How Accurate is a PrintNC / 3D printed CNC ?

[@InheritanceMachining]

I christen thee, CNC King Arthur! That is surprisingly good accuracy for the build. Even with the flex, you have the advantage of climb milling so you can worst case take several spring passes and never remove more material than you intend

[@Joe_Bandit]

Wow, you seem to have a KZbin habit nearly as bad as mine 😅

[@ChristophLehner]

Appreciate it man ❤️

[@marcus_w0]

Wow, two of my new favorite maker subscriptions are interacting. I feel like a father!

[@kitawithsomenumbers]

"woopsie" I feel that pain. Appreciated the honest insight on what we can expect regarding accuracy. Pretty sure I'm going to go the printnc route but it's hard not to have anxiety about capacity before dropping all the money!

[@ChristophLehner]

Yeah, for what it's worth it is a very capable machine. You can't expect a Haas for that price, and imo the biggest obstacle in all this is still in front of the machine, there is so much to learn and optimize.

[@julieta203]

This is actually very impressive for what is essential a DIY cnc and such a simple and economic build. I would be more than happy with these tolerances on Aluminium and imagine it would be near perfect routing mdf/timber.

[@jamesrowlands8971]

I'm able to get ~.01mm in mild steel with my NC on trimming passes, if I measure and adjust the gcode. But that's a semi-automatic process rather than fully automatic CNC'ing.

[@StemmProductions]

Man that are good news. Thanks to your videos i got the motivation to start my build :D

[@tuberroot1112]

There is a classic test piece a bit like this but with a circle ( cylinder ) between the square and the diamond. If you machine is accurate the circle touches the points of the diamond and is tangential to the faces of the square. I think there is a NIST diagram for this standard test piece. Hint: steel stuck in ali, use heat. Thermal expansion of ali is twice that of steel. To avoid damaging the tool , use vice grips but wrap the tool shank with some shim sheet ( or cut open a steel food or drink container ).

[@Dwarfgrinder]

I don't know what software you are using but if you are using linux cnc you can use backlash compensation to solve the 0.01mm ballscew lash.

[@josephjones4293]

On my mpcnc in al i would run pockets in 4 passes to clean up deflection. Sadly i broke my 1/8 carbides trying to cut some soft al sheet and havent been back to al since moving to the pnc.

[@ChristophLehner]

On my setup 2 times was definetly not enough

[@xyzspec82]

I still think that the bad surface finish is mostly due to the unsupported MDF area where the vise is clamped to. Maybe some loose components on the X and Z assembly also... Even the old version of my machine had better finish on "loominum" parts and it wasn't that sturdy. Keep them coming 👍🏼😎

[@ChristophLehner]

I measured it, since you last commented, and yeah you were right, the mdf flexes like a wet noodle, despite being 38mm thick

[@xyzspec82]

Cutting forces on aluminum are wild, even if it doesn't look that hard... Try to add more steel beams on the base, maybe it will flex a bit less. Or... Build a concrete base like mine 🤓

[@russell26596]

@@ChristophLehner A ply base with many layers will be a lot stiffer (like 2 x 18mm baltic birch) and still somewhat affordable. I think I'll be using a single sheet of 25mm for mine. Love your videos, they keep my project moving.

[@ladedk]

It even seems to be visible on the video, looks like the vise is shifting quite a bit.

[@thepichner]

Just ordered our parts kit a week ago and will be starting a build in the next week or so, very excited. Great videos!

