DIY CNC Design. criticism wanted.

[joel0407]

Second advise taken.

Someone suggested the motors may get too hot being mounted directly to the wood. I've also taken this on board. I would prefer the motors to be mounted the other way around and drive the ball screws via a toothed belt. I'm thinking from their advise I'll mount the motors on an aluminum plate when I get to this stage of evolution of the machine. I am not doing it straight up as I want to see how the motors perform and then I can tune the speed and torque by the ratio of the pulleys.

Happy Days.

[joel0407]

The advise I'm choking on the the twist thing. I thought I'd made a fair effort at mitigating twist.

I think I'm getting to the point though that my expectation is less than others that have suggested it will be a problem.

Happy Days

[TumbiUmbiTek]

Hi joel0407.
Have you your motors/electronics up and running yet? I gather you bought to 5 axis brrakout board from steppersonline? Any good?
Cheers

Sent from my LG-H870DS using Tapatalk

[joel0407]

Hi joel0407.
Have you your motors/electronics up and running yet? I gather you bought to 5 axis brrakout board from steppersonline? Any good?
Cheers

Sent from my LG-H870DS using Tapatalk

Nah, Mate. Motors should have arrived Friday. Toll stuffed it up and made bull **** excuse for not delivering. They were sorted for delivery on Thursday so I made sure I was around to take delivery. Waited again on Friday until I finally called about midday. They read the address to me fine and said they would investigate. Then they called at 4:30pm (too late to collect from depot) and asked to confirm my address. They said a unit number and wrong spelling of my street. Funny thing was it was already confirmed by the first call. Now I wont be able to pick them up until Tuesday.

To be honest, I'm not overly worried. My Drivers wont arrive for another couple of week coming from OS. It's more the principle that they stuffed up and tried to cover it.

Yes, 5 axis break out. Not from Stepper but eBay special. It was only something like $15. I figure it's not too much to loose if it's too slow or I want to change to a USB or Ethernet type later on.

Don't worry. I'll be posting updates as it comes along. I'm really hoping I can prove these guys wrong about the twist thing. I was watching a home build machine on facebook only yesterday. Very similar size to mine but only supported rails on the X axis. Y and Z were unsupported. With unsupported rails, there is basically nothing but rods to stop twist. Something I doubt they do at all. This machine was hammering along at 140IPM. Actually working at 140IPM, not just rapid.

Happy Days.

[TumbiUmbiTek]

Yep..always waiting for bits. I'll be making first gantry from ply (till i can cut in alloy)... Unfortunately I will not know till I have a go about how everything will...thats half the fun I spose!

Sent from my LG-H870DS using Tapatalk

[ger21]

I'm really hoping I can prove these guys wrong about the twist thing.

It's all about your expectations. If it's good enough for you, then that's all that matters.

[joel0407]

It's all about your expectations. If it's good enough for you, then that's all that matters.

Dead right there mate.

Happy Days

[joel0407]

Yep..always waiting for bits. I'll be making first gantry from ply (till i can cut in alloy)... Unfortunately I will not know till I have a go about how everything will...thats half the fun I spose!

Sent from my LG-H870DS using Tapatalk

I guess if some alloy comes along then I may just upgrade later but I'm sure I'll be happy with ply in the beginning.

Happy Days

[louieatienza]

Joel, it's not about proving anyone right or wrong, it's about making a machine that you use to create what you envision and design. You asked for criticism, not affirmation.

As to expectations, they should always be tempered. More often than not, our expectations don't fall in line with the choices we make, and most of these choices are made due to economy, availability, or workability...

[joel0407]

Joel, it's not about proving anyone right or wrong, it's about making a machine that you use to create what you envision and design. You asked for criticism, not affirmation.

As to expectations, they should always be tempered. More often than not, our expectations don't fall in line with the choices we make, and most of these choices are made due to economy, availability, or workability...

Sorry mate. Maybe a bad choice of word on my part.

[TumbiUmbiTek]

Hay Joel0407... any news, did you get your motors going?

[joel0407]

Hay Joel0407... any news, did you get your motors going?

Hey mate,

Not yet. I'm still working on it every chance I get.

I fully built the base to make sure it all fitted and all the drilling and cutting was complete. Now I have pulled it apart and put a few coats of polyurethane. I've done 3 coats now but it's slow since I coat one side, wait 2 hours, turn and paint to other. Next day, sand, paint one side, repeat. I have my drivers, power supplies, motors and a parallel port break out board. I have some short temporary cables for bench testing. I'll should have the proximity sensors and limit switches tomorrow or Friday. E-Stop button turned up this morning.

