[justCNCit]

It is very cheaply built en regards to purchased parts. As you can see its mostly square tubing from the back yard. The rails are made of plain cold rolled 1" rod. The machine works ok but is noisy, as it is hardly aligned.

I used a milled steel block about 6" wide and long and 2" high to rough tune the bars into being square. However there seems to be no way to align the rods to their plane of movement. While I made tools to measure the distance between the rods very accurately, without putting the entire machine on a layout table it seems impossible to do it without having to eyeball it.

I tried lasers but they seem to be inaccurate. Plus there is no guarantee of reliability using them.

So this machine rubs all the time because nothing is aligned. The bearings rub and don't roll, which cuts down severely on top end cutting/rapid speed. Plus the work ends up being a bit off.

Any ideas?

[solidwoods]

Congrats on you machine.

To make a flat true frame or adjust frame/rail twist.

You need a straight edge that is long enough to reach rail corner to the opposite rail corner.

Use a level to find the elevation between the rails ,, (found across the rails closest point). Level the level/rails (while its across the rails closest point).
Make a point there. This point is just a spot to set the straight edge on.


Put the straight edge on that elevation point. Use the straight edge to make the center and the 4 rail corners "on the same plane".

Thats "in theory just the corners", If your rails flex/bend then you'll have to level there whole length.

Odds are when you get near the final level, the center point will have to be adjusted by a few thousandths, then continue leveling each corner.

Keep in mind- move the machine and it may be twisted again.

jim
no cnc yet.

[NickAshton]

Just an observation, your X axis is driven with a single screw.

With the X bearings being so close together on each rail, I would guess most of your problems will be the Y carriage scewing as you turn the X screw.

I would suggest a second X screw turned off the first one with a timimg belt, or a second stepper motor to keep the Y carriage running true.

Otherwise making the bearings further appart on the screw side of the X axis would help but not be ideal as you would loose travel.

Nick

[Bobbyr70]

Well you have enough ambition to at least build it. there are a few suggestions, the other replies have offered. Alignment is often overlooked and is the most important. Dual lead screws with a belt drive would solve the twisted racking issue. If you build things too cheap you tend to learn a lot of things the hard way. Trust me I a have done so many things the hard way I managed to get quite an expensive education.

[justCNCit]

Thanks guys,

It's really not that bad, it has always been obvious that the whole gantry is going to twist when under a load because of the screw only being on one side. Knowing this I already purchased the 4 axis drive to mount a second stepper on the Y axis when the machine gets rebuilt. The big moving gantry is what I refer to as the Y axis and I may be incorrect.

When its rebuilt it will probably have precision rod rather than cold rolled. However we can see that the machine works now without sinking $$$ into it. I was in a rush to see if the thing would even work. The alignment issue is by far the main issue as the thrust bearings on the screws rub rather than roll so now and then they bind (luckily enough on the last job after cutting was done).

I was not aware you can pick up a straight edge big enough to tune this in.

Does anyone have a picture of one? I figured they were for squaring something to a 90 degree angle, not for aligning two rails.

[devildog]

Thanks guys,

It's really not that bad, it has always been obvious that the whole gantry is going to twist when under a load because of the screw only being on one side. Knowing this I already purchased the 4 axis drive to mount a second stepper on the Y axis when the machine gets rebuilt. The big moving gantry is what I refer to as the Y axis and I may be incorrect.

When its rebuilt it will probably have precision rod rather than cold rolled. However we can see that the machine works now without sinking $$$ into it. I was in a rush to see if the thing would even work. The alignment issue is by far the main issue as the thrust bearings on the screws rub rather than roll so now and then they bind (luckily enough on the last job after cutting was done).

I was not aware you can pick up a straight edge big enough to tune this in.

Does anyone have a picture of one? I figured they were for squaring something to a 90 degree angle, not for aligning two rails.

I know this is a do it yourself project and it seems to work but I was wondering the purpose of you building this thing. If you are working with wood or plastic, with big tolerances, perhaps this will be good enough. Most machines are made out of cast irons because they don't warp as much. Steel has a memory and usually goes back to it even after you machine it. The only way for that not to happen as quick is if you heat treat it and then grind it to the flatness that you would need. But since the cost of that would be too high, a ground cast iron with a higher steel content is usually used so that you can do some case hardening. With regards to alignment, you will need ground rods with very close tolerance if you want a more accurate machine. If you can hold your machine and shake it and it moves or twist, then it isn't stable to machine. Square tubes without the proper support will not work out. Try to add more support if you want more stability but if you ask me, it is always cheaper to buy a machine than to make one. Without the expertise and technology, you will end up with a project that you will always keep improving and without the results that others are doing. If you are machining steel, then all the savings you have will just go to the toolings you will have to keep buying from all the vibrations you will encounter.

