The chase for the best possible CNC surface finish! - Mirror finish up close, really close. 

Uh oh, now that you’ve got it milling steel, which aluminum parts of the mill have too many nanometers of flex and need to be replaced with stiffer steel parts? 2022 upgrade series in the works already! :D

Very cool stuff, impressive flatness on the facing pass being able to wring those two parts together!

You literally embody the spirit of what Im chasing in cnc. No compromises, just precision, regidity, and glass finishes.. Id love to get a rough number on the overall cost of the components, id like to build something similar scaled down a bit to a more desktop sort of size configuration if possible. Otherwise Im absolutely smitten with this build and your level of attention to detail here sir. Unreal.

Well done. Out of experience the quality of the endmill makes a huge difference. My personal preference is YG1 endmills. Slightly expensive but well worth every cent.

That spark show is normal on hard steels or just with certain cutters on steel. As long as there's no terrible vibrations or extremely short tool life, send it. There's a few parts I run at work that looks like a volcano erupting and those end mills removes atleast 50 lbs of steel before the possible need for change.

Very nicely done ! it behave very good. For the ball endmill, for what is working for me is have a constant or almost constant depth of metal to remove, like have a first pass with the ball endmill and then the finer step over, it help. also having something to cut, I mean something substantial help to have the cutter engaging and not rubbing so much. Awesome primer results, impresives i would say !

Glad to see you pushing out videos again, keep up the great content! this machine is one off the best DIY machines! top 3 for sure!

Benchmark to judge my own work? Dear God, please no! Anything I have done, in comparison, is like juxtaposing a couple of amino acids, bubbling around in the primordial soup, with Richard Feynman. But the thing that really gets me is that it not only does beautiful work, it looks incredible itself. You, sir, are an artist!

Just i quick thought from someone with no milling or metalworking background what so ever. Maybe you can "dress" your endmill. Just like you would do with a surface grinder. Use a ground wet stone and "mill" very, very lightly over it. With that you should be able to get any added up error from spindle, toolholder and tool out of the equation. And have a tool exaclty perpendicular to your machine axis. Of course this result in a some amount of rubbing because you ruin the back relieve. But depending on how minimal the dressing is, it might not matter. I really enjoyed this series and the result is just awesome!

"I still have some room for improvement" - and Deming wheel keeps rolling :) Amazing machine! PS. Geetings from Poland ;)

Regarding your wringing tests- to achieve proper gauge block wringing the flatness tolerances that are necessary are generally not achievable with milling machines, as the flatness of the wrung surfaces needs to be well under 1um flatness, but also in surface finish in the low double digit manometer ra range. It's much easier to lap a part into this condition than any other method. And of course, adding water is not a true test ;)

that sticking actually looks good, 5 micron or something like that. what I expected instead of sticking is just check surface flatness, mill 2 opposite sides, and check it on a surface plate I guess, but overall looks great - good result. Making thing stick like gauge blocks out of a mill - eeeehh, it is a dream, I think, starting from good mills you need even, not necessarily your equipment precision, and all other factors. check repeatability, two similar blocks dimensions faces flatness and just flatness

impressive. i really like granit and big rails :)

Very nice proof of the level of effort you have put into this machine. I'm guessing that the upgrade was not a small expense. The difference in finish is quite impressive.

your a legend Sir, what an amazing machine, im stunned by the results!

Try Destiny end mill for aluminum. They leave a mirror finish. MA Ford 6-7 flute great on 4140 and stainless.

Adding water will make it a hydrostatic(?) ring. It can be achieved with far less finish than what you have achieved. But what you get for a little bit dry, you are close, but you ask waaay to much of your CNC machine to achieve that :-) Nice work!

You went with the surface finish video;-) lovely! Maybe at some point, you could show us how you anodize those wonderful surfaces;-)

Hi, Fox!! You are my inspiration, i`m building my first cnc router now. Not out from granite, but with steel welded frame (it`s more easier in home conditions than work with granite)

Glad to see you pushing out videos again, keep up the great content

Impressive! Especially since (in my limited understanding) surface-ground parts are generally milled first then hardened and ground to final size and shape; the usual practice is to only mill softer steel (before hardening or after annealing to soften the material).

What an amazing achievement! Congratulations!!

I build my own shitty cnc for woodworking years ago but this is insane.. super nice build and pretty helpfull to get my next machine a little nicer ! and a joy to watch to ! i sub for sure !

