China CNC routers - 3020 oldschool vs. 3018 GRBL control

[Koen]

Have you finished your build Kjelt ?

[Kjelt]

Have you finished your build Kjelt ?

No thats going to take some more time, I got ill (flue) for the last two weeks, still in the process of getting better.
But almost all the parts are now in tha house, it is now making time to assemble and adjusting.
I am currently placing the electronics in the cabinet, since the cabinet is steel drilling and cleaning the holes for the connectors is a PITA.

[TinkeringSteve]

Kjelt:

Hey, I'm not asking you to "shut up" ;-)
I am about as mechanical as your average software dev. I am glad of being warned here about some things.
After watching dozens of youtube videos and what people were able to do with those chinese things, reading them being called "useless" seemed unduly harsh.

Guess I have read (more than) enough of the ramblings of experienced, older pro EEs on forums, telling the kiddies who try to get into their new hobby to either spend $ $ $ $ on their gear - or die.

Slightly bemused after watching the recent discussions on how many 100 bucks to spend on a device that heats a metal rod with the mass of a penny (soldering iron with heating *inside* the tip?),
I continue to work with the 70 € hobby-Weller, with the tip I bought 2011, which doesn't look like I need to replace it anytime soon.
For some weird reason, it seems to do the job. So someone doing repair work every day may have a different experience, but I'm not that someone.
My whole setup here probably costs less than 1500€. Including the DIY workbench, lol.
Bought used, older lab PSUs, granted.

I'm not sure I'd dare to buy a used "real" CNC mill, though. Not to mention I couldn't buy a big one that is going to displease the neighbors more, I also probably couldn't tell if there's something slightly wrong immediately, and there's no warranty...

And what you spent on your machine is almost what my Hyundai car cost *new*, incl. climate control! ;-)
A more feasible option than that would be to have everything made by others, but that quickly adds up also to a high cost, with another set of drawbacks.
So I will try with the cheap stuff.
Should I learn that it really does not perform to my (not hugely high) expectations, and I do not manage to break them, I can still sell them and go other routes.

You know. To be able to get prototype  front panels accurately drilled, even if it was just plastic, not aluminium, would be quite awesome. As one of many examples of the more "light uses" I'd have for that.
This would beat the hell out of me sitting at the desk with the battery drill, going hole by hole, with roughly 1mm and 5° accuracy (no I don't have a proper drill, for critical stuff I need to visit my father, the distance is inconvenient )

[TinkeringSteve]

"Supported rails" were mentioned.

Could you retrofit that? Is it difficult / expensive?

[Kjelt]

You're correct, every application and budget has its valid domain.
I did react from my own experience having bought a rather decent 4,5k$ german cnc that promised all kinds of high accuracy specs that in the end was unable to deliver and ending up with lots of frustration has coloured my experience and response.
That machine is highly featured on YT and some people make incredible things with them, but overall it is just not what I really would like, namely a machine where the operator is the weakest link 

The same with soldering irons, if it can not transfer the heat quickly enough in the solder joint you will fry off the copper traces from the pcb.
Also the 3D printer that makes a vase which takes 4 hours, due to a small deviation it results in the tower of Pisa instead of a nice vase.
In that case those companies that let you use their (better) equipment, the makerfairs etc are really a good alternative, although only found in larger cities.

So hope you will enjoy your machine and that you are not getting frustrated by it as I did with my previous machine.

[Kjelt]

"Supported rails" were mentioned.

Could you retrofit that? Is it difficult / expensive?

You have round guidance bars that are only supported on the ends, those are the worst since they bent but for short distances (small machines) they are okish.
You have the same round bars but at the bottom they are connected to a metal bar, these are better since they have a support under the entire round bar, so less bending.
These both do have a drawback being "play" on the construction, you can wiggle them they are not very sturdy.
The best are linear rail guideways, these are hardened steel rods in a special shape where special matching block carriages fit onto. These have almost no play and are made to last also with high mechanical stress.

Can you retrofit? In theory yes. In practice all the mechanical distances esp in vertical direction should match exactly. With tolerances of <0,1 mm that can be tricky.

[Kjelt]

Is it expensive?
Depends on the quality.
You can buy chinese linear guiderails with carriages for <$100 for a meter.
If you buy HiWin Taiwanese brand linear guiderails you pay $200 for a meter with two carriages.
If you buy Rexroth you pay $1000 a meter and two carriages.

Now these three are not the same, but the Chinese you need luck for a decent manufacturer, the HWin are very good and good enough for my machine, the Rexroth is for the Dubai people 

[flolic]

"Supported rails" were mentioned.

Could you retrofit that? Is it difficult / expensive?

