[Acebull]

G'day all as the title says, the truth and nothing but the truth. I've roamed the forums on the internet for the answer and all are different. My question is, what to use, rack and pinion, lead screw or the most expencive, ballscrew. I know ballscrew is good but you need Bill Gates pockets to pay for them. So that leaves rack and pinion and lead screw. A lot of forums say rack and pinion is not good for accuracy but good for speed but wears out easier. Lead screw has problems with whipping and backlash but better for accuracy but slower. So what the hell is the truth, please help me out.

Ted.

[Andrew2111111]

I would use rack pinion.
the trumpf lasers at work use rack and pinion.
but it would depend on the quality of the rack and pinion.

[CarveOne]

Rack and pinion is not as inaccurate as a lot of folks claim. Poor grade cheap parts, incorrect adjustment of the tensioner and lack of maintenance will not give the accurate results one might want. Many of the commercial CNC machines over 48" (1219.2 mm) travels use rack and pinion. One of my two DIY machines has 12' (3657.6 mm) long travel. I don't expect it to maintain really high accuracy over the full length. Over shorter distances anywhere along the travel it does quite well. If I want to cut aluminum with 0.001" (0.0254 mm) accuracy I would not expect to use something like a CNC wood router machine to do it, athough it can be dialed in to that in Mach3 and LinuxCNC at any one point along the full travel. For making large signs on this machine the high accuracy is not really a requirement. Once installed and adjusted there is not much maintenance required if you keep dust and debris off the drives and racks. Same for any other drive system. It's a good idea to install the gear rack teeth facing downward to help keep them clean.

There are machines with 20' and 30' travels, and rack and pinion is about the only way to do that.

[Acebull]

Hi sorry I did not put the size I wish to build it will be 900mmX1500mm and hope to do 2d/3d carving and scale model plane plan cut out

[ger21]

You can get cheap, average rack and pinion or very expensive, very good rack and pinion.
You can get cheap, not so accurate ballscrews or very expensive, very accurate ballscrews.
You can get acme leadscrews for a few dollars/ft, or you can get very expensive, precision ground acme screws with anti backlash nuts.

What is the intended use? How much travel? How much weight are you moving? How fast do you want to move it? How much precision or accuracy are you looking for? What is your budget?

The best option for one machine might be totally inappropriate for another.

[louieatienza]

It all depends on the final use of your machine. You would pick the drive mechanism that suits your accuracy, speed, and power requirements. Precision ground ballscrews are extremely expensive, but extremely efficient at converting rotary motion to linear motion; though that precision is wasted if you do not have precision linear guides (also very expensive), zero flex in the machine (very expensive to impement), and usually paired with servos (also expensive.) Gear reduction may also be needed further adding to the cost. But they offer the smoothest, most accurate and most efficient movement. It's getting harder to find these things on eBay cheap. You can get rolled ballscrews, though their accuracy is not as good. Still, they may be good enough for the application.

You can get high accuracy rack-and-pinion, ground and hardened. This requires extra care in installation, and may also require expensive gearboxes, servos, and anti-backlash mechanism to take advantage of its accuracy - all expensive. Again the more affordable systems like CNC Router Parts are usually good enough for most CNC work, expecially if you're working mostly in wood. Steppers can provide decent resolution, and again, it depends on your accuracy requiements.

The proper sized leadscrews can provide the best bang for the buck. High lead leadscrews can provide good speed and prevent whip. Parts like anti-backlash nuts, collars, couplers, and other parts are more easily available and inexpensive. It maybe overkill to use servos with them unless gear reduction is used, but then you end up with a mechanical resolution far greater than the accuracy of the screw itself.

The size of your machine may ultimately determine what you use. Ballscrews thick enough for a 4x8 table are ridiculously expensive and require powerful servos. Smaller lengths can be found surplus on eBay cheaply for a smaller machine, though all the components needed for them, from end bearings to couplers to gear reduction to servos can add up. You can direct-drive a ballscrew with a stepper, though your speed (and resolution) will ultimately be limited by the pitch of the ballscrew.

