View Full Version : BUILD LOG: Time for mk2 upgrade!

31-05-2010, 02:47 PM
Hi all, :tongue:

The time has come to retire mk1 cnc router and build a better one! Mk1 used mainly wood for the frame and gantry, and had too much flex. Photo of mk1 attached, which is now dismantled.

Attached is a render of the mk2 machine. Spec is as follows:
system3 all in one board
33v power supply
3Nm steppers on X and Y
1.8Nm stepper on Z
ball screws all axes, dia 16mm with 5mm pitch
X axis ball bearings running on base frame, with adjustable side preload
Y axis rail and carraige (20mm)
Z axis rail and carraige (15mm)
1.5Kw water cooled spindle motor (Chinese)
1.5Kw VFD from Moeller

Photos show the mk2 progress on the base frame, gantry, and Z axis. Tomorrow the ball screws are being machined to fit into angular contact bearings :tongue:
The aluminium side pieces need to be replaced with longer ones because during the build I decided to move the side rails out to make the gantry bearing system work better. Another job to do !

Photo of my spindle and VFD unit, to replace noisy router, including mains filter and 200W braking resistor. I intend to use a ThermalTake 850i water cooling unit (for PC CPUs), as I've seen this used elsewhere for this application.

08-06-2010, 10:00 PM
Bit more progress. The render, which I forgot to attach last time(!), shows the design I've settled on for mk2. I've decided to completely seperate the power signals from the data signals, using seperate energy chains running under the bedplate. Power will go up one side of the gantry, and signals up the other.

The render shows large energy chain, but I've actually source a much smaller version to allow me to add support under the cutting bed (which is a bit open at the moment).

The ballscrew ends have all been machined (Arc Euro website has a link to this machinist based in Long Eaton, just outside Derby). They fit into the angular contact bearing blocks perfectly, however, one of the 12x1mm nuts (which I bought off ebay from China) was not tapped straight and binds. Anyone know where I can buy quality nuts in the UK? They are square with chamfer edges containing grub screws.

I changed direction on the way the gantry mounts to the baseframe, and am running bearings on the top and underside, with bearings on preloading plates in the lateral direction to hug the baseframe. Glides well when pushed by hand, and now that the ballscrew is attached I can feel that it is unlikely to rack when cutting thin play, liteply, and balsa.

Mounting bracket for the spindle is complete and looks like it will work well (not shown).

Still lots of small jobs to do like the stepper mounts, the Z axis ballscrew lower bearing plate etc. Plus, as mentioned elsewhere on the forum, new bathroom fever has struck this house too! More when I have it . . .

03-07-2010, 10:35 PM
Brief update on progress, with more photos ....

Ballscrews machined and fitted. Spindle bracket complete and fitted. Stepper mounts complete and fitted. Energy chain support rails still to do, plus end pieces of baseframe, limit switches, and base plate to fit. Also water cooling system to buy and install.

Jogged the axes around yesterday and all move well :tongue:

May upgrade gantry sides to aluminium, but I'm mainly cutting balsa and lightply so we'll see how it goes.

Already thinking about version3 :exclaim:

12-07-2010, 05:58 PM
Do you get any gantry side shake with using a roller bearing slide like that?

I was told that the lower the gantry the less side movement and a more precise cut using the linear rail system.

12-07-2010, 10:23 PM
Hi George,

I've not finished it yet so not cut anything, but . . .

If you mean translating sideways, then none at all. What I've tried to photograph (but it's not eay to see) if that there are a pair of roller balls in plates on the side of the gantry legs, which run against the side wall of the aluminium extrusion. These are the types used to support big sheets of wood or steel and provide a friction free rolling surface, or they are used to push cargo boxes onto planes. On one side they are fixed rigidly to the leg, and one the other side they are pre-loaded by means of two bolts so that the base frame is hugged by the legs. Does this make sense?

If you mean flexing, then even though the sides are 2 x 18mm ply, I can flex them by pushing quite hard. But given that I'm cutting liteply and balsa, I think this will be OK. I've done some calculations to show that if you want to reduce the sideways flexing (due to gantry legs bending), without upgrading or changing anything else, you can significantly reduce it by simply lowering the ballscrew in Z. If you put it down at the same level as the cutting tool, there is no bending load on the gantry sides at all, only translation loads which are reacted by the side bearings. I've got it all on a spreadsheet which I may post when I'm happy with it.

