Scratch Build – Complete Project QUAD - a Quick and Dirty PC/Spindle cooling with CNC Router attachment

This week some collet chucks and collets arrived from China.
I also ordered some plastic chuck holders for an ATC rack with them.

Unfortunately, they are incompatible with my spindle.
The arms which actually hold the chuck sit some millimeter lower than the top surface and this edge would hit the bottom of my spindle during tool pick-up




Now I can obviously flip it around and try to press the dowel pins in the other direction.
But first of all I think it looks kinda wierd and this side would quickly accumulate all sorts of debris...




So I decided to do my own ones. I still have some leftover stock of 8mm acetal, which I used for that.
With the width of the stock I can do 7 at a time (which is a shame, because I need 8 ) and first op was to drill the holes and fix the stock through them to a sacrificial board.
I refrained from double sided tape, because this would gump up my endmil in op no 3...




Op 2 was to mill the top seat for the chuck. Op 3 is to mill the bottom recess.
I did not want to do this in a seperate clamping with the consequent indicating nightmare (or an extra fixture, using the holes for the dowel pins).
Instead I used a slot mill.




That worked reasonably well.




After milling the outer shape I cut off a bar which connected the arms and reduced flimsiness during milling by approx. 1.3%
With hindsight I should have taken larger stock, so that I could have fixed that area better...




It is a miracle none of them broke and all 7 came out intact.




Now the chuck sits flush in the holder and I am good to go.

 

You know that acetal wouldn't break. -Maybe flop around like a flipped door stop, but not break.

 

Those arms flexed about 2mm when the 4mm endmill went through (at just 1mm DoC).
At the final cut, after one side was already loose the other one flexed 3-4mm before the endmill cut through and the arm snapped back.
Kinda nerve wrecking. But I expected something like that and set the feed to just 300mm/min..

 

Verry nice and detailed build, keep up

 

Thanks, mate

Actually I am almost running out of time. When I ordered the router end of April, the estimated delivery date was week 32.
But I got a payment request this week from the supplier, so it will be delivered probably next week, or the week after latest.

Time to step up the lazy pace and do the final parts...

 

What always bothered me in the design of this router are the gantry connections.
They use these zink die casting angles, which are made to connect aluminum profiles.




These have some alignment pegs at the rear, which fit into the slot of the profiles.
But the instruction is to pry them off, because these angles need to sit flush with the plates...



In any case, I am not really convinced about these angles. Not with their squareness, not with their strength.
Especially the ones at the side I feel are undersized, as they are a major element to reduce lateral flex.
So l will add a little bit more beef...

Starting with some high strength aluminum stock.




First a moderate roughing with 2mm ap and 6mm ae.




Haven't made pictures of the following ops, but after finishing this already goes in the right direction.




I cut the pocket with straight runs, 5mm ap and 12mm ae and that was borderline. These hobby mills are just not sturdy enough for this kind of load.

 

So for the 2 side angles, I chose a different approach.

Again some beefy aluminum, which destiny is to be converted into tiny chips...




Same roughing strategy for the outer shape, as this went quite well.




But for the finishing I went 4 times with 10mm ap and 0.25mm ae, rather than 40mm ap and 0.1 ae as I did on the smaller angles.
And, as expected, the surface quality is somewhat better.




I cut off the surplus material with an angle grinder and decked the part to the correct width with a face mill.
Fixing a triangle is rather tricky, as the moving jaw of the vise only pushes against a sharp corner.
Took off only 0.33mm per pass, not to risk any movement of the part in the vise.




For the pocket I changed this time to a somewhat adaptive clearing strategy. Still programming the code by hand and this was more than 1000 lines of code
It is really time I dig into CAM, once the router is up and running.
Basically, I cut repeated half circles with 13.5mm ap and 0.4mm ae with a feed rate of 1200 mm/min, which is the max the stepper motors can tolerate anyway.




4 layers to a total depth of 54mm...

