Tutorial on feeds and speeds

(William Adams) #1

I’m afraid that this will be a bit of a cheat — I’m going to try to crowdsource an understanding of this. Let’s start by looking at all the variables:

Material being cut — this should be groupable into a couple of categories, ideally matched up with typical endmill manufacturer recommendations:

  • Hardwood
  • Soft plywood
  • MDF/Particle board
  • Plastic (hard)
  • Plastic (soft)
  • Acrylic
  • Metal (ferrous/non-ferrous), specific alloys and elements (are there any elements in their native form which are usefully machined?)

(from: http://www.pdsspindles.com/engineering-speeds more-or-less — other useful suggestions would be welcome)

Machine type:

  • Shapeoko 3/XL/XXL — note the sub-variant issue of attached spindle (Dewalt, Makita, other) and possible speed range
  • Nomad 883

(going to limit it to current production Carbide 3D machines)

Endmill diameter:

  • 0.03125"
  • 0.0625"
  • 0.125"
  • 0.25"

Number of flutes and flute length and geometry

Endmill material (carbide/HSS)

Endmill coating

Spindle RPM — as part of this, filled in the chart: https://www.shapeoko.com/wiki/index.php/Materials#Router_Speed_Chart

Machine Feed rate (and plunge rate, which is often half or one-fourth of that) — One thing I’m surprised at is that there isn’t a specific inter-relation between SFM and the feed rate and climb / conventional milling. Logically, when doing conventional milling, the feed rate is effectively added to the effective speed of movement of the circumference of the endmill against the material, while in climb milling, it should be subtracted (and if one could move the machine at a feed rate equal to the SFM, then the rotation and movement would cancel out and the endmill would just revolve around the perimeter of where it should be cutting, as if it were a wheel).

Chip Load — this creates something of a circular effect, since it is derived from a number of values, but the desired chip load is determined by the endmill, its diameter, number of flutes, and material being cut per the endmill manufacturer’s recommendations.

Additional information which may be calculated:

I would like to fill in this post with additional information as it is found, puzzled out, derived or asked about.

Okay, I’ve gotten everything filled in and calculated for the official feeds and speeds at: https://www.shapeoko.com/wiki/index.php/Materials#Shapeoko_3


Harvey Tool has a PDF on climb vs. conventional milling: http://www.harveytool.com/secure/Content/Documents/Tech_ConventionalMillingVsClimbMilling.pdf and of course, there’s a wiki page on it. https://www.shapeoko.com/wiki/index.php?title=Climb_vs._Conventional_Milling

Feed/speed adjust 1/4 to 1/8" bit
Shapeoko Feeds and Speeds Chart - Reformatted
Ridiculous speed upgrade: collaboration
Tutorials - give me some ideas!
King Color Core Set Up - Plunge Rate? - Shapeoko XL Newbie
Understanding Manual Cutting Speeds/ FeedRates
(Phil Thien) #2

This is very helpful, thank you for all your work on this!

(Jude Marleau) #5

Seems like you are calculating the data? I gave up on G-Wizard also, way too complicated for my lack of machinist ability. I did load all the data from your sortable table into my Estlcam setup so that I have a starting point from a verified source (1 I trust and can be called a friend). As I proceed, I’ll post the results of my cuts, if you’d like? Thanks Will
I think it would be nice if the other experienced machinists members joined your efforts, My opinions don’t carry much weight. I also saved “startup” files (in Estlcam) for each material I’ll be using, that way I have tools set up with the data and don’t have to edit my tool list for each different material. 1st Open HDPE startup dwg with tools data for HDPE 2nd Open SVG file to start a new design for tool paths and the tool list is aligned for HDPE.

(William Adams) #6

I suspect that this is going to be like one of those college courses where I struggled to understand things, finally achieved an understanding of things, wrote out my understanding, and found that it then pretty much matched the initial text which I found so confusing.

Okay, cleared out my latest post, and merged it with the first — it’ll probably need some additional editing, but that should stand for now.

The thing I’m surprised at is that the calculations aren’t more physics based — why not start w/ a list of material specifications:

  • hardness
  • ease of starting a cut
  • bend rate / flexibility of the material when making the cut and taking the chip off
  • melting point
  • friction coefficient

and when calculating things such as SFM, calculate the energies involved and compare / correlate.

And, as I noted in another thread, go started w/ adding the Carbide Create values into the Nomad chart — if nothing else, they provide some initial values for stepover.

Just out of curiousity, does anyone find the sortability of the tables in the wiki useful? I’d give my interest in hell to have a copy of Lotus Improv to use for this, but can’t convince them to buy into Quantrix Financial Modeler at work, and Flexisheet is still in limbo. I suppose I should try to clear off my desk and try to get my NeXT Cube booting again.

(Jude Marleau) #7

I don’t sort it at all, I simply scroll along.

What data defines this? or is it a known factor published by material mfg’rs? I would like a physics based chart like you suggest, I’m although not a “physics guy”. I went to college once, I delivered a helium tank to a frat house. Wish I could help you more Will. One note about combining the Nomad data to the Shapeoko data is the differences of the available torc of the spindles (300 watt “?” Nomad and 900 watt makita on Shapeoko. Physics physically persuades personal preferences per particular pretenses.

(William Adams) #8

Because of the differences between the spindles, the Nomad and Shapeoko charts will stay separate — I toyed with the idea of combining them and adding a column for machine types, but my screen isn’t that wide, and I was concerned about increasing the likelihood of mis-entering feeds for one machine into a job intended for the other.

EDIT: Ease of starting a cut should be defined by the surface hardness, and the surface friction, the energy required to start the deformation, and whether or no the material cleaves along internal structures. I don’t know of a single value / rating which covers it — I guess it’s encompassed by the “machinability” rating one sees for some materials, but it seems pretty subjective, and isn’t a numerical value one can easily calculate based on.

(Phil Thien) #9

I found the sorting (on columns) helpful, if it isn’t too difficult to maintain, I’d do so.

(William Adams) #10

Thanks! Yes, we’ll definitely be keeping it — I was just curious if anyone else found it helpful.

I believe I’m now ready to move this to the next level — I’ve pretty much decided that I’ll put this into a spreadsheet, and have selected pyspread:


We’ll use it for the master copy, and copy-paste from it into the wiki for display / access / reference purposes.

(William Adams) #11

Actually there is some public / readily available discussion of this:


(William Adams) #12

Okay, here’s the beginning of a spreadsheat using pyspread.

Imperial, since that’s the measure used in the source data (the Carbide 3D feeds and speeds charts) — also want to work up a formula for displaying the Dewalt and Makita dial settings based on the RPM column.

Got a little farther on this — only Imperial for now (and yes, my current project is brass) — seem to have come across some bugs / display issues in PySpread — if someone is using a version other than the pre-built 1.02 for Windows, I’d be curious if the calculations work there (see the pyspread mailing list for details).

EDIT: Fixed the calculations to use float so that those which had integer only inputs wouldn’t round down to 0

feedsandspeeds.zip (1.1 KB)

(Jude Marleau) #13

Sorry Will, I don’t have a clue about simply viewing pyspread. I would simply like a resource that I can view, copy and paste from would be wonderful, I really don’t want to load a program I have never seen that runs a bat file. It is obviously very useful to you and many others, but I just want something I can view hopefully on the Shapeoko Materials wiki page. But here’s a real question: The Other Mill https://othermachine.co/support/materials/hdpe/ has a suggested feed and speed reference for their machine, for HDPE (for instance) they suggest Tool: 1/8" flat endmill Feed rate: 23.622 in/min (600 mm/min) Plunge rate: 1.575 in/min (40 mm/min) Spindle speed: 12,000 RPM Max pass depth: 0.008" (0.21 mm.
Now their machine seems to have a 50 watt spindle, so their suggestions would have no relavance for the 900 watt (makita’d) Shapeoko. Correct??? I think I should disregard this reference source because of that ? Thanks Will

(William Adams) #14

Wrapping my mind around pyspread, and it’s looking quite useful — reminds me a lot of LyX.

The OtherMill information is useful, especially in terms of the Nomad — given the spindle speed will have to look into it, may be useful for low energy materials such as HDPE for the Shapeoko (at least for those w/ a Makita) — will have to look into it. Thanks!

Did come across a reference which has some numbers and formulae as related to cooling:


(Jude Marleau) #15

I got a successful cuts done today in HDPE. Parts for my dust boot. The 34 mm/sec with a 1/4" flat cutter at 3.17 mm step was way too fast. Router speed makita at 3. Had a perfect cut at 12 mm/sec, 3.17 step and router at 4 makita with 1/4 " flat carbide3d cutter. 1/2" hdpe in 4 passes. The side shields and pulley covers worked fine also. Put the boot together tomorrow and no more mess, I hope. Oh yeah, the cutter stayed cool to touch for every cut, not touching it running but when changing tools during the run. Good shaving a little thinner than rice about the size of short grain rice

(William Adams) #16

Found a site w/ some additional formulas and terminology:


(Jude Marleau) #17

I appreciation it Will, I hope others can gain from the link,looked at it before. Formulas using formulas bring my brain to the brink. I’m sticking with router bits and cutting wood now and I’ll keep pecking along best I can. This speed and feed stuff is way out of my league. I still give a “like” to keep you ahead in the polls.

(William Adams) #18

Okay, to start, let’s try to minimize variables, so we start w/ SFM, since it is calculated using only two values, diameter (invariant in this case since we’re curious about 1/8" endmills) and RPM (which has a pair of finite ranges).

SFM = .262 x D x RPM

(per http://www.productivity.com/resources/calculators/ )

w/ speeds ranging from 9,470 RPM through 29,800 RPM we get a possible range of SFM from a low of 310.1425 (Makita on 1) up through 975.95 (Makita on 6) — this points up an advantage of the Dewalt where it allows somewhat finer-grained speed control.

Next installment, puzzling out how to match up SFM and chipload per tooth.

(Jude Marleau) #19

step by step, one step at a time, we will proceed. Thank you Will Great start.

(William Adams) #20

Okay, the variables which contribute to Chipload per Tooth are:

  • endmill diameter
  • of flutes

  • feed rate
  • spindle speed

Since we’re curious about the #101 and #102 1/8" 2-flute carbide endmills, the first two are invariant for this (but should be kept in mind if willing to choose other endmills), and as noted previously, the spindle speed can range from 9,470 to 29,800 as noted at: https://www.shapeoko.com/wiki/index.php/Materials#Router_Speed_Chart and would set by the selected SFM as noted in the previous post.

For now, we want to work out a suitable feed rate for hardwoods, in particular red oak, but we have some sapele and rosewood, and eventually we’ll get back to Ipé.

For the Nomad, these endmills have:

  • RPM: 9200
  • Feed: 1905 mm (75″)
  • Plunge: 482.6 mm (19″)
  • SFM: 301.3 (metric calculation is 91.83624)
  • Chipload: 0.0040760″ (metric: 0.10353261)

(Jude Marleau) #21

Trying to follow this…
(.262 x .125) x 9200 = 301.3 surface feet per minute
Is this the distance traveled by a flute in a circumference every minute? This not the straight line distance traveled along the axis.
Are we calculating the Surface feet per minute to qualify the RPM ?
I don’t see where in the productivity calculators where the surface feet per minute gets used.[quote=“WillAdams, post:18, topic:5114”]
Next installment, puzzling out how to match up SFM and chipload per tooth.

I see the puzzle but not the matching up yet. Maybe you’re not done though, I’m trying to patiently follow.

(Carl Hilinski) #22

Just for comparison purposes, here is a piece of documentation that Freud provides with its bits as a starting point for determining settings. http://www.freudtools.com/admin/manuals/SolidCarbideSpeedsinParticleBoard.pdf . This comes with a “your mileage may vary” caution from Freud.