It is possible. But there is no sense buying something that you already know is not good enough, than investing more money to make it better. You know what? It will be better, but not good enough.
Those machines are good enough to do cuts in plastic and light cuts in aluminium. But ask yourself, are you buying CNC machine only for cutting holes in front panels? If yes, you'll be ok with it.
But soon you will want to machine some heavier part, like custom Al housing, and will find out that machine is simply not rigid enough for that job.

The best way into CNC world is to build your own machine. But that can be easy or hard way, depending on your knowledge and possibilities. I have built my own machine few years ago and it cost me
around 2000 to 2500 EUR to the point I can fully use it (it's not 100% finished). Critical parts are sourced from ebay, like (used) NSK linear rails, Kuroda grounded ballscrews, stepper motors and drivers,
USB motion controller, water cooled 2.2kW spindle... Other mechanical parts (T-slot profiles, plates, fasteners...) are sourced locally or within EU. Machine working envelope is 300x300x150mm.

I can cut in aluminium just as easy as in plastic with final part tolerances of 0.01mm. And machining speed can be insane

Check this video:

[flolic]

And finished parts:

[TinkeringSteve]

flolic:

As for retrofitting - yeah, I was expecting that it would probably not be worth it.
But better ask than just assume

Sure, being able to do that (like your custom alu case) would be nice.
But it's not what I started out with as requirements, and I should slowly but surely put a stop to the scope creep that has already infected this endeavor

The more of an own project the CNC becomes, the more useless it is to me - because it takes time off the very reason why I want it in the first place.
I do not plan to start what I actually want to do only in 3 years or so...

I have no mechanical workshop. I will not buy some 1000s worth of tools anytime soon. And no room to have an actual work shop either.
For those who haven't guessed by now - I do not have a house, I live in an appartment with 2 rooms  (not counting the bathroom).
So forget about any monstrous "real CNC". It will have to be the 3020/40 or so. (plus enclosure I build).

I will note things mentioned here about better, somewhat costlier options, though.
My father may actually be interested. He has more room than I do.
But there's the (not so) little problem that he doesn't understand English, so most of the software (or information about hardware modding) will be difficult to use for him. I guess it's easier if he has roughly the same gear I have and I can always help him with stuff.

[ironcurtain]

I don't want to hijack the thread but it's interesting. I have been looking into small CNCs that can work with 3mm aluminum and smaller gauge soft steel for front panels and such... tired of drilling my hand though it works and I have a vertical drill press. So far nothing clear as to what is available that can work with plastic/ABS/PP, aluminum and soft steel at a small scale. I got puff already from Chinese industrial rigs when I saw the IR reflow ovens they sell cheaply...

[TinkeringSteve]

As for Air vs. Water cooling:

So this means for air cooling, high pressured air is blown onto the spot where the routing is taking place, like it would be with water?

For what materials is that necessary?
Would just spraying the surface with cutting oil help much?

[Kjelt]

As for Air vs. Water cooling:

So this means for air cooling, high pressured air is blown onto the spot where the routing is taking place, like it would be with water?

For what materials is that necessary?
Would just spraying the surface with cutting oil help much?

AFAIK air only is solely used in milling plasticks/acryl/polycarbonate that kind of material.
For light aluminium you usually use a mist being generated by compressed air blowing small amounts of water mixed with some additive, aka minimalschmierung.
For steel you use a flood what looks like a waterfall.

[TinkeringSteve]

But you could use the water thing for alu? It looks like that's what the somewhat more expensive (for China kit ) have. If that's less loud than air pump, as was hinted at earlier in the thread, I would prefer it.

I have seen some people just spray "cutting oil" onto a PCB before milling, no other means of lowering temperature was used.

Btw, are there tools for, and would it be feasible, to use the router also for polishing a piece of wood? Like, before you apply linseed oil treatment to darken wood, for instance, it needs to be reasonably finely polished. Heh, I guess to polish the edges at 45°, 5 axes would be nice, but that'd be overkill :-) Hmm, or are there maybe tools that are angled 45° themselves...

(please help me find more types of work I could push onto the machine, to help me justify to myself  getting the bigger model, ok? LOL! And again, it may take hours, if I'm not doing it by hand, I don't care)

[Kjelt]

You mean a waterflood for alu?
Yes could be done but you need some form of tray to collect the water, filter it and pump back, recycle.
I never saw that on affordable machines only at the pro cnc stations >$60k but I haven't seen everything of course, Perhaps someone working in cnc could advise.

Polishing, that is a different ballgame AFAIK, milling is high speed small surfaces, polishing is low speed big surfaces. Never saw it check youtube if you see something like that. Polishing needs a steady downforce, I would prefer to do it by hand or polishing machine by hand, more control.

More types of work, you could put a pencil in it and make drawings, put a rotating cutting tool in it and cut plastick stickers, perhaps some kind of laser for etching   

[langwadt]

Is it expensive?
Depends on the quality.
You can buy chinese linear guiderails with carriages for <$100 for a meter.
If you buy HiWin Taiwanese brand linear guiderails you pay $200 for a meter with two carriages.
If you buy Rexroth you pay $1000 a meter and two carriages.

