How to find the error of my CNC

[feldah]

I am in the process to build my CNC. Now I am wondering how I could find the error of my CNC when finished. I had the idea to use a steady and leveled laser and a moving light Sensor clamped in my Router. But I guess there are existing ways to do it. I assume that my machine is leveled and squared after I finish building it. So that I only expect an error from irregularities in my steel beams.
Size of machine will be 4000x x 1400y x 250z

Sent from my MI 6X using Tapatalk

[MikeyC38]

Hi feldah
I'm not quite sure what you mean by your question but what I did was to draw and cut a number of simple shapes (e.g. Square and Circle) of around 100mm and then check with a pair of accurate digital calipers what the final object size was compared to the drawings. At the end of the day, it is the cut accuracy which is the most important. BTW on my machine it was 0.01mm which is good enough for my uses..

[paulus.v]

Here (http://www.mycncuk.com/threads/11987-Help-please-Dip-switches-and-mach-3-settings?p=106415#post106415) I wrote a few things about alignment. Maybe it helps.

[feldah]

Oh both are great ideas. Thanks guys. I will definitely use them. What I was referring to specifically is the z deviation. Sorry for my bad introduction. My idea was to take a long enough spirit level and put a small laser on it. The spirit level would be placed at one end of the table in parallel to my y axis. Now I would draw a line on the wall which lays behind the other end of the table. The further away the better. To find the exact same height I would use a long hose filled with water. Then I would make a mark every inch on that line.
Finally I would adjust the laser so that it's placed on one end of the level and pointing to the first mark on the wall. When that's setup I would attach a light sensor to my end mill. Beginning at one end of the x axis I would lower it until I get a signal from my light sensor. Now moving on I take several data points running down the x axis. When it finishes I would place the laser an Inch further on the spirit level and let it point to the second mark on the line on the wall. I hope that explains it better. What are you guys thinking about it?

Sent from my MI 6X using Tapatalk

[mekanik]

I have no experiance with lazer levels. but from what i have seen you get a rather large red dot, and i think it would be difficult to acurately determin the centre of this. I do however have an optical level that is very easy to read your target (could be a rule on a magnetic base) the level will only focus if it is about 1Meter or so from the target but it does offer good magnification on the target, ideally you need a precision straight edge and precision level.
Regards
Mike

[routercnc]

I bought a precision level to do this. The 300 mm square frame type and accurate to something like 0.02 mm per meter?. Cost about £100.
Rest it on any piece of wood or box section that is long enough to touch both surfaces and measure any relative angle differences at any position you like along the beams where the rails will sit. Or start at one point and shim the level until it reads zero and check relative angles at other positions.
You need to do a bit of maths to convert the angle into the exact height differences based on your rail spacing but nothing too complicated. Then you can shim the rails level to each other and on the same flat plane etc.
I did this I one of my videos installing the Y axis onto the gantry. See YouTube channel.

[feldah]

Ok have a new idea. But maybe first why I want to know my error. I want to use a software like Autolevel which can adapt my gcode according to a error matrix.
What about putting a water filled basin under my accessible area which would be leveled by nature. Now I would put a voltage source in my water. Then I could clamp a simple voltage sensor in my router. And now I could take measurements when my sensor hits the water.
Simple and easy. What do you think

Sent from my MI 6X using Tapatalk

[mekanik]

I don't have a cnc machine but have been following the forum for a couple of years now, my understanding is that there is software available that will probe your workpiece (bed)on a user defined grid, and calculate any offset required for the z axis to machine thin components(PCB's ect) that are not totally level when secured to you bed. This should be able to machine a flat surface ?
Hopefully someone with experience of same may be able to offer better advice.
Mike

[feldah]

Yes I think we are talking about the same software Autolevel. But my problem is that I wether have a leveled bed nor do I have leveled axis. So my approach would be find the error of my axis then install the table and mill it with corrected gcode

Sent from my MI 6X using Tapatalk
https://uploads.tapatalk-cdn.com/20181220/4c922901f186c7fdde7ac9dca099637a.jpg

[mekanik]

Right now i can see what you are trying to do, i would go with the automatic level option i mentioned, should be able to hire one for a day, something like this.https://www.ebay.co.uk/itm/Leica-NA-820-Automatic-Optical-level-Site-Levelling-Tool-with-Case-USED-14-2/113335989640?hash=item1a6359f588:g:ERwAAOSwfPlb0wI ~:rk:7:pf:0
magnetic base with steel rule and sweep your X axis profile.
Mike

[paulus.v]

Why don't you just epoxy level both axis and you are done? Here (http://www.mycncuk.com/threads/9369-Here-we-go-again-MK4?p=102016#post102016) is a good example and there's a lot more information about it on this forum.

[driftspin]

Why don't you just epoxy level both axis and you are done? Here (http://www.mycncuk.com/threads/9369-Here-we-go-again-MK4?p=102016#post102016) is a good example and there's a lot more information about it on this forum.Epoxy yes.

But... a bolted frame. Is it to be moved a lot?

So is it rigid and tight fit so to be able to take apart and together exactly the same ?

If there is any room for error .. it is no use, not really

Maybe a dowelpin system might work to get the same fit every time.

Btw my welded steel frame had +/- 1.5mm @ 1850mm distortion after welding.
Warping and twisting in all directions.
Needed about 7mm epoxy thickness to level out, and have at least 4mm at the high spots.

Grtz Bert.

Verstuurd vanaf mijn SM-A320FL met Tapatalk

[paulus.v]

I cannot see why and how would you disassemble and/or move a 4 meter long cnc. I see it bolted down on a thick concrete floor in its final position and then epoxy levelled.

Grtz Bert, in your place I would have poured two layers of resin. For let's say a minimum 2% of resin curing shrinkage at your 3 mm thickness difference you've got a 0.06 mm final level error. It is not a lot but it could be avoided.

[driftspin]

I cannot see why and how would you disassemble and/or move a 4 meter long cnc. I see it bolted down on a thick concrete floor in its final position and then epoxy levelled.

Grtz Bert, in your place I would have poured two layers of resin. For let's say a minimum 2% of resin curing shrinkage at your 3 mm thickness difference you've got a 0.06 mm final level error. It is not a lot but it could be avoided.Hi paulus.v

yes. i agree.. 0.06 .. it might need another pour..
I hope i can work at that precision level and size any time soon and notice problems in my build.

I am gonna do a laser beam check in a few weeks..


On topic.
If it is at its final position and all diagonals and braces are welded ... go for it. Epoxy level.

Grtz.. Bert.


Verstuurd vanaf mijn SM-A320FL met Tapatalk

[feldah]

Hi,
Yes a 4m CNC is not really the size of Carry on luggage. But it's meant to be shipped in a container. In near future. Yes and I am planning to use dowle to reassemble it. On it's next destination I plan to use Epoxy to get a better precision. You think it will be a big issue once I applied Epoxy to reship it in three years?

Sent from my MI 6X using Tapatalk

[routerdriver]

If the machine is a router,the first cut will be skimming the spoilboard flat.The surface will be determined by whether the tracks are straight and in the same plane when viewed from the side.If your concern is about the amount that the beam of the gantry sags by as the weight of the Z axis moves,build some extra support.If you truly need accuracy,as opposed to resolution and repeatability then your project is probably outside the realms of what a homebuilt machine will ever be capable of.As an example;I have known an apparent increase of 0.006" of tool length as the machine went from workshop temperature to constant operating temperature.The high end machining centres will deal with this sort of thing.The price reflects the fact.