Dell Optiplex 745, test axis fails!

lledee wrote:

I put together a google site, it's definitely WIP but you can check it out if you'd like, it should be access to public...Here is the link .

All I see is an empty site.

My WIP is the overweight brother of yours:


(old picture, is not 100% accurate anymore. Sometimes you have to adjust the plans)

There is a kind-of build log on a local (Dutch) forum. The interesting part starts about here: www.cnczone.nl/viewtopic.php?p=86358#p86358

I thought I had a project up there but I guess I didn't. sorry. Nice machine you've got there What's your primary use for it?

Your project is the same. Linear support here, ball/acme screw there, etc. It's not rocket science.
I only have to make more parts and they are heavier.

My primary use is 'I hate going to four stores to find out the parts they stock are all the same and not exactly what I want'. Thus, general hobby use. Parts for broken kids toys, motorcycle parts, household parts, etcetera. Once you have a CNC mill up and running you will quickly find out that it is a very powerful tool for all sorts of things. Of course you can do all sorts of complicated stuff and artistic stuff, but I even find it useful for simple tasks such as cutting out a square metal plate with 4 holes in it, for example.

Machine is still 'under construction' BTW. All I have now is various parts, but nothing that resembles a mill yet. That will take another couple of months (building from the ground up takes an awful lot of time, but it is pure fun)

Why are these parts so heavy that you are working on?

I'll be using my machine mainly like i said for PCB design and my wife's hobbies, possibly even a little knick knack here and there, but what i'd really like to see is the wood working and carving abilities.

How experienced are you with the entire machine design? What's your background? Hobby only?

I bought the Shapeoko 2 kit because it seemed simply like the best solution for what I was looking. I didn't want to pay an arm and a leg and didn't have the skills do build the machine yet either so i just decided to jump in with this kit. Finally i got the LinuxCNC working with your help, now it's time to calibrate the electronics to the mechanics. I've ordered some pulleys for the belts and they should be in today. Next is to splice in the motors to their harnesses and route them temporarily. Then I'll really start the calibrating.

Have you completed your machine yet? You must be using another machine in the meantime right?!

I've upgraded a few pieces on my machine already and have taken it all apart many times. It was quite fin so i can just about imagine your project to do it from scratch. I'll like it when i can cut my own pieces with my machine instead of hand

What softwares do you use in your toolchain? (CAD-CAM-G-code) any good freeware you've seen? I LOVE sketchup but i'm still trying to work with it...

lledee wrote:

Why are these parts so heavy that you are working on?

Stiffness and vibration dampening mainly. 90% of my milling is in aluminium, and I don't want to lose the ability to mill steel. No need to hog with a large indexed mill, but it must be capable of using a 6-8mm AlTiN coated endmill in steel without resorting to taking off 0,5mm thick layers only.
If you want to do that and a reasonable accuracy (say <0,03mm with some effort) you end up with heavy members with large outer diameters.

My machine will be more on the light side than on the heavy side.

How experienced are you with the entire machine design? What's your background? Hobby only?

Not experienced. I am an engineer, but in electronics. Mechanical engineering is new to me.
Luckily, there are many similarities between electronics and mechanics. Mechanical resonance and electrical resonance are not that different. Volts and amperes vs kilos and Newtons is not that different either. And neither is designing with nuts and bolts vs resistors and capacitors.

I also have a few good mechanical engineers watching my back and the local fora are a great resource of information and help too. One of those guys borrowed me his surface plate, for example. Now I know I could not have done what I did without one.

I bought the Shapeoko 2 kit because it seemed simply like the best solution for what I was looking. I didn't want to pay an arm and a leg and didn't have the skills do build the machine yet either so i just decided to jump in with this kit.

If it does what you want then that's the best solution. Building from the ground up is very, very time consuming. I expect to need 9-12 months to reach 95% completion. But it is fun to do. I did check out used 'real' machines, but they are either too expensive, too large, too bad shape, etc. A gantry style mill provides a lot of working area for the footprint.

