Blade Sharpness Research Project

[helicopter]

Published blade sharpness measurements leave something to be desired. I couldn't find anything on a bunch of current blades, so I decided I want to be able to measure sharpness myself. An edge on up industrial tester with BESS certified test media takes more commitment than I have to this, so I bought some 6lb test fishing line and a shahe force gauge and made myself a fixture with a piece of scrap wood and some m6 hardware from true value for about $40 total.

I cut up a few fathoms of line on it today. It is pretty consistent so far. I might make an aluminum fixture when I find the chuck for my drill press.

I figured I would start with a new Astra SP green India blade as a baseline, and measure it after each shave until it gets dull. I figure 7 measurements on each side is good enough with how consistent this seems. I am not sure what the best single number Stat is, mean, median, mean of middle 5 from each edge... so I will just collect data for a bit and see what seems to make sense. I don't think these measurements are interchangeable with BESS measurements, but the 30 grams or so on the Astra does seem to be in the same ballpark of what I would expect on BESS, based on the published data I have seem.

I guess this is going to take a long time to collect data because because I have to grow more hair every time I take a measurement.

 

[Eben Stone]

I guess this is going to take a long time to collect data because because I have to grow more hair every time I take a measurement.

Does the testing procedure damage the edge?

mean of middle 5

 

[helicopter]

I am sure it has some negligible effect. It is a 0.15mm nylon monofilament, so under ten cuts on each side shouldn't have a measurable effect with this setup, when my precision is one gram, and I don't expect thousands of beard hairs to dull the blade even 10 grams. I would probably have to cut 1000-10,000 pieces to measure it.

I am certainly leaning toward average of the 5 middle values from each side. I will probably do it like that.

 

[geneaut]

I am looking forward to this

 

[Space_Cadet]

Looking forward to the results too.

 

[helicopter]

I am running into some issues as I shave this week with the Indian Astra SP. First, is that the blade is dulling much slower than expected. Second is some statistical noise.

I changed my testing technique to sample different points along the edge of the blade. I don't know if that makes a difference. Intuitively, it might reduce precision while increasing accuracy, leading to noisier, but more relevant data.

I also figured out how to be a little more consistent with tension and started looking at the deflection of the monofilament to make sure it is reasonably consistent. I don't see any difference from this, but it seems like a good practice regardless. Intuitively, it should increase precision and comparability.

I may need more than 7 measurements per side to reduce noise a bit. O will probably step it up with the next blade.

I am not sure if it makes a difference whether I use traditional soap, or creams, particularly a favorite, Cremo. Cremo may reduce blade wear necause it uses polymer microbeads to increase lubrication. Cremo was correlated with reduced blade wear measurements in my small sample so far, but I would need to shave with and without it for at least a week to start drawing any conclusions.

The slant on the Rex Konsul I have been using this week may reduce blade wear. I am getting an Ambassador (not slanted) this week, and I will probably throw in a different blade and try it in the Ambassador after it arrives.

If Ambassador is still slow to show wear, I may try some single edge razors. I don't want to go there yet.

The biggest increase I have observed in cutting force was from (0) to (1). Normally people observe a drop in cutting force as the blade coating is worn away, exposing the sharper bare metal below, and then cutting force increases thereafter. I am not sure why I didn't observe this. It could be an error introduced by straightening the fixture on the meter between these measurements. It could be a change in my technique or tensioning as I got some practice. It could be that the blade didn't follow the ordinarily trend, but started out as sharp as it was going to be. It could be the blade did become sharper easly in the first shave, but dulled after that. It was probably something with the measurement.

I will keep shaving with Astra SP India for a few more days and when I toss the blade, I will report the measurements and comment on them.

At some point, I may come up with some sort of artificial blade wearing process. That would allow me to get a lot of data in a much shorter time, and it wouldn't stop me from shaving with the blade to provide some subjective impressions to go with the measurements. I will keep thinking this over. I almost certain manufacturers would do it this way because it could be much more consistent. I wonder if anyone has published or knows of a good process to wear a blade artificially.

 

[FarmerTan]

I am running into some issues as I shave this week with the Indian Astra SP. First, is that the blade is dulling much slower than expected. Second is some statistical noise.

I changed my testing technique to sample different points along the edge of the blade. I don't know if that makes a difference. Intuitively, it might reduce precision while increasing accuracy, leading to noisier, but more relevant data.

