Hi, I’m using .5 mm endmills to cut .15" deep, .5 mm wide slots in hardwood. I’ve seen a few recommendations for spindle rates, depths and feeds here and there but still I snap these things off almost immediately. Can anyone give me a good starting point for best practice on these?
What sort of runout does your spindle have?
What feeds and speeds are you running at?
I don’t know the runout. I am leaving as little of the endmill out of the collet as possible. I have tried combinations from 10K to twice that, as slow as .1 ips, ramping plunges, down to .5 mm cuts, to faster feeds. Not finding any pattern. These are indeed cheap endmills but brand new.
The adapter fits in my collet suspiciously tightly, I am trying to figure out how to explore that possibility. Just trying to get a recommendation for a starting point that ought to work.
I’m checking into vibration in my router as well.
Are you using a Carbide Compact Router?
Check in w/ @TDA and
Write in to sales@carbide3d.com and see when we’ll have a spindle available? Not sure if we’ll support endmills that small.
No, sorry, I should have indicated that. It’s a Makita in a Shapeoko XXL. Original Z mechanism. I am using a new and trusted 1/4" to 1/8" adapter with it.
Collet adapters typically are not the best choice when you need to worry about runout. If you can, try using a dedicated collet rather than an adapter.
2 Likes
First thing to do is to replace the adapter w/ a precision 1/8" collet:
or
https://www.elairecorp.com/makitaroutercollets.html
Might want to replace the bearings — folks have mentioned doing this w/ ceramic bearings which have worked out well — or, buy a couple of Makitas, test the runout on them, and keep only the best one.
Then, start w/ the smallest endmill we sell:
and gradually work your way down in size from there.
Usual preface, I’m with PreciseBits so while I try to only post general information take everything I say with the understanding that I have a bias. Going to be a little harder here as there is some specific quality issues but I’ll do my best.
This is probably one of the main issues. The old rule of thumb was that 10% of the diameter of the cutter in runout would break the tool. This isn’t completely true anymore and there are ways around it. The real problem here is that the runout is added to the chipload (for worse case in multi-flute cutters). So an easy way to look at this is that the runout is adding a chipload (feed) even if you are only plunging. If that forced chipload is too much for the tool it will break the tool regardless of your set chipload/feed.
There are 2 types of runout, angular (skew) and radial (offset). Easy way to think of it is that angular would be if the bore of the collet was “diagonally” bored. Radial would be where the bore is straight but not centered. If you have radial runout it won’t make any difference how much is sticking out the collet.
Primarily what you are dealing with is cubic material removed per flute. The more this is the stronger the tool must be to resist the cut. So the typical ways to do this are lower chipload/feed, lower cutting pass depth. The issue is that if you have enough runout you don’t actually completely control your feed/chipload so all you can do is try to take shallower passes and hope that the runout is low enough to have a workable chipload somewhere.
This may or may not be fine but the primary things that cheaper endmills have is generic geometries and cheaper carbide. e.g. the lower the rake and helix the less shear and potentially larger the “edge” of the endmill is. This mean that it will increase the cutting forces required for the chip to cut. The helix can actually go either way as the tighter it is the more of the cutting force goes up the tool but the closer the flutes are to each other which makes the tool weaker as there’s less mass in the tool.
The carbide quality difference is huge. You have the basic size grades (micron, submicron, etc), then breakdown by ISO. However, it goes even further than ISO grade. Let’s use 3 carbide examples in ISO K20:
Chinese YG7
T.R.S= 1.9GPa HRA=90
Sandvik DH20
T.R.S.=2.8GPa HRA=92.3
Mitsubishi MF20
T.R.S.=4.4GPa HRA=92.8
T.R.S. is transverse rupture strength and HRA is a Rockwell A test. Clearly all these carbides are different even though they share the same ISO grade. Just the differences in these 2 properties will absolutely change the tool life, and stress the tool can handle. They can also greatly effect the cost of the blank and grinding. While there are other things that effect it, the TRS here is one of the main pieces of data that will predict when a tool will break due to the forces of the cut. The Rockwell hardness will effect how quickly the tool will dull and therefore how long until it takes more cutting force to make the same cut.
