have...
I still have a few parts to place on the Z axis, but I believe I have drawn all the parts I need to cut.
I will be using my old flimzy CNC to cut the aluminum parts and drill the holes in the Steel tubes.
I am considering either this 3/4 X 0.5 high lead acme srew, or this 5/8 X 0.2 ballscrew , both from Roton. Considering the advantages of ballscrew, the 5/8 ballscrew option seems attractive, but at over 42 inches, I am worried the ballscrew might whip at higher speeds. I hope to get rapids of over 300 IPM.
With the acme screw being thicker and having a higher lead, I am leaning towards the acme screw with the threaded mount nut. The thicker ballscrews are out of my price range.
Any input here is appreciated.

[wizard]

Your friction goes up considerably with acme screws and as a result that impact speeds, wear and noise. If positioning accuracy isn't that important to you, you could look at higher lead ball screws.

Frankly I consider 0.010" positioning accuracy to be pretty terrible. Maybe that is all you need for now but I'd strive for better. It appears that you already have parts to make such accuracy possible. I just don't see you having much luck with acme screws and the high speed rapids when it comes to repeatability.

[louieatienza]

The ballnuts from Roton are not anti-backlash, so you'll need two for each screw to "preload" them to remove play. Also if you look up the lead accuracy of some of their ballscrews, it's no better than some of ther ACME screws.

As to efficiency, some ACME screws can achieve 40-70% which isn't horrible, considering the availability of AB nuts. There's less noise since there is no recirculating balls, and AB nuts are self-compensating for wear.

As for your motors for the Y, they seem ridiculously large. Their torque drops off pretty dramatically at higher speeds compared to smaller steppers, so you should factor that into your design as well.

If you want to reach 300ipm rapids, you'll need to spin the motor at 600rpm for the .5" lead and 1500ipm for the .2" lead. There are very few steppers that are capable of providing enough torque at 1500rpm. Also you'd need a very high voltage power supply (and drives that can handle it) to take full advantage of the 640in-oz NEMA34 steppers. If you use ballscrews with .2" lead you may want to consider servos, if you want those rapids.

As an example, I use 1/2"-8, 8 start ACME screws on my machine, direct-driven with 425in-oz NEMA23 steppers with 48V power supply. The steppers could handle more voltage, but I still manage 500ipm rapids on my x (running 2 screws via belt, and a 90lb gantry) and 750ipm on the y (kernel limited). I cut aluminum plate and G10 regularly and the steppers don't bog down. When I read stories of folks doing 750ipm rapids with .5" lead ACME screws I laugh, since their steppers are nowhere near capable of torque at those speeds, and the delrin AB nut would probably start melting! But despite the high lead, I still manage overall about .004"/ft accuracy and less than .002" repeatability; I rarely miss steps, and my machine holds and keeps position well. My X screws are 42" and the y 32"; one X screw whips slightly because it has a slight bend that refuses to go away, so I probably could spin them faster without whipping.

As for accuracy, it's easy to get enamored with very high resolution, but with steppers that comes at a price of speed, unless you use servos. And all that resolution doesn't mean a thing if your leadscrew accuracy surpasses your mechanical resolution (can be .003"-.009" per FOOT for rolled ACME) or even your backlash or machine deflection. Repeatability of position to me is more important, so having anti-backlash leadnuts and a stiffer stucture would trump rolled ballscrews/non AB ballnuts. Related to this, you might want to consider ACME on the Z, depending on the weight of the Z, since the ballscrew could backdrive and drop the carriage when the machine is off (depending on the lead of the ballscrew).

[leix_99]

If you want to use 4row recirculating hiwin rails and blocks keep in mind that you'll need to carefuly tap those steel profiles because rail paralelism should be within 8-10 microns.
If i were you, i'd consider thet gantry design, soldering or whatever you're up to make it sould be very precise if you pretend to drill the steel tubes before soldering them, you'll have no margin to align everything up.

[007CNC]

Thanks guys! these are the types of feedback I was hoping for coming to this forum.
I am currently going with the acme high lead screws as I have those in hand at the moment. I will see how they perfrom, and if I am not happy with them, I will save up and go for a 3/4 ballscrew.
The nema 34 motors I already have, though I suspect they are overkill for this project.

The ballnuts from Roton are not anti-backlash, so you'll need two for each screw to "preload" them to remove play. Also if you look up the lead accuracy of some of their ballscrews, it's no better than some of ther ACME screws.
.

I searched for info on the lead accuracy of their ballscres and acme screws, but I couldnt find it on their website. They are both rolled, so I think the accuracy will be similar in both cases.
The AB nut is where I am really scratching my head. I will really like to have 0 backlash in this machine, but am not sure what type of AB to implement.
My nut is plastic, and am worried that preloading might lead to premature wear on the nut. Anyone have any info/suggestion here?

As for accuracy, it's easy to get enamored with very high resolution, but with steppers that comes at a price of speed, unless you use servos. And all that resolution doesn't mean a thing if your leadscrew accuracy surpasses your mechanical resolution (can be .003"-.009" per FOOT for rolled ACME) or even your backlash or machine deflection. Repeatability of position to me is more important, so having anti-backlash leadnuts and a stiffer stucture would trump rolled ballscrews/non AB ballnuts. Related to this, you might want to consider ACME on the Z, depending on the weight of the Z, since the ballscrew could backdrive and drop the carriage when the machine is off (depending on the lead of the ballscrew).

