As long as your linear guides are straight, the amount of run out you see in the ballscrew will be a non-issue with them spanned unsupported at that length. Ultimately you really only need the pitch to be consistent along the length, reasonable straightness, and mounting parallel to the linear guides. Interesting design with the rotating ball nut. What sort of positional repeatability are you aiming for? For an approximately 6 foot span like that I think you shouldn't look at aiming tighter than a couple thou. At that size, thermal effects, and machine rigidity are going to be big factors limiting you from getting any more accurate.

In high precision machines the ball screws are stretched between the bearing housings. Helping keep them straight. Also make sure they are aligned end to end withe the ball nut. Get it in place and tight leave the ball nut on one end and put an indicator on the axis moving it back and forth checking it for straightness to the axis.

You're ALWAYS going to get deflection on something that long Chinese or otherwise, if you're worried you could add a simple sliding support (E.G. a ball bearing wheel running on a flat steel track with a threaded rod for height adjustment is a cheap simple option) to the ballnut housing to take some of the strain/pressure off the nut itself but I doubt it's strictly necessary.

Impressive Work.. To measure backlash, put indicator on nut or nut housing/carriage. Place indicator tip in thread of ballscrew. Turn screw back and forth over travel. Any needle deflection is backlash. This isolates movement to screw/nut. The straightest screws that length will whip at higher speeds. Driven nut is another solution.Screw is fixed, screw support bearings and servo are moved with timing pulley and drive nut. One more bit of axis cabling that moves, but on a serial cartesian machine you have at least two sets already.

Its better to avoid binding if the ballscrew is as straight as possible, but just getting it close, is good enough. Ballscrew is the drive and the linear rails are the guides for straightness and proper operation, and yes a ballscrew can be so badly bent that it will bind the entire axis, but you really need a great deal of runout to create a problem. I can say from dialing in my big CNC, the ballscrew has runout, but just running it back and forth on the axis - -get it to find it's comfortable home at each end -- it will basically dictate it's own location to avoid binding, and then just shim the bearing blocks appropriately. If you over constrain it, it will bind, but a bit of runout isn't an issue.

Hi friend, thanks for watching. Would really appreciate any advice on how straight these screws should be over this sort of length.

Hi. Having not built a CNC, but planning to and therefore gathering info. Also being a Machinest with quite the workshop, so no complete punter. I like your second purchased set of ball screws much more than the first. Both for the pitch and the quality of manufacture, as demonstrated by the quietness in operation. Imho, the slight woof in the shaft won't be a problem in a spinning nut setup so long as the constrained ends are at the same (correct) hight. My reasoning; the deflection demonstrated with light pressure was greater than the measured woof in the shaft and therefore cannot exert enough pressure to upset anything! I'd probably have the woof high point upside as gravity will tend to straighten it. Like the spinning nut though. Much less inertia/momentum and no whip. I'll be watching.

Aaah the classic Mitutogo dial test indicator

Great Idea rotating nut - make sure the screw is not under compression when installed - the bearing blocks on the rail should reduce any deflection from the screw - just be realistic on the accuracy you will achieve. when you measure the machine accuracy / backlash try measure over the entire distance as that will give you a good average / indicator.

As long as you are not Prototyping for NASA or DARPA, and are not going to do high production applications, you will be fine i think with those you got now. They can bend a little, over the length you got without feasible support, it's almost expected, alone from it's own weight of the component. Beside that a C5 would cost a fortune, you come already with your actual C7 to a point, where inaccuracy is cause be noise of other components and setup where, the run out is only one factor. I think, if you need to go more precise, you would need a complete different machine design. My advice is to continue with your build, without worrying too much, look how it performs, and troubleshoot later. For example, before worrying about about the screws and actuators, you would not put your rail on painted base surface. You would use ground (and possibly scraped or epoxied) base without paint on it. But again, they equal out too on a machine of that size.

Can you see how you cast an aluminum portal? Greetings from Russia

I guess, the magic word is deflection here. if you have a body that long unsupported, it will hang through. and steel is rather one of the stffer materials already. though its also heavy. there are plenty of people who turn to rack and pinion solutions therefore. I have no idea if that threatens steel cutting procedures then, but cutting steel with a cnc that size does not really sound approachable and reasonable anyway. maybe at the front and the end of the screw. but in the middle...? even 12080 aluminium bars have a deflection in milimeter rangee only after 1 m. so rack and pinion is not so gravity dependent. and pulling the screw.... i do not know how. propably some freaks do that. but that needs forces actiing constantly.....

Hi friend... I don't know if you read these comments... I actually enjoyed your build.... although, I am a designer, and I want to put this into CAD... specifically SolidWorks V2022. And I have a few changes I want to make. I would enjoy to benefit from some of your experience... such as what size your ball screw is, and where you got your parts. What size Servo Motors, and where did you get your controllers and other electronics. I am a mechanical designer, not an electronics guy. That is why I am asking for such information. I hope you can help me. Have you put together a materials list and where to get the materials? Also some tips about setting up and software you are using for your setup. I do have access to CNC Mill, and a nice lathe and surface grinder and other equipment so I can do this project. So... if you would be inclined to assist me I would really appreciate it. If your willing to assist me... I am willing to provide you with my designs, and also the plans I make. P.S. setting up the electronics is going to be an issue for me. I can put together components to build computers... so I do have some experience with playing with electronics. But I have never worked with controllers and servos. I can be reached at... I am Leslieallen inventorshaven@proton.me Please contact me. Friend... if you have Signal Messenger it would be a big help, that way we can easily communicate using video or messaging. I prefer video actually. Anyway, write to us if your willing. Thank you. I will post this on each of your pages... so hopefully you will see it.

This doesn't have a matter, because other axis uses ball bearing blocks and they extinguishes this bending.

This is the worryest way to measure. The ball screw is hanging down from theyr one weight.. I build lathe's since 35 year's and the ball screws have specific geometry when you mount them. Sorry my English isn't good enough to tell you some of my experience

ball screw doesn't meant to be straight. it's not a structural element. ball screw must have precise pitch and be concentric by mounting surfaces.

Hello Dear Friend, we are original ball screw factory, can we have chance to cooperate? Thanks.