Wobbling curve-cuts

Hey guys,

I recently finished my primo-build and hope to get some advice with a problem that shows up from time to time.

About the machine:
The work-area is quite large, roughly 700 by 500mm. But as recommended I kept Z as low as possible with 81mm. Also I have a 19mm MDF Plate, that serves as a protection for the table and also reduces Z-height. For electronics I use the SKR-Board, a 55mm Daedalus-Spindle and quite big NEMA17, which run on 2.1A, respectively 65Ncm.

My preferred material is MDF and as a tool I mostly use a 6mm wide, 21mm long 2-flute bit from Sorotec. It runs on 12.000rpm with a feed of 20mm/s and a depth per cut of 3mm.

The problem:
Right now, I’m trying to mill some clocks and sometimes, when making the last ciruclar cut, the tool begins to wobble, resulting in a lot of noise, an uneven cut and me hitting the emergency stop.
Yet, this occured only in deep, curved, single-line cuts and only in one of the later passes (i.e. 15mm deep on a 19mm MDF, so 5th pass). Neither engraving, nor pockets or shallower (<10mm) single-line-cuts have been a problem so far.

Has anyone of you encountered a similar issue? I’d be happy if someone had a workaround!

As I, being new here am only allowed to post one pic per post, here one of the whole build

1 Like

Climb or conventional milling? I would see that when i went too deep on a pass in conventional milling. Also check core is not loose and the grub screws are tight.

1 Like

Thank you for the hint!

I tightened the gantry rail tension bolts at the core and managed to at least cut out the two clocks. Not without wobbling, but at least with less enough to not hit the emergency stop again - so i guess that was part of the problem. A partly uneven surface still remained…

So far, I only used conventional milling, because Estlcam recommends it for hobby-machines in its presets. I’ve checked ryans milling basics and he recommends the opposite, so I gave it a try.
Test-setups were two circles, both 19mm deep and defined as parts in Estlcam. The upper one was milled conventional and started to wobble at 15mm depth, but I let it finish.
The lower one was set to climb milling and startet to wobble much more than the other (also at 15mm) so I had to hit the brake.

Setup:

The wobbling occured 5 times somewhere in between 7 and 8 o’clock (…literally :frowning: ). Once it occured in between 4 and 5, but to me it seems that it has something to do with this curve movement.

Still new, so here the details…
Conventional:

Climb:

When i have to similar problems it was a the set screws in one of the motors on the rails that worked lose so that at random times it skipped steps, did you check the set screws on the gears connect to d xy motors shaft. U can take a flat blade screw driver and carefully try moving them to see if they are loose. Or not that check d clamps the hold the X & Y gantries to the X & Y ralis

1 Like

Double check belt tension is not too loose or to tight. Roll the core around by hand a check it is smooth and not binding or loose anywhere. I had a warped rail when i first assembled mine.

1 Like

Well, as tightening the core-bolts seemed to help, i tightened even a bit more. Before, it was possible to tilt the core a bit towards each gantry rail, as well as diagonally to both. Not much, but enough to make me think, this might be the reason…
As you can see on the top circle, it was not - the vibrating noise was also in the upper passes and it got worse and worse the more the job passed on…

After that, I read the posts of @Hollee and @Mattk331. I re-loosened the core-bolts I tightened before. Also, I checked for belt tension, grub screws and tightened the gantry rails to the trolleys. On the 2nd job (bottom circle) I had less noise in the upper passes, but in the lower it started to wobble again.

Is it possible, that all my screws are too tight, causing the machine to block itself in movement? In the assembly-instructions we were told not to screw them too tight. I was fully aware of this and thought I managed to keep them quite loose - I started tightening them this morning. But maybe my wrist is stronger than I know :smiley:
Are there recommendations on how much Ncm to apply on the screws? I guess, something like this might be easier to follow, than “1/16 turn past snug”

More than 99% of the torque applied to the bolts is simply to overcome the resistance that the nylock nuts provide.

I have some very high precision torque wrenches, and almost every bolt gives me a different result based on how the threads cut into the nylon ring. Also, because of the nature of what it takes to make the bearings contact properly, it changes, because we are flexing the printed plastic piece, and something that is already somewhat loose will need more flexing (More torque) than something that is already snug. It’s a tough call for how to torque these up, but there should be a somewhat wide range of motion acceptable.

I can say that I’ve used both climb and conventional milling, and my Primo (25" by 37" work area) performs quite well at 1800mm/min with a 3.5mm DOC.

All of my bearings in the core and trucks are tensioned to have contact with the steel. I use the guage of finger pressure on the bearings trying to make them spin against the steel. (Usually with the fitting pushed a little to reduce the friction of the bearing against the steel.)

