Hmm, do you have a concrete proposal? For a second I thought maybe ‘500-(83/2)’ but if the harmonic is at 1/6 and I decrease the frequency, won’t the harmonic just move with it, to 76Hz?
1 Like
450Hz ? That would be 27kRPM for the spindle I think. If it still resonates there then I’m talking rubbish as that would be about 5.4 times the 80-83Hz resonance.
Looking back up at your previous data it seems there was a strong resonance mode at 24k when testing the spindle nuts which would be about 400Hz so best stay away from there.
1 Like
Okay, put it at 27k RPM and it looks like a definite improvement. Results.
There was still a little bit of vibration around 80-90Hz but this time it only really seemed to happen when moving along the Y axis and only on the right hand side of the pocket, not the left. I speculated that it could be because my air nozzle is pointing at the endmill from the bottom right corner, and it’s slightly more right than it is bottom, so it could be chip clearing, as it would be difficult for the nozzle to get air in there.
So I decided to test it again but after moving the nozzle to be positioned towards the bottom. Results.
That also left me confused, as now the vibration had moved to the top of the pocket but not the bottom (and the bottom is what I would have most expected if chip clearing was the issue.
So I ran it again, this time with the nozzle pointing from the bottom left towards the endmill. Results. Same result, vibration along the top of the pocket, so I think nozzle placement isn’t the problem.
Then I moved it back over to the right side of the machine, leaving the nozzle as it was. Results. Once again, vibration seemed to be on the right edge of the pocket.
So I feel thoroughly confused but the vibration seems to have something to do with the position.
Anyhow, the vibration no longer seems too substantial as long as I’m doing dynamic tool paths, so I think I might try to continue using them. The machine seems to survive so worst case, I break a few endmills.
1 Like
Well,
It sounds like you’ve found a spindle frequency that’s less effective at exciting the major vibration mode(s) of the machine which is good.
As for different vibrations at different positions on the machine, that’s exactly what I see on the Shapeoko and what my simple models suggest. The effective stiffness of your X rails for example, changes rapidly as you move from the center towards either end. Are you seeing more vibration when the Z carriage or the Y carriage are towards the middle of their stroke and furthest from solid supports?
1 Like
And just to make you feel better, my 2.2kW high quality Chinese spindle has started to make an awful screeching sound thanks to the ‘lubricant’ in the bearings all going dry and crunchy. I’ve had to try to squirt new grease into them to keep it going now.
Fortunately some helpful person has put a video on YT of exactly this problem (beware the horrible noise at the beginning which is what I heard when I shut off the dust extraction)
Although his spindle seems to have two fewer bearings than mine which are, amusingly, marked for alignment like proper SKF angular contact bearings would be.
1 Like
Ugh, so I spent an hour this morning testing 10 endmills but when I came to analyze the data, turns out the accelerometer had written all zeroes for some reason 
Still, I did take photos with the microscope and that’s still interesting (open these in a new tab to see full-size).
| Endmill | Before | After | Pocket bottom | Pocket corner |
|---|---|---|---|---|
| CNCFraises FC1DA3008 | ||||
| CNCFraises FC1D3008EVO | ||||
| Carbide 3D ZrN Single-Flute | (Missing) | |||
| DIXI 7561 | ||||
| Sorotec Alu | ||||
| Sorotec ProAlu | ||||
| Sorotec Alu-B | ||||
| CNCPlus 102-30 | ||||
| Janpo 1MEM | (Missing) |
I’ll have to re-test most of these except of course the Carbide 3D endmill, which is the only one that broke. Here are some more shots of the broken endmill:
The only conclusions I feel I can make right now are:
- The CNCFraises, DIXI, Sorotec Alu and Sorotec ProAlu endmills all left pretty decent surface finishes as-is
- The CNCPlus 102-30, Sorotec Alu-B, Janpo 1MEM all left pretty meh surface finishes
To save on time, I might only bother with the first set of endmills when I re-test.
4 Likes
Great pictures and interesting results.