Now these three are not the same, but the Chinese you need luck for a decent manufacturer, the HWin are very good and good enough for my machine, the Rexroth is for the Dubai people 

you can get mgn12 much cheaper, I think since they begun to use them in some 3D printers, e.g.

https://www.robotdigg.com/product/493/GCr15-MGN12-1H-500,-MGN12-1H-600-or-MGN12-1H-1000-Linear-Rail-and-Carriage

$54 for 3*500mm with carriages

[TinkeringSteve]

Hehe, "more types of work", it wasn't fully meant serious, I already have some ideas

About the spindle:

I found out that the china spindles, if you apply enough sideways force, e.g. 18N, can have play of up to 0.2mm, as there apparently is some rubber dampening that holds things inside.
I guess for PCBs, the effect would maybe not there, as there isn't much sideways force(?).
For milling plastic, maybe more so. Then again, I perhaps won't need super fine accuracy for plastic parts.

But just playing this in my mind: Say I was to buy the Kress spindle, which is said to have max 0.015 mm play: It apparently only  has a few fixed speed settings which cannot be controlled via software.
While the china box does not come ready-to-go in that regard, I saw someone open the controller box - and their PCB is prepared for that, you just need to solder some additional parts and wires, and snap, you have PWM control via software.

In what scenarios would it be advantageous for the milling software to control the spindle speed?

(P.S. as I saw some DIYers do buy such china spindles, so *if* I was to get the Kress, I guess I could sell the china spindle to get some money back. Just would suck if it later turned out something not immediately obvious is wrong with the kit and I need to return it Damn it, there's always something...)

[Kjelt]

I used the Mk1 Kress for many years, it is ok for starter. Disadvantage was the incredible noise and that the rpm started at 10k-30k while esp for wood and plastick you like lower speeds like 3k or the wood can burn or you have to increase the feed.
Current Kress can be bought with 5k start rpm which is better.

Most persons on the cncforum buy a chinese spindle with vfd, same price but way better than the Kress, lower noise, lower runout.
But I hear the quality does vary so you have to be lucky.
Second is the watercooling in a few years leads to leakages in the o-rings, rusting the bearings.
So I would advise an aircooled spindle but this has disadvantage of overheating at lower rpm since the fan is mounted on the axis. So if you go to lower rpm often you need to mount a high air force fan ontop of the spindle and ofcourse extra dustcover since it attracts huge amounts of dust.

I also heard of EU "companies" that buy the chinese spindles and then modify them with good western bearings even ceramic bearings and reduce the play. But quite costly almost doubles the price.

If you go china spindle you need to know just as with any other pro spindle that you can not run the spindle at 24k rpm from the start.
You first need to warm it up, starting at low rpm increasing slowly in a few minutes till 20krpm then you are set to go. Otherwise the bearings will not appreciate it.

[flolic]

I have water cooled 2.2kW chinese spindle on my machine, but currently I am running it dry. For majority of jobs in alu it only needs few hundred watts of power, so 2.2kW spindle is really an overkill.
But that capacity helps when you reduce the speed, because those spindles have low torque. I am running it from my LS industries (LG) vector drive VFD. It is running warm to the touch after a hour of use,
so no problem with (lack of) cooling. Spindle is very quiet, on pair with the steppers. What's really making a noise is end mill cutting action.

[TinkeringSteve]

Ah! Some important newbie tips, thanks guys, very useful, e.g. letting it run warm before high RPM, maybe there's a chance now I won't destroy my gear in a month *g*

Rusty bearings? Lol. Damn. Could you pump oil through it instead water? :-D Probably dissolves their O-rings.
But I guess I might try air cooled.
I have found ebay sellers who sell apparently better chinese spindles with "4 bearings" and claimed < 0.01 mm runout. They are called "ER11". (or is that a size? lol. Chinese things not always have product names...)

They also come with "VFD inverter" control box, I guess if I buy a kit that already has VFD for the crappier motors, I don't need to buy that also, are those usually compatible so I can use the kit's own one?

EDIT:
Btw, what kind of tasks would you need low RPM for, so that air cooling may be a problem (unmodded)?

[langwadt]

Ah! Some important newbie tips, thanks guys, very useful, e.g. letting it run warm before high RPM, maybe there's a chance now I won't destroy my gear in a month *g*

Rusty bearings? Lol. Damn. Could you pump oil through it instead water? :-D Probably dissolves their O-rings.
But I guess I might try air cooled.
I have found ebay sellers who sell apparently better chinese spindles with "4 bearings" and claimed < 0.01 mm runout. They are called "ER11". (or is that a size? lol. Chinese things not always have product names...)

