My screwed insert came today. So I did let the CNC drill all 309 holes with a 6mm drill bit.

Then I started to drill the holes for the screwed inserts with 8mm by hand. I noticed that the wood starts to split when I used the 8mm drill bit directly. So I used countersink drill bit first with the diameter of the head of the inserts.

So every 2nd hole (82 holes) has an insert that provides an M5 thread to bolt down things. And the other holes will be used with simple wood dowels to provide a fixture for x and y axis.

I guess when I am done with the inserts I will use some sandpaper to smooth the surface again.

The taps for the M5 threads came and as I knew myself I ordered 2. And as I already knew that I am not patient I broke one at the 86 hole. I cannot get it out, but that is not such a big deal as I have plenty of those holes to mount things.
When I was done I mounted a wooden board on the steel that will be my sacrificial fixture plate.
As it is wood it should not make too much damage to the bits or the whole CNC when I make a mistake with the G-Code. And I will make mistakes.

I have used 6 M5 countersunk socket screws and I may add another 3 in the center of the x-axis, but for now, it should be fine.

The next step was to face mill the board so that I have a flat surface, at least to the geometry of the CNC gantry. I used a cheap 6mm 4 flute bit from China. It did the job.
Of course, I would like to have used something with a bigger radius, but as the shaft of the tools in an ER11 collet is limited to around 7.5mm I did not find anything better. At least for wood.

I am satisfied with how the surface came out. I started with a feed rate of 400mm/m,  and then increased it on the fly. I shot this video around 600mm/m. Later I pushed it to 1000mm/m.
I did not try more as I remembered that there was a limit listed for around 1000mm/m in the item description when I bought the CNC on eBay.

As you can see from the colour change the surface was not really flat to the ganrty geometry. Those are the different layer in the multiplex board.

When this was done I went back to my Mac to create a new g-code file in Fusion360 to mark holes in a grid with 25mm spacing.

My plan is to use screwed inserts with an M5 thread inside and a self-cutting outside thread that can be screwed into the wood. They have a hexagon socket that makes it easy to install them with a cordless drill.

I have ordered a pack of 100 pieces at eBay for a good price. They will be installed every 50mm, just like the holes in the steel plate. The holes in between will be just plain holes for simple wood dowels.

This is how far I came today. I wish those inserts would have arrived in time, then I could have finished the fixture board today. The scratch you see right next to the drill bit was from a mistake I made in the G-code. I moved the spindle a 0 z height and scratched the surface.

Finally, I have all 96 holes drilled in the steel plate. I also started to tap the M5 threads.
As I am lazy I have used my cordless drill and when I got too excited I broke the tap bit.
I ordered new ones from amazon and while I need to wait for them I thought I take a look into the control box.

So this is my control box and it looks nice from the outside.

Also, the back side is nice and all the connectors properly marked.

When I looked inside I was surprised in a positive way. I have seen some other Chinese CNC control boxes on youtube and in CNC forums. They looked like someone did just throw in all parts.

This one looks really good. All the components are mounted nicely and properly. The cabling is done well. That looks clean and professional. 

Tha Mach3 USB Board is the BSMCE04U-PP from Bitsensor.com. I tried to look up their website, but you get redirected to a Chinese shopping portal. The board name did not bring any really useful hits on google either.

But this board is sold on eBay as "MACH3 4 Axis 100Khz CNC USB Breakout Board" for around 20€ including free shipping.

There are 4 single motor drivers with the Toshiba TB6560 chip, each has its own heatsink. You can find similar boards on eBay between 5-20€ rated up to 3A.

And then there is the VFD that controls the spindle motor, which is just the board and not a complete VFD wich chassis. The ribbon cable connects to the control unit in the front of the control box.

At least this is rated to 1.5KW according to the stickers.

And here is the backside of the VFD control unit.

And last the power supply. I did not check if there are any marks to what power it is rated.

Maybe I do that to a later time. But there is a raiser shield to have the VFD board mounted on top of the power supply and a think aluminum block for the VFD board.

For sure not highly sophisticated components, but the work.

One thing I had on my mind for two days now is to replace the Nema 23 stepper motors on the machine. It still gives me sleepless nights that the z-axis lost steps when I tried to drill the holes in the steel plate.
The easiest way would be to upgrade to close loop steppers. I found the JMC iHSS57-36-20 closed-loop stepper for around 90€ and the driver is mounted directly to the top of the motor.

Compared to a driver/motor bundle from Leadshine which costs around 150€.

In theory, I would just need to replace the original Nema 23 with them and it should work fine.
But for 130€ you can also get the JMC iHSV57-30-18-36 180W integrated servo motor.

