Modifying a Proxxon milling machine Y axis

[e100]

Many years ago, prior to the advent of affordable 3D printing, I bought a tiny Proxxon milling machine.
I also bought a generic stepper motor kit to make the axes controllable via a computer. I added a custom frame to attach the motors and got as far as doing some test runs to see how well it worked.

Unfortunately at this point I discovered that the relatively small 70mm front to back Y axis movement was a problem in real life, which pretty well scuppered my plans to mill PCBs etc. The X axis travel at 150mm was fine. I spent a considerable amount of time trying to figure out a way of increasing the motion but finally gave up and put it to one side.

Did anyone else make the same mistake as me, but actually find a way of converting their milling machine to use an XY table with a more sensible movement range? Effectively it would need the vertical pillar to be magically disappeared so there would be nothing in the way of a longer replacement Y axis.

[beanflying]

Even on the larger Home mills throat travel is really limited. A lot of those are 130 to under 200mm for the 1.1-1.5kW ones and even less when you fit DRO slides depending on the setup.

The only real option is to build a really solid square gantry to mount the column to it (over part of the table) or leave the ability to tram a larger XY table under it. The downside to this obviously then becomes the reach and stiffness of the spindle down to that table.

If all you wanted to do was larger PCB's then you might still be ok depending on that travel as you don't need a lot of Z? With my small Bosch Drill to get over the throat depth as it reached down low to the table I was able to make an on workpiece base for it but I could I guess make a gantry to go over a Table too. And it just so happens I have one in the projects list 

I went over this one on the Granite and it measured up well and the travel was solid outside the claimed figures (125x160). $0.50C review here https://www.eevblog.com/forum/testgear/test-equipment-anonymous-(tea)-group-therapy-thread/msg3625712/#msg3625712



Bosch in Over the Table mode

[coppercone2]

you can put the PCB on mounting pins so you can shift it once its done with a sector but thats really bootleg. Stick index pins in there and drill holes in the PCB.. semi automatic, fully ghetto.. I think that might double the work area

[e100]

The only real option is to build a really solid square gantry to mount the column to it (over part of the table) or leave the ability to tram a larger XY table under it. The downside to this obviously then becomes the reach and stiffness of the spindle down to that table.

My first thought was to disconnect the column from the base then float it over the table on a platform so the cross table goes underneath.

[Kjelt]

Your milling accuracy is directly depending on the fixed C connection (base and column and milling head)
Making the milling head movable instead of the base requires a complete redesign in which case it is easier to build from scratch with those requirements.

[beanflying]

Yeah that is the sort of thing I was badly describing

The Z issue and reaching to the table with a drill tip is your main issue (apart from stiffness). That Bosch drill above the chuck nose will actually go below the bottom of the column so I will be able to set it up with the table under the column for bigger low stock (maybe 30-40mm). Anything taller I will just live with the reduced throat and likely clearing a small Vice holding whatever I am working on.

About 120mm is what I have if I go with the column behind a table.

[Doctorandus_P]

I would not bother with trying to modify it.
At first sight the idea seems attractive, but if you think a bit more about it, you have to modify almost everything to make it bigger.
Instead, I would just simply build a completely new machine.
The picture below is from a router you can quite easily build yourself.
guide rails was one of the biggest hurdles for a DIY machine, but these days you can get very affordable guide rails from China in different sizes.
For a lightweight machine I would use MGN12 rails (I would not go smaller)
For a medium sized machine, for about EUR 450 you can buy a set consisting of:
HGR20 Rails 2x400, 2x700 2x1000mm
HGH20 Bearing blocks 12 pieces.
3x 1605 ball spindles 3x.
BK12 bearing blocks for all three ball spindles.

You probably won't find EUR450 very "cheap", but consider the sizes in that set. It's a meter long. (So approx 800mm travel in the longest axis).
As shown in the picture, this machine is quite easy to build from wood panels.
It is also an almost direct copy of a Sorotec CNC router. But Sorotec makes pretty good qualtity machines, makes them completely out of Aluminium and also uses higher quality rails / bearings and bearing blocks. (Sorotec is a german company). A similar sized router from Sorotec costs around EUR 4000. The prices on their website are without motors.

About the motors...
Get the long Closed Loop Nema23 motor sets. A motor + driver + cabling costs around EUR 100 per set.

GRBL has been splintered over many 32bit uC's after it ran out of gas on the atmega328, but since a few years a lot of the 32 bit variants have come together in the grblHAL project on github. You can choose what brand of uC you prefer. ESP32 for Wifi, Teensy for wired Ethernet, several variants of STM32.

If you're serious about the EUR 350 rails set, then also buy 6 angular contact bearings (I think size 7002). The BK12 blocks (axle diameter 12mm) in the cheap variant only has deep groove ball bearings, and the better quality angular contact bearings only cost a few EUR each.

[Kjelt]

From experience first get your requirements straight.
PCB milling is a total different game then wood milling or aluminium milling which is another league.
You get the highest accuracy without exploding the costs you need to make the machine as small and stiff as possible matching the required accuracy and speed.
It is not a pick and place machine where you have no milling forces pushing against the mill as you go.
So get the required accuracy and max. size first then look at designs.

If you look at the pro pcb milling machines under $10k you see they are very limited in size have an iron heavybase and spindles with small pitch. The size of the linear guides and the diameter of the spindle together with the baseframe design are key factors for stiffness and accuracy.
So step for step.

[bsdphk]

I invented the "height probing" technique (http://phk.freebsd.dk/CncPcb/) which made routing PCB's usable for modern SMD component pitches.

As you can see on the picture if you follow that link, I did so on a MF70, modified by Usovo.de to get eurocard sized travel.

If you look closely on the picture, you can see they removed part of the column to increase Y-travel.

I have seen no ill effects from this for my uses.

However, getting the larger travel is the least of your troubles, whereas holding larger PCBs firmly down will be.

Height probing takes care of the uncontrolled thickness of the fiberglass substrate, but it does not handle stress-relief as you partition the copper surface(s), and unless you hold the board down really firmly, it will buckle when the copper surface is fragmented and that invalidates your height probing.

I found a small vacuum-board on eBay and that has generally worked OK for me, but more than doubles the noise, because now I need two vacuum's running throughout.  (If you go that route, make sure the one you use for the vacuum-board is designed to run with no airflow.  Dont ask me how I know this is important :-)

On the balance of probability of success, I still only make relatively smaller PCB's on my MF70, say 10x10cm, and mostly for "hackish" experiments.  Things I want to last, and larger PCB's I send out and pay for "real" PCBs.

I'm very happy that the community out there has taken height-probing and run with it, and the algorithms used today are way better than my hacks 12 years ago, but I am somewhat disappointed that nobody has found a good way to apply a soldermask, because that would really improve the result.

[Fred27]

A long time ago I managed to modify an MF70 to iincrease the Y travel from 48mm to 82mm.
https://0xfred.wordpress.com/2012/04/17/extending-the-mf70s-y-axis/

[peter-h]

For any accurate work and especially PCBs you need a "zero backlash" machine: one with recirculating ball guides. They cost quite a bit more. I have a Wabeco F1410 which cost about 8k. I was going to mod it with x and y servos although nowadays I would use this way
https://www.eevblog.com/forum/mechanical-engineering/stepper-motor-to-drive-a-lathe-tailstock/msg4202455/#msg4202455

[Kjelt]

Or milling software that can compensate the backlash which you should measure first ofcourse on each axis.