Disaster!
Iāve got the spoil board and extrusions assembled on the torsion box bench and Iām trying to align the X-Carve Y and X rails. The right hand Y carriage is about 1cm closer to the front plate than on the left side. After much tightening and loosening of allen screws, I put an engineers square on the Y plate to measure how square the Y plate was to the Makerslide and it isnāt - square that is!
This could happen if I had debris from the tapping of the makerslide caught between the plate and the end of the Makerslide but I was careful to de-burr the tapped holes and clean up afterwards.
Could it be that the two X axis makerslides are different lengths? I checked the other 4 pieces of Makerslide that were bought from the same supplier at the same time and found that the length varied by up to 1.5mm over 1 metre.
A bit more measuring showed that the right hand Y plate is not vertical to the table surface which could indicate that ends of the makerslide were not cut at 90Ā°.(EDIT: Just checked the spare pieces and they are not cut at exactly 90Ā°)
Iād love someone to tell me that it doesnāt matter if things are out of whack a little but Iām assuming that the accuracy of the finished product depends a lot of each part being accurate in itself.
I have a mill so I can machine the remaining makerslide to the correct length and make sure ends are square but thatās going to be a right pain.
I can see a disassembly video coming up 
I should point out that the Makerslide I have did not come from the current Australian supplier, 3dtek. Mine came from an Indiegogo campaign about 2 years ago to produce Makerslide in Australia. Only one batch was made before the producer lost interest in it all.
My mill is quite small and getting it set up to handle a 1 metre length of stock will be a challenge so Iāve chosen the quick solution which is to buy some more MS.
For a while, Iāve wanted to build a rostock style 3d printer and the MS currently in the X-Carve will be fine for that as the squareness of the ends is not an issue.
Sounds like you have a solution, but I would think a good cut-off, radial arm, or ban saw with a good blade should be able to handle the ms aluminum, and make it truer, then what you have. To be honest Iād be surprised if Inventables supplier used anything besides a saw of some kind.
I have a metal cutting saw designed for aluminium but itās not really suitable for taking a very small amount from each end of the MS. TOouse it properly, Iād have to shorten the MS by at least a few millimetres , which Iād prefer not to do.
The mill can be set up to shave off less than 1 mm to square up the ends - assuming I get it trammed/adjusted properly.
I have no doubt the supplier of the original MS used something like a mitre saw and the problem was with technique rather than equipment. Some of the extra MS I have, has one end cut perfectly but the other end is out of square.
Iām confident that the current supplier will provide a square cut product.
Iāve tested machining the ends of the spare MS and it wasnāt as hard as I expected. Ends are now perfectly square, pieces are the same length and I didnāt lose more that 0.5mm in the process. My order of new MS is already on its way so looks like Iāll have some spares.
While things are apart, Iām going to replace all of the eccentric nuts with eccentric spacers. Iāve already done this for some but ran out of longer bolts. I then remembered that I have the eccentric nuts I got as part of the aforementioned indiegogo campaign. These spacers are much thinner than the standard ones, about 2mm, so the original bolts will still work.
EDIT: just noticed that Inventables sell these - doh!
A tomorrow job, too hot today (33 Ā°)
Removed the MS and machined the ends. I didnāt measure exactly how far out it was but enough to be very noticeable when using a square. I hope this solves the problem.
I started to replace the eccentric nuts with the low profile spacers and struck another problem. The boss that extends through the Y axis plate is too thick. The Y axis plate is 3mm and the boss on the eccentric nuts is slightly over 3mm. No problem for the X carriage as its 5mm thick. Fortunately, the low profile eccentric spacers I have are made of brass so will be easy to file/grind/mill down.
And if that wasnāt bad enough, the rabbits have started digging in my veggie patch! The fight is onā¦
Received and fitted new makerslide. Big improvement in the left/right Y carriage offset problem but not completely removed.
I got sidetracked playing with some new woodworking tools so back to CNC tomorrow.
After loosening all of the bolts holding the spoil board and base extrusions together and fitting the new MS, everything lines up!
Went to run a few tests before trying a first cut and it seems I have a homing problem, the axis directions have reversed. So far one dead micro-switch and one faulty wire have been replaced. I really need to get some better quality micro-switches.
Itās alive! Although not well.
After sorting through some issues with cambam (and spiders, donāt ask), I made my first sawdust!
Just a simple rectangle and circle - yes that second part IS a circle or at least would be if everything was done up tight.
The belts are really loose and possible some of the wheels as well. Iāll go over all the bits that need tightening, including the stepper pulleys, and try again.
Tightened up everything and this is the result, a 100mm circle that measures 100mm.
Another test after this failed because the X axis belt slipped. Time for some heat shrink on those I think.
X axis slippage problem.
When I cut pockets (second image) thereās unwanted movement in the X axis, Y direction seems OK. If I cut profiles (first image), then they come out correctly - X and Y measurements are identical on the square in this case, circle is as round as youād like a circle to be.