Rack and pinion would the the least expensive way to build a large machine, with good speed. Careful installation and design can lead to components that won't wear as fast, though that can be said for any design. I think it would be impractical for a tabletop machine, as the speed advantage may not be realized, and the initial expense higher than ACME leadscrews.

ACME leadscrews in my opinion give the best price/perfomance ratio for a small to medium sized machine. Again, sizing the steppers and using the correct lead for the application are very important to get the best performance. ACME leadscrews also come in different acuracies, even precision ground, though that can be expensive as well. Most anti-backlash nuts are made of delrin or other thermoplastics which limit the max load of the nut, so a heavier machine may requre bronze nuts with some anti-backlash mechanism.

Again, the higher the accuracy of your drie components, the higher the degree of accuracy you'll likely need in other components, requiring a stiffer more expensive machine base, otherwise you're just wasting money. Once you figure out what you want your machine to cut, you need to find out the speed you need to cut it at, and the accuracy you want to cut it to, which should help you decide what drive mechanisn to use.

[Acebull]

G'Day All thanks for your input what do you think of going 3/4 .500 4start lead screw for X&Y and 1/2 100 lead screw for the Z and Linear Rail for the motion the machine size that I want to build is 900mm X 1500mm I leave you to convert that as in Australia we are in metric I hope to do 2D/3D carving and scale model plane plan cut out

Ted

[louieatienza]

If you are doing 3D carving you may have problems with the different lead on the Z since it won't be able to move as fast as the x and y axes.

[ger21]

Yes, 1/2-10 single start is too slow for the Z axis. You'll want something faster.

[coherent]

The accuracy of rack and pinion should not be an issue for your intended use. Like a previous poster stated, you can get cheap or quality parts no matter which design option. I don't agree that the different size screw or pitch for the Z will cause any problems. The speed of the axis also depend on your settings, motor capabilities etc. Most Z limits are minimal (as compared to the X/Y) unless you're building a huge machine to manufacture big parts. If that was the case you wouldn't be asking what you did. Rack and pinion lends itself well to Laser and Plasma machine because the Z axis of the machine doesn't come in direct contact at the cutting point like a router or mill does with the material. This is important because with many rack and pinion systems the gantry (for the Y and Z axis's) "floats" on the rack. That doesn't mean there aren't CNC routers that don't use a rack an pinion system. If the design is right and the Y and Z axis isn't affected by the downward or linear pressures of the Z axis they work well. The first question should what you need to cut, then what size machine needed to cut it, then your budget. With those basics in mind, find something on the forums someone else has built/designed that fits within your parameters and go from there! Look into ball screws on Ebay from china. They work well for some projects and are actually pretty reasonably priced for us budget home builders and definately won't take a Bill Gates budget. There's a lot to be said for someone already working out the bugs and design issues of a build, or showing what works well. You'll find some builds overkill and some under built. Most build threads identify these issues. The happy medium you want depends on your specific needs, available tools/building skills and how much you want to spend. And, that's "the truth and nothing but the truth"!
Best of luck

[louieatienza]

...I don't agree that the different size screw or pitch for the Z will cause any problems. The speed of the axis also depend on your settings, motor capabilities etc....

Let's say optimistically that the OPs Z axis stepper can reach 1000rpm. Using a 10tpi leadscrew would give you 100ipm no load, so I'd expect quite a bit less true speed pulling the weight of the z carriage. This wouldn't necessarily be bad for profile cutting, 2.5d, and even 3d z-level (waterline) machining. But if you had to do a full 3d finish pass on a surface, this will limit the max speed you can machine at. And with the usual smaller stepovers of finish operations, this can really add up, especially when machining wood and foam which can be cut at faster speeds (over 150+ipm...)

[coherent]

But if you had to do a full 3d finish pass on a surface, this will limit the max speed you can machine at. And with the usual smaller stepovers of finish operations, this can really add up, especially when machining wood and foam which can be cut at faster speeds (over 150+ipm...)