They also come with "VFD inverter" control box, I guess if I buy a kit that already has VFD for the crappier motors, I don't need to buy that also, are those usually compatible so I can use the kit's own one?

EDIT:
Btw, what kind of tasks would you need low RPM for, so that air cooling may be a problem (unmodded)?

ER11 is the size of the the collet chuck

[Kjelt]

Yes ER11 can handle shafts upto 8mm diameter (sorotec.de has a 8mm collet) so that is the limit.

Slow rpm jobs : wood if it burns at higher rpm but you can increase the feedrate also.
Plasticks, they are tricky, if the  milling cutter runs hot the plastick melts sticks to the cutting surface and bam your cutter breaks off. The same can happen with aluminium if it sticks to the cutter that is why for alu you want some kind of cooling spray with some additive and for plastick forced air to blow away chips and keep the cutter cool.

The chinese sets with vfd are nice esp for their price. The vfd takes some time to program and understand, i have to start with that also soon.

[TinkeringSteve]

One thing I don't understand yet (at least ;-) ;-) )

Earlier in the thread, Magnum mentioned getting a better "base plate", linking to one at vakuumtisch.de.
With lots of holes. Because, it's actually a vacuum table, i.e. a table that holds something on it with vacuum?
But if you need some sacrifice plate anyway, under your actual material (at least if you need to drill / route through it), then you could just clamp it on the base plate/table (whatever that's really called) and be done?
Why vacuum?

I guess the recommendation was maybe not mainly due to the vacuum, but because it looks way more massive than the one that comes with the kit, right? That one looks like "more air, less metal"
But if I put some MDF (or plastic?) sacrificing plate on the table, won't that give the whole setup more stability than a flimsy "alu profile" table alone?

EDIT:
Oh and by the way, I saw the CNC kit comes also with a Lathe - an extra stepper and a thing to hold tools that are then rotated.
Does someone know what that is *practically* good for, in this china CNC edition?
One thing would be just for positioning something round to then drill holes into or cut something out.
But could you actually do lathe work with that, i.e. get a cutting tool for the Z-axis that, say, cuts away the 90° edge of a cylindrical piece of soft alu or plastic to 45° and stuff like it?

EDIT #2:
Hah! If the "lathe" motor is too weak, why not do it the other way around - fixed blade, rotating workpiece?
Just that you probably cannot use a very long cylinder material, as the forces on the spindle would get too high... But funny idea ^^

[Kjelt]

Normally you mount a t nut alu plate as a first base.
Then you can:
Or place mdf plate on it and mill it level with your machine
Or place 4mm or so PVC strips with thin two sided sticky tape and mill it level with your machine.
     Place and mount mdf on the pvc

After that you can put a vacuumplate on it, mostly used for items that can not be easily clamped down or need to be very flush with the machine such as pcbs, thin plasticks.

You need a high volume vacuumpump for it not a deep vacuuum low volume since each leak will blow out the vacuum
If you need to cut through the object then place Between the vacuumplate and the object some special rubber sheet that passes the vacuum and to preserve the vacuum plate.

[TinkeringSteve]

Hey,

so I ordered one with the ball screws, it came with water cooled, but I bought an additional air cooled 800W spindle with 4 ball bearings.
I need to assemble it now ;-) (and first, build a damn table that's big enough )

Seeing in a YT assembly vid that the side motors are already on there, although with some protection material around them, I can imagine that if the transport service ignored the "this side UP" arrows, there could have been angular strain on those...

So:
After assembly, what initial tests should I do to make sure everything is allright, including accuracy, nothing bent or whatever?


Btw, I ordered a beginner CNC book now, so I don't need to bother you all with more n00b questions
Just that those books don't talk about specifics of the china cncXXYYt models. (there is one book that's more a stepcraft manual than generic cnc router book, that's not the one I ordered...)

For now, I ordered some cheap mechanical switches which will arrive soon, which should work for limit switches for now, but I don't want to drill threaded holes yet to mount anything until I know I don't need to return it... maybe wire straps or something will do...
I also found cheap (< 2 € !) chinese inductive sensors (shipping 2 months ) which trigger ~ at 4mm. Sounds nicer (no bounce?) and more long living - good idea, or do they have their own problems?

Some ppl on YT said they don't use limit switches for Z axis, they use soft limits. Really? Why? The motor is not homed then?

Also, from what I gathered, the controller expects normally-open switches for limit switches.
I do not like that. Broken cable -> crash. How about inverting it 1) at the sensor, and 2) again in the controller box, to have normally-closed logic? Makes sense?

Anyway, thanks a lot to you few guys who have been doing all the answering here!
Is there a CNC forum where there are people with specific china CNC router experience, and where they don't chase you away if you mention you have a china one? (i.e., not like going to HAM radio meeting and shouting "BAOFENG!"...)