Jim's corexy

Building a 3d printer for my home. I have access to several at my brothers workplace so it hasn't been a priority to have one at home, until now. 400x450mm travel using 20mm nsk/THK linear rails. 80/20 1020 extrusion. I have built several CNC machines so no stranger to linear motion and automation. I can machine any parts on my lathe or mill that is required. My question is that I have this ballscrew driven 400mm travel THK linear actuator that I was thinking about using for the z axis. The actuator is laying on the extrusion pictured here.

[imgur.com]

Testing linear rail movement
[youtu.be]

The actuator is rather long and will make the printer around 750mm tall. Seems like there is not much liking to large cantilevered beds but this actuator is pretty rigid and strong. The bed will be made out of leftover 1020 extrusion topped with a aluminum build plate. Do you think this will be a problem???

I can also design the z axis using two 20mm THK rails and two kerk lead screws i.e. Lifting the bed on both sides of the build plate. That was my initial design until I found this linear actuator for cheap on ebay. The linear rails are shorter so I will have much less z travel, probably about 200mm

Pros/cons. Thanks...

Edited 1 time(s). Last edit at 02/22/2017 06:53AM by Jim Fong.

It might sound like a sacrilege, but is it possible to shorten the beam of the actuator to a convenient length?

The biggest problem with cantilevered beds is that the edge opposite the side with the guide rails tends to bounce when the XY mechanism is flying around causing the machine to shake. Flexibility of the guide rails, the printer's frame, and the cantilever itself all contribute. If you put a lead screw or two under that floppy edge the bouncing can be almost completely eliminated.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

DD,
I have to defer to your experience than. The single linear actuator would of made it really simple, just mount it to a vertical piece of extrusion and you're basicly done. I now have to machine two identical linear actuators using the shorter linear rails and antibacklash kerk lead screws. It's going to take longer now, plus ordering couple more pulleys/belts to drive it with a single motor. Similar to your printer but with screws instead.

The ballscrew actuator had essentially zero backlash. Atleast nothing measurable with my Browne sharp test indicator.

I wouldn't want to cut down the linear actuator since I could sell it for more than I paid for......or use it for some other project instead. They are very very expensive new. The spec sheet says it rated for 800mm/sec travel speed!!

Edited 1 time(s). Last edit at 02/22/2017 11:42AM by Jim Fong.

You shouldn't have to worry about backlash at all in the Z axis- the weight of the bed and it's support will keep the screws and nuts in contact at all times.

The belt driven Z axis I built uses two lifting points and is an experiment. I've done two major redesigns and now feel like it's ready for print testing.

I think that it would be better to have three lifting points to minimize the possibility of the bed wobbling as is moves down the Z axis, but for now, using two is a lower cost and mechanically simpler approach. If I see any hint of print quality problems in the Z axis, I may add a third lifting point or maybe even scrap the whole belt drive system completely. I have not fully priced it out but I don't think my Z axis design is any cheaper than using two screws to lift the bed with a single motor. I used a pretty expensive motor gear-box combo for which there may be lower cost substitutes.

I think that three screws driven by a single motor with 2 rails to control lateral motion is probably optimal in terms of performance.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Three lifting points would be nice but I don't think I can find another matching Kerk leadscrew cheaply. I bought both for $30 or so a long time ago. They are pretty expensive new,

Hopefully using profile linear rails will keep it from wobbling if properly mounted. I actually have a choice between 15mm wide Star/Rexroth linear rails which are longer than the 20mm THK SR rails. The Star/Rexroth have very very high preload bearing carriages. They will not fall down if held vertically. They run very smooth but requires much more force to get them to move. They were sent to me by mistake and the seller told me to keep them instead of returning so they were free. About 350mm long. I think the high preload may cause problems.


Regarding ballscews, I just like using them. My CNC gantry and lathe are using ground ballscrews and retrofitting my bench mill with them too. The repeatability is nice.
[youtu.be]

Edited 1 time(s). Last edit at 02/22/2017 02:35PM by Jim Fong.

The high preload linear guides could cause backlash problems if the weight of the bed and support aren't sufficient to overcome their stiction.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Stiction. That's the term I was trying to remember. i think it would be a issue.

Since your linear actuator is so long,
a long triangular brace below the table
may take the jiggles out of the front of bed
(no longer a cantalever but a frame braced)

Sacrifice printing height
and printer base taller to accommodate brace

[youtu.be]

First movement, about 600mm/sec.

Using grbl1.1e to test until I figure out how to configure the cohesion3d smoothie board for corexy.

This is the max speed steps per second for grbl and Arduino uno board.

Edit.

4000mm/sec^2 accell
10microstep
24volt PS
Arduino UNO
Grbl 1.1e

Edited 1 time(s). Last edit at 02/23/2017 04:19PM by Jim Fong.

When traveling parallel to the axes, both motors are contributing torque to the motion. If you really want to test speed, move at 45 or 135 degrees when only one motor is providing all torque. That is going to be the limiting factor when you're actually printing.

Over the length you are moving, and with that acceleration, does the extruder carriage ever actually reach 600 mm/sec?

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

It does 45 and 135 at the same speed too. Hard to say if it is actually reaching those speeds, I did double the acceleration and it looks the same. I may be hitting a ceiling with the grbl software and it's step pulse output capability. Smoothie firmware will do 99khz step pulse rate vs 30k that grbl does. I'll have to install the Cohesion3d board later and do some more bench testing. Really this is for fun and games, since printing speeds are under 150mm/sec anyway.

These motors will spin several thousand rpm with pretty decent torque. Applied Motion makes pretty good stuff. The drivers are programmable with a bunch of settings that can be tuned to make them faster/Better. I haven't done the tuning yet.

[youtu.be]

Another video, this is a little faster. Ran the software config and changed some settings. What a pain that was. None of my laptops had a DB9 serial. Had to dig out a old desktop to connect and run the config program. Found out the programmed current was set at 60%. Bump it up to max. This gave the motors higher acceleration without stalling.

Grbl1.1e
40,000mm/min (666mm/sec) speed
10,000mm/sec^2 acceleration
10 microstep (2000steps per rev)
45volt power supply
Applied Motion STM17s motors

If I set the speed any higher, grbl locks up. It is max step pulse limited.

I was looking at the video and noticed when the X axis moves toward the right, the lower belt segment between the right corner pulley and the pulley on the end of the X axis does not appear to stay parallel to the Y axis guide rail. If that is actually happening, the belt tension is going to vary with the extruder carriage XY position, and the prints will be distorted.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Ya I know, I messed up the when I CNC machined the part. Wrong measurement. Easy fix to move the pulley over more when I get a chance. The plates need to come off and have the corner chopped off too.

Edited 1 time(s). Last edit at 02/24/2017 07:28AM by Jim Fong.

Bad news My house took a direct lightening strike today. Multiple computers dead, power supplies, multiple stepper and servo drives for the all the cnc's all dead. Even the 3d printer electronics is dead. Stuff plugged into network like hubs and raspberry PI's all got zapped. Cable modem dead, furnace heater dead. I don't even know where to begin. Atleast the house is still standing. Oh yea even my 40 watt CO2 laser cutter died.

It will be awhile before I get back to the 3d printer.

Well atleast all the large woodworking power tools are good.

Ouch! That sucks! I hope you get it all sorted out.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Just a little overwhelmed. Sitting in my upstairs electronics lab checking things over. Most of the lab equipment turns on.

The lightening hit the rooftop furnace exhaust stack which is metal. Blew that to pieces. The energy traveled down to the basement furnace room area which happens to be where all my CNC machines/workshop lives. Most everything in that room is dead. Even some stuff that wasn't plugged in. Must of been a huge EMP like surge. A lot of stuff i haven't even looked at yet. I hope the dozen brushless servo and stepper drives sitting on the shelf survived. Lots of expensive stuff was in that room. There is a smell of burnt electronics down there.

It's been a couple of weeks since the lightening strike. Still dealing with insurance but had some time to put in new stepper motors and drivers on the corexy frame. Haven't done much more than that yet.


[youtu.be]


Another video moving a 635gram steel 123 block around.

[youtu.be]

Edited 1 time(s). Last edit at 03/21/2017 01:09AM by Jim Fong.

Glad to see you are up and running, although I hate to point it out but your belts arn't parallel.... The belt from the Y axis to the motor/idler needs to be straight, otherwise the belt tension will change with the Y axis movement and cause distorted prints. Only the stationary belts (eg, from the motor to the upper idlers) can be non-parallel, although they must still be symmetrical with the other belt.

It was already pointed out to me, mistake when I measured wrong. It's a easy fix by moving the position of the bearings but I haven't done it yet. I'm just glad to get it working again. Corner brackets need to be cut down. It will be done the next time I remove them.

CNC is still down. I sent all the bad servo drives back to Geckodrives for repair. Any work on the 3d printer we have to wait until the mill is up and running again.

Edited 1 time(s). Last edit at 03/21/2017 04:13PM by Jim Fong.