PDA

View Full Version : BUILD LOG: A brand new machine build. 2x4 size or something in that region. :)



IN-WondeR
19-01-2013, 03:03 PM
I have recently started designing a few bits and pieces for my next CNC build.

I have opted, for ballscrews on all axis.
Hiwin style linear guides all around.
Servoes from DMM-Tech.com on all axis.
And a heavy duty z axis, for optimal stability.

I will be using a UNP steel beam for my Gantry so it should be very sturdy.
All flat surfaces for mounting will be machined so everything will be perfectly angled.

I have for now, only made a few renderings of the gantry main beam, including the servo and the hiwins.
This will be mostly a design build, and after complete design has been made, then I will start building it. ENJOY

79847985

Jonathan
19-01-2013, 03:52 PM
When the machine is cutting parallel to X (with Y being the gantry axis) the cutting force applies a moment to the gantry, which will twist it, hence the gantry must be a structure with good torsional stiffness (in the XZ plane). The torsional stiffness of a beam is only good if the section is closed, so for a C section it is very poor. The same is true for an I-beam or plate, so I suggest changing it for RHS or multiple pieces of box section.

Torsion Equations (http://www.roymech.co.uk/Useful_Tables/Torsion/Torsion.html)

IN-WondeR
19-01-2013, 03:58 PM
Since the machine is no more than 1200mm wide, then the stiffness of a UNP 160 beam should be more than enough.
If I were to change to RHS tubes instead, then I'm gonne sacrifice something else..

So if you think that you can bend a 1200mm long 160mm UNP beam, over a max length of 1000mm, then go ahead, I believe that deflection will be set at a minimum.

But then again, as I complete the design, I will have a friend of mine, do some stress and deflection calculations on it. :)

Jonathan
19-01-2013, 04:25 PM
Since the machine is no more than 1200mm wide, then the stiffness of a UNP 160 beam should be more than enough.


Have you calculated this to know for certain? If not, then this spreadsheet (http://www.mycncuk.com/forums/linear-rotary-motion/2214-cnc-machine-stiffness-calculator.html) may be of assistance. You just need to know the length of the Z-axis, to work out the torque applied to the beam and the force which can be approximated from knowing what materials you're intending to cut. Also, even if it does turn out that the UNP 160 beam is strong 'enough' that doesn't mean you should definitely use it since the formula's for torsional stiffness make it obvious that using a closed section will require a lower volume of material than a C shape.


If I were to change to RHS tubes instead, then I'm gonne sacrifice something else..

RHS is just one idea, I'm sure you can come up with others with good torsional stiffness.


So if you think that you can bend a 1200mm long 160mm UNP beam, over a max length of 1000mm, then go ahead, I believe that deflection will be set at a minimum.

No, I don't believe I can bend it - it's maths not belief, it's twisting not bending and it's not a question of if it will bend, it's a question of how much. With a CNC-Router the structure needs to be very stiff since although the magnitude of the forces is quite small, even a very small deflection can cause problems (poor finish, tool wear, limited MRR), unlike most applications where we're worried about the material failing which is not a problem here.

m.marino
19-01-2013, 04:43 PM
Interesting design,

How do you plan on attaching it to the up-rights? Do you plan on mounting it directly to the plates for a axis?

Also Jonathan I know you have FEA software so if you have the time run realistic loads through and see what you get. Remember to add the stiffness that the Hiwin will add. I am probably missing something and hope some one point it out. Also what thickness UNP steel beam are you planning on using OP?

Michael

AdCNC
19-01-2013, 05:16 PM
Id recommend machining a datum edge in on of the rail pads so that you can true up the other rail! Thats how i have done my other machine but have not posted anything on this build!

IN-WondeR
19-01-2013, 05:29 PM
The connection to the uprights have not been decided yet. As I said, this is a design build.
By the way, then this is my fourth machine I will build. This time a smaller machine so I can have it in my garage, and still do other kinds of work in it.. :P

AdCNC
The rails are already mounted on top of a welded on piece of steel, which will be machined flat to true up the rails.
A Datum edge, is that a straight edge I can mount one rail along, so it will be true to the surface of the machine?

JAZZCNC
19-01-2013, 05:43 PM
When the machine is cutting parallel to X (with Y being the gantry axis) the cutting force applies a moment to the gantry, which will twist it, hence the gantry must be a structure with good torsional stiffness (in the XZ plane). The torsional stiffness of a beam is only good if the section is closed, so for a C section it is very poor. The same is true for an I-beam or plate, so I suggest changing it for RHS or multiple pieces of box section.

