Possibly dumb question... upgrades


(Luke) #1

The mrs is out and I’ve been racking my brains trying to consider more modifications to do my shapeoko… I know I said I was done but no one really believed that…

For me the logical next stage is improve the X/Y axis. The problem I’m facing is that unless I just chuck in the X/Y there isn’t really a way to improve them, I don’t want to build a whole new machine if I can help it. Then I got thinking about where the weak point as I see it is and it’s the belts. It would be great to swap these to ball screw, but if you do that you might as well do the whole machine.

Then I really thought about it, I’ve seen cnc machines use rack-and-pinion style axis. Why not just lob on some of that? It’s essentially a reverse of the belt method.

I can’t think why it wouldn’t work, but it seems too simple…


(William Adams) #2

It was explored a bit in the SO1/2 days — and one clone machine used rack and pinion. It works, the problem is backlash, and the expense of the rack.


(Jose Prieto) #3

Belts . In the market I have seen belts with much more stability .


(William Adams) #4

Yes, but all the stronger belts have much greater backlash.


(Jose Prieto) #5

Only the ball screws remain !.
Like Z axis to find the stability .
Only the ball screws to look for stability in the measurements !.


(Pierre LaFromboise) #6

Forget the x/y upgrade, add a fourth axis instead :smiling_imp:


(Luke) #7

Thanks, as always I’m staggered by your knowledge


(Luke) #8

I’ve been looking into this and the cheapest solution is to buy a 6040 machine and implant a shapeoko control/brain.

to build a rail kit is around $400, the aluminium would be in excess of 200, bearings time etc will put it all at over $100.

I wonder what the maximum size motor a shapeoko would drive would be?


(Adam X) #9

The belt and pulley system driving the X/Y is really very good. I’m not sure how much improvement there is going to be in that (I’ve expressed this before).
If you’re going to put your brain at something, how about an improvement/replacement for the vwheel guide system? I’d wager there is more slop caused by those than there is in belt-related foo. I’d thought about machining some AL vwheels, or a conversion to linear-guide rails which could be done, though it’d require a full disassembly of the machine.

Also, I’ve been sketching through ways to reduce the entire Z/X carriage’s twist along the Y/Z axis’… Possible horizontal extensions, extending behind gantry and pressing bearing against back of the rail to resist the twist incurred when making heavy cuts along the Y direction. Going to integrate this when I build a new Z axis.

The shapeoko’s Brain is just an arduino stamped onto a 4axis stepper control board (essentially). GRBL is great, but it’s inferior to a true parallel-driven motor control system (think mach3, linuxcnc, etc). GRBL can only move in increments, there is no true jogging. This is by virtue of how a usb connection works. Another mod on my stack of things is to swap out the C3D grbl board for a full blown 4 axis stepper controller and a PC running LinuxCNC to drive it.


(Ben) #10

So is a 4’ expansion out of the question? :wink:


(Jose Prieto) #11

I believe that it is important to have the prescision that we require in our work, if a person who works in large wooden objects does not require it, but a manufacturer of machines or a jeweler if they require it. Without offending anyone personally I like a more stable and precise machine because of the type of work I do.
I thing ball screw is a good alternative . The belts in four months they vary and you have to adjust again, without checking every day of work and not being sure of their behavior.
I have also thought about the lack of the 4 axis and putting a new brain, but little by little this opportunity that Carbide 3D has given us is very interesting and at a reasonable cost.


