After doing quite a lot of carpet cutting recently, repeating it with the same knife from four to six times, I have learned two fairly critical aspects, one obvious and one not :
1) There is a lot of variance in the results. To get a decent estimate of the performance you will want to repeat it several times. Rope cutting is very consistent, from roll to roll, from day to day. However used carpet by its very nature can produce large variations. For example one of the recent runs on a SAK hit a large rock on the first cut which mashed in a section of edge about one cm long, for the rest of the cutting all the work and edge retention was basically done on 4 cm of edge (cutting is restricted to about 2"). If this happened on the first run I probably would have thought it was due to the soft steel, but it didn't happen on the other three runs, and using median based statistics it doesn't significantly skew the results. So in short, do it as many times as you patience allows, and use median based statistics to determine performance estimates.
2) Checking the sharpness by slicing (I was using 1/4" poly) is problematic due to the way in which carpet effects an edge. The edge shows damage usually from 0.1 to 0.2 mm deep, if the damage is circular in nature it isn't very aggressive, however if it is triangular, this can be a saw like edge. On some rare cases I have seen actual hooks form on the edge which are massively aggressive, double that of the circular teeth. These fractures can thus elevate the slicing aggression, however they don't increase the push cutting performance. Therefore you will get a difference measure of sharpness by push cutting. This means that both should ideally be done. I was only doing slicing mainly because I can do it much faster. Jeffs suggestion of ribbion cutting is a solid idea here. Note the nature of the breaks was not well correlated to blade material. All types could be seen across various steels.
Now as to what does it show. You can clearly seen the influence of hardness by the amount of impaction. Using a 62/63 (CPM-10V) knife, 57/59 (52100 MUEK), and ~55 (SAK), I had friends rank the blades by visual inspection of the edges under light after cutting. All would do independent rankings exactly matching the hardness orders (all edges had the same angles). The edges also suffered less blunting in the same order (wear resistance probably is a factor here), and required less work to resharpen. The latter is however also influenced by machinability and the SAK rebounded faster than the others, which almost compensated enough to overcome its much greater blunting.
The main downside is that it makes a huge mess, and it *really* wears on a knife. Sharpen 0.1 - 0.2 mm off a knife six times (you have to sharpen past the visible damage to remove all the weakened metal) and you can readily see the edge has moved down from the choil, not to mention the time it takes to grind all that metal off. On the positive my freehand sharpening skills have never been so good.
-Cliff
