ABRASIVES COMPARISON - multiple types, grits compared by 4 testers

[Bushmiller]

One problem I see with the testing, in fact any testing, is that is is besides being relatively tedious it is also very demanding on time. It is why there are probably relatively few instances of extensive testing being performed outside of commercial or sponsored studies.

It is for that reason I am quite happy to encourage anybody prepared to commit time to such activities. It seems to me they get relatively little out of it compared to the rest of us, who tend to sit back and criticise. Such tests can be expensive as well as time consuming. There is another factor that is too easily forgotten and it is that in a forum such as this there is much more transparency than say a company commissioning a test. If it doesn't suit them, they don't have to publish. It didn't come up with a result that suits their agenda so that is the last that is heard.

Now having said all that I am going to make an observation (I hesitate to call it criticism) and it as to the use of the pencil and the length of time it take to sand a pencil mark away.

The pressure exerted on the pencil is critical to my mind: In fact more so than the sanding pressure on the ROS. Any pressure on an ROS beyond the weight of the machine will be detrimental to the life of the sandpaper. consequently I would suggest that merely the ROS weight is utilised. It will be constant and allows just for sideways motion. If you have to exert pressure the sanding disc has lost it's cutting power.

So back to the pencil mark. If pressure is exerted, the pencil will leave a deep mark in the timber with more sanding required to erase it. Consequently a consistent method of marking the timber is required. I don't see that a hole in a piece of timber will work to hold it at a constant height as the pencil point will be worn down as you go. Clearly the pressure at the point will change. I don't think the width of pencil matters much other than to slow down the wearing down of the point. On that basis a carpenter's pencil would probably be better.

I don't know the solution: Only the problem.

Regards
Paul

[FenceFurniture]

It's starting to sound like it may not be a good idea to publish the results.

[FenceFurniture]

One problem I see with the testing, in fact any testing, is that is is besides being relatively tedious it is also very demanding on time.

So far I have spent:

• researching abrasives and their availability, about 6-8 hours
• Purchasing, organising freight (including one lost package for two weeks, and the time involved in tracking it down), assembling the various packages to go to the testers, emails of general instructions, liaising with another supplier and getting them to produce a 17 hole die for an even playing field about 6 hours
• Preparing spreadsheets, preparing timber about 1½ hours
• Sanding with just one grit in the range, taking about 50 different photographs, about 6 hours and significant fatigue to go with it (sanding 1.8 kilometres is pretty taxing)
• Writing the report on just that one grit, uploading photos to the computer, cropping, resizing etc in Photoshop, uploading to the forum about 9-10 hours
• Writing posts about 3 hours

So that's a total of about 33 hours, and there are 5 grits to go. I have tested 5 out of 23 discs, but there may be a few more to arrive from Klingspor following my conversation today (sounds like about 6-8 different discs).

Such tests can be expensive as well as time consuming.

So far I have spent about $140 on purchases and postage. I will have some discs leftover from that.

It seems to me they (the testers) get relatively little out of it compared to the rest of us, who tend to sit back and criticise.

There is absolutely nil benefit to me in publishing these results on the forum. None whatsoever. I started this thread thinking that the results might be useful to those who are capable of reading without prejudice, and who may have enjoyed the spirit that the thread was started in.

Any pressure on an ROS beyond the weight of the machine will be detrimental to the life of the sandpaper. consequently I would suggest that merely the ROS weight is utilised.

This is already going to take long enough by exerting the usual amount of pressure that I would on a sander (which is apparently about 9kgs, for coarse work anyway). Maybe I need a lesson on how to sand. Using only the weight of the sander will add a prohibitive amount of time and strokes, and as such, is not going to happen.


There is a reason why I have pointed these things out, and I'll come back to that.


So far, there is one poster that just wants to criticise for the sake of it, and tell me how wrong (again) and biased the whole exercise is. This is to be expected. There are two other occasions where methodology suggestions have been made, that have already been put in place (using the same piece of timber, and a larger pencil). I had previously stated that I was going to be using the same piece of timber, and that the pencil I used for marking was 5mm thick (and that's a cylinder, btw).

My point is that people are already not taking in the detail that has been provided, and suggesting I do what has already been done. If this is the way that it is before any results have even been published, what is it going to be like when all the test data has to be digested and comprehended? Should I really set myself up for that???