[@Duraltia]

Two things that kinda come to mind ( as an Amateur myself ): 1. @01:38 The Z-Axis _almost_ touching the Stock Material *_WHILE_*_ barely_ skimming over the Vise was way too close for comfort... Maybe consider making a recess in the Waste Board for that Vise to fit in so you regain more clearance? I was sweating like a pig watching this 🔞Video. 2. Maybe consider using different feed rates for your Spring Passes? Keep the fast ones for the Roughing Operation to get things done but once you do the Final / Spring Pass reduce the feed rate to like ½ or ⅓, of that? Pretty sure the time you spent on doing additional fast Spring Passes could have been saved by doing a single slower one 🤔

[@ChristophLehner]

Hahaha yeah, i need to do something about the Vice, it was too close for comfort. You are absolutely right👌. For the finishing passes i should use some lower feed rates

[@axelSixtySix]

The engineer's two cents: if you're looking for "optimal tool paths on a semi rigid machine" that means the machine is not accurate itself: you duck parameters to get the expected accuracy allowed by the machine itself. Experience showed me the greatest handicap is always the machine itself. If you want a 200µ accuracy, the machine itself needs to be built with 20µ accurate tools and use at least 100µ capable components. Simple as that. If you did not learn to study a structure by simulation, a proper way to test a machine is to set the feed and speed at the bit specified abilities. Say the biggest collet of a machine is 12mm. You want to test the machine with a 12mm bit, at the feed and speed required by the tool. Into hard wood, for example, 11.000mm.mn @ 24.000 RPM with a depth of cut of 12mm: nothing fancy for a 12mm bit, just the bits requirement. Machine a part with these parameters and measure the machining quality on the part. If the cut quality is not good, or something went wrong, you don't have the right spindle for the structure of the machine. The standards always comes from the cutting tool. If the machine is not capable to run properly a 12mm bit, downsize the bit for a 8mm and run it at its feed and speed requirements, something like 8.000mm.mn @ 24.000 RPM with a depth of cut of 8mm. And so on until you get perfect results with at a bit's required feed and speed. For example, a Chinese 3040 cannot run at proper feed and speed a bit over 3mm in diameter. So there's absolutely no need to put a 1500w spindle on a 3040 as a 500w spindle is powerful enough for the structure. It can already holds 7mm bits which is more than two times what the structure and components can really hold prior to deflection or vibrations. People need to keep in mind that running a bit under its required feed and speed is wrong for the bit which overheats and get dull too fast, wrong for the machining time and the overall machine lifespan, and could be wrong for the material itself as it can overheat the material, add some burn marks and even bounce on the surface during very thin passes. What people believe as "good parameters" because the cut quality looks good, are mostly wrong parameters for the bit. That's why the real standard measure is the bit itself. When an engineer study a new machine, he always start from the biggest cutting tool the machine is meant to hold. The notebook can be as simple as "The machine should take 20mm passes in stainless steel with a tolerance of 50µ". All is said: you can deduce the required bit, the power of the spindle, then what is left is the structure and components to hold the forces corresponding of the resistance of the material against the cutting tool. Do your calculations, add a safety margin, select your components and the job is done. Hobbyists do exactly the opposite: select reasonably priced components (biggest allowed by budget), select a structure presumed strong enough, design parts to put all that stuff together, done. Then the question/problems waltz begins: what the machine is capable of ? Which parameters to machine acceptable parts ? And so on. Any hobbyist can learn how to study a machine, this is not difficult at all, you just want to start from the bit.

[@jordymoors]

So the bit says 24.000rpm... Many industrial, pricey ($150000) machines have spindles that don't even go up to 6000rpm. Now what? It's ALWAYS a matter of budget, as well as technical limitations, whether hobbyist or professional.

[@xyzspec82]

That gave me something useful 👌🏻

[@axelSixtySix]

@@xyzspec82 You're welcome.

[@Jimunu]

Didn't know Matt Hardy did cnc learning as a hobby between acting gigs.

[@justinwalker8364]

I think the weight and height of your vise is wobbling as you cut, you can see it in the sped up portion of the video in the beginning.

[@clayflannery2608]

I bet you can greatly improve this with some better workholding! you can see ringing even in your chamfer

[@ChristophLehner]

Yeah the mdf spoilboard is quite flimsy and the vice could be mounted more secure....

