[kostas1]

Hi guys,

I am having a problem regarding the cutting accuracy of the piece from wood (linden).
At the begininning of my tests, I had a lot of tool deflection and the resulted piece was 1mm smaller from what I had in my design (Vectric Aspire). The stickout was 30mm, quite a big one. Then, I decided to reduce the stickout and the quality improved, but not as much as I wanted.

After that, I used the following settings, but the resulted item is 0.5mm smaller than this in the design:

3.175mm spiral carbide end mill 2 flutes,
Stickout:24mm
Cut pass depth:3mm
Feed Rate 1900 mm/min
RPM: 24000

Please take a look at the 1st attached photo in the red rectangle area. This area is 8.5mm width in the design but the resulted piece is ~8.0mm.
The tool I used is from China and is carbide end mill. The material is soft wood (european linden) and it is 18mm thick.

Except from the fact that the resulted item deviates almost 0.5mm from the design in Aspire, in 2nd and 3rd attached images you can see that the cutter leaves a line per pass along the whole toolpath (6 passes of 3mm depth of cut). Even if I decrease my feed rate to 1000mm/min, I keep getting this line. Even more, if I replace the spiral router bit with a straight one, the problem is minimized but it is still there. These lines are like steps and I can feel this "stepping" with my finger.

Could you please advise to where I need to pay attention for these two problems? Could this be caused by material's vibration? I have to tell that I mount my material with screws in my mdf table and also apply hot glue. Also, I measured my machine's accuracy and it is OK (~0.03mm repeatability).

Thanks in advance,
Kostas.

[bhurts]

The lines are most likely caused by spindle not being parallel to the z axis. The steps from this also may be part of the part size coming out incorrectly. Also measure your cutter to make sure it is actually the listed size.

Ben

[kostas1]

The lines are most likely caused by spindle not being parallel to the z axis. Also measure your cutter to make sure it is actually the listed size.

Ben

Hi Ben and thanks for the reply,
if spindle was not parallel to z axis, these lines would only exists in 1 or two sides of the part and not in all around this, right?
The cutter size is what is listed and what I put in the Aspire tool database.

The steps from this also may be part of the part size coming out incorrectly.
Ben

Sorry but how this can help?

Thanks again

[ger21]

if spindle was not parallel to z axis, these lines would only exists in 1 or two sides of the part and not in all around this, right?

What type of machine do you have? Because the lines go all the way around the part, my guess would be tool deflection, or the entire Z axis deflecting. This would explain the lines, as well as the parts being smaller. If your conventional cutting, the tool will be pulled into the material, resulting in a smaller part. But with a small tool like that, your machine must have a lot of flex to get what you're seeing.
If you try climb cutting, you should end up with larger sized parts, without the lines.

[kostas1]

What type of machine do you have? Because the lines go all the way around the part, my guess would be tool deflection, or the entire Z axis deflecting. This would explain the lines, as well as the parts being smaller. If your conventional cutting, the tool will be pulled into the material, resulting in a smaller part. But with a small tool like that, your machine must have a lot of flex to get what you're seeing.
If you try climb cutting, you should end up with larger sized parts, without the lines.

Hi Gerry,this is my homemade cnc machine

Do you think my z mounting brackets are not enough for this purpose?

Video.

[IN-WondeR]

When you cut your piece, do you only cut the 3mm steps down, and don't do a finish pass??

I would after you have done your rough cut, do a finish pass, which should (theoretically) make the lines disappear. But it's not for certain. You might have tool deflection.

Have you checked up on the fact that you might have deflection of the gantry itself? It looks quite light, and not very sturdy. That could be your problem.

And last but not least, check up on your tool. It might not be the size you think it is. Some of those cheap chinese cutting tools are not what they should be.

[ger21]

What does your machine use for linear bearings? I can't see in the video.

[IN-WondeR]

They look like square type linear rails, so they are probably not to blame...
I do however think the stiffness of the gantry is a bigger issue, atleast in one direction...
It looks to be a 40x120mm which is not all that stiff.

The Z axis looks to extend well below the gantry, which means there's a lot of deflection in the backplate when it's under pressure. This is a design flaw seen on many builds sadly. Unless the Z axis moving plate is 25mm thick then there will be a chance of deflection in the plate it self.

[kostas1]

When you cut your piece, do you only cut the 3mm steps down, and don't do a finish pass??