Given that there is no benefit to me in publishing the results here, and given that I am quite clearly just going to open myself up to whatever grief in whatever form it may come, and given that the spirit of this thread is already tainted, I think the prudent thing to do will be to keep the results private.

I'll sleep on that and make a decision tomorrow.

[Bushmiller]

Brett

I hope that after a sleep you will feel it is worthwhile to continue:Personally I think it is important that you continue. As I mentioned before, it will provide a good supply of hitherto unavailable information. Sanding is a big part of woodworking but arguably does not carry the "glory" of some other operations (planing, cutting etc).

From my previous post:

"It is for that reason I am quite happy to encourage anybody prepared to commit time to such activities. It seems to me they get relatively little out of it compared to the rest of us, who tend to sit back and criticise. "

The work you put in is appreciated both in this thread and others such as the saw files, which was/is an epic. There is always a detractor somewhere, but plenty of others who appreciate the effort and the results.

Go for it!

Regards
Paul

[FenceFurniture]

Paul, to be clear the testing will continue - it's a matter of whether or not I put the results up here, for the reasons described. It may be that I'll leave it at if someone wants to see the report then they can send me a PM and request it.

[BobL]

RE: Time and effort needed to do testing.
Welcome to the world of testing

All the testing I have done (both as part of my old day job and associated with this forum) has been " time wise" in a similar vein.
I now only do it because I'm interested and post what I think some folks might find useful
Constructive criticism (not all seen by the tester as constructive) can be very irritating but is an essential part of the process because it does improve things.
Non-constructive criticism is inevitable and can drive experimenters around the bend so it's far easier to ignore than to fight it - that is if you want to stay sane.
It can get so bad that you can spend all of your time dealing with critics and get little else done.
just draw a line somewhere and don't cross it and get back to testing.

[FenceFurniture]

It can get so bad that you can spend all of your time dealing with critics and get little else done.

And I have little interest in dealing with critics when they are ctitical of a technique that they have either read incorrectly or are suggesting it be done a certain way, when it already has. Unfortunately though, the effect of this unfounded criticism is unknown upon other readers, and one doesn't know if it's just a waste of time. Then there is the wasted time correcting the critics unfounded criticism, which also just clogs up the thread.

just draw a line somewhere and don't cross it and get back to testing.

And that line could be getting the report sent in private with the caveat that the tests have been done to the best standard can manage with the non-lab equipment available. The readers could either accept that the tests have been conducted in good faith, and for example the pressure on the pencil is reasonably consistent, and averaged over several repetitions to iron out any inconsistencies, or not.

At least that way the prospect of a public spectacle is obviated, and currently it's difficult to see this not turning into a spectacle.

[Simplicity]

Brett
Keep going there is probely a tone of us reading this and not saying much.
But enjoying it and gaining from your hard work.And thank you.
In my own business 95% of my clients are fantastic
5% of my clients take up quite a bit off time
I focus on the 95 percent the five percent I as a rule generally only work for once.
At a guess I would say 95% percent off us here can't wait for the finally conclusion
We all know how it's being done
And we can then decide what we think.
Keep up the great sanding
Matt

[LGS]

Here are some things that might help.

1. Pencil mark and scratches

It was my experience in setting up a protocol for hard burnishing, that it was better to run a set of sanding strokes rather than sand to "what felt right".
I use 10 up and back strokes (20 total) with each pad, then move to the next. It may be better to use something like this in your study.
Define an area of the board, say 1 metre x (x) width and keep your answers to the results you get from this area.
For scratches and pencil line, you can award a point system.
Mark 1 for no change from one grit to the other, 2 for some removal of scratches and 3 for total removal of scratches. By using 10 up and backs, you can display results as % efficiency by comparing the grit's value against the mean. It is important not to deliberately focus on areas which show severe scratching or try to erase the pencil mark.

While you are doing this stock removal, you can set a Vernier Caliper to the starting depth of the piece of timber, then measure the final depth for each grit you use and record it. Repeat for each grit.