Your machine and work are tremendous. Since you have the microscope view at 3:14, it looks like there is a dimpled finish that you mention. They are ~150 um in diameter, based on the hair. They are not regularly spaced. You mentioned that you thought that this was the vertices of the line segments that make the sphere. In my experience, if your machine is truly doing this because of the vertices, your servos could use a retune. If you are using acceleration feedforward, it can make the system "jump" on changed in trajectory like moving from one line segment to the next, if it is set too high. Also tuning "too tight" or to too high a bandwidth can do a similar thing. Backing off on your tuning can make following paths like this smoother, if maybe not as accurate. It would be pretty straightforward to set Fusion to a much looser tolerance so it uses less points. Then you could actually see if the dimples match the cam vertices. Having said that, I wonder if this is chip recutting? The dimples look a lot like chip recutting issues I have seen before. It seems like if you could diagnose this and fix it, you are looking at a near optical finish on the sphere, which would really be amazing. I have never achieved a surface finish where there was "one last" thing like you have. Amazing work!

The sparks are a normal part of hard milling. Looked like you were going a tad fast but you mentioned your cutter held up. End mills do not last long hard milling with poor settings

very good video..thanks for your time

Piotr, your videos are great! Now that you are done chasing microns, maybe next you can chase speed. (Not saying your machine is slow) I am just excited to see what you come up with. Keep up the great work.

Nice & Congrats Piotr! Tho, I would love to see some dimensional control. For example different heavier cuts & measuring the amount of "spring--back" towards the CAD dimensions. I think that would be a nice way to get some "true virtual" feeling of the machine rigidity. Kudos & looking forward to the next video! AS

oh wow, so amazing! looks like all the work was worth it. congratulations Piotr, this is so impressive. i'm not sure what you want to improve, maybe gcode that supports arcs? whatever else you come up with, i look forward to hearing more about it!

Wow, getting so close to being able to wring parts dry is insanely impressive. Get yourself an MCD fly cutter or touch them up with precision ground flat stones, I bet they would wring dry then!

Hats off to you that is very inpressive!

Impressive. I think I will make my first CNC to mill / turn wood where lack of accuracy can be described as, "character". What a wonderful result for you - hope you can share some of the products / prototypes you produce on it.

Hey Fox, like so many others I am inspired by your work here, thanks for sharing. I am part way through building an 2.4m x 1.2m CNC, mainly for wood - but of course, the dream is for it to do a respectable job on aluminium, and even look mild steel in the eye without peeing its pants. I have used composite materials - steel & concrete frame, epoxy levelling, the gantry is aluminium, steel and epoxy-granite. It's quite the mash up, from bits n bobs I had available. I'll make a video for my channel when done :) Thanks again.

Strictly speaking.. (correct me if I am wrong) but parallels are not mild steel. They are generally made from cast iron, stainless or tool steel. If they are stainless, then what you did there was impressive.

That was very impressive!

Fantastic work. Many congratulations. Apologies if I have missed this answer already but have you tested any of the Datron tooling?

Easy way to get a very good surface finish in aluminum is using an uncoated YG Alu-Power 3 flute 45 degree helix solid carbide endmill in a heat-shrink holder.

very cool build. I would like to see you mill som regular cold rolled steel that is NOT hardened. Milling hardened steel is the exception for most people, but milling mild steel is, for me at least, a requirement. Thank you for an excellent series.

Thanks Piotr the quality is pretty high. Nice work ! What's next high end tool upgrade ? A bit surprised there was no arc generation from F360, may be it influences the smoothness of the toolpath (depends on the post processor and the controller)

Think you need some diamond tools to test now

Excellent stuff! You've convinced me to restart my CNC improvements, not because I need it, just because I have to have it :-) Awesome! Footnote: Why don't you take the 5um passes just for the hell of it? Maybe on a smaller part?

Well done man. Great stuff as usual!! You'll to redo the torture test you did a year or so back.

cooling like crazy!

Hi Fox, Absolut great work, thanks for uploading... Please would you be so kind to share the F360-file for this halfball. I think, that there are some details that I don't understand. Do you think that there is a chance to mill it with not so much cooling water; to better see the real cutting process. Over all, I think this piece is a good method to check the accuracy of the machine Thanks.

Is it really a problem that the chips did glow? I’m not sure but I think thats about to be expected when milling hardened steel? Surface looks nice after all and you mentioned that the endmill also still looks happy.. Super great machine by the way! Very inspiring all the work and effort and thought you put into it!

Surface speed was way too high on the hardened steel, try around 2000 rpm on that size endmill with same feed. Aluminum finish is decent but polished endmill from a high tier manufacturer would give you excellent results. I like those from YG-1 which are very reasonably priced.