I also figured out how to be a little more consistent with tension and started looking at the deflection of the monofilament to make sure it is reasonably consistent. I don't see any difference from this, but it seems like a good practice regardless. Intuitively, it should increase precision and comparability.

I may need more than 7 measurements per side to reduce noise a bit. O will probably step it up with the next blade.

I am not sure if it makes a difference whether I use traditional soap, or creams, particularly a favorite, Cremo. Cremo may reduce blade wear necause it uses polymer microbeads to increase lubrication. Cremo was correlated with reduced blade wear measurements in my small sample so far, but I would need to shave with and without it for at least a week to start drawing any conclusions.

The slant on the Rex Konsul I have been using this week may reduce blade wear. I am getting an Ambassador (not slanted) this week, and I will probably throw in a different blade and try it in the Ambassador after it arrives.

If Ambassador is still slow to show wear, I may try some single edge razors. I don't want to go there yet.

The biggest increase I have observed in cutting force was from (0) to (1). Normally people observe a drop in cutting force as the blade coating is worn away, exposing the sharper bare metal below, and then cutting force increases thereafter. I am not sure why I didn't observe this. It could be an error introduced by straightening the fixture on the meter between these measurements. It could be a change in my technique or tensioning as I got some practice. It could be that the blade didn't follow the ordinarily trend, but started out as sharp as it was going to be. It could be the blade did become sharper easly in the first shave, but dulled after that. It was probably something with the measurement.

I will keep shaving with Astra SP India for a few more days and when I toss the blade, I will report the measurements and comment on them.

At some point, I may come up with some sort of artificial blade wearing process. That would allow me to get a lot of data in a much shorter time, and it wouldn't stop me from shaving with the blade to provide some subjective impressions to go with the measurements. I will keep thinking this over. I almost certain manufacturers would do it this way because it could be much more consistent. I wonder if anyone has published or knows of a good process to wear a blade artificially.

Good science! I wonder what @ShavingByTheNumbers would say?

 

[DrakeMegrim]

Good science! I wonder what @ShavingByTheNumbers would say?

Just what I was thinking. I went to reference his chart a few weeks ago and had to use internet archive to look at it.

 

[AndyPanda]

I'd be curious if your rig can detect a change before and after "corking" a blade like Personna Comfort Coated or Nacet that seem to be sharper on 2nd shave. If a pass or two thru the cork can wear off enough of the coating to show up on your apparatus.

Or can you slice a few ribbons from a sheet of paper and detect a change in the edge?

 

[AndyPanda]

Also curious (just call me Mister Curious ) how are you holding the blade when testing? Blades are so flexible I would think it makes a difference to the measurement by the way the edge is held and whether or not it from flexes sideways as it cuts thru the filament

 

[helicopter]

Just what I was thinking. I went to reference his chart a few weeks ago and had to use internet archive to look at it.

I went on there too. I was researching his methodology. He did 2 shaves, face and head, with 48 hours growth. I think this explains why he was wearing blades so much faster than I am, but the slant of my razor probably isn't helping. He was using a Merkur 34.

He also made a cool fixture to keep the blade straight, flat, and perpendicular, but it also meant he was testing the same spot on the blade with every cut in a set.

He also used an edge on up tester with those pre tensioned clips.

I don't actually think I have a problem with my fixture or blade angle, as I have identified some other trends that explain some of my observations.

 

[helicopter]

Also curious (just call me Mister Curious ) how are you holding the blade when testing? Blades are so flexible I would think it makes a difference to the measurement by the way the edge is held and whether or not it from flexes sideways as it cuts thru the filament

They are thin, but even 50 grams force isn't much. I don't think they are slicing or flexing much. I did order some bread lames. Maybe I will try those on the fixture. They won't prevent slicing, but the ones I ordered should mitigate twisting and flexing. I actually ordered them to assist in artificial wearing.

 

[helicopter]

I'd be curious if your rig can detect a change before and after "corking" a blade like Personna Comfort Coated or Nacet that seem to be sharper on 2nd shave. If a pass or two thru the cork can wear off enough of the coating to show up on your apparatus.

Or can you slice a few ribbons from a sheet of paper and detect a change in the edge?

