Acorn6!!!???

[tblough]

And, scales dither. So, at best, you'll get 2X the scale resolution amount of dither. If you have 5 micron scales, your machine will be CONSTANTLY moving +- 5 microns around your commanded position. It does affect surface finish on a mill (I can easily see a 0.0005" step) and to a lesser extent on a lathe.

To minimize this, you really need 0.1 micron scales which cost over $1K per scale. Shoose your poision carefully.

[Gary Campbell]

Setting the display abilities aside, providing motion from encoders or scales is a product of motion, full feedback controllers. Aren't position (step & direction) controllers (by definition) not capable of this feature? You are "wish listing" Oak features onto the AcornSIX.

[Muzzer]

There's no reason why a step/dir interface isn't capable of being used in a closed loop system - that's what I've got on my machine. Most (all!) systems are digital at their core, so they work with discrete increments of distance. Going to analog controls (+/-10V etc) doesn't actually help anything in terms of functionality and if anything introduces noise.

As for "constant dithering", the solution that is usually employed is to incorporate a deadband that is similar to the minimum resolution. So my machine has 5um scales and after taking account of the deadband, we can get at best +/-10um or so. My machine is perhaps "past its best", so attempting to improve beyond that would be trying to slap lipstick on a pig.

Yes, we are indeed wishlisting some of the Oak features but the Centroid team are already some way down that road and I doubt most Acorn users would be in the market for an Oak. I expect they have a product roadmap and it possibly accounts for the fact that the Oak has been around a few years now, while technology marches on. There are clearly more innovative and exciting products coming our way!

[suntravel]

scales for high precision 3µm have a resolution of 0.001µm

https://www.heidenhain.com/fileadmin/pd ... 049_en.pdf

and if your pockets are deep enough for them you don´t care about the pricetag for an Oak system

On a DIY lathe I am sure on X you loose precision with cheap 5mµ resolution scales and cheap ballscrews compared to a well build CL-Stepper or Servo system with Acorn and preloaded HiWin ballscrews.

Uwe

[Muzzer]

Yes, it's about making coherent choices - no point overspecifying components in one area and cutting corners in another. On my machine I fitted NOS Korean ground preloaded ballscrews for both axes and (closed loop) AC servos, then spent ages refitting and adjusting the various gibs to minimise slop and backlash. However, it's still a worn out old machine that wasn't designed for CNC control. Having done all that, I then had a serious accident involving a mouse and a crate of beer, so now I have a pukka CNC lathe that is full of HSK bearings and ballscrews.

I have no means of checking the absolute accuracy of the Chinese glass scale used on the Z axis but LinuxCNC allows you to examine the error between the expected and actual position, which reported pretty good correlation. For the X axis I have a miniature European scale that may or may not be better but also appears to give good results.

Must admit, I don't understand the datasheet you linked - it talks of accuracy of "+/-3um in 3040mm" but a measuring step of "0.001um for +/-3um". I guess you may argue they are of a very similar order (~0.1%) but I will never have the means to know what I am dealing with!

[rk9268vc]

and if your pockets are deep enough for them you don´t care about the pricetag for an Oak system

Can OAK close loop between servos and scales?
if so, that would be nice for converting a manual bridgeport knee mill for 2.5 axis CNC using the stock leadscrews.
You would still have to mind the backlash for finish passes, but for position that would be nice

[tblough]

You can use scales with both Oak and AllIn1DC but you're wasting time and money doing it with leadscrews.