I love your work, but here are my two cents on comparing surface finish: don't compare a surface ground a long time ago with a surface just milled, because any surface finish/mirror dulls in time. Compare instead two surfaces equally freshly worked (freshly ground against freshly milled)

Playing around with a new toy is always fun . Enjoy! !

How a flat surface is obtained is compared to a high accuracy class surface plate. With an oil paint, e.g. Prussian blue

Just ... Awesome!

Very impressive, I can't imagine the videos really show the amount of effort you have put into that machine. Are you working in a climate controlled area? I guess you need to send it to someone who can measure the surface finish. Ra, Rvk, Rpk and Rk value. I always found dull looking to have a better surface finish. What is your profession btw ? Amazing videos. Keep it up.

La maggior precisione che io ho visto da sempre

Great Video!

I think your surface cuting speed is way to fast how it looks, I usualy cut steel at 150-200m. And second paralells are usualy heat treated. We do hardmilling at work as well. But try some mild steel next time.

Awesome machine you have! Can you maybe build a finishing/polishing endmill head?

Piotr please send the link for you cutting bits supplier

Hi Fox, one question. In the beginning of your epic upgrade you were talking about stepper and change to servo for a better surface quality. What are your findigs about this, is it an improvement for the surface quality to change from stepper to servo? - provided the same mechanical stability of the machine.

Try PCD fly cutter plz

Spindle speed and end mill diameter when you cut Steel? Spindle speed seemed high because the sparks

This having only 20k views is funny part..

Any reason you didn’t use flood coolant for the steel cuts? It would be interesting to see the difference in surface finish. Love your machine!

Hello Fox! Could you tell me where you buy your standard endmills? They seem to perform quite well for chinese ones.

Nice Piotr- I would be interested in what CNC controls / servos etc you decided on using...

is it possible to make a hammer with this CNC ?

So, you are using 2 servo's for the Y right? does that sync up pretty well? Really nice result! really want to get different parts for my build now I have seen this result.

How would the steel cut have come out with coolant, as was done with the aluminum?

Te stalowe kawałki stali były pewnie hartowane więc stąd iskry. S=4200obr/min f=850mm/min i chłodziwo wtedy Ae spokojnie 0.5mm. Widać to po powierzchni i dźwięku. Zrobiłbyś jeszcze test sztywności śrub i prowadnic. Zatrzymać silniki i działać siłą wzdłuż śrub napędowych mierząc ugięcia czujnikiem zegarowym na elementach osi maszyny.

for cutting steel you need less cutting speed. i mean speed at which cutting edge travels trough material. in this try you have sparks from too much rpm and too less of a chip load. your machine is absolutely capable of cutting steel with 6mm end mills (or 8mm roughing end mill). all you need is some research on speeds and feeds. good way to start is to look at cutting charts from tool manufacturers. here you can choose similar tool and download some cutting data www.helicaltool.com/products/tool-type/steels/4-flute-steels

Room for improvement? Lol. Impressive!

Cool! Which tool did you use for facing here?

Is the steal flakes burning off really a problem? If the heat is removed with the chips and there is no chattering, it is perfect, isn't it? As long as the endmill and the part are good, what more do you want? The surface looked great.

Tha looks really shiny 👍

I would like to see your machine to test some cheap DLC endmills, i do get great surface finish from those endmills, but i belive you will get mirror surfaces for sure! 8mm 3 flute is the one i have, and i use that one as a finisher. Also, you went way too slow on the steel ;) way lower rpm or faster feedrate :) look up FSwizard as an app on your phone, that one is great for getting a ballparked feeds and speeds, and you get quite abit of materials on the free version.

better tools definitely!!!

It looks like the tormach vrs datron video from nyc cnc? If it is similar challenge it would be great to compare the times! Of course you would nedd same tooling and fusion setup with i believe they go over in the video? Btw what axis controller do you use? Is it pwm or step n direction yet? I dont believe its as much software that upgrades the machine as much as it is the axis controller n drive n servos n yours looks pretty darn good. What is your max feedrate and rapid? Nice video!

Hello Fox. Since you are in Europe where do you buy your measurement granite plates? Are they in-size? Is there any online shop I can have a look? Thank's my friend. you are being awesome. :)

What RPM are you using for steel ? thanks

Perfection. o/

what is the repeatable accuracy of your homebuilt CNC?

Is that the Chinese 2.2kw spindle?

i cant believe you bailed when the sparks started :(

Are you from Poland?

Do you not own a surface roughness tester? Lol

How does this compare to Kern?

climb cuts! = bad

Information. When you dint try Toolox33+44 try that steel. Almost perfect steel to mill shiny.