[@tuberroot1112]

IIRC he used 10mm pitch ball screws, so what is the step accuracy of the drive. A 200 step drive on a 10mm pitch gives 0.05mm or 2.5 thou steps, before you even look at mechanical rigidity. 10mm is great for rapids but you pay on movement precision.

[@W0LLL3]

Do you think it is possible to machine mild steel/stainless steel sheets (1-5mm)? Thanks!

[@ChristophLehner]

Are you talking about one time, or permanently? For the odd occasion you can make it work, but in reality its a compromise, the spindle is too fast, and the machine not rigid enough, as its a gantry style router

[@W0LLL3]

@@ChristophLehner Only occasionally. Thanks for the answer 😊

[@daniels4338]

How accurate are your calipers?

[@shrimpinpat]

im wanting to make my own cnc like this. how do you go about a pass with a different tool? how do you realign it all?

[@perw12345]

Nächstes Mal, wenn du so einen Excalibur hast, probier's mal mit einem Gleithammer. Nimm eine ER-Aufnahme, klemm den Fräser, steck ein wenig Gewindestange(20-30cm) ins andere Ende, mit zwei gekonterten Muttern am anderen Ende der Stange. Ein Stück Ausschuss, da ein Loch durch, an der Gewindestange auffädeln, und dann (kontrolliert) mit Schmackes auf die Muttern. Zieht sowas ganz entspannt und zerstörungsfrei raus.

[@ChristophLehner]

Klasse Idee Danke 😃

[@jannik3556]

Ein Zehntelmillimeter Abweichung ist doch gar nicht schlecht! Wo siehst du noch verbesserungspotenzial?

[@ChristophLehner]

Die 3d gedruckten faceplates die die spindelmutter halten, könnte ich aus alu machen, das sollte einiges bringen

[@jannik3556]

@@ChristophLehner Klingt auch nach einem schönen Projekt!

[@marcus_w0]

If I was you, I'd try to fill up the beams with concrete or sand. This should add a ton of rigidity. And weight. And weight is a good thing.

[@ilikewhenitgoeswub]

How much of the XYZ accuracy loss do you think was due to backlash in the ball screws?

[@ChristophLehner]

0.01mm is you typical import ballscrew accuracy, Biggest issue is the operator, then feeds &speeds and rigidity.

[@rafal7217]

Hi Christoph, You are the author of PrintNC?

[@dermeister314p]

This is not really a base case for the PrintNC, since you used advanced and expensive equipment to level some parts, like the one metal tubing.

[@RoboArc]

This gives me a little hope. Your machine isn't that stable, it's about as stable as my bug ass cnc. It's called a super heavy cnc. You csn find me on reddit

[@digitalghosts4599]

The way you're measuring it with the calipers is very inaccurate, so are the calipers :P If you want a good measuring tool, throw those toy calipers away - they might have 2 decimal points, but the accuracy is at best to 1. If you are into machining, getting good calipers, like those from Mitutoyo will really save you a lot of headache. Just make sure you get the original ones, because there are countless fake Mitutoyo tools. Get something for $100, ideally not digital, and it will last you reliably for a lifetime. Honestly though, for $2k machine it looks awesome! Especially how effortlessly it cuts into aluminum!

[@erwinz5926]

with "excalibur" you got me. thats a 20th of a mm. id call that accurate. depends on what you are doing. maybe not making front teeth or so with it. but essentially. booom mf. high five. what would be the spot to raise precision? filling the gantry with concrete or sand?! better rails or screws?? the first thing, that i would build with such machine would be replacement for most pla parts. but those are minor c0ncerns. actually.

[@user-xb5zu6zu7j]

Lol

[@MauricioHernandez-de8is]

Hi, great video, where did get the dual carriages on all axes design? Is very cool, Is your design?

[@ChristophLehner]

i have only dual carriages on the y axis. check out the website of the creator: threedesign.store/

[@MauricioHernandez-de8is]

@@ChristophLehner thanks, i wish you success in your channel