I will give it a shot at some point. I want to re-measure the Astra SP India new to see if there was a problem with the consistency of my very first set of measurements. Of course a different individual blade does add another variable, but they will be from the same tuck, and the changes made to the fixture and technique between (0) and (1) are more likely to have changed the measurements.

 

[BigJ]

Following this with considerable interest!!

 

[gpjoe]

I also figured out how to be a little more consistent with tension and started looking at the deflection of the monofilament to make sure it is reasonably consistent. I don't see any difference from this, but it seems like a good practice regardless. Intuitively, it should increase precision and comparability.

That was an issue which I thought might affect your results.

The tension must be consistent for meaningful data. I would think that varying tension would require varying force, all else being equal.

 

[DrakeMegrim]

I went on there too. I was researching his methodology. He did 2 shaves, face and head, with 48 hours growth. I think this explains why he was wearing blades so much faster than I am, but the slant of my razor probably isn't helping. He was using a Merkur 34.

He also made a cool fixture to keep the blade straight, flat, and perpendicular, but it also meant he was testing the same spot on the blade with every cut in a set.

He also used an edge on up tester with those pre tensioned clips.

I don't actually think I have a problem with my fixture or blade angle, as I have identified some other trends that explain some of my observations.

Agree on everything. You should be good if you're doing multiple cuts to test for consistency and listing standard deviation.

Being consistent will help make this data internally consistent and comparable. That's probably going to be what's hardest for you.

 

[tejolot3]

Normally people observe a drop in cutting force as the blade coating is worn away, exposing the sharper bare metal below, and then cutting force increases thereafter. I am not sure why I didn't observe this.

One possibility: what people subjectively observe as 'sharpness' doesn't actually correlate to scientifically measured 'sharpness'.

 

[helicopter]

It looks like we are around 0.8-1g change in cutting force per shave with this combination of face, blade, and razor, maybe 1.2g at the middle of the blade. Perhaps I should just tape a few pennies to the head of my razor and keep going. There are probably some blades that start out duller than this blade is now, but I am not sure I would want to use them. I am going to throw this one out now, and think about what's next over the next 22 hours or so. The Rex ambassador won't be here until Saturday, but I can always skip a day. I have a Yaqi single edge razor, and today I ordered a Leaf Thorn from TRC with the 11% off holiday promotion. I will probably try the Ambassador for the next set of shaves with measurements.

Blade	Astra SP India
Date	5/13/24	5/14/24	5/15/24	5/16/24	5/17/24	5/18/24	5/19/24	5/20/24	5/21/24	5/22/24	5/23/24
Razor	Rex Konsul
Soap	NA	B&M	Proraso Pre&Cream	Cremo	Cremo	A&E	Denali Dreams	TOBS Cream	Cella Green	Fine	Captain Choice
Shave	0	1	2	3	4	5	6	7	8	9	10
Top	29	33	33	31	35	34	34	31	36	35	39
	31	35	37	27	49	32	33	45	37	39	41
	30	33	39	35	42	34	36	43	39	40	45
	28	40	37	30	28	44	37	40	36	47	40
	32	35	40	33	31	31	35	44	41	46	53
	28	33	35	36	34	36	35	41	35	37	44
	31	36	34	34	32	45	38	38	33	35	39
Center 5 Mean	29.8	34.4	36.4	32.6	34.8	36	35.4	41.2	36.6	39.4	41.8
Bottom	32	38	35	34	34	36	39	37	35	31	48
	33	33	32	36	37	38	32	42	38	43	46
	30	34	32	35	33	38	34	38	44	41	46
	30	49	39	40	31	43	45	49	42	43	46
	29	26	36	31	40	45	41	47	39	50	46
	28	36	33	40	35	39	39	45	37	46	49
	29	30	35	37	35	37	37	37	40	40	43
Center 5 Mean	30	34.2	34.2	35.5	34.8	39	38	41.8	39.2	43.3	46.4
Center Mean Total	29.9	34.3	35.3	34.1	34.8	37.5	36.7	41.5	37.9	41.3	44.1

After the first set of measurements, I tried sampling multiple positions on each blade and became more methodical as I went. With the 5/18 shave, I started going from one end of each edge to the other and recording the measurements along the edge in order. sampling one zone of the edge gave more consistent results. Sampling from one end to the other revealed that blade wear was much more severe at the center of the blade than at the ends. Intuitively, this makes sense, and it seems like, since the middle of the blade is doing more work, its sharpness is probably more important than other zones on the edge.

