The Triple or Check Haul

[Merlin]

Hi Paul

Here is a run performed with my last fly line flight. As usual there are simplifying assumptions (horizontal parallel legs, non morphing half circle loop). I illustrated the “WEE” model (Work Energy Equation) which was published by Noel and Caroline years ago. That is an analytical model (exact mathematical solution for the assumptions), and mine is numerical (e.g. 1500 calculation steps) but allows calculating tension at the ends of the loop.
At this time it only handles level lines but the next version is under work, which will allow introducing a taper.

Checked cast 1.JPG (26.04 KiB) Viewed 3261 times

Check 1 tension.JPG (20.6 KiB) Viewed 3261 times

Fine checking the model made me realize that there are unfortunately issues in three papers from Caroline & Noel

Merlin

[Dirk le Roux]

Would you agree that there is a noticeable delay for the fly leg's velocity to start increasing relative to when the rod leg's velocity starts going negative?

Hi Gordy

I understand what you were driving at. I agree that there is a noticeable lag between when the rod leg starts going negative and when the fly leg actually starts positively accelerating. On the rough velocity slopes drawn over your data, the rod leg velocity goes negative at 0.27s and the fly leg starts speeding up at 0.47s. A lag of 0.2s. On my zoomed-in measurement, the lag is 0.15s between pullback start at 0.36s to fly leg positive acceleration at 0.51s.

With your earlier pullback video, that lag was about 0.125s and with Graeme's vertical snap, the lag was around 0.12s.

You stated that you do not think the fly leg velocity will necessarily change instantaneously with changes in the rod leg velocity . A change in velocity is not necessarily a positive acceleration. It can also be a less negative acceleration. Here, I believe the evidence supports that I am still in the clear in saying that a change in rod leg velocity is surprisingly instantaneously accompanied by a change in the fly leg velocity.

I did err in using speed up in a later reply. I apologise for the confusion and correct that now. I began to confuse acceleration with jerk.

My charts show that every time that pullback intensifies, there is an accompanying positive jerk (velocity slope changes to less negative, more positive) readable in the fly leg velocity profile.

All the best,
Dirk

[gordonjudd]

Merlin,
What is the mechanism that causes your model to start deviating from the WEE model at about the time the tension start to have different values?

Fine checking the model made me realize that there are unfortunately issues in three papers from Caroline & Noel

I have a limited number of fly velocity history measurements that fit quite well with the Gatti-Bono closed formed expression for a level line. Should the errors you found affect the results she gives for a level line?

Gordy

[gordonjudd]

I understand what you were driving at. I agree that there is a noticeable lag between when the rod leg starts going negative and when the fly leg actually starts positively accelerating.

Dirk,
Its good that we now agree (at least I think we do) that it takes some time for the increased acceleration (that is associated with the higher tension associated produced by the increased tangential velocity) due to pull back to produce a noticeable increase in the fly leg velocity.

Thus I hope you can see why I questioned your previous assertions:

I have already been told the answer, which repeatedly was surprisingly instantaneous, by measuring properly and by trusting my eyes. It is for models to match reality, not act as constructs to tell us what reality should have been.

Rod pulls back, fly simultaneously speeds up. As expected. Moving on

The fly leg's speed up reaction time upon the rod leg being pulled need not beg agreement. The evidence speaks for itself once you have less broken data.

Those assertions arguing for a "instantaneous" speed of the fly leg once the rod is pulled back did not make sense to me because for that to happen the loop would have to be acting like a stationary peg.

It also did not fit with the results discussed in an earlier thread http://www.sexyloops.co.uk/theboard/vie ... ain#p34970.

How quickly we forget, I just came across that old discussion. At least we now have some fly cast data to support what was concluded from the bead chain test three years ago.

Gordy

[Dirk le Roux]

Hi Gordy

That complicates the timing for how the velocity of the fly leg is affected by a change in the rod leg velocity. It could well take a model to give some insight as to the actual timing for the increase in the fly leg velocity for a given pull back velocity variation. My measured data says the timing is not instantaneous.

I said I shouldn't have used "speed up" for all instances of affecting.

