Hi team - I’m still chasing this forward leaning nod I had on the Z - here’s where I’m at:
I removed the whole z assembly, and then I re-tightened all of the fasteners on the x gantry and riser plates - after that tightening, the orientation of the X changed, but it’s still not square. It leans back on the left side and it leans forward on the right side.
I’m thinking I need to take the x gantry all the way off and the disassemble and reassemble everything down to the bearing blocks.
Thoughts? Any gotchas I need to prepared for? I didn’t build this machine, so this is all new to me.
See pictures for context on the amount and direction of the lean.
Reference surface is the aluminum rails - I couldn’t get the z in good enough tram to run any decent surfacing passes on the spoilboard, so I’m using the outside aluminum rails to test for square.
What really matters is your spoilboard being square to your spindle. The goal is to adjust your way to that outcome…
As you build your machine you want to try and get things as square and level as you can reasonably get them, but at the end of the day after you surface your spoilboard the first time you’re creating the surface that you’re actually going to be cutting on, so that’s the thing you want your spindle to be plumb square and level to.
Good question and solid recommendation - the rails are level and seem square, but I don’t have anything to reference that to.
As for the surfacing, I got to this point after spending a couple of hours trying to shim the tramming mount to eliminate the forward nod - you can see the amount of nod with no shim in the attached picture, and that’s on a freshly surfaced small piece of mdf screwed into the old spoilboard from the previous owner. I just couldn’t get it exactly trammed in front to back, which is what led me to the further tear down effort to find the source.
Do you suggest that now that everything is tightened back up that I throw the Z back on, surface the spoilboard, then use that as a reference to test for plumb?
I’m not opposed to that, so let me know if you think that’s a solid plan or wasted effort. Since I’m about halfway through the teardown process it seems like getting it down to the plates on the bearing blocks and reseating/tightening everything from there back up might make sense?
I have a support video call scheduled for tomorrow, but I was planning to work on it again tonight so that either all of my ideas for alignment are exhausted or I break something else that the team can help me fix…
It’s hard to say exactly what to do because every machine is different, but here’s one recomendation:
Yank one of those dial indicators off of that tramming tool. I get why they made it that way, but honestly it makes it more confusing to deal with.
With only one on there put it parallel with one axis, zero it, then swing it 180 degrees. If it’s higher or lower that will tell you witch way to adjust the tram… but I don’t think you need it at all, read on:
If your machine is reasonably well built I’d just get the spoilboard on there and surface it. If you’re out of nod/tram don’t worry about it too much at this point, just get the spoilboard leveled.
Even if you’re out of tram/nod the machine should be moving over the table just fine. (assuming your table is level, which at this point it should be)
What you can do now is get a big square (like a 2-3 foot construction square) and put it on your spoilboard again, then put it against the body of the spindle. (do this parallel with X and Y)
The spindle body is a machined surface. Using the square will tell you if it’s square to the table. Since you’re using a large square you’re bridging over any ridges created by a potentially out of tram/nod spindle. Adjust the tram/nod as needed. Then re-surface.
That tramming tool, while neat is really not a good tool to use in this scenario because if you surfaced and you’re out of tram or nod (which you likely are if you’re reaching for this tool) you could get the dial indicator into one of the many ridges that are created and you’ll never get a reliable reading.
You CAN lay something that’s dead flat OVER the spoilboard to compensate for this, but that’s just adding complexity. Just use a big square. Simple and easy.
I can do all of that except the square on the spindle - I have a round 80mm spindle from pwncnc, so the mounting collar is the only thing that the square would rest on - which might be enough to check that squareness. The previous owner had a bosch router mounted, which I did a quick test surfacing with while I was waiting on my 80mm mount to get delivered, and that’s where I first noticed the nod - huge ridges in the surfaced mdf. Since then, I’ve replaced the Z moving plate (ball end of my hex wrench sheared off in one of the bolts) and I’ve installed the new tramming plate and 80mm mount.
I just ran out to the shop and grabbed a 24" construction square and tested it against the gantry rails and the x plate using the sacrificial piece of mdf I surface when I got the z almost trammed. It’s plumb on the left side, but leans forward on the right side - which might also explain some of the hard time I had getting it trammed to begin with.
You’ll have to get creative… Maybe some 123 blocks could help you here, or some setup blocks if you have them.
Not sure if I totally follow you here, but I think you might be focusing on slightly the wrong stuff here:
Your spindle needs to be have correct tram and nod to the surface you’re cutting onto, not necessarily to the other mechanical components on the machine itself. (not to say that you can have a totally whacked out machine, but the endgame here is to get the spindle to track to that spoilboard properly)
Your spindle also needs to be moving straight up and down relaltive to the spoilboard too. You can have a spindle with perfect tram and nod, but it might move up and down at an angle. You can check this with the square against the spindle body as well.
Makes sense, but my worry is around your last point - if the gantry/rails/x plate is not close to plumb even with a trammed spindle I’ll get that ‘escalator’ effect. My great worry - and maybe a better explanation from me here - is the difference in angle of the x axis across the machine. So what I just did:
Yesterday I had the z trammed to about as close I could get it and I was still getting ridges in a raster surfacing path with a 2.5" surfacing bit - that was done on a 24x24" piece of mdf screwed into the spoilboard for testing.
I moved the x plate back over that 24" piece from left to right - on the left side of the piece, it’s plumb - as i move across the 24" of flat mdf it starts to lean or nod out on the top, so there’s a gap at the bottom between the face of the x plate and the square.
