Image by RobertW posted Jan 21, 2017 at 8:41 PM

I found a tapped hole in the cbeam end plate for the limit switches.



Here's a pic of the working limit switch plate.

 

Hi All,

I've completed my C-beam machine build and just completed my first aluminum plate cut. The machine has a few changes to it from the original intention - I added side rails for the work surface so it would be more stable and could be widened to about 14 inches. I also used larger and more corner braces that I got from 8020.net... just a little paranoid about making it strong and square. I did the test pattern design with Sketchup and Sketchucam, and used the universal G-code sender to run the machine. The first cut did not finish because my laptop went to sleep (doh!) so I started it over with the laptop plugged into the wall and it went fine. I'm pretty happy with the quality of the cut. I set the router (Dewalt 611) to full speed (about 26000 rpm), set the feed rate to 3000mm/min and the plunge rate to 1500mm/min, turned on ramping and multipass with a depth of 0.25mm. The bit I used was a 3mm solid carbide two flute mill from MSC - Accupro ALTIN coated. I have to say that I'm pretty happy with the result and am very grateful for all the online resources I studied before determining a set of cutting parameters.

Also, if anyone is interested, I've designed and printed micro-switch endstop mounts for X-min, Ymin, and Z-max.

Pictures...

 

Very nice. I really appreciate that you made the extra effort to design and print your endstop mounts

I'm hoping that some body can chime in about the speeds and feeds. I generally cut AL at around 12-13k RPM with a max feed rate of 70 inches per minute (sorry, I'm an old guy).

 

I've had a 3D printer of one kind or another for 4 or 5 years now... I rarely pass up a chance to design and print a custom bracket or some such thing!

 

Wow, those cut speeds are nice. I too am only cutting aluminum around 70 in/min like stargeezer.

I have a very similar setup as yours, I increased my spoiler board size to 14 x 14 inches and added wheels on the outside of the C-Beam. I am interested in you design for the side rails under the spoiler your added.

Could you provide a link to the end mill you are using, I think one of my biggest problems is using cheap end mills.

 

The side rails are 500mm 40x20 Vslot rails mounted on their edges on either side of the screw drive C-beam - 110mm is the space between them. Each of the side rails has two Mini V Gantry plates that attach underneath the corners of the MDF build surface. The original build plate attached to the nut block was replaced with a C-beam gantry plate. That way the moving parts were all in the same plane and it made it easy to mount the MDF plate to them. A second MDF plate is screwed on to the first one to act as the spoiler board. I hope that is clear!

Here is the link to the mill: 0.118" Diam, 0.354" Length of Cut, 2 Flute, 62773601 - MSC Less than $10 each and pretty quick delivery since they ship from Reno NV and I live near Phoenix AZ.

When I did the cutting I did not use any lubrication or coolant, but I did hold the nozzle of a vacuum cleaner near the cutter to keep chips from building up.

 

Thanks for the info. I have been thinking about adding some rails as I do get some side to side flex on the spoiler board, and your design looks like the solution, thanks.

I'm going to order a couple of those mills and see if I can get my cut speeds up.

Thanks again.

 

I arrived at my numbers using some on-line formulas I found... pleased it worked the first time - I've seen lots of horror stories of people dodging broken mills and bad finishes...

 

can any one tell me what step res. they run on xy and z axis . I have set mine to 1/8 but seams slow so i change it to 1/4 seams much faster but moves are as fine ..? is there a torque lose through a finer Res. such as 1/8

 

I've seen this article before, but most people don't fully understand what it means. You DO NOT lose torque or accuracy by increasing the number of steps...

The article says "the INCREMENTAL torque per microstep drops off drastically".

Well of course it does! But that doesn't mean the torque per degree has changed. In fact it doesn't. Let's try to understand this better...

