What do you mean?

 

Everything you say is true, but it is still a case of mistaking stiffness for strength. When bending/molding a material to permanently deform it you are pushing it beyond it's yield strength thus you see a clear difference between the alloys at condition O and T6. If you flexed the two alloys to just before the point of bending (of the 6061-O) they would deflect about the same amount, the 6061-T6 will require more force to bend because of the higher yield strength.

To try and visualise this, if you had two gantry plates (@6mm) clamped in a vise at the base, one 6061-O and one 6061-T6 and applied a load at the top (in a way to try and fold it over) the T6 part will undoubtedly require a higher force to bend (permanently deform) compared to the one at condition O, but, the force required to bend the part at Condition O would still be significantly higher that the force the OX would be able to exert on them. Also, upto the load where the 6061-O gives way, both materials would deflect the same amount. I think this is similar concept to what Rick 2.0 was saying above.

The short story is, use whatever you can get hold of but thickness should take priority over the material strength. A 6mm 5083 (or whatever) would perform better than a 5mm 6061-T6.

Chris, your formed parts look very nice, very impressive.

 

Sorry guys...i think it was my fault.

 

Ahh... the good old days. My very first machine shop class five years ago when they actually let me use the machines.
It was also my first "customer"... my next door neighbor was put in charge of a new boy scout troop and the boy scout store was out of stock of the tie clips, so he asked me to make some for him after I told him I was taking a machining class. I tied up one of their HAAS machines for 3 full days to machine the dies. Then I had to learn how to operate the 150 ton press to stamp the parts. I remember I had to go to about 140 tons to get a good impression with the 6061-T6 and the brass ones I also made. 6061-0 only took about 70 tons.

 

I stand corrected, thickness is more important the type of aluminum when it comes to rigidity of the OX plates, but if in a bout of frustration you decide to take a sledge hammer to your OX, it would survive a lot better if it was 6061-T6.

 

Ok, so question on the electronics if I may.....
I am just about ready to start getting the electronics portion of my OX going... I bought the gshield and the Seeduino v4 blue that was identified in the BOM. The boards I ordered (I had to order from different locations) did not come with any preinstalled software. I bought BOBCAM-CAD v26 PRO software and will have that running on the tower computer. So here are my questions:
1) what software will I need (and where can I get it - preferably already configured for the most part) for the gshield
2) what software will I need (and where can I get it - preferably already configured for the most part) for the Seeduino v4

This was taken a few days ago..I have since painted the table top, attached the gantry and z-axis. Still awaiting a few parts from OpenBuilds and casters (please disregard the garage mess)

I will upload a newer pic once I pull the photos from my phone....

 

Here's an idea for running cable chain along the y-axis, I'm going to try and do something similar on the x.

I used some standoffs and a bit of Aluminium angle to support the chain. I made a simple bracket to fix the chain to the Gantry plate.

Machine is coming together...Slowwwwly

 

No expert, and a Mach person besides but sounds like all you need is Grbl.

And my question:
Decided I wanted to build a router table, after much browsing and discovering this forum/OpenBuilds and decided an Ox would be a fun starting point, ordered the parts and am now more or less patiently (or not) waiting on a minor V-rail backup giving me lots of time to review video's, posts, etc...

Anyway so I started thinking about the Z-axis and wondering why a 4 start-8mm per revolution lead screw? Was it readily available/on-hand, or was there other reasoning for the choice?

Obviously from the videos it works, and with micro stepping 0.04mm per step should be plenty of resolution, but it seems a bit steep for such a short axis to me so I was wondering if anyone had any thoughts/experiences/ideas???

 

Cool. Thanks Philip UD. Is there a modified Grbl out there that is BASICALLY already set up for the OX? I know I will need to make some adjustments....but since I am new, the less I need to modify in the beginning the better.

Also, Will Arduino Sketch upload it?

 

Sorry that I don't know.

I use Mach3 for a retrofitted micro-mill pretty easy to drop code into from just about any source with only minor modifications when needed.
Not sure what I'll finally use for my router build, to start I'll just create a profile on Mach3 for the Ox and use the mills controller.

