Digital BW, The Print

My information is about twenty years old, so I'm just trying to 
recollect from memory.  And I never studied statistics, I had a 
statistician at HP I always worked with. With that in mind...



It's called "design of experiments". At HP, we'd set up  15-30 
samples, as statisically that's about what you need to prove 
probability that you don't have a stray data point. In other words to 
predict what the entire population will do, subsample and the math 
works out to this many samples.

That's 15-30 for each variation. So if you had one variable, let's say 
 it's temperature effect on fade, you keep humidity, amount of light 
exposure, pollution, yada yada...EVERYthing else the same....but you 
have 15 samples at say 100 degrees f and 15 at normal temperature as 
"control". 

A lot of samples. So before you come up with a hypothesis you might 
take three or four samples and do stuff to them varying more things. 
When you see what seems to you to have the biggest impact, you have to 
design a way to do it consistantly to your real test samples. And 
CONTROL everything else but one variable if possible (there's a way to 
vary two variables+ at a time but it makes certain assumptions and 
makes my head spin too much<g>)

So, what I might do here is say that physical stroking of paper causes 
 the problem. But I'd try 3 or four samples and see if it gets worse 
when humidity changes say or type of paper or ink. You will probably 
have to restrict yourself to just one or two inks and paper types as 
variables multiply the number of tests needed to be run. Find out 
whether say paper change or ink change has the biggest effect this way 
and change just that one variable. Knowing that Hahnemule papers are 
more fragile than say Museo, you might use Museo as a control. 
Actually choosing enhanced matte might be better, as it is the old 
standby and everyone is familair with it- so it can be a "yardstick". 

Find a uniform way to abrade the paper. You need to control two 
variables here: co-effecient of friction and pressure applied which is 
composed further of force x area. For example a pencil eraser has a 
higher co-effecient of friction than a polished ball bearing. If you 
apply 1 ounce of force downward, perpendicular to the paper, but use a 
2 inch diameter ball bearing it has less pressure than a 1/8 inch 
diameter ball bearing.

So I would say here find somthing of constant area that applies a 
constant force and keep angle of force consistant. You could rely on a 
spring loaded something  to apply a constant pressure on the paper. Or 
just keep it simple and use the weight of a 2 inch ball bearing or 
even a cube of steel or plastic with polished corners - glue a thread 
to it and drag it across the papers- simple. No spring needed. This is 
where you experiment with a few samples first to find what material to 
use and weight- and the machined finish of the steel or plastic is a 
variable also of course.

In your case with a book format, why not have a method of 
non-permanently gluing another piece of the same paper to the bottom 
of the weight and dragging it across? Replace the paper often so it 
doesn't vary IT's characteristics either...

And eliminate all other variables. Even subtle or unlikely ones. If I 
stroked the paper with my thumb, you might say it's non-uniform 
surface area, force, etc. The co-effiecent might change with how much 
oil or callus is on my thumb that day...and further the oil might have 
an unknown effect on the ink/paper...how do you really know? Well 
anyway, streching the point but you get the idea.
JIm H. 



--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "Cleavis 
<lyonscox@c...>" <lyonscox@c...> wrote:
> I'm not much of a scientist, at least in developing a procedure.
> 
> How to set up a theory and test parameters for abrasion resistance 
of 
> inkjet prints, particularly as they would be applied in a book 
format?
> 
> Ideas?
> 
> Thanks,
> Cleavis