As you can see on my current design, with the ballscrew right at the top, I worked this out after much of the construction was done! Next time though . . .

Lowering the gantry height would help, but I want the clearance to the workpiece if I want to do other jobs.

If you can afford it profile linear rail is about the best there is I believe. Next best is probably supported rail, and way down is running bearings onto the ali section. But it is a big step up from the unsupported 1200mm rail that I had before, so making progress.

12-07-2010, 10:31 PM
How much did you pay for the Ally extrusions supporting the Y axis,if you don't mind me asking and where?

I'm making my own foundry to melt and cast ally parts for the CNC build.


13-07-2010, 12:36 AM

You need to start a build log of your own... two actually - one on the CNC machine and one on the foundry (also on my list of things to try out when I get the time) :smile:

13-07-2010, 01:05 AM
Irving,been watching youtube vids on melting and casting Ally,very instuctional.

The best homemade foundry or melting pot is the Propane bottles,take the tap off and cut the lid off,line it with fire cement,dig a hole in the side for an air to propane torch and the melting is in no time..

Gotta find a round steel or cast iron tube for the crucible though.

Put "casting aluminium in sand" without the quotes in google and the youtube will come up.

13-07-2010, 01:09 AM
Speaking of foundry.. heres a link you might find usefull http://metalcast.boorman.us/index.html

13-07-2010, 01:15 AM
Nice one AdCNC. :clap: got about 5 kilo's upto now.

I'm told the best ally to melt and cast is the type thats already been casted before.


13-07-2010, 01:02 PM
Hi George,

I got the aluminium profile for free, as surplus to requirements. It was well used, but I selected the good bits to use on the machine.

To buy something similar would be about 40 per meter. E.g.

Good luck with casting parts. Another route to making bits, depending upon what you want to make, is posting a drawing on this site and getting a quote. Put RFQ in the title and see what happens.

If you are starting out on a new design, then you can always post a sketch on this site for constructive critism and comments. I would suggest adding as much detail about your intentions and limitations.

13-07-2010, 04:06 PM
I was hoping to buy from here....

But then I seen the clincher Miniumum order 100 :(
This site has far more profiles than I have see elswhere.

14-07-2010, 11:16 AM
Oo. I know very little about CNC or anything else for that matter, but the Series 8 60x60 is doing wonders for my aluminium profile fetish.
ty for the link :)

14-07-2010, 04:57 PM
Oo. I know very little about CNC or anything else for that matter, but the Series 8 60x60 is doing wonders for my aluminium profile fetish.
ty for the link :)Glad someone got a buzz out of the link. :wink:

This is one Extrusion site where a group buy would benefit as discount would be apparent on bulk buy.

Thing that also drew my attention was a 4% surcharge on the CC transaction,never known that to happen in any purchases over the net I've done and I have done many,most odd.

13-09-2010, 09:04 PM
It's been a while, with bathroom projects and other distractions, but managed to do a bit more recently. Here are the updates with the some photos -

Re-wired all the stepper motors with 4 core shielded (let's not kick off that old debate!) from ebay. X and Y went well and the machine jogged around nicely. Went to wire the Z axis, and get a quick hum from the stepper then bam - all dead. Despite my checks I had managed to wire it up wrong. Very annoyed since it was the last step to get it all running around. I've blown the driver board (system3), probably the Z axis chip. Email to Roy asking about my options, he said if it was the chip he'd repair it for 10. What a nice man! Board will be posted tomorrow, will have to wait and see how it goes.

Limit switches
The microswitches are all in place on mounting brackets, ready for wiring. I bought some twin core shielded signal cable from my local cable shop (0.30 a meter) and am intending to wire it different to mk1. Will put the home X,Y,Z on seperate pins, and the limits chained together with the e-stop on pin 13. All connections will be normally closed for safety reasons (plus I think this helps false triggers).

IGUS cable
This is now all in place with the motor cables inside, and working well. I've split my wiring down the middle with the power on the left, and the signals on the right.