I get that all of the bolt tightening I did could have moved the x one way or the other and so now that mdf I surface yesterday wouldn’t be square/plumb today - BUT, would hope that it would be the same all the way across and not good on one side and out of plumb on the other - again, back to your original question of the x axis being twisted.
I’m not trying to let perfect be the enemy of good here, but the x axis is at least 1mm off between left and right sides, which I think will be really hard to overcome with spindle tramming only - but I’m happy to be wrong on that.
This is a bit of a Catch 22 situation… It’s hard to check for plumb without a machined surface, and it’s a pain to get a machined surface without the whole machine being together.
You’ve got these two buckets of things to deal with: Tram and nod, and plumb.
What I do when I assemble machines is this: Do the best job you can getting everything square, plumb and level. (machine frame wise)
Surface your spoilboard.
Use the square to see what your issues are. Check for plumb first on the spindle itself in X and Y using the square against the spoilboard. If you have issues with the spindle not moving straight up and down you may have to pull the spindle off and adjust the plate.
Once you get the plate moving straight up and down (keeping in mind that you can dial SPINDLE tram and nod later)
You put the spindle back on and check for nod first (since that’s the hard one to dial out)
Then adjust tram (because it’s the easist, so save that for last)
You can do all of this with the square.
NOW you should have a machine that’s probably more than good enough. If you want to to dial it further you could get something really flat (like a glass or granite plate) and put it across the machined spoilboard. A big piece that’s flat and has consisten thickenss will bridge any ridges you have. You can then use that fancy tram tool to dial out the last few thou if you want.
And then the humidity or temperature in the shop changes…
@timbo This is all good exercise, for sure. I don’t recall if you’ve used a CNC machine before this one. I don’t recall if you’ve specified what you hope to cut out with it. But here’s my $0.02:
All this calibration, recalibration, tramming, etc. can be done later, after you’ve become a capable, safe operator.
What are you cutting? What’s your plan for a spoilboard material? If “wood” for either, good luck chasing micrometers friend! Darn close on a CNC router is way more accurate than most other woodshop tool applications.
Understand that pretty much anything you cut will require at least some hand finishing.
Hand finishing is often where you save LOTS of overall project time, versus a tiny tapered endmill and 15 hours of finishing passes. On the flip side, the project sometimes dictates the approach!
Hey @DavidL and @PaulC thanks for the responses in addition to @Eric great contributions to the thread - here’s some answers (and why I’m still chasing) -
This is my second and a half machine - I bought one of the original xcarve machines off FB marketplace about 3 years ago, then converted that to a NewCarve from CNC4Newbies and ran that for awhile - it was too small and the patchwork of parts was a pain (mechanics were cnc4newbies, controller was Sienci, spindle was PWNCNC
I cut almost exclusively hardwood and some plywood, although looking forward to cutting more plywood with this machine. Spoilboard is mdf, using the matchfit work holding system
The best I could get this current setup trammed still left visible and raised ridges on surfacing passes, and out of round small holes (assume from the magnified effects of the forward nod on a longer 1/4" bit)
Understand about the hand finishing - but I would hope this machine would be capable of getting something surfaced without scallops or ridges and cutting holes that aren’t so out of round that you can see it without even measuring
I did get the right side all the way down to the bearing blocks and then reassembled it and that seemed to really help. I’ll probably eat these words here in a few minutes, but for now, the x plate where the z mounts now seems to be plumb across the 24" square of previously surfaced mdf. We’ll see if can get the z axis assembly back on - I seem to have extra bolts somehow that I of course didn’t bag and label when I took them off, so this should fun…
Thanks again for all of the responses - I’ll be back later with an update!
After reseating the riser plate on the back right bearing block, everything was plumb on the gantry - so I added the z and all its components and got the spindle back in. Tested movement and proximity sensors and they all worked so I think I got all the parts and wires back.
With all of that, my fancy little tramming tool is reading almost dead flat for the z - a couple of thousandths lean to the back now, but 1000% better. I didn’t have time to run another surfacing pass on my test mdf, but I’ll do that tomorrow and hope it’s ridge free - if it is, i’ll test circles again since those are a big part of my spoilboard work holding setup and if that passes, I think we can move on - or maybe shim the top with foil to get rid of those couple of thousandths…
I bought three of those leaf gauges that I can take apart and use for tramming they’re only about 10 bucks a piece there’s a couple of guys on YouTube that demonstrate how to use that SST gauge.
On my Bench top pro, I have to tram the X and Y occasionally, especially when I run it hard or have a collision. No doubt the tramming mount is junk IMO. Has too much wiggle. 2 123 blocks and a good tramming tool, shim stock and a small mallet for love taps is all you need to tram the spindle. put some copper anti-seize on the concentric nut and shoulder bolt. Shim stock on the bottom to bring the spindle forward in Y. K&S Shim stock verity pack from your local hardware store, hobby store or Amazon. If you spin a large surface bit, you will probably have to re-tram the spindle. I found 2" face mill is about the max. haha
In my experience, yes. Pushing it or plowing it through material too hard will offset your tram. The tram mounting system is less than desirable. If your trying to hold a tolerance of .010 to .005 large face mills will pull or push it out of tram. I’m considering using epoxy behind my tram plate to hold it in place after I tram it. Typically once its set, it shouldn’t change.