 

I'm running the latest version of GRBL from the source on the edge branch.
https://github.com/grbl/grbl/tree/edge
I've attached a compiled version from this commit
https://github.com/grbl/grbl/commit/f298cd925db3c8cce5c8a74fd6f26249773b422d

Here are the GRBL settings I use. You can (and should) adjust feed rates and max travel to your machine. To upload all settings, I connect to the serial port with Putty and paste them all at once.

Spoiler: Grbl 0.9 Settings

Modify RED values for your machine
$0=10 (step pulse, usec)
$1=25 (step idle delay, msec)
$2=0 (step port invert mask:00000000)
$3=0 (dir port invert mask:00000000)
$4=0 (step enable invert, bool)
$5=0 (limit pins invert, bool)
$6=1 (probe pin invert, bool)
$10=3 (status report mask:00000011)
$11=3.000 (junction deviation, mm)
$12=10.000 (arc tolerance, mm)
$13=0 (report inches, bool)
$14=1 (auto start, bool)
$20=1 (soft limits, bool)
$21=1 (hard limits, bool)
$22=1 (homing cycle, bool)
$23=0 (homing dir invert mask:00000000)
$24=20.000 (homing feed, mm/min)
$25=500.000 (homing seek, mm/min)
$26=224 (homing debounce, msec)
$27=1.000 (homing pull-off, mm)
$100=26.670 (x, step/mm)
$101=26.670 (y, step/mm)
$102=50.000 (z, step/mm)
$110=1000.000 (x max rate, mm/min)
$111=1000.000 (y max rate, mm/min)
$112=500.000 (z max rate, mm/min)
$120=100.000 (x accel, mm/sec^2)
$121=100.000 (y accel, mm/sec^2)
$122=30.000 (z accel, mm/sec^2)
$130=500.000 (x max travel, mm)
$131=750.000 (y max travel, mm)
$132=80.000 (z max travel, mm)

Special note on settings, I am using much higher feedrates, but you should TEST your machine to make sure it can move reliably at these speeds. Heat, belts, motors, bearings and many other things can have an impact on your speed. These speeds should also only be used for SEEKING. I really doubt you could cut anything going this fast, so if your CAM program doesn't have good speed control you should limit your GRBL speed for safety.

Spoiler: Danger Settings

$25=2250.000 (homing seek, mm/min)
$110=15000.000 (x max rate, mm/min)
$111=15000.000 (y max rate, mm/min)
$112=750.000 (z max rate, mm/min)
$120=125.000 (x accel, mm/sec^2)
$121=125.000 (y accel, mm/sec^2)
$122=30.000 (z accel, mm/sec^2)

For host software I'm using Universal GCode Sender 1.08 compiled with netbeans, as it works with Grbl 0.9. You could also try http://chilipeppr.com/, but I think it is still on GRBL 0.8.

I use Xloader to upload to my Arduino UNO. http://russemotto.com/xloader/
Modify your devices.txt to include UNO if that is what you are using.

Spoiler: XLoader devices.txt

Mega(ATMEGA1280);m1280;stk500;57600;
Duemilanove/Nano(ATmega328);m328p;stk500;57600;
Duemilanove/Nano(ATmega168);m168;stk500;19200;
Uno(ATmega328);m328p;stk500;115200;
Mega(ATMEGA2560);atmega2560;stk500v2;115200;

I haven't had any issues with line noise using Grbl 0.9. I'm using inverted logic, so the switches send a GND to the board when they trigger. The UNO has internal pull-up resistors and seems pretty stable. I have had issues with some cheap and or smashed switches giving false triggers on other machines in the past. Occasionally a wire can come lose due to vibration if they aren't connected very well, this can also give false triggers. I don't have an oscope hooked up to the wires, but I use this same setup on a couple RepRaps and they don't have any issues either.

If you are having issues, you could build filters and debounce circuitry. However, that seems overkill if you aren't having any problems.

One last note: If anyone is cutting ovals and you can't figure out why, you might check your setscrews and add some threadlock to them. They like to come loose on my machine.