The loss of the accelerometer data might not be so tragic. It might have been hard to tell if any differences were a cause of wear or an effect of wear.
The main thing I’m hoping for with the accelerometer data is that some endmills will require less force than others to make the same cut, due to sharper edges or less friction. That would then show less vibration. Or maybe rigidity of the endmill could be a factor.
2 Likes
What are you cutting and how?
So I tested a bunch of endmills:
“Why are they all Cncfraises”, you might ask? That’s because right after I started cutting with a DIXI Polytool 7561 endmill, my accelerometer died. I saw that the light was off, so opened it up and discovered that the cavity that it sits in was filled with chips, which presumably had shorted something. I’ll need to solder up another one or replace my Teensy (in case the short broke it somehow).
Good news is that it looks like there is a point to all my testing. If I look at the last loop of the spiral path, I can see that the level of vibrations are pretty different for all the endmills:
- EVOMAX
- FC1DA3008
- FDC1DA3008EVO
However we know that some of this is up to where the machine happens to be at the time, so to exclude that, I compare the lowest levels of vibration on Z only, in which case the results are:
- EVOMAX: 7.26
- FDC1DA3008EVO: 12.01
- FC1DA3008: 13.68
So the EVOMAX endmill has substantially lower vibration than the other two, indicating that the endmill does make a difference and making me feel much better about all the effort I’m going to.
25mm square rounded pockets in 6062 or 6081 Aluminium (can’t remember the alloy) using a Solidworks CAM Volumill toolpath, 5mm ap, 3.175 ae, 27k RPM, ~1130 mm/min feed rate.
2 Likes
Are those all the same helix angle?
1 Like
The helix angles aren’t specified but eyeballing them next to each other, the EVOMAX has a steeper angle than the other two. Both of the other two have the same angle.
2 Likes
So the Evomax is spreading out it’s pulse of cutting force over slightly more time?
1 Like
Hmm, I’m trying to wrap my head around the question but it doesn’t make sense to me. The amount of time spent in the cut is defined by the RPM and chipload, no? With 27k RPM, it takes 2.2ms to make a rotation, regardless of the helix angle. Then, how much of that 2.2ms rotation is actually spent cutting is defined by the angle of engagement, which is a function of the endmill diameter and feed rate.
My guess is that the helix angle would only affect the velocity at which chips are evacuated?
1 Like
I think the difference is that, over the depth of the cut a more helixed cutter will be engaged somewhere in the cut for a larger angle of rotation, starting with the tip at the bottom and ending with the last bit of flute in use at the top of the cut. The ideal with a deep cut being to have some bit of one or more flutes engaged all the time.
2 Likes
Tested 3 more single-flute endmills:
Looking at the same vibration troughs and combining with the ones above:
- Cncfraises EVOMAX: 7.26
- Sorotec Alu: 8.02
- Sorotec ProAlu: 8.23
- DIXI 7561: 8.96
- Cncfraises FDC1DA3008EVO: 12.01
- Cncfraises FC1DA3008: 13.68
I don’t think I’m going to continue using these toolpaths though. The chatter seems okay when I’m in the room but that high-intensity ~85Hz vibration travels really well through the walls and I don’t want to bother the neighbours too much. I’m going to be using multi-flute 6mm tools for most of the grunt work anyway.
That would require an unlikely 63 degree helix angle. With a more reasonable 25 degrees it probably looks like this.
@Moded1952 Is it possible that the cutter impacts are acting as impulses exciting workpiece/machine resonances (“ringing the big bell”)? Maybe measure the machine’s impulse response?
@spargeltarzan’s calculator predicts “High tool stress”. Do you think that overheating is also causing and/or contributing to the endmill damage?
Earlier, at 30k, it was doing it even more, so I dropped the spindle RPM to 27k and it’s still a problem. So I don’t think it’s that.
Sorry, I derped when I gave you the AE earlier. It’s 0.3175, not 3.175. Divide by 10.
It only takes one impact to ring a bell.
Yep, but hit it 27,000 times and that bell knows it’s been rung 