Cost is an issue too. It is for hobby and anything decent quicky goes above $5000 once you have all the this and that's. Too much for something that only costs money; the round-breasted financial manager at home would kill me.
So far I spent about $1000 on parts (bought almost everything used; Ebay is my friend). When finished total cost is somewhere in the $2000-$3000 range.

now it's time to calibrate the electronics to the mechanics. I've ordered some pulleys for the belts and they should be in today. Next is to splice in the motors to their harnesses and route them temporarily. Then I'll really start the calibrating.

There should not be much to calibrate... It is more like math: X travel per revolution times Y step pulses per revolution. If that does not match, re-check your bearings etc.

Have you completed your machine yet? You must be using another machine in the meantime right?!

No, I am currently working on the gantry. Finished adjusting the planarity of the rails to within +/- 0,01mm deviation on the dial indicator yesterday.
Took me 7 hours of work to adjust and secure them within this tolerance. And that is 'only' mounting two pieces of linear rail... You pay for accuracy, either a lot of time or a lot of money. Once the gantry is finished, I can assemble the table and gantry together so I have movement in X, Y and Z. Getting all that square will take another 10+ hours of measuring, minute adjustments, leaving it alone for a while so it can settle, and measure/adjust again. Then I will start worrying about ballscrews etc.

I currently do have another machine: a small CNC-ed Chinese mill known as BF20 or G0704. It is a nice machine for the money it costs, but in the end there are so many little irritating things and inaccuracies. I read around on various fora, and saw people spending an insane amount of time and money resolving all those, and in the end the only remains of the original machine is 2 pieces of cast iron or so. So I decided not to go that route, but build something else instead and sell the BF20 afterwards to recover part of the cost.

I'll like it when i can cut my own pieces with my machine instead of hand

'like' will soon become 'love'
Take this for example : just a few holes and saw cuts. Nothing special. But I drilled and cut off excess stock on the CNC. Doing it manually is quicker, but now these holes are a very exact fit to the linear rail carriages.

What softwares do you use in your toolchain? (CAD-CAM-G-code) any good freeware you've seen? I LOVE sketchup but i'm still trying to work with it...

I do have access to an older version of Solidworks. Licensing is not the 'crack.exe' type, but not entirely white either. So I need to search for a replacement, but I am not in a hurry. I am not really good in using it, but it is a very un-intimidating piece of software to learn. I think it is possible to download a 30-day evaluation, you might want to check it out.
With Sketchup I didn't get very far; it just doesn't work the same way my brain does.
I stil have to checkout FreeCAD. Did not try to design something in it yet, but it manages to open a 5000+ components assembly just fine. Solidworks needs half an hour of thinking to conclude that it's out of memory. Thus so far I am impressed by it, and it is worth a good look.
Designspark Mechanical is also a possible option for you. Quite powerful, and free. (as in: doesn't cost money)

For 2D drawing (which is usually sufficient) I use Solid Edge 2D. That's free to use. QCad is another option, but the Windows version is not free.

The CAM-part is handled with CamBam+. Not free software, but not expensive either and worth every cent. Not click-and-ready, no wizards, but very powerful. Can also import Gerber and drill data. The trial license is very generous (40 starts) thus you might want to check that one too. I did toy around with PyCAM, GCam, dxf2gcode, etc. They all sort of work, but I like CamBam better.

I also do many things without CAD/CAM. Pocket here, contour there, if it is simple enough it is faster to just type in the G-code directly and do a few subroutine (O-word) calls..

I'm new to this whole CNC thing also but i guess i may be like you, mechanically inclined to an extent...

If it does what you want then that's the best solution. Building from the ground up is very, very time consuming. I expect to need 9-12 months to reach 95% completion. But it is fun to do. I did check out used 'real' machines, but they are either too expensive, too large, too bad shape, etc. A gantry style mill provides a lot of working area for the footprint.

Cost is an issue too. It is for hobby and anything decent quicky goes above $5000 once you have all the this and that's. Too much for something that only costs money; the round-breasted financial manager at home would kill me.
So far I spent about $1000 on parts (bought almost everything used; Ebay is my friend). When finished total cost is somewhere in the $2000-$3000 range.