 

I find these far sturdier and they should improve rigidity a little bit...




Need to cut 4 more for the X-axis, but I ran out off stock...

 

Heavy metal!

 

You milled a whole bracket's worth of material away just because of clamping. Yep, Turning into a real machinist.
The cheap party would have killed their arm hack sawing triangles, and surface milled the rough cut after eyeballing the angle in the clamp.
I do like your mill enclosure, though. straight out of the bin.
Jokes aside, the brackets are beautiful, and far better than they need to be.

 

Not quite right. It is actually straight before the bin

Yeah, you are right. 60% waste to do these parts
I perfectly understand why the manufacturer of the router kit uses the cheap zink castings...
But I wanted them to be beefy and dead square and I couldn't think of a different clamping strategy

Nevertheless, I need to become more cautious with optimizing stock.
I paid my local metal dealer a visit this afternoon. They are halfway back to normal after the Covid situation calmed down over here.
So you can go there and buy spontaniously, instead of only ordering online. But you still cannot walk into the warehouse and browse their scrap bin.

I replenished my square stock, 50x50x500mm and paid a shocking 40 EUR
Aluminum prices up 60% in the last 12 months




Still, I need to do 4 more...

 

I just checked the latest price of the router and it is 10% more expensive than it was in April
The current economic situation sucks...

 

I am done with those brackets for the time being.



The thick side will get an M8 thread and the thin side a Ø8,4mm hole.
But I need to measure their exact location at the machine, where the mating threads and holes are.
I assume they are 25mm away from the corner each, but this is just guesswork, looking at some YT assembly vids.
Don't want to f..k them up just now...

This part of the design with the flange nuts and washers to barely cover the T-shaped hole in the cast bracket I find particularily shitty.
There is almost no contact at all. Having a thread directly in the bracket is imho the far better option...

 

I set-up the mist cooling controller today.

First off I milled and drilled the cutouts for the display and the switches into the box.




The display comes with a 3D-printed frame/spacer, so it can sit flush on the lid.




Next came a mounting board for the controller out of some scrap acetal.




And the wiring. This will become eventually a little bit more tidy, after the box is finally fitted to the enclosure (a proper one, made out of alu profiles and plexi - not waste bin paper )
For this I need to disconnect the cables anyway, in order to pass through the enclosure.
While the machine control cabinet is inside of the enclosure, this box will be obviously be outside of the enclosure, same like a box with a push button for manual tool change and the E-Stop button...

 

The controller has 3 switches.

One to activate air.
One to activate the coolant pump. This one is in parallel to a relay, which is activated by M7 G-Code, so I can switch on mist either manually or by the milling program.
One to activate a fast pump mode, in which the pump spins at max rev (while you press the button)
This is usefull when the 5 meter coolant line from the pump to the nozzle has fallen dry.
Or when you quickly want to flood your endmill/work piece during operation.




Then there are 2 potentiometers.

One regulates the pump speed, means coolant flow.
From 0 to 240 ml/h. Whether these values are actually accurate or not I cannot say.
I will not run the thing for an hour to see if the pre-set amount of liquid is missing in the tank.
I'd rather trust my eye sight and stomach to set the proper flow rate.
But you can adjust the mist ultra fine and precise and more importantly ultra convinient from outside of the enclosure, without having to fiddle around with needle valves next to the running endmill during operation.

The second poti regulates a spit mode, from 0 to 8 seconds in 1 second intervalls.
Spit mode means when you activate the coolant, the pump spins at max rev for the preset time.
Like when you want to soak the endmill before the first cut.

I made a first test run with the cheap and cheerful nozzle from Amazon and it worked just fine. So I will reserve the other one for...whatever




I think this is an excellent product and much better than what else you can find on the market at this price range.
It is made by a Youtuber, who started a side hustle with this.
I love the modularity. You can get the stuff from a single piece to a complete kit, whatever serves your need best.
In my case, with the pump inside and the control box outside of the mill enclosure, buying individual components rather than a complete kit was the better option...