Been here before haven't we Boy wonder.!! . . . . It depends on the Job it's going to do.?

If it's mainly wood or plastics then it will be perfectly fine.

If it's for meaningful cutting of Aluminium and steel then Yes it's much less than ideal, Has is a Gantry style machine really.
So with that I'd suspect it's main use will be softer materials in which case then it will be perfectly fine. RHS would also be unsuitable for Good Aluminium/Steel cutting machine so both are unsuitable in this case.

Another thing with UNP beam is that it doesn't take much to minimise any twist and few bracing plates will sort it.! The advantages with ease of build for a wood cutting machine are obvious and I know for a fact that it works on a Wood router with no detriment what so ever because I built my mates Panel router using one.! This worked 10hrs day 6 days week and did so for 2yrs producing 1000's of panels until recently when the building caught fire.!:sorrow:. . . MK2 will be bigger and better but will still probably use UNP beam.

JAZZCNC
19-01-2013, 05:53 PM
Thats how i have done my other machine but have not posted anything on this build!

Why NOT.? . . . Comon man get ya finger out. .:whistle:

AdCNC
19-01-2013, 05:57 PM
The two rails you have welded on will be skimmed flat as you planned. But if when u have them milled flat then leave a lip on say the bottom rail that you can but your bottom rail up agains so its true if you get me!

AdCNC
19-01-2013, 06:01 PM
Jazz! Lol will do matey I've been promising you some pics for donkeys now time i pulled me finger out.

AdCNC
19-01-2013, 06:03 PM
In wonder if you didn't have anyone to do your machining for you then i could perhaps work something out with you.

Web Goblin
19-01-2013, 06:17 PM
If you are worried about any twisting forces you can weld some stiffeners ( viagra as they are called ) inside the "c" section of the beam as Jazz says. This will stiffen it up nicely.

IN-WondeR
19-01-2013, 06:17 PM
AdCNC

Got it... Will make the change as soon as I get the program started up again.. :)

IN-WondeR
19-01-2013, 06:18 PM
AdCNC

Hmmm, I'm not sure having you doing my machining will be a cheap thing. I live in Denmark, so it's a bit of shipping back and forth... :P

Web Goblin
Yeah, I have thought of adding som braces to the back of the UNP profil. There's enough room for it..

AdCNC
19-01-2013, 06:21 PM
Lol lets rule that one out then haha

Web Goblin
19-01-2013, 06:23 PM
Dont you deliver Ad?

AdCNC
19-01-2013, 06:27 PM
I can do yeh but it would cost a fair bit for a bit that weight, all for it tho.

Jonathan
19-01-2013, 06:28 PM
Also Jonathan I know you have FEA software so if you have the time run realistic loads through and see what you get.


Ok, I've done a very quick simulation as I don't think it adds much to the discussion since I'm only comparing the stiffness of sections for which the equations are already well documented. Granted the rails add a bit of stiffness, but it's not that great in the whole scheme of things.

79867987

That's run with a 20N force, so nothing serious, at 200mm from the bottom of the gantry. Deflection at this distance (i.e. the tool) is 8.3um for channel and 1.5um for box section.

Before everyone starts pointing out flaws - this model is just a quick simulation to demonstrate the principle, not the exact machine, except for the UNP160 dimensions which I got here, so the exact numbers are not important. What's important is to note that just using the same size RHS made the deflection over 5 times less than with UNP160. This model assumes the ends are perfectly rigidly fixed, so in reality it will be worse.



The rails are already mounted on top of a welded on piece of steel, which will be machined flat to true up the rails.
A Datum edge, is that a straight edge I can mount one rail along, so it will be true to the surface of the machine?


Partly that, but also to help mount the rails straight and parallel.

JAZZCNC
19-01-2013, 06:34 PM
What's important is to note that just using the same size RHS made the deflection over 5 times less than with UNP160. This model assumes the ends are perfectly rigidly fixed, so in reality it will be worse.

Ok then now do it agian with 2 Viagra braces has Web goblin calls them equally spaced and see what happens.?

AdCNC
19-01-2013, 07:07 PM
All that aside 8 micron is bloody good for an home made jobbie. I know i know its only a quick sim and there are a lot of variables that that can influence the end results but it goes to show that it will be more than enough for cutting most things a hobbyist will want.