And I guess I might try the servos rather than the closed loop steppers. Even when it seems that there more difficult to tune to the system.
Either way, with a closed loop system, stepper or servo, there should be no more lost steps and those models have much more torque than the original Nema23 steppers.

As the router could not drill the 4mm holes in the steel plate I used a cordless drill for a few holes.
I recognized that my cordless drill was getting warm and that means the motor is working hard. An overheated motor reduces the lifetime of an electric tool drastically.
I did not want to damage my cordless drill and looked for other options with the tools I had available.
The perfect tool would be a magnetic drill that is held down to the steel plate by the magnetic feet.
But I don´t have such a tool and renting one is quite expensive.
So I went with a poor man´s choice and used some clamps to fix my drill press to the steel plate and started to drill the holes.
That went surprisingly well and quick, even I had to move the drill press and clamp it again for every hole.
I managed to drill half of the 96 holes before I had to stop. Yes, there is just a little time between the end of work and the time you don´t want to bother your neighbors with the noise of your drill press.

My holes are 50mm apart from each other and the idea is that those M5 threaded holes are used to bolt down other fixture plates like wood, aluminum and whatever can be milled flat by the router and in some cases could be sacrificed during the milling of parts.

This is the setup I used for the attempt to drill M4 holes with the router. I did not work, but I learned a  lot about the Fusion 360 CAM module again, especially about the drill operations.

I can only strongly recommend using the simulation function before sending any G code to your CNC. It will show you exactly what the CNC will do. It is much easier to fix any issues on the computer than on a pile of ruined parts.

So meanwhile I bolted the 10mm steel plate to the frame of the 6040. I have used the holes in the frame that have been used to bolt down the extrusion profiles.

I had to redrill 2 holes a bit bigger as the screws did not fit. I guess my fault as I did not have a good workspace for the marking and drilling of the holes. But overall I was happy with the result and the 6040 around 30kg heavier as before.

I already feared that the plate might not be perfectly straight and it turned out that there was some bending. But I don´t have the tools to straighten out a 10mm steel plate.
If you look from the side you can see the bending. It does not look that much, but from the front to the middle(high spot) of the plate it is nearly 1mm difference.

So I had the silly idea that the router could simply mill the surface flat.
Well, there are some problems with that:

1. The work area according to Mach3 is not 600x400mm. It is 566x385mm. So you have to live with the fact the everything outside is not flat and cannot be used to clamp something down properly.
2. This is a router, not a mill. A mill runs much slower and usually has a gear that provides enough torque to cut steel. The router spindle is a direct drive that has nearly no torque at speed below 5000rpm.
3. Which leads to the next problem. The High Speed Steel(HSS) mill bits are not designed to run with 5000 or 10000 or more rpm. As the ER11 collect cannot take any bit with more than 7,5mm shaft, I had only a 6mm flat end mill bit. The optimal speed for the size of the bit to cut steel is between 1000 and 1500 rpm depending on the stell and the lubricant/cooling. So this cannot work. Yes, there is carbide and others, but nothing really designed for cutting steel at 10000rpm and more in this size. 
4. There is a lot of force required to cut steel. I tried it anyway and thought I could take off maybe 0,1mm at a time until I am done. But the machine started to rattle as soon as the mill bit caught on the steel. The spindle was moving like crazy and I don´t mean the collet, the whole spindle.

So the result was not very good and I tried different speeds and feeds. But as the first screws dropped down from the z-axis assembly I knew that this was going nowhere. You can see the tool marks the wobbling spindle left on the surface.

I gave the idea up for now and I going to try to mill a flat surface on the fixture plates that will be bolted to the steel plate.

The holes you can see are actually drilled by the router. I used a center drill with a 2,5mm tip to drill those holes. And surprisingly that worked well with the help of WD40 as a lubricant.

In the next step, I tried to drill 4mm holes(at the time I had no 4,2mm) for preparation for tapping M5 threads. This was when I realized that not only drill bits are not designed to drill with that high rpm´s in steel and the spindle was stuck sometimes as even with 5000rpm there was not enough torque to drill holes in steel.
Another problem was the feed from the z-axis. The z-axis stepper motors lost always a lot of steps, so that mach3 was thinking the drill bit is at -10mm, but it was not even at -2mm.

I stopped the drilling and tried to think of a better way with the tools that I got. Also gave me time to order some 4,2mm drill bits.

After the disappointment with the "Liquimaster Pro" I gave it another try to mill the same part.
This time the workpiece was a little big bigger so I could clamp it better to the T-nut table.

I applied the coolant manually and again everything looked very promising. As before I milled very slowly and as everything seemed to work fine I went away to do something else.

But as you can see here, the same issue again:

As soon as the mill bit milled the last thin layer, the material caught by the flutes and lifted the workpiece.
And again the bit took some chunks and broke.

This is not as bad as my first try, but still a total failure.