Yep, OK I agree, but there are obviously other factors that are important. I reread the thread and noticed he's looking to build a machine that is capable of over 3 ft x 5ft and cut 3d. I saw model planes and was thinking small. A direct drive stepper machine is going to have limitations no matter what. The available torque from the stepper motors, being speed related (more speed - less torque) sets the practical limit for the fast traverse (and cut) speeds of the finished router. While the no load rotational speed, under Mach3 control, could be as high as 1300 rpm (and since he's in AU, I'll use metrics... ) which equates to a fast feed rate of around 6.5 Metres per minute. And yes, the screw/tpi/ etc will be a factor. But at these motor RPMs (as you pointed out) the torque is so low that any reasonable resistance to rotation will stall the motor. It would be extremely difficult and very costly to build a machine with low enough friction and big enough stepper motors for this to be a practical feed rate to aim for. It is this very limitation which leads professionals, where machine time is money, to chose servo motors as the driving force, but hey, that is another subject. A practical speed of around half the maximum attainable would be a good starting point to aim at. I have found that a maximum motor speed of 600 rpm giving a fast feed rate of 3 Metres per minute is easily achieved and 100% reliable under most circumstances with a well designed stepper machine with half way decent parts. The screw pitch will definitely make a difference, I'm sorry if I made it sound like it was not a factor at all, but my thoughts were that it's going to depend as much or more on other factors... direct driven? gear/belt ratios?, or if it's rack and pinion, their gear ratios etc. Then add in slide & screw quality etc (and of course the screw size) to determine maximum attainable speeds. All of this is well and good (and gave me a chance to ramble needlessly), especially for the X & Y, but my thoughts were that were talking the Z axis screw here where travel distances are much shorter overall (a few inches max?, and the bottom line speed/time to complete a cut isn't affected any where near as much as what the X & Y speeds do. Hence my comment about the different Z screw size being acceptable. But you're right... I've cut 2.5 D finish cuts with a slow stepper driven machine (slow X & Y, much less Z) and it has literally taken hours. So other questions unanswered become pretty important also: Is this for production or hobby use? How much is the budget? I likely wrongly assumed that just because of the forum section this was is that we're talking a hobby build where $ is a major factor and speed of cut (as affected by the Z travel speeds) was secondary. I should ask more questions before I try to offer advice.

[ger21]

With the screws he's proposing (1/2" lead), 600rpm will give you 300ipm, or about 7.5m/min.

This is easily attainable with acme screws and Nema 23 steppers. Most use 1/2" acme, but you have to be careful of whipping at longer lengths. Going to 3/4" acme will greatly increase the inertia, resulting in lower acceleration. Also be aware that 1/2" screws are more efficient than 3/4" screws, which will probably affect top speed. You can probably go with a bigger stepper, but as steppers get bigger, they tend to spin slower. So you reach a point of no return. For this reason, most people these days choose rack and pinion over acme when lengths get over 4ft (1200mm). Rack and pinion is much more efficient, and can give much higher speeds. But you'll trade off some resolution, unless you do a 2 stage belt reduction, or use some type of gearbox reduction.

[Acebull]

G'Day All And thank you for your input but let me put it to you this way if all of you would like to build a CNC Router with a work area of 3'(900mm)X5'(1500mm) to do 2D/3D carving (on cupboard and pictures and the like) and as I am in to scaled RC model planes up to about 4 meters cutting the wood out to build them so back to the CNC router how would you go about driving this beast

Thanks Ted
PS. But you don't have a open check to do this if I did have a open check I would have just bought one from the CNC shop LOL

[ger21]

Look into getting chinese 1610 (or 2010 if you can find them) ballscrews from Ebay. That's probably the best method for the price.

[coherent]