Cooling system
I finally decided on a cooling system, based on more research. It turns out that these spindles need very little flow and cooling. Pump spec is 5 gall/hr which is about 20 litre/hour. Windscreen washer pump is about 200 litre/hour and PC cooling pumps are about 500 litre/hour. No wonder a windscreen washer pump gets hot trying to force all that flow through the narrow cooling channels. The PC ones may be getting away with it because they are designed to run for long periods, and have ceramic bearings apparently. I bought a copper radiator rated for PCs of 300W (although the CPU is probably putting out 50-100W), which meant using a special coolant to reduce corrosion (spindle channels are probably ali). The pump I had lying around is around is 500 litre / hour so I'm thinking of using this one at a reduced voltage (5-6v) if it will run down there.
Some people are using a 6 gall / hour pump into a 5 litre steel bucket of water and running just fine, relying on the bucket of water to cool down without a radiator, so this is probably going to be a case of giving it a go.
I've got a 120mm fan (not pictured) somewhere which will bolt to the radiator. The whole cooling system will probably mount somewhere above the spindle. All for now. . .

28-01-2011, 04:11 PM
Been away for a while, but here are some updates.

1. Water cooling system
Installed and working well. I went with a 12cm radiator, cooling fan, 12v pump, 3/8" hose and coolant used on PC CPU cooling systems. I bored out the small barbs on the spindle to help with flow, but having run the system for some time now it probably isn't required. Cutting wood barely raises the temperature (on the outer case anyway) and it is only very slightly warm to the touch after prolonged operation. In fact I recently ran it for 10 minutes cutting and forgot to start the pump. It was still barely warm to the touch.

2. 1.5kW water cooled spindle
This was a great upgrade from the router. Much, much quieter - you can easily hold a conversation next to the machine with an air cut. There is some noise from the cutter when actually cutting, but overall a great improvement and very smooth.
The other advantage is the lack of air blowing down over the work piece and throwing dust everywhere.

3. Power supply
Last year I had problems with glitching stepper motors. I had a 30-36v variable power supply turned down to 30v (switched mode), and found that replacing this with a 24v DC supply cured it. But this was not a long term solution so I bought another power supply, but this time 24v-30v. I tried various voltages and did not have a single glitch with it. Maybe the other supply was faulty. Anyway, sorted that one out.

4. X axis bearing arrangement
The X axis bearings run along the top, bottom and sides (for pre-loading) of the main aluminium sections. After maybe 10 hours of cutting there is a slight witness mark appearing on the profile where the bearings run. This was expected but the long axis is the most expensive to add linear rails to. I'm considering either adding a wear strip to the top surface, or adding supported rails.

5. Control panel
Made a new control panel with more inputs (option for touch probe etc), see photo.

6. Dust extractor
After several aborted ideas, I'm making a floating dust extractor foot. More photos when complete.

Overall I've cut out a number of items and been very pleased with the results. The machine is a massive improvement over mk1, which is progress.

26-02-2011, 10:38 PM
Whilst this new machine performs very well, there were 2 features which were on the list to upgrade when time and finance allowed. First was to replace the x axis bearings running on the aluminium profile, and the second was to replace the wooden side members of the gantry.

So I'm planning my next moves. The X axis will be changed to 16mm supported rail and I have ordered the parts. The gantry sides will be RHS, probably 6" x 2" sort of size, with spacers inside where bolts pass through to avoid crushing the section and maintain the bolt torque. For the top and bottom gantry rails these spacers will be a slice of extrusion, just big enough to slide inside the RHS. For the bearing mount they will be some large turned tubes.

Whilst waiting from the rail to arrive I've been sketching some ideas, and settled on the ones attached. I've yet to see a really neat way of adding a rail connection to a full picture frame style gantry with single leadscrew drive. Anyone seen anything better?

So considering the forces acting I think this method will work ok. The only significant load on this cantilever is the weight of the main gantry and the z plunge of the spindle into the work. This is resisted by a 30mm thick ali block bolted to the RHS with 3off M12 bolts. The clamp load will therefore be very high, and since the block is full width across the RHS it will be able to transfer this load nicely into the stiff side walls of the RHS.

I could add corner brackets to add extra support, but I'm going to try it like this first and see what happens.

In the X direction there is no load because the bearing can slide, and in the Y direction the weak link will be at the top of the gantry side where it joins the horizontal part, so this stubby cantilever shouldn't make any difference.

That's the plan anyway, here are some preliminary sketches showing the current machine but with new RHS ali sides and 16mm supported X rail.