So the root cause of all this is that you cannot clamp down anything really on this mill as the T-nut table is not ridged at all.
As you can see the thin, hollow profiles have no support over the whole distance form the front to the back of the frame. They are just screwed to the frame in the fron and the back of the machine.
You can wobble the profile with just a little force.

The next issue with those profiles is that they bended when I clamped something with a little force.

So I decided to change something.

A T-nut plate of massive aluminum in the size of the machine costs between 350-500€. That is quite a lot, but don't ask for one made out of steel. You would be surprised how much they charge you for a used one from an old mill.
So I thought a massive 10mm steel plate would be the best as table and then put something on top of it.
I managed to find a supplier that agreed to laser cut a 700x490x10mm steel plate for me and charged my around 90€ including the shipping. As a side-note, they payed the delivery service for 8kg, but the plate is more like 30kg ;-).
I also found a cut off end from a massive aluminium profile in 250x400x20mm for 40€ that I am going to mount on that steel plate. If this will work I might think about covering the whole area with a t-nut profile in the right size.

The next step is to mark the holes and drill holes in the steel plate that allow me to mount the steel plate with new M4 countersunk screw to the frame. I am using the same holes that where use to mount the aluminum extrusions. I was thinking to re-drill and re-tap the holes in the frame to M5, but I don't think that is really necessary. If it turns our to be, I can still do that later.

No updates for a while, but I learned a lot about the CAM Module in Fusion 360. I watched many youtube video, some good and a lot of bad ones.

Let´s get to the first problems. The T-Nut table is made of hollow 16mm extrusion profiles, not one plate. Those are mounted with M4 screws only on the front and the back of the mill frame.
There is no additional support for those over the whole distance from front to back, about 700mm.
Not really a ridget construction.

Meanwhile, I ordered some carbide 2 flute mills on Amazon. According to some other bloggers and YouTubers opinion, they are best suited for aluminum as the aluminum does not stick in the flutes.

Then I saw a video from CNC Kitchen in which Stefan used an attachment from GoCNC for his CNC router that did come really close to the idea that I had in my mind for a while.
Basically, it is a tray that has some T-Nut profiles glued in and allows you to mill with coolant and have a closed loop without the coolant running all over your router.
The "Liquimaster Pro" is a metal tray (345/530/45 mm) with the T-Nut profiles, a small pump for room fountains, a hose with a nozzle and you get a bottle of coolant.
As this looked good in the video I ordered it for 235,19€ including shipping for 12,90€(usually you can send big and heavy parcels here for around 6€!).

Well, I am not so sure what is "Pro" about this thing. The painted tray need to be clamped on the existing T-nut table with M4 screws. The little pump is also glued to the tray, not sure why.
There is a piece of foam that is put on the hose that sucks in the coolant from the tray, to prevent chips from being sucked into the pump. Well, that does not work at all.
Like in Stefan's video the nozzle was clogged with chips after a short time.
Also, there is nothing to mount the nozzle, so I used a piece of art made out of cable ties. I planned to 3D print something for that.

Maybe I should also mention that I asked per email how the profiles are mounted to the tray before I ordered. The answer was with "special" glue!

So I installed the "Liquimaster Pro" and prepared to mill my first part. The hose from the pump pushed the foam pice up every few minutes. So I had to put something on it.
And the coolant did not come out in a constant flow. I added some Water to the coolant to increase the volume to get a constant flow on the workpiece.

So far I looked promising:

Here you can see the cheap clamps and how they bent. Under the workpiece, I had a piece of thin wood as I did not want to mill in the profiles.

Here I already removed the nozzle from the hose as it was clogged with chips.

Until then I was really amazed that I spent the money to good use and made already plans in my head how to design a 3D printed mount for the nozzle. As this was my first part and also in aluminum I milled very slow with slow feed rates and 0.2mm layers.
As this seemed to take forever I left the mill to do something else in the meantime.
But when I came back to ma garage, the Liquimaster Pro was already torn into pieces.

What happened? As the mill bit was milling thru the bottom of the part, the thin remains caught on the flutes and lifted the part up. But it was clamped down? Yes, but the "special" glue did not withstand the force and both profiles came loose.
The bit milled a little in one of the profiles, took some bites out of the workpiece and then finally broke.
The small pump did not work anymore as well. I guess some chips are in the pump, and as it is glued to the tray I did not found an easy way to clean it.

Here are some picture from the workpiece:

Here you can see the big bites the mill bit took out of the workpiece and bent the edge.


The "special" glue is just some sparly attached silikon:

I guess I might be made some mistakes with the milling parameters for sure. But the "liquimaster Pro" is a piece of toy that does not hold up to the forces when milling metal and something goes wrong.
The whole construction is a joke and charing more than 200€ is extreme.