I assembled my machine in about 3 late nights after getting off from work. I had the flu and was sick at home while completing it and learning the software! I had a blast, i couldn't stop thinking about it, my wife ties to understand... Like today i told her i got a package in, her response was, i guess i know what you'll be doing when you get home today!

There should not be much to calibrate... It is more like math: X travel per revolution times Y step pulses per revolution. If that does not match, re-check your bearings etc.

By calibrate I meant that i need to make the machine match the software's parameters. Also i need to adjust the current settings of the driver and all. It shouldn't be too much work but what i need to do is research and find good details of the individuals that make up the whole package and also learn the definitions of the parameters that LinuxCNC is asking for. This shouldn't be too hard, maybe just a bit more reading

No, I am currently working on the gantry. Finished adjusting the planarity of the rails to within +/- 0,01mm deviation on the dial indicator yesterday.

Took me 7 hours of work to adjust and secure them within this tolerance....

All i can say is DAMN! I tell my wife that I need to more projects so I can learn more patience, and it's true, just look at your situation!

like' will soon become 'love'

I already have a love, maybe now she will take me serious about that threesome idea I had??? This one has a motor on it...

I do have access to an older version of Solidworks. Licensing is not the 'crack.exe' type, but not entirely white either. So I need to search for a replacement, but I am not in a hurry. I am not really good in using it, but it is a very un-intimidating piece of software to learn. I think it is possible to download a 30-day evaluation, you might want to check it out.
With Sketchup I didn't get very far; it just doesn't work the same way my brain does.
I stil have to checkout FreeCAD. Did not try to design something in it yet, but it manages to open a 5000+ components assembly just fine. Solidworks needs half an hour of thinking to conclude that it's out of memory. Thus so far I am impressed by it, and it is worth a good look.
Designspark Mechanical is also a possible option for you. Quite powerful, and free. (as in: doesn't cost money)

For 2D drawing (which is usually sufficient) I use Solid Edge 2D. That's free to use. QCad is another option, but the Windows version is not free.

The CAM-part is handled with CamBam+. Not free software, but not expensive either and worth every cent. Not click-and-ready, no wizards, but very powerful. Can also import Gerber and drill data. The trial license is very generous (40 starts) thus you might want to check that one too. I did toy around with PyCAM, GCam, dxf2gcode, etc. They all sort of work, but I like CamBam better.

I also do many things without CAD/CAM. Pocket here, contour there, if it is simple enough it is faster to just type in the G-code directly and do a few subroutine (O-word) calls..

I've just downloaded FreeCAD yesterday and need to give it a fair shot since it works on all OS's.
Sketchup is qutie amazing, one day i just sort of told myself i wanted to use it... so i just picked it up and went with it! (I'll need to do the same with another software it seems like unfortunately)
I just downloaded the DesignSpark Mechanical software and it seems to be friendly so far, we are getting along very well and it is playing fairly, so far so good. Thanks for the advice.
I've seen the Solid Edge program a while back and downloaded it, just never installed it yet.

I really appreciate all of the advice, but i must ask, what are you doing on this site if you don't run LinuxCNC?

What would you recommend for limit\homing switches? What types of switches etc? Preferably on the cheap side!

I've learned so much so far, thanks for sharing what you know

lledee wrote:

I already have a love, maybe now she will take me serious about that threesome idea I had??? This one has a motor on it...

One motor? At least four of them!

I really appreciate all of the advice, but i must ask, what are you doing on this site if you don't run LinuxCNC?

I do run LinuxCNC to control the current mill. And the new one. I love it.

I have yet to see another control that offers the same level of flexibility that LinuxCNC does. I can try out every single weird idea that I can cook up, implement every feature i think I need, and I love LinuxCNC for that. Not because it doesn't cost money, not because it is not Windows. I left the Tux vs. Redmont war a long time ago and compared to the cost of building and running a CNC mill a solution like KFlop, Mach3, PlanetCNC, etc. is just a fraction.

What would you recommend for limit\homing switches? What types of switches etc? Preferably on the cheap side!

Ebayed inductive proximity switches. Repeatability is about 0,02mm. I have 8mm diameter ones, the 12mm ones are even a bit cheaper.