My brain is hurting trying to understand LinuxCNC at the moment. I don't know what you know about it but 2.8 uses joints linked to axis (roughly). As I have 2 motors on 1 axis, it's complicated. Once working right, it will automatically square the gantry. I have 2 motors, one either side of the gantry. If one motor was to loose step count either by missing a step or being moved by hand, it can just be homed. Once one motor hits home, it will stop and the other motor will continue until it too hits home then both motors will reverse past home until the home switch releases and then they should be square. If your home sensors are out of alignment then you can further use software to advance one side more than the other after the home switch releases. I think part of the problem is that 2.8 isn't official yet, just pre release. Some of the documentation needs some work, especially the parallel port stuff.

Happy Days.

[TumbiUmbiTek]

I've gone Arduino and GRBL for first build. Design is C Beam style with moving cutting bed and fixed gantry.(3xNema23 all up)..that way I can make the gantry strong and hopfully cut out the aluminium plates for 2nd build. I went Arduino with CNC shield cause I didn't want to go down the software rabbit hole and loose my drive to make the first machine (I don't have bugger all spare time). I'll use "x carve's" "Easel" first, while learning Fusion 360. I want to graduate to a much bigger machine but will let my mind create the paths of understanding with these first cnc baby steps. Ha!
Cheers

Sent from my LG-H870DS using Tapatalk

[TumbiUmbiTek]

CNC NUTZ, yotube says some interesting things about limit switches. On our "mat cutter" at work (cuts cardboard mats, belt driven cnc) we bump the gantry up hard against the "fixed metal" end stops to square the machine almost everytime we use it.

Sent from my LG-H870DS using Tapatalk

[joel0407]

Well I got my first stepper motor to run from my computer late Thursday night. LinuxCNC will run a Axis test during setup. Entering steps/rev, drive ratio and screw lead, you can just simply test speed until it stalls or missus steps. Holding the motor in my hand got pretty intense but I managed up to get it to stall at 400mm/second with 1500mm/second acceleration. My 16mm ball screws wont handle anything close to that so it's good to know that neither my motors, drivers or computer wont be the limiting factors in my gantry speed. It appears I might be at a bit of over kill actually. LOL I was running effectively 15mm back and forwards.

Happy Days

[rodw]

Joel, you won't have any reliability issues with LinuxCNC Master branch. Stick with it and it will all be worth it in the end. Good to see your machine is taking shape.

Here is a bit of inspiration for you from another Aussie with a gantry running LinuxCNC Master branch. First cuts today after 12 months of effort!



Not perfect first go, but I'm sure after a bit of tweaking it will work perfectly. This part was originally cut out of 10mm plate by a laser cutter and after a bit of hand coding, I got it pretty much exactly the same size.

I had a late regression with my home switches and I sorted it after slowing the homing velocities right down. One of these days I'll do a video demonstrating LinuxCNC's Joint axis homing. Its new and I don't think there are that many machines out there using it yet.

[joel0407]

I had a late regression with my home switches and I sorted it after slowing the homing velocities right down. One of these days I'll do a video demonstrating LinuxCNC's Joint axis homing. Its new and I don't think there are that many machines out there using it yet.

Saw your video on your LinuxCNC thread. Posted there.

I think the Joint axis thing needs complete instructions. I'm getting bits from all over. It might be easier for you running Mesa (don't know). I'm struggling to find things for Parallel Port stuff and joints.

I'm super impressed by the speed I'm able to run from the parallel port. There is no way my ball screws can handle that RPM. For a less than $15 break out board on a $25 PC, it's going to be hard to justify anything more. I have run out of inputs. E-stop, XYYZ homing and I'm out. It'll still be cheaper just to add a parallel port for auto touch-off that going down the Mesa path.

Happy Days

[joel0407]

Oh, I have to fully work out the homing sequence too. I'm hoping I can, hit home with some pace, Back off until sensor unlatches, Approach again slowly until latches more accurately. Then this is the questionable part, I am hoping rather than backing off a set distance, I can continue on at a set distance. My home sensors are about 35mm from the limit switch so it has time to stop after the sensor is tripped when approaching with some pace. If it was to back off after it approached slowly to get an accurate sense then I'd be wasting maybe 40mm of gantry. If I'm able to continue on after that slow approach, I'll be able to go right up to my limit switch and use the full gantry.

Happy Days

[joel0407]

Well the base has turned out super rigid. It was about 2mm out of square but the turn buckles straightened that out.

And I figure that'll be neat enough for my 8 wire motors.

[rodw]

My home sensors are about 35mm from the limit switch so it has time to stop after the sensor is tripped when approaching with some pace. If it was to back off after it approached slowly to get an accurate sense then I'd be wasting maybe 40mm of gantry. If I'm able to continue on after that slow approach, I'll be able to go right up to my limit switch and use the full gantry.

You can do that using HOME_OFFSET in the .ini file. Personally, I thought the homing sequence docs for LinuxCNC was very well written. I move 45mm closer to the limits after I home. I end up being about 1-2mm from the limit switches. My HOME_OFFSETs are 2mm different to square the gantry.

Good to see you've made a start.