IN-WondeR
19-01-2013, 07:19 PM
just to show you all what I have made before, then here's a pic of my last machine.. ;)
7988

Jonathan
19-01-2013, 07:34 PM
On the next one, put the Z-axis rails on the plate that the spindle is attached to since that way is stronger as you're reducing the overhang.

JAZZCNC
19-01-2013, 07:44 PM
On the next one, put the Z-axis rails on the plate that the spindle is attached to since that way is stronger as you're reducing the overhang.

Yes what he said and paint it black because white gets dirty too quick.!!. . :playful:

IN-WondeR
19-01-2013, 08:08 PM
Yeah, the rails will be mounted on the moving plate on the z axis this time. I also found the other way to be less effective, although it did the trick as it should.

IN-WondeR
09-11-2014, 11:12 PM
Well, it's been some time since I updated this thread. And A LOT has happened... :)

The machine is in fact ready to go. All I need is dust extraction. But that will get there when it gets there.

And I have done some cuts on it already, so here's a few photos for your enjoyment. :)

EddyCurrent
10-11-2014, 08:45 AM
Excellent looking machine, I like the simple base design and how it can be adjusted at each corner but does it vibrate, resonate ?

IN-WondeR
10-11-2014, 08:55 AM
There's very little vibration in it. Especially after I added the 41mm MDF top to it. 22+19mm MDF.
Plus I added extra stiffness on the crossmembers with 60x60x4mm steel tube.

I have done a cut test in Aluminum, with a rather agressive cutting parameter, and it came out lke this, with a single flute cutter. You need to look at the area within the red rectangle.
The cut was done rather agressively, and with a 3mm single flute cutter at high RPM. Will go for a slower RPM next time I try it. It looks bad looking at the picture, but the cut is actually very flat and smooth.

EddyCurrent
10-11-2014, 09:21 AM
I once looked deeply into making loudspeaker cabinets (big ones) and it was deemed a good idea to use a glued sandwich of MDF and plywood to reduce cabinet resonance.

IN-WondeR
10-11-2014, 09:59 AM
Laminating sheets is always a good thing, it does remove alot of resonanse and vibrations all together... I had hopes to find another 40mm bottom sheet this time, but sadly I did not find one for cheap this time around.
But the overall look of the machine and the precision is more than ok for my uses. Which mainly will be woods and plastics, and perhaps a bit of aluminum.

Blackrat
10-11-2014, 04:11 PM
Looks good , which dmm kit did you opt for ?

And what happened to the first machine ? That frame looked pretty awesome

IN-WondeR
10-11-2014, 04:13 PM
Did not have the funds to opt for a DMM kit this time... Sadly...

So I opted for Leadshine AM882 drives, and 1600oz-in steppers (low cost ones) Will be switched out at some time in the future, with Leadshines own stepper motors instead.
But for now, they run ok.

The old machine was sold. A bit to big for my garage. It took up almost 6m2.

Boyan Silyavski
10-11-2014, 05:00 PM
How you did the T slot?

IN-WondeR
10-11-2014, 05:02 PM
The T-slots I bought from www.motedis.com

I let the machine route out the recesses for the t-slot rails, and just screwed them into the table top.

Boyan Silyavski
10-11-2014, 05:14 PM
Now i have the question here if the result on aluminum is a bit poor cause the aluminum t-slots would not hold it strong enough/ the alu plate/ . cause i have to decide now for my machine and still in doubt. I have been searching for the same but in steel.

What do you think? Is the slot strong enough? Any vibrations from the work holding?

IN-WondeR
10-11-2014, 05:16 PM
The aluminum was made on the old tabletop. Since then the crossmembers have been reinforced with extra steel, plus the tabletop itself has been changed from 22mm OSB to 22+19mm MDF.

If the T-slots will hold the aluminum in place I don't yet know, but I'm gonna make a test soon, where I'll cut some hold downs for the T-slots...

IN-WondeR
16-12-2014, 03:14 PM
Well, the machine is up and running as it should be. And I have already cut a few bits and pieces in Aluminum on it. Recently I cut 16 plates in 5mm aluminum sheet.. Ended up looking really good... They are 121x156mm and tolerances are within 1-2/10 of a millimeter. Not the best tolerances, but I only had a single flute 3.175mm cutter, so it came out as it came out.. But the finish was great...