Here is a link to the stuff, in case someone is interested...
https://www.end-cnc-shop.de/geloetetes/26/pumpensteuerung-pro-1.0-inkl.-kabelsatz-und-1.3-display

 

Hmm...I don't know exactly why, but somehow I have the feeling this is going to be a busy weekend...




But before final assy, I have to modify a few parts, make a completely new cross plate and, unfortunately, have to wait for a new X-axis BS, because the one they sent has the wrong pitch...

 

NEAT!
Couldn't you just change the tuning for the odd screw?

 

I could, but I am using servos and in contrast to steppers I do not want them to run too low rpm.
The machine is originally equipped with 10mm pitch on X and Y and 5mm pitch on Z.
But this considers steppers. The faster they turn the less torque they have.

The servos max rpm is around 3000. With a 10mm pitch this is 30m/min.
Impossible to mill at that speed and even too much imho for G0. It would accelerate and deccelerate more than it actually moves...

The X-axis BS is 16mm in dia and 1m in length. So because of bending forces I will not go higher than 2000 rpm anyway, but even 20m/min is too much for my taste.

Realistically, I will mill at 4-5 m/min, depending on the material sometimes maybe even less.
And I find 500 rpm or less is too low for the servos.

So I specifically ordered 5mm pitch for X and Y.
It is more precise, better for acceleration, has a more reasonable rpm during milling and 10m/min G0 ist ok for me. 5 seconds for the entire +20kg portal to go from left to right is already pretty fast.
At the end, this is still a hobby machine.

Besides, they already gave me a warning by email yesterday that the spindle may not be correct and unfortunately they were right.
I confirmed it today and the replacement is already under way...

 

Spindle #3. Post money collecting customer service though? That's awesome.

 

It is in fact one of the reasons I went with this company.

The design of this router itself has some pros and cons. Since I have a second mill at my disposal, I will try to eliminate the cons as best as possible (at least as far as I see them...)
So when I checked various options before I made my decision, the actual router design was important, but not the key decision making factor.
From a pricing point the possible candidates were all within a 10% range, so no real factor either.

One factor was their online shop. For almost every singe item they sell you can download the spec sheet, drawings, sometimes 3D step files, etc... Top notch documentation up front.
Others believe they can sell a +3000 EUR piece of milling equipment and the only thing they need to provide upfront is a sketchy 500x500px picture...

Second factor were up front phone calls. Now while I agree that many shops have different sides pre- and post-purchase, the one I have a bad feeling with already pre-purchase is certainly not the winner...
While they offer some 50+ different routers, with different grades of quality and rigidity and then again different sizes for each grade, they take a lot of time trying to understand your real needs.
When you have certain preset preferences they actually question them, but not to sell you more expensive stuff, but to rightsize it to your needs.

And finally the overall rating in different forums.
CNC forums in Germany generally suck (like so many things over here )
They are full of attention whores, who boast about how fast their router is, in which micro tolerances they can work, and so on...
Professional CNC operaters dissing the average hobby Joe, who is just seeking some advise for his current problem. Telling him his stuff is total crap which he better throws away and are happy to start a pissing contest at any time in his own thread.
And when it comes to different makers or brands, local or China, it is more like talking about religion...
After browsing some of them for general information, I chose to stay away from that.
Despite this overall poor source of information, however, you can derive a trend, when you are able to read between the lines and exclude the opinions/posts of the obvious trolls.
And here I found Sorotec had reasonably decent comments about their overall service.

Routers are expensive and unless you do not have to care about your dollar you want to have a good feeling about your purchase (at least I want)
The product itself, but almost equally important, about the service you can expect from the dealer/manufacturer.

Setting up these things, mechanically, electrically, control software wise is challenging (at least for me).
So to know that you can call the people who sold it to you any time afterwards is a comforting thought.

Wether I actually need this service, and if how good it will be in my case...we will see