Oh, add some slide way oil lines in all your slides just in case you don't have one yet. If your base is more rigid, I will tell you to use linear guides but I don't think that is an economy way of doing it. Good luck and hope you get this to run well.

[justCNCit]

It's meant for wood/plastic. Problem is when you look at the rails fitting against the roller bearings, the rails are warping so that they aren't contacting at all times, there is about a .050" gap now between one of the bearings and the rail.

If only there was an accurate way to measure straightness over that length, I could at least contemplate putting in a precision rod. Otherwise without a way to measure it nothing works at all. The 2" square tubing doesn't work because that is warped as well. Lasers from the hardware store come with a huge tolerance too so they aren't any good, and I don't trust bubble levels for this.

I see machines built all the time on a budget, sure they use precision rod and better bearing systems but how do those people measure for straightness/level guide rods?

[PlasticWorker]

Hi,

Starrett makes a 72" steel straight edge that is +\- .0002 (two tenths) per foot for $ 400.00. Or a four ft one for $250.

Barwalt has an 8' aluminum one for $ 60.00 but I don't know the tolerances.

[sawmiller]

i've built several routers, and i understand your pain.
are those 425 oz/in motors ? there are supposed to be resonance issues with those.
the best i have gotten using mine is 50 in/min.
thats with 2 y ( long) screws. i am investigating the resonance issue right now with mine, also i found that i had enough misalignment between the motor and the screw that i was getting binding from that, have to look close if your using lovejoy connectors.it was compressing the rubber on one side more than the other, causing binding. $400 for a straight edge sounds pricy for a hobby.
if it's just for wood, i would think that a 4 foot level would be ok.
if you read the posts on the 2006 cnc and such, they mention one roller sometimes not touching, yet still seem to cut ok.
i would try simple first, get another screw and motor to do the other side. you can always reuse them on building another machine.
dan
ps hope this helps
btw if you get 50 in/min with 10tpi screws ($6) i figure you can later change them to multistart 2 tpi and get 5x that.

[LeeWay]

First off, congrats on getting the machine up and running.
Things I see that are contributing to your issues, you are mostly aware of already.
You have some nice material there, but chose a design better suited to an MDF or wood build rather than a nice steel build.
You already see some of the issues you can have with the round bar as guides. The only way I would ever use anything similar is if they are fully supported. Simply stated, there is no other way to get the flex out. The gantry being driven by one screw is even worsened by placing that screw on one side high and in the way rather than under in a central point.
I think most guys that use this type design do so because it's cheap and fairly easy to build.

My router design uses basically the same materials as your's but it is extremely rigid. I always design in adjustments. There is no way to get something set perfect when building and expect it to stay that way indefinitely.
My design uses flat bar bolted through the tube. This makes it a solid member at least in one direction. I used braces in between on the X and nothing way needed on the Y. The Z carriage actually helps on it. I do drive mine with a single screw on the X. Because of the adjustments built in, I was able to eliminate racking by preloading the gantry to the X frame and rails.

I used roller skate bearings on X and Y and some small linears on the Z. These really save real estate.

Take a look here at my build for some ideas on where you might strengthen yours up.

http://www.cnczone.com/forums/showth...2&page=1&pp=40

I'm not trying to say that my design is the best by no means, just that I was able to produce a very strong and fast little machine using some of the same parts you already have. Good luck with this and I hope you get er running to your needs.

[justCNCit]

Well thanks for the input.

Yes, those are 425 oz in motors from Xylotex with their 4 axis drive box, the plan is to add the 4th motor to the Y at some point.

The motors seem to be performing fine, I've noticed no missed steps. The box is configured for 1/8" step so hopefully if it does miss steps it won't be noticed.

I was thinking about this problem today and maybe have a solution. I will bring in the parts of the frame under the Y axis slider rod to the machine shop, and have them machine them flat as far as their mill is able.

Then it should be only a matter of dialing the height of the bars in, ofc, using precision rod this time, and on both sides, then aligning the rod in X and Y to the milled bar and to its opposite rail.

I figure this should fix that problem, which seems to be the most obvious and important problem at this time.

I guess the second problem is to figure out how to gear down the router, but one thing at a time.

[justCNCit]

My uncle had the idea of using rails in this way. After that I went to school and took machining then was able to build this in my garage. The garage was put up in order to work on jeeps and build a CNC and then produce designs using Rhino 3D.

Since my access to precise machines is very limited everything has been drilled to .5" holes and 10mm bolts pretty much make up the whole thing, till the Z axis which is comprised of alot of 1/4" screws.

The only components made in the machine shop were the milled faces of the aluminum bearing blocks