(William Adams) #12

My thoughts on low-hanging fruit for hobbycnc upgrades:

  • the SO3 Z-axis is obvious to replace and folks have been very successful with it
  • some sort of indexed 4th axis (I’ve got a sketch of a design and have posted a link on it in the past)

Everything else gets pricey / invasive and involves replacing enough structure or electronics that it may be cost prohibitive:

  • larger structure — there have been statements in the help support e-mails that Carbide 3D will work up a larger design, eventually, but no idea when that will be (I really want a 3’ or better still 4’ X-axis and that seems the likeliest thing to get done first)
  • different drive mechanism — lots of tradeoffs here, I agree w/ @Adam_Xett that belts are likely the sweet spot for low-end hobby machines
  • positional encoding — this entails replacing the electronics, but could pay off handsomely — certainly it worked well for the now discontinued LoboCNC

(Adam X) #13

Absolute positioning… now there’s an idea… I bet one could adapt the cheap-ish optical scales sold in DRO kits for metal lathes. Something like… http://r.ebay.com/WIcVft

Now interfacing any output from that to a machine controller is going to go beyond grbl (I think), but might be interesting as a manual check/reset/offset utility.


(Luke) #14

It’s a fair point, I mean the belts are good, and if you want MM precision I have no worries there, my wood projects are on point. The parts I’m making are accurate to 0.1mm but my repeatability is far from that when re-starting a job.

If you were to improve the V wheels/guides then I think you’d be better off doing both guides and drive at the same time. I know Openbuilds do a ball screw and v wheel combination all in one unit which looks interesting. If I were to do it I do think you’d be better off doing both, I thought the gearing would be a good half way house but @WillAdams makes some great points.

I too thought about the twist, but you’d need whole new wrap around section and it could get messy, I think replacing the X for a custom setup would be easier with guide rails and ball drive.

The thing for me is a nomad is incredibly accurate, repeatable and gives a great finish (on metal), something I think could be improved on the S3. I have had fantastic results and the S3 has exceeded my expectation, but I’d love to see what a ball drive setup with linear guides would do. I Nomad on steroids so to speak.

I love my new Z - it’s so much more repeatable and dangerous than the belt setup - there is no more jamming or sticking which was causing me many problems.

I also have no desire any more for a bigger machine, I looking at 99% of my projects they will always be less than 50x50cm, and 90% smaller than 10 x 10cm.

What I’m keen to do is have a really solid structure I can build from and upgrade as feel I need to. This is why I’m considering giving the shapeoko brains and guts to the bare bones 6040 machine, it has the size I need but a more rigid structure and it way cheaper than building my own machine from scratch.

Ages ago, Carbide 3d sent out a survey to ask what they should work on, do we know if they looked at pumping out a new machine? I feel it’s time for a Shapeoko 4 - a DIY kit we can build, upgrade and work with but driven by ball screw and guides and maybe a slightly bigger work area?


(Adam X) #15

I looked hard at a 6040 as a small machine to supplement my xxl, but ended up turning away from them for a few reasons:
First, on every 6040 variant I could find the Z carriage rides on unsupported guide rods, and there are numerous videos/reports of these having bad sag/twist in them. Most people replace them with tower-support ball guides.
Second, most folks have serious complaints about the quality of the steppers and their wiring - basically that it all needs to be replaced to get a reliable machine.

AliX is always getting new variants of the 3020/6040 kits in, hopefully somebody will move to the supported rails on the X gantry, that’d make it a contender for a very stout machine, IMHO.


(Luke) #16

I suppose I’m just spitballing on the 6040 - I have seen reviews and they look pretty shocking and as you say lots of variations out there with unsupported rails, different Z axis etc. If you watch a twist video it’s poor form.

There is a company I’m talking to in china that is looking at square rails but I think they are a while away from having something.

The thing with china is there is limited support and quality control issues all over the shop which is why I’d be keen to see what Carbide 3d can do for me/us.


(Jose Prieto) #17

Luke Stepcrafts is a machine like ours , this is the screw and not have screw balls nut , only a bronce nut control , Y axis . Whit a balls nut maybe go better !.

Anti backlash nut.


(Luke) #18

I can’t say I’ve heard of them but interesting. It would be cool to see how rigid their machines are…

They look a little delicate, small motors, 8mm lead screws.


(Jose Prieto) #19

But in a S3 you can work whit out limits , we have more space .


(Jose Prieto) #20