Every step between the router/spindle and the tool can add runout. Trying to use tooling of this size with multiple consumer use tool holding will lead to problems unless you are the luckiest person in the world. From our experience for stock router collets you are usually looking at a minimum of a couple thou runout. The best 1/4" to 1/8" reducers that we have seen are at least 2 thou. So lets say that you have 2 thou in the router collets and 2 thou in the reducer. That’s 0.004" potential runout which means that at that 10K RPM on a 2 flute cutter you are going to be functionally running the tool at 80IPM (2,000mm/m) when you plunge it.
Short version:
Get your runout down.
Take as small a pass depth as possible to reduce cutting forces. Increase feed/cutting depth if successful.
Look at the tooling. One other thing than the above is remember that every part of the tool will be used, either for you, or against you. So as an example don’t get tooling with longer cutting lengths if you don’t need it.
Hopefully that’s helpful. This was written with tons of assumptions and skipping specifics to more easily make the points. So I can go into more detail if anyone wants it. This is probably long enough for now though…
18 Likes
All this is very useful information and I am sure will help me run down my problem. But can no one answer my simple basic question? Where is a good starting point that, if it doesn’t work, I KNOW I have a problem?
I can’t start larger and go smaller. The part requires this size cut. Thanks for all the suggestions and I will first try the custom collet.
The short answer is not without more information. There are too many variables. Need to know the tool geometry, material being cut (specific wood) at a minimum to even take a guess at it. The material will have minimum chiploads and the geometry will help determine safe feeds.
8 Likes
Many thanks to all of you!
1 Like
I just love when you chip in like this every now and then! Your posts are very informative John!
Thank you!
4 Likes
For anyone else coming across this who simply wants an answer to my question instead of advice to buy a half dozen things first, you will find this discussion helpful:
1 Like
Was this PUN intended?
1 Like
Thanks, I appreciate the comment. I’m glad that it’s useful.
If not it should be.
Please don’t take the following the wrong way. I’m primarily addressing this as I don’t want others to be mislead.
There are no easy answer to your question without more data (need specific material and basic tool geometry at a minimum). You might find something similar that will work for you but that again depends on the variables not listed here being similar to the ones you find elsewhere.
Normally you have a big enough range that you can run with a conservative estimate/example. When you get into micro tools that range is so much smaller that it becomes an issue. It’s also why everyone starts to jump at the runout numbers as you no longer have as much room for them to be absorbed. I’ll give some examples below but they are by no means a comprehensive list. In the following when I say “changes chipload” that means potentially both the minimum and maximum chipload before breaking the tool.
Tool geometry:
Number of flutes
Obviously a 3 flute cutter vs a 1 flute cutter is going to cut at very different feed. Additional to this all things being equal there will be a large change in flute volume and core potentially changing the chipload of the tool.
Aspect ratio
This is the ratio of the length of cut to the diameter of the tool. The greater the ratio the more fragile the tool. So someone using something like a stub endmill where the length of cut is 1.5x the diameter will have a different experience then someone using a deep reach that is 10x.
Upcut, downcut, 0°
A upcut tool will have the least load and the 0° will have arguably the greatest. A downcut tends to pack the flutes and has more force issues due chip flow and tip geometry. These can all change the chipload.
Helix
As previously stated this can change the direction of force and strength of the cutter. This again will change your chiploads.
Rake
Rake is basically the angle of attack of the flute. The more positive rake you have the less force to make a cut but the thinner and weaker the edge gets. So if you are using a more generic cutter the rake will probably be causing more load and again change your chipload. This doesn’t address neutral or negative rake which change things even more.
