On quantifying feathering -- am I on the right track?

[A Smug Dill]

6 minutes ago, wbpeoria said:

I had hoped to come up with a relatively quick method of testing multiple inks, faster than filling and cleaning pens.

 

My personal ‘insight‘ from several years of delving deeply in the fountain pen hobby, and trying to systematise, simplify, and streamline testing of inks and such (and cataloguing one's findings for future reference), is that such an approach or ‘endeavour’ is folly, especially if the underlying motivation is to minimise one's own expenses with regard to time, effort, and consumables in acquiring the desired information.

 

The ‘only’ plausible way to reduce one's costs is to leverage the output of someone else's work (i.e. at their expense of time, effort, and consumables); but they will have designed their methods and practices differently from what is ideal for you (or me), and there needs to be work done in reconciling and/or mapping between their framework and yours, as well as gap analysis to identify and acknowledge what will not be covered (and evaluate the importance of what is absent).

 

18 minutes ago, wbpeoria said:

My single figure of merit method might be overly simplified, even if it does work.

 

Reduction of information (or knowledge) down to simple, easy-to-understand metrics is a very different matter from making the acquisition or generation of that information easy and ‘cheap’. I was once asked by management, at the behemoth of a corporation for which I worked, to find a way to “stop ourselves from signing up to impossible(-to-fulfil) contracts” on account of the required service levels. (How the supposedly critical task trickled down to a nobody like me five levels below the CEO still bemuses me.) I ended up creating a mathematical model that reduced the inputs to three numeric parameters, just so executive managers could wrap their time-poor heads around it, to deliver on my assignment. It may have been “overly simplified”, but it should have been good enough, as my model would spit out a single figure of probability of how likely the company could meet the specified service level targets each month going forward, based on ‘known’ performance history over the past 12 months on something comparable.

 

The calculation mechanism itself was so huge and clunky that it would take more than the duration of a meeting for my work-issue laptop to process a single modification in any of the three input parameters; and one of the input parameters was (conceivably, from the very start) nigh impossible to extract from middle managers in the organisation reliably, for political and arse-covering reasons.

 

End result: I was highly commended for the work, and the whole thing was swept under the office carpet, so to speak. It really wasn't what executive managers wanted to know after all; the upper echelon wanted to be able to sign up to these “impossible” contracts (and bump their bonuses), without risk of recrimination when the company copped financial penalties arising from service level breaches month after month. (To be fair, I already knew that from Day Zero.) My mathematical model essentially told them, in “overly simplified” terms, it wasn't going to happen unless they greatly improved actual operational performance of the supporting processes, and there was no relatively quick and/or cheap method for doing that.

 

[wbpeoria]

19 hours ago, A Smug Dill said:

My personal ‘insight‘ from several years of delving deeply in the fountain pen hobby, and trying to systematise, simplify, and streamline testing of inks and such (and cataloguing one's findings for future reference), is that such an approach or ‘endeavour’ is folly, especially if the underlying motivation is to minimise one's own expenses with regard to time, effort, and consumables in acquiring the desired information.

Thank you for the additional reference. 

 

"Folly" seems a bit harsh, in my opinion.  At some level, one might achieve an adequate, albeit less than ideal, utility from such an endeavor. 

 

My plan (or perhaps "hope", if I'm being honest) was to group the bottles in my ink collection by color, and, armed with a dip pen and a capillary tube, write the name of each ink on both Rhodia paper and office paper, and use the capillary tube to do a set of dots on the office paper.  Then the ink bottle gets a wrap of parafilm, unless I know I'm planning to use it soon.  Efficient, and useful (with the goal, of course, to buy more ink in colors that I don't have 🙂).

 

On 9/28/2022 at 10:40 PM, A Smug Dill said:

Reduction of information (or knowledge) down to simple, easy-to-understand metrics is a very different matter from making the acquisition or generation of that information easy and ‘cheap’. I was once asked by management, at the behemoth of a corporation for which I worked, to find a way to “stop ourselves from signing up to impossible(-to-fulfil) contracts” on account of the required service levels . . . .

 

This is a great story.  I would expect something like that in academe, too, by the way.

[txomsy]

I can't remember who said this, I think I got it from "Programming pearls" but could have been any other Software or Project Management book, or even have been me:

 

"There is always some way to measure something that is better than no measure at all".

 

It is not that any measure is valid, but there is always at least one measure that can increase, however minimally, our knowledge. The value of  simplification and standardizing is that the measure may be used by others as reference.

 

That looking for metrics that will give direct answers may not be not feasible without a lot of resources, I will not deny as I fully agree. But it is always possible to find indirect or 'back of the envelope' (rough estimations) measures that may be good enough for a given purpose.

 

Actually, we cannot measure anything precisely and all our "understanding" of Nature is based on models that may be utterly wrong. But the measures and models we use are often good enough to allow us to do what we want.

 

In the case of feathering, I think that while it is true that it is a combination of factors, one may take each one at a time and process them separately, then attempt combinations and with any luck at some point one may know enough to make educated decisions.

 

The experiment in this thread only addresses ink, by trying to isolate it from pen and (supposedly) paper. Once the pen and measuring approach has been refined (I agree it still needs some rounding) it won't answer how a given pen/ink/paper combination will behave, but will give an idea about which inks are inherently and in isolation more prone to feathering.