I think that each users normal routine for sanding should be adequate for a study like this. You can include these as "acceptable" normal variations. The same is true also for how the sander is passed over the surface. (Though any serious outliers from the means you generate should be seriously questioned)

To make for consistent measuring, reset the vernier for each grit. You can then add the values for each run and define a mean. Anything that looks way out should not be used to establish the mean. Once you have the mean you can derive a % varianace from the mean. Either that or you can graph the mean versus the value of the stock removed and directly compare efficiency.

You may be able to measure longevity of the pads by running each again after the initial run and measuring again the amount of stock removed and quality of the finish.

It is very important that you all use the same timber, preferably from the one board. If you can't do that, then set the type of timber to be used. I would suggest either "Tas" Oak or radiata Pine. You will get more data probably using the pine (it would probably show greater variation between operators too.)

Values for dust produced and where it lands don't really have a place in this study. What they tell you is really how efficient your D/E is.

[shanesmith80]

What ever way you decide to go FF, kudos for giving it a go. Sounds like a lot of hard work and I would be very appreciative of knowing how things pan out.

[FenceFurniture]

I have come to a conclusion on publishing the results, and have taken a few things into consideration to arrive at this conclusion:

1. The other testers don't seem to have concluded (not even sure if they have started), and I wouldn't publish any of the results until they are all finished
2. There are some people who would like to see the results, and I already have some results from my testing, and should have finished in about 36 hours.
3. As I have previously described, in my opinion there are two test sets: one for the coarse starting grit, which in this case is 80 (and is also my normal starting grit), and another test set for the grits finer than 80. The first test set revolves around sheer material removal (not particularly interested in surface finish), and the second is purely refinement of finish (not particularly interested in material removal quantity).
4. I do not wish this thread to degenerate as some other threads have, in terms of "it should have been done this way" or "this is rubbish because everything you do is rubbish in my opinion, regardless of how good it is, or how useful it might be to other people". And that second one is pretty damn common from one quarter. That means that the report will not necessarily be made available to everyone who asks.

So, with all that in mind, I am going to make the report available to people who request it by PM.

The report on 80g is finished, and BobL has already read it two days ago. Therefore there can be no changes to that report, and there certainly won't be any changes to the methodology.

Whether or not the final report (with all tester's notes) is published on the forum remains to be seen, and so that decision can be made later. The testers may simply decide to publish their reports on here, and that is most certainly their prerogative if they wish to.

As for my methodology, the readers can make up their own minds as to whether there is any bias or incompetence. Bob is most welcome to post his thoughts on the competence of the methodology and the fairness of the 80g report.

For those who receive the report: you are welcome to comment in general in the thread at this stage, but remember that the other testers have not finished so don't spill the results in your posts.

At the moment the 80g report is in a word document with BBcodes, ready for posting on to here, so it will take me a little while (tomorrow) to replace the BBcodes with the actual images, and then turn it into a PDF. The people who are on my email list will receive this report in due course anyway (at the end of all testing), but you may wish to have a look sooner than that, so send a PM.

[FenceFurniture]

I believe I have solved the consistent pressure problem for the pencil line in the second group of tests. Following Bob's suggestion I mounted (glued into a hole) a pencil in a small block and then dragged that in a wave motion down the board with the heel of the block always in contact with the main board (i.e. not lifting the heel which would vary the pressure), and noted how many strokes it took to remove it. I purchased a 0.7 mechanical clutch pencil with some 2B leads. I would have preferred a 1mm lead pencil but there was not one available.

As it turns out, the 0.7 is absolutely ok, and is producing results consistent with predictions*. Furthermore it is consistent with the results I got from the 80g test where I was drawing the wave freehand. When the lead becomes too short it is simply a matter of a click on the button to show some new lead. My original concern with this technique turned out to be quite unfounded - I thought that a fine pencil might dig in to form a scratch. Using this type of pencil ensures that a consistent diameter of graphite is in contact with the board.

*There was one aspect the went exactly the opposite of what I had thought, but when I thought about it, it made perfect sense. That is, I had figured that the number of sanding strokes to remove the line would increase as the grits got finer (because the grits are becoming less aggressive). Exactly the opposite was the case, which created a furrowed brow and narrowed eyes until I figured out why. Pretty simple really. For this second round of tests where I am really only interested in how long it takes to get the finish of the board up to the next level of smoothness, I am preceding each different disc with 10 strokes of the grit below it. That is too say that for testing each of the four disc in the 120g range, i do 10 strokes with a 100g disc before each new disc. For 180g they are preceded by a 150g disc, and so on.