If you apply a steady current to each of the two motor windings and there is no load on the shaft, the motor will turn until it finds a resting position based on the magnitudes and polarity of the currents applied. At this position there is ZERO torque. If you try to turn the shaft by hand it will move slightly off of the resting position and fight you. It's now producing torque because it's not in it's happy position. So how much torque is it generating? Well that depends on how far you pushed it off its resting position. The greater the angle the more it's fighting to push it back. For small angles the torque is roughly proportional to the angle. If the angle becomes too great it will break free. At that point you've hit the holding torque limit.

Table 1 tells you the amount of torque produced for various angles - the angles resulting from being 1 step off from the resting position. Well obviously if you use 256 microstepping and then apply 1 pulse, the new angle is tiny so the off angle torque is tiny.

Let's say you are currently using 8-step microstepping, and lets say you apply 1 pulse. The incremental torque will be a certain amount.

Now let's say you change to 256-step microstepping. But instead of applying 1 pulse you apply 32. Why 32? Because that will try to move the shaft to the same new position as a single step in the 8-step example. Take a guess how much torque will be produced. It's exactly the same as 1 pulse using the 8-step microstepping!!! No torque loss!

If you had applied only 1 pulse with 256-step microstepping then you would expect the torque to be approximately 1/32 that of the 8-step example. Look in table 1 at the 8-step and it reads 19.51%, and for 256 it reads 0.61%. The ratio of the 2 numbers is 19.51/0.61= about 32 which agrees with what we expect.

Also realise when you increase the stepping on your system (let's say from 8 to 256) the microcontroller that pulses the motor driver will increase the number of pulses it sends by the same factor - in this case by 32. Since it increases the number of pulses sent then in the end there isn't any torque lost. You simply have more resolution.

There is a problem however with increasing the number of pulses that has nothing to do with torque. The stepper driver you are using has an upper limit on how quickly it can receive consecutive pulses. Additionally the microcontroller outputting the pulses has an upper limit as well. These upper limits will determine the maximum rotational speed of your motor.

If you've increased the microstepping in your system and noticed a reduction in the maximum speed, it's got nothing to do with torque, but rather the upper frequency limit of your stepper system, which could be the driver itself, or the microcontroller that's pulsing it depending on your system.The software I use to control my stepper board is Grbl Panel and unfortunately it doesn't warn that you've exceeded the frequency limit on your system.

He're another argument that the absolute (not incremental) torque doesn't change. If you think the torque drops with an increase in microstepping, then if there were a way to increase the stepping to be extremely high (like infinite) then there would be zero torque. There is a way to simulate infinite microstepping. Toss the motor driver and instead connect the two winding to two current regulated power supplies that you can adjust to any value you want. So you have infinite resolution. Does that mean the torque goes to zero? Of course not! For example if you connect a fixed current of 1 amp to each winding I can guarantee there will be torque if you try to rotate the shaft by hand. Yet this is equivalent to infinite microstepping since you have infinite resolution.

There is no loss of torque by increasing the microstepping. Period. So crank up the microstepping and enjoy the smoother ride!

If you find your maximum velocity is now lower than you require, then either reduce the microstepping or get a better stepper controller board which can handle the higher frequency requirement.

 

I've been trying to figure out how to do the exact same thing with out riggers on the cutting board. A few pictures are worth a thousand words. Can you post some more detail pictures to go with your description.

Thanks a bunch

Bruce

 

Finished my rebuild. For some reason I couldn't get the gantry to work correctly, so I swapped out to linear guide rails. As soon as I cut my plates, I'll add X linear guide rails and make the machine bigger.

But for now, can I PLEASE get some help on getting started? I have tried the pen draw and it was half successful. The gcode for sample plate ran which is good. I'm a newbee and some help would be much appreciated. I can be found on facebook by: [email protected]

 

i like the micro-switch endstop mounts,

can you share them with me , thx

 

I like the idea of swaping out the current plate the table sits on for the x-large c-beam gantry plate. Would give a few extra inches of travel and seems like even though a little smaller would be a more sturdy base for the table. Anyone see any reason not too?

 

Hello All.