Yes, I make 6 passes of 3mm each in the 18mm thick material wood. No finish pass

I would after you have done your rough cut, do a finish pass, which should (theoretically) make the lines disappear. But it's not for certain. You might have tool deflection.

I had for sure tool deflection at the beginning and after that I decided to use the GWizard calculator, from which I understood that I had to minimise the bit stickout. So, I respected fully the results from GWizard and the quality improved but not as much as I wanted to (0.5mm smaller piece and the pass lines were still there but lit

Have you checked up on the fact that you might have deflection of the gantry itself? It looks quite light, and not very sturdy. That could be your problem.

No I haven't done so far. I am not sure how to do so? Could you please suggest something or some link from other case?

And last but not least, check up on your tool. It might not be the size you think it is. Some of those cheap chinese cutting tools are not what they should be.

If you mean the tool diameter I have tripled check this and found it correct.

Thanks for your reply.

[kostas1]

What does your machine use for linear bearings? I can't see in the video.

These are from LSK company, here is the manual from Chai
and the page from the company

Product | Deans Manufacturing Group

[kostas1]

They look like square type linear rails, so they are probably not to blame...
I do however think the stiffness of the gantry is a bigger issue, atleast in one direction...
It looks to be a 40x120mm which is not all that stiff.

The Z axis looks to extend well below the gantry, which means there's a lot of deflection in the backplate when it's under pressure. This is a design flaw seen on many builds sadly. Unless the Z axis moving plate is 25mm thick then there will be a chance of deflection in the plate it self.

It is 40x160mm.

The Z axis plate is 20mm thick and the tool's diameter is 3mm. So small tool can cause such a gantry deflection? Please remember that I get these lines and the inaccuracy even if the cut pass is 2mm deep.

Thanks.

[kostas1]

Here are some photos from my cnc

[IN-WondeR]

The problem with your gantry is the thickness of it.

40mm is not a lot, so deflection is definetly possible in this.
The z plate looks to me, to sit way down from the bearings on the Z acis rails. This (even though it's 20mm thick) can also cause deflection.

What is your tabletop made from?? I see no tabletop on your router in that cideo.

[kostas1]

The problem with your gantry is the thickness of it.

40mm is not a lot, so deflection is definetly possible in this.
The z plate looks to me, to sit way down from the bearings on the Z acis rails. This (even though it's 20mm thick) can also cause deflection.

What is your tabletop made from?? I see no tabletop on your router in that cideo.

So you say that I need to lower my gantry and replace the 40x160 profile with a stiffer one?

The tabletop is from one sheet of 16mm mdf above 25mm of another mdf, so totally 41mm mdf table.

[IN-WondeR]

A stiffer Gantry will definetely give you less deflection.

Regarding your Z axis. If you don't need all that clearance you have when the z axis is in top, then yes, remake your z axis, and make it higher, so you can move your z axis all the way up, even with a long routerbit mounted.
I sadly did a mistake on my own, not making it possible to clear more than a 50mm long router bit from the gantry bottom.

Your table top is a good setup. I run the same thickness, although I use 22+19mm.

What do you use as hold downs??

[ger21]

Are you climb or conventional cutting? You really shouldn't be seeing much deflection with a small bit, and shallow passes in a soft wood. Although that tool is very long for such a small diameter. Have you tried with a 6mm tool? I'd also recommend trying a downcut bit.
From my experience, softer woods are much more difficult to cut cleanly than harder woods.
I was cutting some poplar today, with a conventional cut, and the scrap side looked very much like your cuts, which leads me to think you were climb cutting. I'd make sure your conventional cutting, and as mentioned earlier, leave a small amount (.25mm or so) for a full depth finish pass.

[kostas1]

Regarding your Z axis. If you don't need all that clearance you have when the z axis is in top, then yes, remake your z axis, and make it higher, so you can move your z axis all the way up, even with a long routerbit mounted.
I sadly did a mistake on my own, not making it possible to clear more than a 50mm long router bit from the gantry bottom.

Don't you think my z axis is not high enough? It has 150mm travel. Sorry, am I missing something?

What do you use as hold downs??

I put few screws and also some hot glue.Then, I put some plastic washers of 1.35mm thickness underneath the material in order for me to take out the finished part and the material too more easily.