 

Very impressive wood working to support your Ox.

 

Very impressive indeed. You have a nice working space for your Ox.

 

Thanks. here are two more pics. One shows the painted top and the Makita RT0701C mounted, the other is a side/rear shot showing my printed cable-chain and v-slot wire-holders. There will be another cable-chain (which I am currently printing) along the x-axis that will sit atop the silver v-slot. I will have to remove the power cord from the Makita and run it thru the cable-chain because the plug is too large to fit thru the slots. Oh, and the OX will be bolted to the table top.

One point of note: you notice (or maybe not) that the silver v-slot is only being held with one screw on each end, that is because if I rotate it 180 degrees, the nuts behind the rear z-plate actually set into the v-slot and I didn't like that idea. There isn't going to be any weight on the solver v-slot, so I figured the single screw on each end would suffice (and it is the perfect width for my cable-chain in the orientation that it is in). Also, the red brackets around the NEMA23s may or may not stay....they might be used to attach the cable-chain ends, but I am looking at another option...

I have 40 feet of stepper motor wire on its way, once that gets here, I will start my wiring task.

Rich

 

Im running a 1/2" lead screw for my z, modified some things to make the ox bit more stronger. Theres alot of force everywhere. I been having an issue with my X skipping steps, and the reason is the GT2 belt keeps slipping loose.. Does anyone have a solution or a trick to keep the belt tight?

 

This weekend I finally ( finally, when did I start this, February??!!) cut the ALU extrusions and set up my mechanical bits. I still have some more fabrication work to do for the X and Z axes but its almost time for electronics!
The table is a torsion box design. It is a bit bigger than 2'x4' so the final cut area will be something like 23"x36" (580x900). This configuration has 4 1/2" of Z travel plus space for a 1" spoilboard. It can be raised another 20mm by going up 1 notch on the angle brackets. I'm a little disappointed with those, they aren't square. I will probably end up bending them into shape.

 

There should be room for 3 X pieces. The plan is to put one in the middle and then put bolts through all 3 to clamp them together. This is the longest span in the machine so it needs to be super rigid.

Oh yeah, that 20mm! Moving up one slot on the Y rails.

 

@JustinTime the motor pulleys on my Y & (slaved) A dont slip at all and those are a 1000mm long. Im looking for a way to keep the belt taunt. Mach 3 gave me a warning saying that the feedrate is be too fast for the pulley. Alot of builders on youtube show that their machine can go "so fast" and its actually not under any load. There has to be some way to keep the two ends of the belt from slipping away from that "t nut".

 

A bigger pulley (with more teeth) will surely help, add a second t-nut so it will limit the slipping of the other one. I think Mach3 means spindle, not pulley.

 

Have been looking at the OX build for some time and would like to build a 2'x4' size one and wonder if anyone sells a kit with all the parts at once deal that way i don't have to buy part by part ?

Thanks

 

I did the same exercise. I took pictures that I magnified and printed. My backlash on the X axis seems to be 0.333 mm and 0.188 on the Y axis.





What could we do to improve this? This is a standard build with GT3 belts.

Frederic

 

Thanks a lot @Frederic Goddeeris

I was thinking it was some my mistake. I realized that different micro-stepping change the backlash. 1/32 and 1/16 I get about 0,2mm or more. Now, with 1/8, the max backlash I can mensure is 0.13mm.

Here's a table about the belts backlash: http://www.sdp-si.com/resources/imgs/belts/fig10.jpg

I will try a reduction system in next weeks and report here about. I don't want to use Mach3. It's the cost of software + PC with DB25. Or "high" cost of a solution to use Mach3 in USB computers. I think the GRBL + some mechanics solution will be cheaper.
I mean (sorry my english, guys):
Money to get Mach3 + breakout usb board + belts setup = Arduino + ballscrew setup.
Even the ballscrew solution would be more expensive, I prefer improve this backlash in mechanics, and keep with GRBL.

Also, I will figure out how to use this (link below) in Gcode that I need reduce backlash, like PCB.
https://github.com/Lenbok/NoLash