Iām cutting in MDF at 1400mm feed with a .7mm DOC, tool is a 2mm endmill (new). Router is a Makita RT0701. Iāve tried dropping feed rate and DOC but no change.
This morning, I put heatshrink on the belt ends and adjusted tension. Iāve checked the wheels and they donāt seem too tight/loose, Iāve checked the set screws in the X axis pulley and they are tight. Iām running a Gecko G540 controller under Linuxcnc so motor current setting should not be an issue.
What software are you using? It seems to me that if it was a machine problem, the profile cut would end up with at least a little bit of offset on itās go-round. If the square and circle are perfect, then your machine is likely dialed in. I would guess your problem is either on the software or the electronic side of the X-Carve rather than on the mechanical side. You may want to start a new forum thread showing this, just so the whole community sees it rather than the few that have followed this thread.
1 Like
Well as reported in another thread, the issue is solved. Turned out to be that PC Iām using canāt quite keep up with the stepper pulse rate required. A reconfiguration has fixed that and I can now cut stuff!
I spent most of today just cutting circles and squares, not very exiting unless youāve previously been unable to do so :grin
I did have to tweak the steps/mm settings to get correct X and Y measurements. The Z axis was WAY off, only moving 1/4 of the distance told. I changed that setting then promptly drove the only small bit I had, into the spoil board. 
I do need to do some more alignment to see if I can get the spoil board a bit more level than it currently is, thereās about a 2mm height difference left to right.
Just to finish this thread up. Hereās a sample job I ran today after levelling the spoil board.
3.175mm single flute, 2400mm/m feed-rate, 1mm DOC.
What this doesnāt show is the second part where I tried to do a 4mm deep, 100mm diameter pocket but I forgot to zero the Z axis and drove the bit through the work piece and the spoil board. Donāt tell anyone.
I think thereās still some fine tuning to do,some of the wheels feel a bit tight, and the spoil board is still not as level as it could be.
1 Like
@GeoffSteer thatās beginning to look pretty good. It looks like thereās still a few bumps here and there but it looks like youāre getting pretty close.
As for the forgetting to reset Z, Man, I know that feeling. I know for a fact that for every job Iāve proudly shown to anyone (or posted here), there are often a good many pieces of wood in my āscrapā or āmaybe I can use them laterā bucket.
We used to call it the āGozunderā ratio.
Or, the ratio of jobs that are successful, compared to the number of jobs that Gozunder the bench. Anything under 1:1 is probably pretty good for some folkā¦
2 Likes
@DavidWestley What stands out in this image as a problem? I must admit that after the issues Iāve had, I probably canāt see anything short of a complete disaster!
I was quite impressed that the system was able to drive a 3mm bit thought the 6mm MDF I was cutting AND the spoil board and still keep cutting - without breaking the bit. Iām just glad it missed the extrusions under the spoil board.
I donāt mind making mistakes as long as theyāre new ones, 
itās nice to have a use for all the MDF offcuts Iāve acquired over the last few years.
Sorry, this may have come across as critical, what I meant was is that it all looks pretty good. When I looked (quickly) it looked like some of the lettering had ābumpsā or irregularities in them - such as the ārā, but looking closer, this appears to be image artifacts either in the compression or in the viewer for my iPad.
When I was still setting mine up, Iād occasionally get ābumpsā or grooves in the side of smooth curves or straight edges, thatās what I thought I was seeing.
Apologies, didnāt mean to sound critical or suggest the lettering wasnāt up to scratch - it was an off the cuff comment, whilst having a glass of red wineā¦
With the issue I had, everything I do at the moment looks perfect in comparison, a critical eye over the results is a great help.
I can see the bumps in the word Hobart, Iāll check what the tools-paths look like (CamBam). I might try printing the same word on the CNC router at the hackerspace tonight - not an X-Carve.
If I look closely into the deeper cutting on Hackerspace, I can see the different layers of the cut. Hopefully some tuning will help with these.
Just when I thought I had things sorted!
Undoing the spoil board to adjust/level it and I found that one of the 20x20 extrusions seemed loose. Removed the spoil board and found that all the screws holding the rear extrusion (X axis direction) were not tightened to the 3, Y axis direction extrusions and one of the right angled brackets had only one screw it! Casting my ageing mind back, I remember being one short of the right length screws and, at the time, thinking āIāll take care of that before I put the spoil board onā
This probably explains some of my problems in getting the whole thing square.
Lining up the 2 captive nuts in the centre extrusion with the holes in the spoil board is going to be funā¦
EDIT: Well that wasnāt as bad as I thought. I did have to move the centre extrusion a little to one side to get the nuts to line up with the holes.
Y axis rails are measured at exactly the same spacing front and rear. Corner to corner measurement within 1mm - without having to apply any force, well not much.
Now i need to re-tension the wheels and belts. This is easier when youāve done it a dozen times. 