Back off the tightening bolt (In the core bearings, those are the ones that go straight down. In the trucks, they’re the ones that go horizontally across the top) until there is space between a bearing and the steel rail. Tighten the bolt very gradually until all of the associated bearings are in contact with the steel, and it is no longer possible to spin any of the bearings with a fingertip.

I tighten the core X and Y bearings first, then the truck ones. A little extra tension as per the squaring instructions on the docs page on the truck bearings can help with keeping the core naturally square, but I don’t stress about that much. The motors will keep my machine square in operation, and 5mm or so (Probably up to 10mm) isn’t a problem.

The Z bearings are a little different. The little wedges allow some variation in tightening the Z bearings. Mine are probably still a little too tight, but I find that the Z bearings are more forgiving that way. These also have to be in good contact with the Z rails. Any slop here will translate into bad cutting. Anything that allows you to move the core without it moving a motor is something that is going to cause you grief, and will result in some bad cuts.

1 Like

Hey Alex, I have got the same problem (not as bad as you though, it does not really show at all, but it is there, in harder wood it sometimes begins to stutter) and after 1.5 years I haven’t found the culprit yet. It only happens when the core is moving “away” from the router in both directions. It might be a problem with my Z-rails, since one nut trap is exactly on top of the weldseam on the inside of the pipe. I tried filing that down, tightening the screws etc, but it keeps happening. Since it does not really change the accuracy of my cuts but just sounds a little stuttery, I have got used to ignoring it.

I am from northern Germany btw… :slight_smile: Sorotec is nice.

I have seen problems like this if you tramming is not square. The deeper you go it is more apparent and as the mill begins to rub and vibrate when it hits harder cross grain in your material. It will really maginfiy the give in your Z axis. Tramming is squaring of you router bit to the x/y table.

Wow, thank you all for your comments!

I disassembled the machine and set back all bolts to the minimum tension, just with contact to the plastic parts. I can’t really get my Y-gantry square, als the top truck is a bit pulled back by a cable chain, but it is within this 5-10mm @SupraGuy mentioned and I have the dual-endstop-feature installed.
To be square while tightenig I started the machine, homed X and Y, let it move around a bit to settle and then turned the tightening-bolts on the core so the bearings were just “…no longer possible to spin any of the bearings with a fingertip…”.

What @larryk stated seems quite reasonable, so I tried to ensure the z-axis sits quite tight. Before, the top-right and bottom-left bolts were quite a bit more loose than the other two. I compared by measuring the thread that sticks out of the nut - first it was 5.4mm (tl&br) respectively 4.6mm (tr&bl), now they are all 5.4mm.

Anyway, the next attempts werent successful…

I am still (respectively again, after loosening th bolts) able to tilt the core a bit towards the gantry rails, as you can see in this video. Is this normal? Also, the second video shows the process and the crash.

As you can hear in the longer video, the “stuttering” @Tokoloshe mentioned appears on my machine too. Usually in first pass, and then again, when the crash happens. I guess, in the beginning its because the “outer layers” of MDF are a bit harder than its core. When the problem occurs, we are in the last passes, so “outer layers” again and maybe this adds up to the problem…
Also checked for the weldseam, but they are both not in contact with it.

Also I’m qute sure, this all somehow is linked to the motion, as the crashes mostly appear in a X-positive/Y-negative set of motion:

1 Like

Is the holder of the router really tightly screwed into Z?

I checked that, the router sits tight to the z-axis.

I also checked tool orientation, and the bit sits rectangular to the surface (at least as precise as the piece of woods 90°-angle is…)




On a closer look, the tool is a bit worn off, so I ordered a new one. But I guess its from the crashes, not from before.

First, very nice video and pics. You are correct, the straight bit is looks to be chipped near the end and 2/3 of the way up I see some nicks. The video’s show way to much movement if you are not putting to
much pressure on it.
@superguy comment on Nylock nuts is so true. My first time setting up a core I didn’t have the right Nylocks and I used regular nuts. It was so easy to set up and square, no slop. The reason was that the
tenison I felt on the bolt and nut was the real presure. When using the correct part (nylock) you really can not tell other than if the tubing is really against all 3 bearing. My second build I put Nylocks on the first time . Cranked down on the nuts to set the bearing on the tubes. They were still sloppy. As a sanity
check I remove the nylocks, and put regular nuts back on. Wow, no effort to tightin up and remove slop and not break the plastic. Put the nylocks back on and tighten the nuts till the bearing were snug again.
The trick is , take a nylock and thread it all the way on the bolt and get the feel of how much pressure it
takes. It will surprise you.
The other thing that will help the life of your bit is to ramp down instead of plunge. End mills really don’t like plunge cuts. Tends to dull the end and create heat, everything else looks good. Get that core slop out and I believe you will have chips everywhere again.
Welcome to the world of V1 cnc

This is what I love about this forum! If someone has an issue people jump in with suggestions from their own experiences, or should i say, challenges?

2 Likes