I may just footnote this and select a zone in the middle of the blade, such as the middle half, to take all measurements, taking them from one end of that zone to the other.

Razor	NA
Soap	NA
Shave	0	Cork 1	Cork 6
Top	31	27	36
	33	35	35
	21	25	36
	39	30	32
	35	24	37
	37	27	43
	40	38	37
	37	35	36
	35	36	39
	27	39	36
Center 8 Mean	34.3	31.6	36.5
Bottom	49	34	34
	32	36	48
	37	33	33
	34	37	32
	40	39	34
	36	34	35
	39	36	35
	35	37	34
	34	36	35
	37	31	36
Center 8 Mean	36.5	35.4	34.5
Center Mean Total	35.4	33.5	35.5

A little cork experiment here, and measurement of a new blade from the original tuck. I think it is safe to say shave 0 measurement on the first blade was flawed somehow, and assume the actual starting sharpness was more like 34-36g.

It also looks like corking a blade is worthwhile if you don't go overboard, and you are someone who only uses the blade for 1-4 shaves or so.

I could wear a blade artificially with the cork, but I think I will start with simulating wear using strips of standard office paper, which seems to be a standard for some such testing. I am not sure how well it will work for razor blades, but it is going to be more consistent than a wine bottle cork. I will think about it more when my bread lames arrive.

 

[helicopter]

I grabbed a Russian Astra SP, measured it, shaved with it, measured it again, and finding the high cutting force observed surprising, went back and re-measured the Astra SP India (10) and Astra SP India (Cork 6).

I found myself disappointed with results too inconsistent to be confident in much of the 50 or so measurements.

So I ordered a TOTL Edge On Up Industrial tester. I was thinking about a blade fulcrum, and tensioners, perhaps pulleys and weights for tension, and decided my expected future time and effort justifies paying $250 for a $50 scale and fixture where someone else has already solved some of these problems.

 

[helicopter]

Here are the rest of the results I got with the home made fixture:

Blade	Astra SP Russia
Date	5/24/24	5/24/24	5/25/24	5/26/24	5/27/24	5/28/24	5/29/24	5/30/24	5/31/24	6/1/24	6/2/24
Razor	Rex Ambassador	Rex Ambassador	Rex Ambassador	Rex Ambassador	Rex Ambassador	Rex Ambassador	37C	Rex Ambassador	Gillette Tech	KCG	Slim
Soap	NA	Arko Stick	VDH Soap	Proraso Cream	Arko Stick	Captain's Choice	Cremo	Cremo	Oil and Barbasol	Oil and Barbasol	Stirling Soap
Shave	0	1	2	3	4	5	6	7	8	9	10
Top	34	40	41	36	37	33	43	31	40	65	90
	39	52	46	36	36	33	39	31	39	48	47
	42	44	37	34	49	34	46	64	66	36	96
	68	45	45	40	72	36	65	80	64	40	24
	70	46	38	41	62	34	47	79	73	37	35
	51	51	45	39	48	42	38	48	53	44	39
	56	46	47	48	52	70	41	42	37	43	48
	54	40	76	53	43	74	30	43	37	65	34
	52	50	37	54	38	36	31	36	40	34	29
	47	43	33	50	38	32	52	39	40	31	32
Center 8 Mean	51.1	45.6	42.0	42.9	45.9	39.8	42.1	47.8	47.4	43.4	44.3
Bottom	40	40	58	52	40	59	46	46	36	50	58
	48	40	56	56	41	64	42	26	45	51	90
	47	36	54	52	41	58	26	40	40	52	54
	60	42	53	40	41	46	30	39	32	27	55
	48	55	36	59	39	37	25	43	36	50	33
	52	60	57	49	66	27	50	63	50	29	32
	48	50	34	36	52	32	48	43	59	48	30
	57	55	40	39	52	33	47	56	63	61	31
	49	61	37	35	65	35	49	57	54	39	29
	38	44	41	40	54	30	53	65	66	60	33
Center 8 Mean	48.6	48.3	46.8	45.5	48.3	41.3	42.3	48.4	47.9	47.4	40.8
Center Mean Total	49.9	46.9	44.4	44.2	47.1	42.1	42.2	48.1	47.6	45.4	42.5
Std. Dev.	9	7	11	8	11	14	10	15	13	12	22