 

Once there is some estimate, one may proceed to measuring different papers with a single ink chosen based on the previous knowledge (or various if one had infinite resources).

 

Etc...

[wbpeoria]

I think that I have to report a negative result here.   I redid the capillary test, and did so several times for the more feather-y ink.  I tried to keep the capillary tube nearly vertical each time, and tried to keep the height in the tube approximately constant, at first.

 

One problem is the contact time -- it's a little more tricky to ensure that the capillary touches the paper for the same amount of time each time.  Another variable has to do with ink on the outside of the capillary -- it's difficult, if not impossible, to wipe that off without touching the drop on the bottom, and absorbing some of the ink inside the capillary tube. I ended up just putting the first drop elsewhere, and then putting subsequent drops in line on the page.  Each line represents a new filling of the capillary.  As you can see, the drop sizes aren't consistent from line to line in the LE.  

 

I've added a close-up of a few of the feathering drops, in case anybody's interested.  The irregularity of the shape means that it would be difficult to get a precise area using ImageJ, as I was approximating the ink spot as an ellipse.

 

Thank you all for your input and insights.  I especially appreciate the references to previous posts that address the issue of feathering.

[txomsy]

Then, may I suggest a different approach?

 

If I understood it well, the reason for using a capillary was to remove pen dependencies. But you might get similar independence by always using the same pen for all measures. That may be more cumbersome, but allows for an extra bonus.

 

If you use a pen, then you can draw a line. Granted this is a "hard" task, for all lines should be written with similar pressure and speed, but that may be addressed.

 

If you draw a line, then you can simply measure extreme deviations from the center. If the line goes over a paper reference line then you know well where to measure from. A line is one dimensional, so you can take several measures or even ask the computer to detect and connect all extreme points and then the coordinates of peaks will give maxima of lateral spread, feathering. Maxima can be detected easily with several tools; I know of some in R.

 

I've got guests for a BBQ today, and a number of appointments at the beginning of the week, but will have a look and give it a try myself. Sorry, must leave.

[InesF]

On 9/29/2022 at 4:45 AM, wbpeoria said:

OK, you want to talk about a rabbit hole . . . .   🙂

 

To be honest: ... better let's talk about good things in life, like nice fountain pens, ink colours and paper ... 😇

 

On 9/29/2022 at 4:45 AM, wbpeoria said:

I really like your work in that thread.   Do you have a reference that explains measurement of surface tension by the drop size method?  I've used a tensiometer for that type of measurement, but that involves larger volumes than I'd care to use, and I don't relish the thought of having to clean ink residue off of the platinum ring.

Thank you!

Yes, I have a textbook about "Physical measurement methods" from 1975 which describes the principles. The capillary method is only one of three there. But you can find (a less detailed) description also in Wikipedia. Take any capillary that is narrow enough to require a good time for one or two milliliter ink passing through due to gravity and calibrate your system with known liquids.

 

Good luck!

[wbpeoria]

8 hours ago, InesF said:

Yes, I have a textbook about "Physical measurement methods" from 1975 which describes the principles.

May I have the full reference, at your convenience?  Mostly curiosity, at this point, but I am curious. 

[LizEF]

53 minutes ago, wbpeoria said:

May I have the full reference, at your convenience?  Mostly curiosity, at this point, but I am curious. 

I suspect it's reference #8 from Ines' paper (referenced in the thread linked earlier):

 

Quote

[8] K.-H. Näser, Physikalisch-chemische Messmethoden (Eng.: physico-chemical
measurement methods). VEB Deutscher Verlag für Grundstoffindustrie, Leipzig
(1976).

The nearest copy appears to be about 1800 miles from me...   http://www.worldcat.org/title/75164866

[dipper]

On 9/29/2022 at 3:45 AM, wbpeoria said:

Do you have a reference that explains measurement of surface tension by the drop size method? 

A brief explanation:

 

Mount a thin-walled tube onto the lower end of a larger diameter tube that has an on/off tap. Mount vertically with the thin tube at the bottom.

 

Add ink into the larger tube.

Open the tap and adjust to a slow drip-drip-drip from the thin tube.

 

Collect some drops of the dripping ink in a small beaker, counting the number of drops as they fall into the beaker.

 

Weigh the collected drops of ink, and calculate the weight of one drop by division.

 

Calibrate the equipment by repeating the drop-weight measurement process with A ) Water, and B ) Ethanol. These are two liquids of known surface tension values, given in reference books.

 

The weight per drop is assumed to be directly proportional to the surface tension of the liquid, all other factors being held constant. So the unknown surface tension of the ink can be calculated by interpolation.

 

If the weight per drop of ink is at 92% of the interval from ethanol to water weights per drop, then also the surface tension of the ink is at 92% of the interval from ethanol to water surface tension values.

 

[InesF]

On 10/2/2022 at 11:48 PM, wbpeoria said:

May I have the full reference, at your convenience?  Mostly curiosity, at this point, but I am curious. 

Yes, you are welcome. It is exactly what @LizEF cited:

Karl-Heinz Näser: Physikalisch-chemische Messmethoden (Eng.: physico-chemical measurement methods). VEB Deutscher Verlag für Grundstoffindustrie, Leipzig (1976). 2^nd edition.

(I was a bit confused about the printing year)

The paper in my book are meanwhile ochre-brownish and smell like old furniture.

 

For your (and mine) amusement, I have made a photo from the page with the capillary measurement method.