Material:
Hardness
To give some quick references wood hardness is usually measured by the Janka scale. Silver maple is 700 lbf, birdseye maple is 1,450 lbf, brazilian rosewood is 2,790. These can’t be cut the same, you will break a cutter at too low of a chipload in the silver maple due to rubbing while maybe getting away with that in the birdseye, and being the true minimum for the rosewood. Alternatively the chipload that will work for the silver maple may break in the rosewood.
Expansion, binding
For lack a another simple explanation this is how much room the cut material takes in the flute before packing and if there are additional components in it that make it “gummy”. Not just chipload in this case but also required changes in geometry.
Material strength
For this purpose this is how much the material is bound to itself. It’s ability to resist the cut and stay together basically. This can greatly change the chipload numbers. This is different from hardness as the grain structure can change it (hard but brittle depending on grain direction and size).
Now lets apply this to the first example in that post. they say they are using a 0.024" cutter at 24KRPM and 40IPM and 0.010" pass depth. So lets go down the list. How many flutes is it? That isn’t in the first post but further down they say they are using a 2 flute and their friend is using a 3 flute. That’s all fine as long as it’s at least stated and you change it by a ratio or chipload per flute.
So let’s calculate the chipload. Chipload = Feed/RPM/Flutes so 40/24000/2 gives us a chipload of 0.00083" (0.021mm). We’ll assume for the time being that this is actually cutting although that’s impossible to tell without the material and runout.
The next thing to look at is the aspect ratio. They list it as 0.12" so 5:1. If ours is the same or shorter then we can maybe use this data. If it’s longer might just snap it right off.
We don’t know what this tool is but most likely an upcut. So again we can assume that if we have the same it should work. However if we have a 0° tool we are engaging the entire flute at once and causing more load that can again break the tool.
Don’t know their helix angle either. Maybe they are using a 20° and we have a 35°. Like for like we again might be breaking the tool.
Rake is also not know. Could be that they have a soft media tool with high rake and we have a lower rake general use tool. That will equal more load and again we are potentially breaking the tool.
Material is a whole mess too. Since they are cutting fingerboards, and looking at their small chipload the are probably cutting rosewood or ebony. This is a very hard wood, that acts almost more like a composite due to the integrated silica in the grain structure. Most of them require little flute volume and don’t bind up like more oily woods. This will work the opposite of most of the above. Using these numbers in too soft of a material will cause so much extra load and heat from rubbing that you will break the tool from too small a chipload. The other possibility is they are cutting a softer wood but have a decent amount of runout effectively moving all the cut to one flute at a much higher chipload.
Some of the above are minor changes and might not have a great effect on their own. However, most likely a few features will be different from any other example and that could change things significantly. The material changes can also be so great that they alone determine a great deal in terms of minimum and breaking chiploads/feeds.
Again, this is isn’t to attack your post. I’m just trying to point out that there is no easy answer here with this number of undefined variables. Even if this one worked for you it wouldn’t for the next person with a different enough tool or material. There are potentially good starting points with more data but not universal ones.
8 Likes
Deflection is greatest when plunging, and endmills are 4 time better at cutting sideways than down.
1 Like
Forgive me. Would certainly need an end-cutting mill / flat bottom drill.
Was spit-balling ways to rough it out to reduce material & make milling pass easier.
Perhaps the very tiny DOC is still the best option??
Would it be possible to just change the geometry? Do you really need a 0.020" wide slot 0.150" deep?
You might be able to predrill both ends of the slot so you are only side cutting but its a PITA to get a drill that sized runout down to acceptable limits unless it’s something like a Harvey with reduced shank.
Come to think of it, I might have some Harvey 0.0157s left over from a Microdrilling video but they aren’t cheap and runout has to be pretty much 0.0000.
Going even to a 1/32 sized endmill will increase tool stiffness quite a bit over a 0.5mm. Also look into Datron single flutes, pretty sure they have a 0.5mm. And yes, shallow with low doc is your best bet.
4 Likes
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.