Rolled Screws with scale feedback - Acorn6

[travasky]

I'm designing a new build. Besides the spindle the single most expensive bought component is the ball screws. Initially I've planned on buying C3 ground screws but then the scout encoder board for Acorn6 caught my eye. Are there any downsides to using a C5 or even C7 rolled screw, for instance, in combination with a linear magnetic encoder like a magnescale to compensate for any lead error in the screw? It seems to me that such a combination is a win/win. I don't believe that rolled screws have any more intrinsic backlash than a ground screw. So tell me what I'm missing!

[johannes]

If you go this route, you need the most expensive Acorn software license to make this work. So instead of Acornsix "Pro" at $230, you need "Ultimate Plus" at $1400. Depends on the price of your ballscrews if it is worth it only for that.

Screenshot 2024-09-02 at 19.08.36.png (8.89 KiB) Viewed 527 times

[travasky]

This machine will eventually have a trunnion so I had baked the price of Ultimate Plus in already.

[suntravel]

Best way to eliminate backlash on the cheap is a spring loaded 2nd nut.

Scales will work only on one of the paired axes if you are building a router,

Glass scales have less temperature deviation than magnetic scales on a piece of aluminum.

Uwe

[travasky]

I’m not worried about backlash as much as lead deviation which is an inherent property of any rolled screw. The scale tapes would be directly mounted to the epoxy granite frame so temp stability should be pretty good.

[Houseman303]

hiwin rolled ball screw are very good quality

[Muzzer]

I’m not worried about backlash as much as lead deviation which is an inherent property of any rolled screw.

It's tempting to think that "backlash compensation" can make up for ballscrew backlash. If only life were that simple. Imagine you are driving a car with half a turn of lost movement in the steering wheel over an undulating, twisty track. Knowing the backlash (half a turn) is only of limited benefit, particularly if you are unable to see the road ahead. Backlash comp doesn't account for the twists and turns of the track pulling at the road wheels - it only "adds" the lost movement on top of the desired position when the direction of movement (made by your hand on the wheel) changes. In some conditions, it can actually double the backlash error.

You are more likely to have a happy result if you minimise the backlash in the first place, rather than expect some magic software to fix it for you.

As Uwe says, you can use a double ballnut to largely eliminate it - aka "zero backlash" ballnuts. These are often designated DFU (double) rather than SFU (single). If you are on a budget, you may find new old stock ground (high spec) ballscrews of high quality that you can machine to your requirement at a similar price to the generic Chinese rolled ballscrews. I went this route on my manual lathe conversion.

Here's DY Global's Industrial Parts Shop that I have used on several occasions https://www.ebay.co.uk/str/industrialpartsshop
Also FA Parts if you are after good quality, used electronics such as servo drives and motors https://www.ebay.co.uk/str/faparts

[tblough]

This isn't about backlash. It's about precision. Rolled thread ballscrews have higher pitch deviations than ground ones. Can scales compensate for this? Sure. But, by the time you add in software upgrade and the cost of the scales, it probably would have been cheaper to get the ground screws.