Also: I am only running home switches and let the soft limits handle the mechanical limits. I used to have limit switches, but hard limit switches cost travel range (when hit at full speed there must be enough travel left to decelerate to a stop), when hit it is a juggle with the 'override' checkbox, machine power and jogging to get it off again, and our low-powered steppers do not inflict any mechanical damage in the rare occasion the soft limit fails.

Hey,

Sorry it's been so long since my last reply. I have a few things going on at home and had to relocate my setup and then other things came along!

So far i am enjoying the little features of LinuxCNC but cannot compare it to anything since I have no experience with any of them.

I still haven't had a chance to check out the switched but maybe i'll get to that soon...
I think i'll go more along the lines of what you are using with your switch setup, only homes.

Last night i finished up the wiring and was trying to input the settings for the distance and man i was overwhelmed! My wife is an engineer and she can do all of the calculations easily and loves to help but I cannot explain everything to her in five minutes.. She did try though.

Here is my situation if you'd like to have some input:
Aluminum GT2 pulley from Inventables link and this pulley has 20 teeth.
GT2 Belting from inventables also link .

The configuration of the drivers has been set to 1/16 step.
i cannot seem to find the right settings to make this thing travel the correct distance.

This post seemed enlightening but i could not get things to work!!! FYI i'm using INCHES. I'll post some pics soon but it looks shitty i'm afraid with the electrical tape and shitty wiring (only until everything is proven then it will look better)

lledee wrote:

Hey,
Sorry it's been so long since my last reply. I have a few things going on at home and had to relocate my setup and then other things came along!

No problem. It is hobby. Sometimes things go quickly, sometimes it is gathering dust for a week or two.

So far i am enjoying the little features of LinuxCNC but cannot compare it to anything since I have no experience with any of them.

Neither do I. I have seen Mach3, EdingCNC and a real control on a Deckel machine. Those 'real' controls are lightyears ahead (and this was an old one with CRT tube and all) of LinuxCNC when it comes to fast part programming on the machine itself but otherwise I still prefer LinuxCNC.

Here is my situation if you'd like to have some input:

The configuration of the drivers has been set to 1/16 step.
i cannot seem to find the right settings to make this thing travel the correct distance.

I suppose the belt moves your machine directly and the pulley is on the stepper?
Ufff, that is more a 3D printer setup than a milling setup.

Your belt's teeth pitch is 2mm. Your pulley has 20 teeth (website says 18, BTW), so for every rotation of the stepper shaft you have 20*2=40mm of motion. In inches that is 40 / 25.4 = 1,5748". The reverse is also true: the stepper shaft must rotate 1/ (40/25,4) = 0,635 revolution for every inch of movement.
One rotation of the shaft is 200 full steps (I assume you use regular 1,8 degree stepper motors). You divide every step into 16 parts, so 16*200=3200 pulses must come from the computer to move 1,5748". Which makes every microstep equal 0,49 mils.

In Stepconf I would try this for a start:

- Motor steps per revolution: 200
- Driver microstepping: 16
- Pulley teeth (Motor:leadscrew): 1:1
- Leadscrew pitch: 0,635 revs/in
- Maximum velocity: 75 in/sec.
- Maximum acceleration: 10 in/s2

Maximum velocity and acceleration a a bit on the low side, but first it has to work stable, then fast.

electrical tape and shitty wiring

I also love ducttape as an engineering tool, don't worrry
But do keep the motor wiring a few inches away from the 'delicate digital signals' coming from the PC and going to the stepper drives.

I suppose the belt moves your machine directly and the pulley is on the stepper?
Ufff, that is more a 3D printer setup than a milling setup.

You are correct, more like a 3d printer, I bought it for softer materials mainly. I don't expect to try and mill aluminum so hopefully I should be good.

Your belt's teeth pitch is 2mm. Your pulley has 20 teeth (website says 18, BTW)...

If you look in the details you will see the ...-03 version, this is the 20 tooth. I had a few emails with inventables about this.

I'll give these details a shot and see where to go from there! THANKS!

I've tried to upload pictures but no luck...