Well of course the surface that I apply the pencil line to is smoother than it used to be, and so the graphite doesn't have as many hills and craters to be erased from. Thus, the number of strokes to remove the line becomes less as the grits get finer because the surface is smoother to start with.

Furthermore, this notion was vindicated when I noticed that the given protrusion of lead between clicks was laster longer as I moved up to finer grits. Obviously the smoother surface of the timber does not abrade the graphite as much.



As for using the same species of timber for all the testers, and preferably from the same board: I disagree. As long as each tester uses the same piece of timber (as I have) then that is all that is required. If the species vary from tester to tester then that just gives us a better idea of how the same abrasives perform on different species, and that of course is very worthwhile information. If there are wild variations in the results, from wildly different species (hardness, abrasion) then we may have to figure out why that might be, but it most certainly WILL NOT be because of bias.

Quite frankly, I think it's an affront to suggest that someone would be biased, particularly as that claim was made before even the methodology had been revealed or reported, let alone the results.

[pmcgee]

I think many people will also appreciate subjective experience ... may even filter out all the details about attempted fairness of comparisons.

I think some people might respond to, eg ... FF's opinions were XYZ. I've seen the sort of work he's interested in ... or I can search back and check some of it out ... and if he thinks ABC is pretty good then that's ok by me.

I assume people would also take into consideration any commercial interest that might or might not exist ... Youtube videos are like that. There are people you get to know over time ... they voice their opinions and we evaluate the opinions based on what we have seen of them as a person. If they work for a company, or benefit from a product we tae that into account ... we've all seen infomercial stuff that would have you saying "Hang on, hang on ..."

I'm all for carefully measured results ... but sometimes a "rough evaluation" (not labelling this effort as that) is quick and sufficient to come up with some indications pretty painlessly. eg we don't need a microscope to tell an aardvark from an artichoke.

In fact you might well possibly think ... if the difference requires careful measurement to detect ... then for many people it would be sufficient to say there is little difference. I think Bob's instinct is to try to be certain there really is little difference, or there really is a large difference ... but you can also modify the aims to suit the methodology ... vs modifying the methodology to suit the aims.

Cheers,
Paul

[FenceFurniture]

That's well said Paul.

Quantifying is important for determining small differences, isolating one aspect that may be the determining factor, and vindicating a previously held "gut" feel. It is also important for determining the relevance of certain observations and whether is worth continuing to note them.

The two different test sets have highlighted much of this, and should help to disparage any prejudices to a fair minded person. In the first test set I made note of everything i could observe, and as others have said, some of them weren't relevant, such as the dust left on the job. Some of these observations were dropped for the second test set to a) de-clutter the reading of the results (which can be tedious for many people, particularly those who have little tolerance for minor detail - that's not a criticism - some just want to cut to the chase) and b) speed up the process.

As I have mentioned before, and this won't give anything away, there has previously been criticism of Jöst from a couple of other people because dust gathers on the back of the disc. Had I not made note of the dust left on the job and showed the pics of the disc backs then those people would no doubt have been baying for blood that I did not take observations of this and would have remained convinced that the dust on the job must be an issue because it can't look like that on the back and have been effective at removing dust from the job.

My observations have now proved that it is just not an issue, although I very much doubt that will be convincing enough for at least one of those people (who has, by the way, never responded to an opportunity to sample the discs for himself - he would rather maintain the prejudice in ignorance - perhaps because he didn't want to run the risk of being proved wrong).

Another very relevant example is starting to emerge from the second test set: one disc is the quickest at removing the pencil, but curiously it also yields the smoothest finish. "How can that be?" one could quite rightly ask. The answer is "I dunno, but them's the facts". It would seem completely counter-intuitive that the smoothest surface comes from the disc that gets the smooth surface quickest, but that's the way it is.

In the first test set I didn't need any measurements at all to work out which was the best disc - the difference was stark, but the numbers will back up my comments.

[FenceFurniture]

The 80 grit report is ready now for those who would like it.