Here we go again . . . MK4

[routercnc]

Hi Clive
Thanks for the encouragement! If I multiplied time taken per piece by number of pieces to make I might get disheartened. So I'm trying not to think about it.
Vice was from Arceuro. It was about £35 and is surprisingly accurate (a DTI along the fixed jaw barely moves) and the base is fairly parallel to the jaw runner surface (therefore workpiece is fairly parallel to bed)
But after first use the thread got swarf on it which went into the nut and it is very hard to turn. I can't get it apart to clean it out so if you get one put a cover on the thread first. It's actually a drill press vice but is working ok for my gentle machining
To be honest I am going to replace it with a precision vice as one of the future jobs needs to be more accurate than this and I'm fed up trying to tighten the screw. Arceuro make some nice ones with ground surfaces all over and antilift jaws. They are also quite low in height which is good. if I buy one and like it I may get another as a pair of vices can be really useful

[routercnc]

It's been a while since I last posted and I'll explain why in a moment. I've made a matching pair to the last part, with a bit of finishing off to do:



Reason for the delay is that I've noticed for some time now that even mild cuts are often generating lots of chatter. I think it has been getting slowly worse, and feeling the Z axis around the rails and bearings there is some relative vibration between rail and bearing, especially at the top left bearing, when the chatter occurs.

Problem is that the chatter can quickly lead to welding, and with so many parts to make for the new machine it was time to check things out. For info I have been running these 15mm linear profile rails (classed as 15mm minature) for about 5 years on 3 successive machines, and that they were bought as used from 'fa-systems'.

I stripped the Z axis down and found significant fretting on the rail, especially around the top left bearing:


Luckily I had a second set of 15mm rails (new) that I'd ordered for the Z axis on the new mk4 machine. However, the hole spacing was 5mm further apart on the bearing carriages. I ended up drilling another set of 4 holes rather than 2 extra holes because the holes would have been too close to the existing ones:


I took the opportunity to drill some access holes in the outer plate so I could assemble the Z axis as a unit, and then bolt it on to the Y axis. It's easier to get everything lined up off the machine rather than trying to assemble it bit by bit on the machine:


I then set the first rail off a side datum, and then clocked in the other rail:


Then assembled the Z axis onto the machine:


I trial cut revealed a much improved machine. A 1mm DOC at 900mm/min 12,000rpm 6mm 2f carbide was fine with no chatter.


So is there a moral to this story? Well my rails were used when I bought them so it's had to be sure but for the occasional poster on here who asks about rail size then I will always recommend 20mm over 15mm.

Back to the new machine. As originally drawn it had 4off 15mm rails around a box Z. Although these are used as 2 pairs loaded against each other I did not feel confident in continuing with 15mm rails all round. Luckily I had left space in the Y axis to upgrade rails or spindle size. So it was a simple matter to add pockets and fit in 20mm rails. I've ordered 20mm rails from Fred, and will use the new 15mm rails just as auxiliaries away from the spindle where the loads are lower.

[routercnc]

Whilst cutting another new part I used 'feedhold' to stop the machine just to check something was not going to collide. Then I hit 'start' but instead of continuing on a +X direction to the end of the cut the machine moved in +Y. It was about 8mm into a profile cut so snapped the cutter.

Luckily this was in the scrap part of the material so no harm done. I changed the bit, homed, ran it all again. Later on in the same profile cut (in the -X direction this time) it was almost to the bottom when I pressed 'feedhold' again to check one of the clamps was not going to get machined through. When I hit 'start' is again went in the +Y direction. Luckily I had my hand over the e-stop but it still took a chunk out of the part!

I've never had this happen before and haven't changed anything in software, just the Z axis bearing change. Any ideas?

[Clive S]

Did you by any chance stop it in a canned cycle I have a feeling that it happened to me once.

[routercnc]

Did you by any chance stop it in a canned cycle I have a feeling that it happened to me once.

I don't think so - I'll check the g-code when I get home as I wrote down the line number I stopped it on. I seem to remember lots of X and Y positional commands so probably not in canned cycle? It was cutting out the outline profile to a part which was almost a rectangle.

The last feedhold problem was done on a pass which had tabs on it, but it was about 50mm away from having gone up, across, and down to create the tab and was going in a straight line to the corner. I was about 20mm from the corner going X- when I pressed feedhold. Upon starting it went Y+ as if it thought it had reached the corner.

Next time I'm in the workshop I'll do some air cuts on the same file to see if I pressed the feedhold twice or something . . .

[routercnc]

OK, looking at the gcode where I had the 2 crashes (when 'starting' after a 'feedhold') and I can see there are several straight line moves (must be geometry nodes all along in a straight line from when I was creating and changing the CAD), followed by a G2 radius arc in both cases.

#1
I've read about problems with feedhold in arcs so always feedhold on a straight edge. But I was definitely on the straight bit about half way along the top edge when I pressed feedhold. About line 930. It was some way off getting to the G2 ARC at the corner. When I hit start the machine went up in +Y and snapped the tool clean off.

I've loaded the file onto my house PC (hence the DROs are all zero!) Here is the first one:


#2
This time, on the same part, I was cutting along a straight edge at the bottom, -X direction, when I hit feedhold. It had just done line 1060. This time cycle start moved the machine in +Y. I hit e-stop straight away this time and recovered it. However, there was a gouge in the work piece (!). It's cosmetic but annoying none-the-less.

Looking at the gcode shows another set of linear moves, followed by a G2 arc radius at the corner:


I'll go back to the CAD and see if there are repeat nodes in the geometry as a start, but I'm confused by this as I've used feedhold for years and not had this problem. Now twice on the same part!

I went to Mach3 website in case there was a later version - but it's still R3.0.43.066 I which is what I am running now.


EDIT:
OK, the reason for the repeat straight moves is that there are tabs on those sides. Vectric obviously adds those moves in on every pass, but only adds the Z up and over move when the required depth is met. However, I was nowhere near the depth to create the tabs (they are only 2mm high) so that might be a coincidence. So, still confused about what is going on . . .



EDIT2 / side note:
I've had to do a few 'run from here' operations after hitting the e-stop recently. It is probably out of position as it could be mid-step so I've found you need to:
1. Raise the tool right out of the workpiece
2. Re-home the machine
3. Press 'run from here' on the line you e-stopped on
4. It will show you the required resume coordinates in a window. Before accepting these you should jog or MDI to the X and Y positions it shows. If you do not it will try to move straight to that location which could be through a clamp or the workpiece etc. if you are low. This has caught me out a couple of times!
5. Press OK and it will rapid down in Z to the position it got to when the e-stop was pressed
6. Start the spindle (or there is a check box if you use spindle control from Mach)
7. Hit cycle start

[routercnc]

Some pictures of the matching 2nd upper end part - cleaned, drilled tapped and finished:


Then onto the next set of parts. Bed cleaned down and a new 10mm plate ready to cut out 4off bracket plates:


Small pockets cut and holes spotted in all 4 parts:


First one cut out:


Last one cut out:


Opening up the light cnc spots with the same 6mm spot drill:


First pair of them cleaned up, drilled out, counterbored:

The counterbore in each part was a bit nerve racking as it is to take an M8 cap head slightly under flush. This means a depth of 8.1mm in a 10mm part, which leaves just 1.9mm left. Although I had my DRO set on the drill press you still drill them blind and need to trust in the gauge not to go all the way through! I guess I could have used the depth stop nuts as a backup.

Second pair straight off the machine ready for drilling: