Mitch Grunes's DrawBlade program
to draw figure skating blades.

This was part of an idea for making one's own blades.

The idea is to attach the printed shapes created using my DrawBlade program to a piece of sheet steel (also called a bar) of the right thickness, cut around it with a jig saw. In later iterations, my idea was to use a strip cutting jig and circle cutting jig, to make the cuts smooth and clean. With a more little work I could move the lines so you could cut on the outside of the line, rather than the center, and I could extend the cut lines, to simplify cutting. Some jigsaws can be adjusted to cut cross picks instead of straight cut picks, but I haven't addressed that.

I guess one needs a milling machine to narrow the working part of the blade to the exact desired thickness?

Early prototypes for individual testing would use cheap steel, final versions would use more expensive steel (like 440C stainless steel), and the edge would be hardened with a torch.

The resulting blade runner (or the runner and toe pick separately) would be bolted using right-angle aluminum channels onto an aluminum mounting plate. I could modify mark the positions of a bunch of (countersink) holes in the runner and toepick so they could be mounted at various heights, and at various forward/backward tilts (or alternately, various forward/backward positions) depending on which holes you use.

I also explored the idea of using other software to generate an image file, a PDF, and a DXF (to import into CAD/CAM software, to guide CNC hardware, like a milling machine, instead of cutting with a jig saw).

This index.html index of contents

My definitions of blade terms, and alternate definitions

There is no standardization of definitions of terminology for blades.

So let me clarify my usage of "rocker", "rocker radius", "sweet spot", and the front and back toepick teeth:

The bottom part of most blades, but not including the toe picks for figure skates, has a long-wise curved shape, or sequence of several segments of curved shape, that I any many other people call the "rocker profile." But some people refer to that entire section as the "rocker". And some people use rocker to refer to the entire curved portion; some to what I call the main rocker segment, others to what I call the spin rocker segment(s). ("Segment" is my terminology for this idea, not used by anyone else I know, so no one else should complain - I hope.) :)

Nordic skates are designed for cross-country skating or for skating on large ponds, lakes, canals and streams. They typicall have no curvature on the bottom of the blade, which is the fastest design for skating in a straight line. This program is not currently designed to draw those blades. There is no "rocker" or sweet spot of any type.

Skating on vertically curved blade bottoms (segments with "rocker curvature") makes it easier to do turns and spins, but pushing and stopping are less effective. The more rocker curvature (which means the smallest rocker radius) glide in straight lines less easily, resulting in slower speeds when skating straight.

Most short and long track speed skating blades have a single fairly long radius of curvature. Typically in the 18-23 foot range. It helps them skate around the curved arcs of the track, which are very gradual curves. Incidentally, those blades are also horizontally warped, or can be adjusted to be, which helps too, because they always skate in a CCW (counterclockwise) direction. They call the radius of curvature the "rocker" or "rocker radius". E.g., some speed skaters say "23' rockers" are best. (I'm not sure if that is specific to long track speed skating??)

Modern hockey blades typically continuously vary the radius of curvature (although older hockey rocker profiling machines varied curvature in many small steps; maybe some blades are still made that way??). The innermost parts are least curved, and are most effective to push, stop, and change direction. One also glides there to minimize friction, and go fast. But the ends are very curved, so they can do fast turns.

Offsensive hockey skaters (who mostly skate forwards) often modify their blades (or have a skate tech modify them - often using a profiling machine such as are made by CAG) so the position the least curved part is a mm or two forwards of the blade center. Defensive hockey skaters (who mostly skate backwards) often position it a mm or two backwards of the blade center. According to Blademaster, the leading manufacturer of powered sharpening machines, when skating on large skating rinks (Olympic size or larger), some agressive skaters chose profiles with a no rocker curvature in the center, which they call a "flat", so they can skate and stop and change directions while going very fast.

All figure skate blades I know of are divided into several segments with different fixed radii of curvature on the bottom part of the blade.

Traditionally MK (Originally made by Mitchel & King Blades), Ultima (made by or for Jackson Skates) blades have 2 segments. Usually JW (originally made by John Wilson Skates) have 3 segments. (MK and JW are now owned by the same company, HD Sports, and I think the blades are manufactured by the same production line in Sheffield, England. But they still use different trade names for different blade models.) Paramount and some other companies tend to roughly imitate both their designs, with some differences. There are other minor brands. The most expensive HD Sports blades tend to dominate high level competitions, in part because if a skater wins a medal at such a competition, and are using something else, they often give them a free MK or JW blade. (E.g., someone told me that Irina Slutskaya won a medal at a World level competition in Club 2000 blades. They are made by MK, but are often considered beginner level blades, and cost more than an order of magnitude less than the top blades. So she was given a more expensive pair of blades free.) Perhaps they sometimes also sponser them. And they occaisionally give away blades to influential coaches and figure skating directors. Their blades are not cheap. I would love to have MK Dance blades, and used to, but at my skating level and budget, I can't justify it. Oh well.

The main length of the blade, in the back, is the least curved, and is what I call the "main rocker" segment. But "main rocker" is also used by some others to describe its radius of curvature. E.g., traditionally, MK blades had a 7' main rocker radius, and JW had an 8' main rocker radius. The part of of the blade forward of what I call the main rocker segment is called the "spin rocker", though that may also refer to the radius(es) of curvature of that segment, or those segments. Because it is more curved, it is used for spins and forward-to-backwards turns. (Where backwards-to-forwards turns are best done varies according to different coaches.)

For 3 segment blades made by JW, there is a complication I don't understand: When skating on the forward transition point between the back and front spin rockers ("forward sweet spot" by my definition), the toe pick touches. Which means you can barely use the forward spin rocker segment at all - and not at all after a very few sharpenings, unless you keep trimming the back toe pick. Thus the forward spin rocker section becomes virtually useless. It only touches during high jumps and jump landings, when the blade is pushed deep into the ice, or on very soft, warm ice.

A few figure skating blades, such as MK Gold Star, have been designed with another spin-rocker segment in the back. MK calls this a "secondary rocker", but that is confusing because many other people use "secondary rocker as a name for all the spin rocker segments. MK says the back secondary rocker is there to create more heel lift, to make it easier to spin. (Of course, some blades only have an accidental such segment caused by poor sharpening techniques.)

I call the point(s) along the length of figure skating blades where the rocker curvature changes "sweet spot(s)".

However, to be consistent with other sources, I think I should instead call my "sweet spots" "intersection points" (though that is a bit confusing too - usually you talk about an intersection as a point where two lines or curves cross, but they are only tangent here). More sources use "sweet spots" to label spin rocker segments. Plus some sources say that spin rocker segments can in fact be used during both jumps and turns, because the toepick digs in a bit during those moves.

Many people in the skating community use "sweet spot" to refer to the entire spin rocker segment, or group of segments, even if some of them are in the section of the blade that usually can't be skated on, because the toe pick gets in the way - which seems to be true of the forward segment of JW blades. And some use "sweet spot" to refer to the entire segment or specific point that a particular skater or coach does a specific spin or turn on. In general, and in sports, "sweet spot" often refers to an optimal point or position.

User tstop4me on skatingforums.com keeps complaining about my usage of "sweet spot", because he says it confuses skaters whose coaches or skate techs use it differently.

The front part of a figure skating blade, the "toepick" (or toe pick in some people's lingo) consists of several sharp teeth. It is sometimes used to do forward-to-backward turns (usually called forward turns), scratch spins, forward jumps (which start moving and facing forwards), backward jumps, and pivots (except some usual and difficult pivots and spins are done off the "tail" (the back part of the blade). What I call the "drag pick" (or back toepick tooth), some people call the "scratch pick", "anchor pick", "master pick" or "drop pick". What I call the "frontmost pick", some people call the "king pick" or "primary pick". Apparently some people can sometimes reach it. In a few cases (specifically certain ice dance blades, like the most popular MK Dance) the drag pick is blunted instead of sharp.

Discussion at https://skatingforums.com/index.php?topic=8922.0

_Skatingforums.comDiscussion.rtf Early messages from that discussion pasted together into one rich text format document.
GoldenSkate.com subforum on Boots and Blades. I also plan to post here.
Some files and links referred to in that discussion. Some have been renamed or modified, as of 10/7/24, and later.
WilsonbladeRadius.JPG An old image (I'm not sure where I got it) showing typical rocker radiuses for MK and JW blades at that time. Note that some sources give a 27 inch instead of the 2 foot back spin rocker for many JW blades that this image lists. In addition, some sources give 12 inches instead of 9 inches for the MK spin rocker. Links to other info on measured rocker radii are
  cdn11.bigcommerce.com/s-u1ecyriv/images/stencil/1280x1280/products/3099/4782/paraCP-3__05219.1580504904.jpg?c=2
  scarletskater.wordpress.com/2017/06/01/choosing-figure-skating-blades
  scarletskater.wordpress.com/2017/06/02/blade-profiles
  figureskatingstore.com/blade-type-information
  ice-blog.riedellskates.com/know-toe-picks (uses a different definition of "sweet spot" than me)
  www.afterness.com/skating/blade_rocker_measurements.html
  SkatingForums.com list of blade manufacturers; they show pictures of their blades

7foot.png image of 7 foot rocker section cropped out of rocker.png.
9inch.png Similar for 9 inch rocker
sweetspot.png Shows how two rocker segment of different rocker radii can be cleanly pasted together. The idea is that the tangent lines are the same at the boundary. Or equivalently that the radial lines are the same.
sweetspot2.png Similar, but labelled.

Files relating to my
DrawBlade program to draw
(mostly figure skating) blades

DrawBlade.f90 My (free format) FORTRAN 90 (or newer) language program that draws creates Postscript output of specified blade profiles.
_ToDo.rtf A list of things about it that could be improved.
DrawBlade.dialog.txt Dialog that ocurred during a sample run.
DrawBlade.txt The parameter file it generated during the sample run. This is a text file that you can edit to make minor modifications, and the program has an option to read from it. Note that I chose an extra wide line width, and to draw radial lines. I also chose to draw 1 inch tic marks along the x and y axis. These are needed because many printer apps don't scale things correctly by default. For my example I chose 8.5x14 inch size paper, because the sample blade shape was too long to fit on letter size paper.
DrawBlade.ps The principle output file, containing Postscript text commands from that sample run. Clearly labelled to be human readable. You could look at it to see parameters that you would need to guide a CAD program.
Program that display it need to be told the paper size is 8.5"x14" or larger. And some programs would be confused because some of the line segments are outside that boundary.
ghostview can display it with
  print*,' gv -media=Legal DrawBlade.ps
Likewise, if you choose 8.5 x 11 inch paper, use
  print*,' gv -media=Letter DrawBlade.ps
Likewise, and if you choose 11x17 inch paper, use
  print*,' gv -media=Tabloid DrawBlade.ps
Other supported paper sizes (in 1/72 inch point sizes) are specified here.
DrawBlade.pdf PDF file generated from DrawBlade.ps by https://cloudconvert.com/ps-to-pdf
. (There are other free ways to do such conversion.) If your CAD program can't input Postscript directly, perhaps it can input PDF files. The conversion technique couldn't quite handle an image this large, so the toe pick is cut off. I could work on that if needed.
DrawBlade.png An image file created by an old version of Paintshop Pro from DrawBlade.ps. Many other programs, like GhostView, can make such conversions.
DrawBlade_labeled.png, similar, but I edited it to add explanatory remarks on the input parameters used by DrawBlade.

However, it turns out that my printer does not scale output correctly. To get accurate scaling, I need to enlarge x-axis coordinates by about 9/8=1.125, and y-axis coordinates by about 12/11=1.09090909... For that correction I created DrawBlade2.txt
DrawBlade2.ps
and
DrawBlade2.png

Of course, you should instead find scaling corrections for your printer.

rocker.png A printable set of circular arcs with typical rocker curvatures generated by my IDL language
  rockerprofiles.pro program. (Note: it was an old version of IDL, and newer versions may not be upwards compatible.)
The actual program output was idl.ps, a Postscript language file, that was converted and cropped. That file is text too, but is not human readable.

I have other rocker curve images that I now prefer produced by my DrawArcs program.

DrawBlade.zip contains a compressed zip archive of the rest of this folder - to simplify transfer.