Rack and pinion would be reasonable as far as price. You're eliminating the need for high cost linear slides/bearings. If you want to stay low cost, do a search on the forums & google for 80/20 cnc machines. You can use either 80/20 slides or mount different linear rails/bearings to the 80/20. I have a 4ftx8ft 80/20 frame machine with rack and pinion on the X/Y that works well and is accurate. I also have a smaller 80/20 machine that uses iges linear rails and bearings with acme screws that has been accurate and reliable for router work. One thing about using aluminum extrusion is that everything is fairly modular and you can change, alter, disassemble, and replace parts easily when you want. I've seen about every drive (direct, belt, R&P etc) and every slide type used with 80/20 because of it's versatility and simplicity. I've not used the the 80/20 linear slides, but the cost is low and I've read on this forum and other searches where folks have used then for build with mixed results. (example is here ----? mikelow ?----). If you have any cad/3d design software that you are proficient with (especially something like inventor etc) you can find most of the 3D 80/20 parts here: 3dContentCentral.com. Seeing how everything fits will help eliminate issues during the build.
If this is your first build, another option is to get a 3 or 4 axis stepper driver (like a gecko g540 or similar) and your stepper motors and start small. Build a machine with a smaller footprint first. You'll learn a lot that way. Then when you feel you have a good understanding of the software, the machines operation, etc. Enlarge it or build a second. Your driver, motors, couplers, cables, software, and other parts etc can always be used on the larger machine and there's always a market for used parts you don't want to keep. That doesn't mean you can't dive right in and build a large machine, but it's just an option. I know the most basic option is wood and drawer slides, and while that's great as a learning tool you're limited to a very small machine and its capabilities are limited due to alignment and fastening problems that are more prevalent due to the nature of the materials used. These are issues that are eliminated with with a more solid, rigid and square structure (i.e metal). An 80/20 frame, chinese ballscrews & slides, (or lower cost slides and acme screws/nuts to save a little more) and stepper motors would be about the lowest cost machine that would be fairly rigid and dependable for hobby use. One of the least expensive rail/slide systems I'm aware of is the redi-rail from pacific bearing (here: http://www.pbclinear.com/Redi-Rail-R...ar-Guide-Rails). I haven't personally used them so can't comment on quality vs price.
If you have access to steel or aluminum tubing and are capable of cutting & welding, that can cut some costs for your machines frame parts, legs etc. Above and beyond that it just depends on your budget and how fast you want to cut (servo motors and higher cost lower friction parts). You just need to start reading build threads and start pricing parts and decide which direction you want to head.

[Acebull]

G'Day Coherent just read what you posted on my thread and you know a lot more about this cnc stuff than I do which is why I'm on the forum to learn and I see that you are using Rack and Pinion on your big machine so could you let me know what R & P set up you are using such as motor size pinion and rail size and if its geard or not as I think R & P might be the way to go

Thanks Ted

[coherent]

My rack and pinion machine is a 4x8ft plasma cutter. Is is an 80/20 frame. the Z is an electronic automatic torch height controller but the X & Y are pretty simple. The X is 2 steel strips mounted to the top of an 80/20 frame. The gantry rides on bearings along the steel plate. A rack is on each length of the X axis. The steppers are fairly small (only about 160oz) and drive a shaft by belt drive that extend the length of the Y and have a pinion gear on each end the ride on the rack. A plasma cutter doesn't operate with very high cut/movement speed and in comparison to a high speed router/spindle CNC machine is pretty slow, so my specific machines design may not work for what you want to do . I don't have any specifics or plans for a rack and pinion build. Here's a video of mine cutting some thin cold roll steel of some cowboys Beaver Creek CNC Plasma cutter in action - YouTube
I'd suggest you start searching build threads for folks who have built something similar to what you want. Be it rack and pinion or other. That's really your next step. Then start scrounging/searching/pricing parts that would work for something similar and go from there. There are some great build threads on this forum. Searching and reading will get you headed in the right direction and give you a better understanding of the how to's.

[louieatienza]

There's no rule that says you need to use the same type of drive for each axis. There are commercial machines out there that use rack and pinion for the long axis and ball screw for the gantry.

Note that a plasma cutter does not need a lot of holding torque to keep position since there are no cutting "forces" against any of the axes. So for routing your needs will be different.

I built a machine with a footprint slightly smaller than your needs, but it seems to work well for me. In my YouTube channel I go through the different iterations.... AtienzaLouie's channel - YouTube

[ssutton]

My setup is R & P for the X and Y and lead screw (single start 1/2-10) for the Z. I Love my machine for the combination of speed and accuracy. The only thing that I would change on it as far as drives go is I would change the Z to a 2 start screw for more speed on the Z. I mostly make aluminum machine parts with my machine from CAD models and it works great for the fit and finish of the made parts.