I purchased a C-BEAM machine a few weeks back and joining the community at this occasion.
The build was quite easy to do and well explained, it was a breeze
After a few "quick and dirty" tests, I am adjusting some elements on the machine, better endstops, enhancing the squareness of the machine, tramming the router bit, etc.

I have pretty good success on all of that except the X/Z squareness. Let me explain:
* I start by flattening the MDF base
* I checked that my router was well aligned with the Z axis (putting a laser pointer instead of a milling bit and moving up/down)
* I used then a homemade "tramming dial indicator" to check that the router was well square to the plane surface.
* for Y/Z squareness, it was not very good, but I could adjust the whole gantry and all is fine now. With all the screws on the side of the gantry "pillars", I could adjust it progressively and secure it tight.
* for X/Z squareness, I can see it is not right, but I don't know where I can "play with" the screws/bold/plates of the machine to fix it

I short,; I have something like that (of course - very exaggerated on the image, I have just something like 1 or 2 degree off)


As I understand it, I can only play with these elements:
* Idea 1 : The way the wheels of the X axis are on the X-rail, moving the eccentric nuts or adding more eccentric on the other side.
* Idea 2 : The way the Z axis 250mm C-BEAM is bolted to the X Plate, perhaps I can move it a few degrees if there is some margin on the Vslots.

Note that for idea 1, the upper holes are 5mm and not 7.2mm - so I will certainly have to modify the gantry plate.

Before starting my "adjustments", I wanted to have some advice from you. How did you correct this kind of misalignment on your machine ? Would you see other options ?

Thanks a lot for you help
see you soon !

 

Thanks for the quick reply.
I'll do a general check of all the assembly. The "degree of freedom" of the wheels has been tested yesterday but not the eccentric nut alignement - I'll check all of that again, thanls

Whey you say "the mount not being square to the Z mounting plate", you mean that my router milling direction is not perfectly parallel to the Z-axis motion ?
I used the "laser point in chuck" hack to check that before all, but perhaps it is not the best method. How could I check that the mount (and the router) are accurately mounted ? just by visual inspection ?

Thanks

 

I found my C-Beams were not perfectly square at the ends so setting a level on top of the z caused me a bit of grief. The beauty of cnc and especially a small format machine...that generally is using a small diameter bit...small errors in square and true will not kill your final results unless you are trying to be accurate to the thousands. I would fun some cuts and see if your results are within your spec.

 

I prefer to use a set or engineers square (but not a cheap combination square) rather than a level (or even the 3 - 4 - 5 measurements with a ruler) for the simple reason that I know that the surface that I have my machine on is flat, but not level in itself.

 

Thanks for all you suggestions.
Having a close look revealed that everything is a little off on my machine I didn't saw that the parts could be moved and adjusted so much in the build process and I was tightening everything like crazy a little too soon and without checking the exact alignments of anything.
Now I'm with my engineering squares looking for things to set square.

I'll tell you how it goes.
Thanks again

[EDIT]
This seems to solve my problem. Now I have around 0.2mm of difference on my "tramming dial indicator" of 160 mm diameter... that would be around 0.1% accuracy on the angle. 0.1 degrees. nice Thanks to all for your help

 

Thank you David. I found those moments after I posted. (stupid me didn't go back to edit my post). I am now back to ground zero with my c-beam from open builds. I finally got the frame setup and am working on the software end of it. I'm such a noob. I am using bCNC as the controller for the machine. Can't seem to get it setup

 

So in desperation to get my machine going I resorted to using the software from the assembly video "Grbl-Panel-1.0.9.14". Using the provided setting's image I attempted to change the settings BUT their isn't a "save" button for me to save changes. Additionally their isn't a $14 (auto start, bool) parameter for me to update. Every time I do anything what so ever in the software the old parameters repopulate. Any help?

 

UPDATE- resolved with uninstall - reinstall.

 

To save you have to double click the parameter you want to change. You can only do one at a time.