I am retrofitting a Chinese Ebay 1m x 1m CNC I obtained from an older gentleman in June who wasn't interested in the CAD part of CNC work and passed it onto me. I will be exclusively doing very lengthy (24-36+ hour) finely detailed 3D finishing cuts in various wood species. At this point six months after I started it's basically a Lead 1010 as I have replaced just about everything with Open Builds components following their excellent build instructions. Notable outstanding differences are: The tool is a 80mm dia. 2.2kw ER20 liquid cooled 24,000 RPM spindle, controlled with a VFD (see this thread) I have a custom high-z configuration to accommodate the spindle on a C-Beam for max Z clearance and strength. The machine is built on a pallet rack. Here are a few photos: I have a few open questions that I am hoping the experts here can help with. Last night I finished the build to the point where I was able to successfully cut my first piece. I then started a longer one and noticed some issues that I would appreciate some help with. Here are my current issues: Issue 1: Whip on Open Builds 8mm Lead Screw. See this video: iCloud My X axis is based after the OB High-Z mod. I have two C-beams, with one Open Builds NEMA-23 high torque motor on the lower beam. I have two nut blocks on the lower gantry cart with the screw. There is a third cart that is attached to the two "travelling" carts at the front, and another cheese plate that attaches them together at the back. The spindle is attached to a 750mm C-Beam that travels up/down in Z. This gives me maximum clearance that I need. Everything has been squared and trammed to the best of my ability. I am getting bad whip on the X direction unless I loosen and move the lock collars in about 0.5mm from both the motor and the end plate, and leave the nut blocks on the gantry cart slightly loose. This allows the bearings at each end to "slop" move in 0.5mm, but the lock collar isn't pushing the bearing inwards when the cart is travelling in that direction. This seems counter intuitive to have sloppy bearings. But it eliminates the whip. Should I: A) Leave the lock collars loose and continue to allow the bearing to travel inwards? B) Switch out the nut blocks with an anti-backlash nut in the middle of the gantry? C) Add a tension nut on the outside of the end-plate and pull tension on the screw? Here are the other issues I need to work out, but I will save those questions for a later post. XYZ Probe is always grounded (green light on) when the magnet is on the spindle nut. Need a reliable way to re-start a lengthy (24H+) 3D cut file after a stoppage as I am able to do in both ShopBot and Mach3. Can't seem to get the mobile Jog to work, it hangs on "Waiting for USB" on both Safari and Chrome Thank you.
Switch to a tensioned leadscrew setup: C-Beam® Tension XL Linear Actuator Bundle as inspiration of the basic assembly. With that big a spindle, it would just be a better option You need - Different end plates - thrust bearings - flange bearings - tension nut Also check that the screw is straight - if it picked up a bend you can try straigtening it, or replacing it. AC Earth connected to DC GND somewhere = bad. See XYZ Probe Plus with VFDs and other ground-loop inducing components Hand-edit your gcode to ensure all modals are in place and you wrote good Entry moves to prevent crashes. As per the thread you commented on earlier - I had a longer writeup in there. Automation cannot ensure safety so we are putting the risk on you Works fine for me - make sure the computer running CONTROL has firewalls open, and nothing else like that (malware scanners etc) blocking delivery of the javascript files used by /jog. Make sure your CONTROL is up to date 1.0.332 at the time of this reply (Release v1.0.332 · OpenBuilds/OpenBuilds-CONTROL) (.335 in my GIF below, internal Beta but no changes to Jog widget so still valid for test)
Cheers for this Peter. Regarding the tensioned leadscrew setup - I already have all of those parts on hand. So the basic idea is just to add a tension nut outboard of the plate, which I replace with another C-Beam motor mount plate with a thrust bearing at the opposite end of the motor, right? Edit: Regarding the ground issue - the XYX Probe is lighting up on the collet, and on the body of the Y axis stepper motors. Here is what I tried unplugging: the spindle unplugged, with the VFD unplugged from the mains. I also pulled out the iOT connection to the power strip The VFD unplugged from TOOLHEAD Edit2: Removing the Z limit switch wire turns the LED off on the probe. So now I have wires to follow.
Process of elimination. Unplug PSU, unplug USB (some PCs have bad PSUs with leaky DC GND to AC EARTH), if its not that, unplug limits (forgot rubber spacer?), etc onwards and etc until the LEDs go out. Something is the guilty culprit The stackup is : Tension Nut > Thrust Bearing > Flanged Bearing 688ZZ 8x16x5 > C-Beam Motor Mount Plate > C-Beam with leadscrew > C-Beam Motor Mount Plate > Flanged Bearing 688ZZ 8x16x5 > Thrust Bearing > 1/4" x 8mm Flexible Coupling See video below from 12min30 onward to about 17min. Also the tensioning process around 19min10
Sweet! Glad you found it. Switch mounted upside down perhaps? The bottom pad is connected to GND, so flipping it around isolates it (screw head on top plate, rubber washer under GND pad - hole edge is isolated from screw shoulders)
Both Z switches are right-side up, with spacers, and all wiring is right. If I unplug the lower one, the Spindle isn't grounded. Could the switch be bad? (NB - this setup isn't permanent - I plan to eventually move the lower switch to the gantry plate).
Adding a rubber washer under the head of the screw securing the limit switch to the c-beam eliminated the issue. Question: The LED is still green if the magnet is touched to any of the stepper motor chasses. Is this still a problem?
Both are upside down in the picture This is right way up: With the nice logo plate facing you: Where the component pins come through is the Bottom, and where the rubber spacer should go Checkout :
Normally stepper motor bodies aren't grounded (by their wiring, as that just goes to the coils) so, check other limits too - perhaps more mounted incorrectly, GNDing the extrusion, then the motor via its connection to the extrusion. Touching the extrusion won't show as anodized (aluminum oxide) is an insulator, but tapped holes, scraped anadozing, etc can complete the circuit
Looking around, sure enough I do have some of the backwards (solder "pyramids" in towards the extrusion.). Should be an easy fix when I disassemble and re-do the X axis beam.
That's right way up. Solder pads under the switch, spaced away from plate/extrusion by the rubber spacer: The others are wrong
Question for you Peter - why does this config use a 4-wheel, one-sided XL Gantry Cart instead of a two-sided six wheel "sandwich?" What is the difference between the two in terms of strength and rigidity?
As its a 100% modular system, it doesn't have right or wrong answer. Economics might play into it (C-Beam plate costs more than wheels). 'Sandwich' looks stronger on paper, but there's no way you are deflecting a wheel on a spacer in a one-sided config either (you'll twist or deflect the extrusion itself long before the wheel itself moves enough). 3 vs 4 wheels - as you move along the length of the extrusion there is some variance - four wheels allows it to ride any low spots a little better
Thanks Peter. I have one more major R&R operation - swap out the metal-wheel gantry carts and 10mm lead screws along the side Y-axis C-Beams with OB components, and then my little machine will be ready for action. Not sure why this design was implemented by whomever made this machine, but without eccentrics it's impossible to tighten and there is a great deal of slop in it.
Peter – when I rebuild, this X axis c-beam with the tension nuts and thrust bearings, is it better to use the two Nut blocks or the one anti-backlash nit block?
The two-nutblocks strategy further reduces whip as you hold the leadscrew at two points some distance apart - we prefer that for longer screws. On a Z axis that's short we'd prefer the Anti Backlash nut (single one)