[kostas1]

Are you climb or conventional cutting? You really shouldn't be seeing much deflection with a small bit, and shallow passes in a soft wood. Although that tool is very long for such a small diameter. Have you tried with a 6mm tool? I'd also recommend trying a downcut bit.
From my experience, softer woods are much more difficult to cut cleanly than harder woods.
I was cutting some poplar today, with a conventional cut, and the scrap side looked very much like your cuts, which leads me to think you were climb cutting. I'd make sure your conventional cutting, and as mentioned earlier, leave a small amount (.25mm or so) for a full depth finish pass.

I haven't tried to climb cut yet at all and haven't tried with 6mm tool yet. I will be able to do so this weekend

As for the shallow passes, I have 0.4-0.5mm deflection in passes of 1.5mm depth and the stickout was pretty high at about 30mm. If I cut with the same straight bit and stickout but depth of cut = 3mm, then the defleciton is getting bigger, and the resulted item is 1mm smaller than this in design.

However, as I said to my first post, when I decided to use GWizard calculator and respect its calculated stickout (24mm), feeds (1800mm/min), depth of cut(3mm) and speeds(24000 RPM) with the 3.175mm spiral carbide bit, the result was quite the same (0.5mm less than this on design) like as I was using the straight bit with stickout 30mm, depth of cut (1.5mm), feeds (1000 mm/min) and speeds (20000 RPM).

So, are these clues an indication of gantry being deflected? How could I measure the machine's deflection?

Does this has to do with the bent ballscrew in gantry's y axis ? I opened a new thread here a week ago. I didn't mention this till now,because I was not sure if it plays so significant role on machine's accuracy and resulted item size.

[kostas1]

Here are some more photos from the output parts and a photo from the Z axis parts.

In two of them (DSC08093_s and DSC08092_s), there is a red rectangle which emphasizes the toolpath starting point. This may be where the deflection is happening.

The last one (DSC08091_s) describes part from my previous post: 12 passes of 1.5mm each (18mm thickness material), straight carbide bit, stickout 30mm,. Please notice that the lines are now only in the circular side of the part and not on straight side. This parts is closer to the Aspire design size and it is about 0.4mm smaller.

[kostas1]

Ok guys, here are my test results from the weekend just past.
Firstly, I measured the deflection in Y axis by getting the gauge probe touch the Z axis along the Y axis and pressing with my fingers the Z axis and then the spindle. Below are the results when Z axis was up and when it was down:

Z Axis Up
Gauge probe touches Z axis side. Force is applies from the other side of the Z axis towards to the gauge probe:

1 finger 0.07mm
2 0.10mm
3 0.14mm
4 0.19mm
hand 0.25mm

Z Axis Up
Gauge probe touches spindle tool bit mount:

1 finger 0.12mm
2 0.18mm
3 0.24mm
4 0.28mm
hand 0.35mm


Z Axis Down
Gauge probe touches Z axis side. Force is applies from the other side of the Z axis towards to the gauge probe:

1 finger 0.10mm
2 0.14mm
3 0.19mm
4 0.22mm
hand 0.38mm

Z Axis Down
Gauge probe touches spindle tool bit mount:

1 finger 0.16mm
2 0.19mm
3 0.22mm
4 0.26mm
hand 0.40mm

After that, I tried some cutting tests with 24mm stickout and 6 passes of 3mm depth each pass in 18mm thick material (wood linden -> soft wood):

The climb cut just resulted in a bigger piece which was far from this in the design (aspire).
Then, I tried a rough cut and then I left 0.5mm (a little more than my max deflection 0.4mm in above test) for the finishing path and the result was almost perfect with the size I want and no 'side step lines'. This led me think that my gantry's vertical aluminium profiles are thin and they are not mounted properly in the horizontal aluminium profiles that run along the X axis. Please see the attached image for better understanding.

My thoughts for reinforcing the gantry are:
1) Replace the vertical and horizontal 40x80mm profiles with 80x80mm alu profiles,
2) Make the vertical profile a bit longer so as it reaches at the very end bottom of the gantry side
3) Replace the little plate with a larger and thicker one.

My questions:
1) What are your thoughts on this please?
2) Is it apparent that my cnc produces these faulty parts because of this deflection?
3) Should I look into something else to find the reason?
4) Is there any setting in aspire that lets you rough cut and then finish cut without requiring from you to draw duplicate designs (one small and the other a bit bigger)?

Thanks again.