Yes I do agree. rigiditgy and repeatability are higher priority than resolution in this build. I think my structure so far is very stiff (will post pictures shortly). AB nut is where I need to come up with something strong and reliable enough. I have a ballscrew with AB nut on the Z of my current machine, and it doesnt drop. I will be using the same ballscrew/nut set for my new machine.

[007CNC]

If you want to use 4row recirculating hiwin rails and blocks keep in mind that you'll need to carefuly tap those steel profiles because rail paralelism should be within 8-10 microns.
If i were you, i'd consider thet gantry design, soldering or whatever you're up to make it sould be very precise if you pretend to drill the steel tubes before soldering them, you'll have no margin to align everything up.

haha.. that's actually what I did... I drilled all the holes first before welding.
I had to drill over 200 holes with my CNC on the steel tubes. There is no way I could have done it accurately after welding.
I never planned on tapping any holes. I will be using nuts and bolts to hold the rails. For the places that will require fine adjustments, I drilled the holes a little larger to have room for adjustments.

[leix_99]

haha.. that's actually what I did... I drilled all the holes first before welding.
I had to drill over 200 holes with my CNC on the steel tubes. There is no way I could have done it accurately after welding.
I never planned on tapping any holes. I will be using nuts and bolts to hold the rails. For the places that will require fine adjustments, I drilled the holes a little larger to have room for adjustments.

My fault, replace tap for drill. Anyway, i'm telling you this because i'm facing the problem now, i made a mistake in my build so rails were unparalell by 0.8 mm and i'm having to redrill my rails to a bigger diameter to get them aligned.

Check it: http://www.cnczone.com/forums/epoxy_..._g_mill-2.html

Cheers

[007CNC]

Welding is not one of my strong points. I own a mig which I pull out when I absolutely have to, but I am yet to get good at laying down the beads. After considering how much time and effort I have already put in, I decidd I will let someone who knows what they are doing to handle the welding. I called up a mobile welder off of craigslist.
I will line up & square the metals while the welder lays the beads.
Prior to welding, I drilled all 200+ holes. This was done over the course of a few days. I picked up a machinist level from e-bay, and I put together some adjustable leveling feet (getho style levelers, but they worked). The levelers were made with half inch bolts, pieces of wood, and threaded rod couplers. I get about 1inch of adjustment from each level.

[007CNC]

Welding day pictures
Not so much comment... just pics.
The mobile welder showed up bright and early at 9:00 in the morning, and we got busy. I think it was past 6:00 pm by the time he left. Most of the time was spent aligning and squaring the structure. Here are some pics.

[louieatienza]

Thanks guys! these are the types of feedback I was hoping for coming to this forum.
I am currently going with the acme high lead screws as I have those in hand at the moment. I will see how they perfrom, and if I am not happy with them, I will save up and go for a 3/4 ballscrew.
The nema 34 motors I already have, though I suspect they are overkill for this project.



I searched for info on the lead accuracy of their ballscres and acme screws, but I couldnt find it on their website. They are both rolled, so I think the accuracy will be similar in both cases.
The AB nut is where I am really scratching my head. I will really like to have 0 backlash in this machine, but am not sure what type of AB to implement.
My nut is plastic, and am worried that preloading might lead to premature wear on the nut. Anyone have any info/suggestion here?




Yes I do agree. rigiditgy and repeatability are higher priority than resolution in this build. I think my structure so far is very stiff (will post pictures shortly). AB nut is where I need to come up with something strong and reliable enough. I have a ballscrew with AB nut on the Z of my current machine, and it doesnt drop. I will be using the same ballscrew/nut set for my new machine.

As for AB nuts, I use the ones from DumpsterCNC.com. Although they're only rated at 25lbs force, I've successfully cut G10, other phenolics, and aluminum with my machine. I'm sure the finish isn't as great as it would be with ground ballscrews but for my needs it more than suffices.

I've seen double ballnuts used with wave washers to provide preload, possibly something similar can be done with bronze ACME nuts? I woould ask in the Mechanical forum. Thomson makes really nice AB nuts, though I don't know of their cost.

[007CNC]

That was the question I pondered on for days.
I had started down the path of acme screws. I bought a set of 3/4 dia 4 start acme screws and plastic nuts from Roton. Felt that the 0.5" lead was a good ballance between speed and resolution for my application.

The point I couldnt get past with acme screw was the issue of backlash and how to deal with it. I wanted a solution outside of compensation in software. most of the solutions I came across required 2 nuts with some kind of spring compression in between. With this setup, I was worried about premature wear on the nut.
The other split nut aproach just doesnt seem that it will hold up well to a lot of axial force while cutting.

Finally, I decided to go for a Chinese anti backlash ballscrew -- 16mm dia with a 10mm lead. I am hoping that when they say anti backlash ballnuts in Chinese, it really means anti backlash ballnuts. The order is probably about a week away (ordered a little over a week ago)

With this change, I had to redo my design and scrap all the alu pieces I had already cut to use with the acme srews.
I will probably throw the 3 acme screws and 6 plastic nuts on ebay whenever I have time.