The charts with proper data show clearly the relation between the rod leg pullback and how "velocity of the fly leg is affected". Look for the correlation in the jerks. Those are where rod leg velocity slope steepens downward and the fly leg velocity slope steepens upward. You will see that with early pullback jerks the fly leg's decreasing velocity slope flattens. With a next pullback jerk the fly leg slope is now changed to increasing velocity and with a next it much steepens.

The earlier pullback spurts where not enough to reverse the fly leg's deceleration, but they sure affected it.

All the best,
Dirk

[gordonjudd]

Here, I believe the evidence supports that I am still in the clear in saying that a change in rod leg velocity is surprisingly instantaneously accompanied by a change in the fly leg velocity.

Dirk,
I am not following your argument. By change in velocity I assume you are talking about acceleration and not the derivative of the acceleration which is jerk. I can produce jerk curves on my smoothed data in Python if that will help illuminate what you are saying.

Here are the acceleration curves that Tracker gives for the time range we are talking about. Can you point to some of the time points in those two curves to explain where you see that changes in the rod leg's acceleration is accompanied by a simultaneous change in the fly leg's acceleration?


Note the fly leg's acceleration stays negative up to .5s because it and the loop are being slowed down by drag losses.

Gordy

[gordonjudd]

Here are the x velocity curves that were used to produce the above acceleration curves. You can see it does not take much of a wiggle in the velocity to produce a relatively big change in the acceleration values.

Gordy

[Merlin]

Gordy

The WEE model only handles tethered casts. The Tension model can handle in addition pull back and untethered ones. I keep the WEE data in the background because it makes the comparison with the same tethered cast easy. This is why it is illustrated in the graph in post #51. It takes 3 columns in a spreadsheet to get the WEE result. My tension model uses 10 times more columns in its current version. And that number will increase to cope with various situations. Both graphics illustrate the effect of the check I described in the post (starting at 0.9s).

Here are the issues I found in publications:
• In the one presenting the WEE model, the illustrations do not correspond to the tables. In fact, the fly leg should always be 4.8m whatever the loop size is, but it is as if the loop had been deducted in a particular case and the corresponding fly leg length kept for all cases. The graphics correspond to a fly leg of 3.65m (approximately). The difference corresponds to a 0.3625 m loop radius which is the value used to compare the WEE model for a level line and for a tapered line. Was that peer reviewed?
• In the paper about drag affecting loop, the expression of the horizontal drag force is wrong since it does not contain the skin drag contribution (skin drag forces are not symmetrical). In the same paper (and not in a third one, I missed the point), the expression of the vertical acceleration is wrong, it should be divided by pi. Same question.
In fact I had to recalculate the horizontal drag forces for a case where leg speeds are different. And now everything is fine in a tethered case comparison. On top of that it is applicable to other casts.

I started playing with the model in the case of an untethered cast. There are a couple of issues coming out:
• Although the Tension model can handle a case when the fly leg is set free at the very beginning, it may well not be the case in real life. I take my own situation when I shoot a cast: I know the loop is shaped as I feel the increase in tension in my line hand, then I release the line. The time for the information “tension is there” comes to my brain is about 40 ms (biomechanics). The time for my hand to open up after my brain decides it is time to do so is also about 40 ms. So there is some 80ms delay during which the rod leg increases. For a 30m/s cast and a one foot loop it corresponds to 1.2 m of rod leg. I use a 9m “head” of level line weighting some 280 grains. That makes 15% of the fly leg length and it has an impact on the fly leg speed history.
• I started testing the model with a lighter head, something like 155 grains, 9m long. Here the fly leg speed story is different as the loop front speed can go to zero even if I release the rod leg after 200 ms. So if anyone has some experience with a light short head casted with reasonable speed (30 m/s), I’m curious to get the feedback. I do not think there is anything on the market below 300 grains so it must be a spare light line cut to the desired length (e.g. 9m).
I think I must add the possibility to use a running line (with the possibility to set its mass to zero). That will better correspond to WF tapers.

Merlin

[Lasse Karlsson]

What kind of experience are you after Daniel?
Got short heads from 3,7 meters and upwards..

Cheers
Lasse

[Dirk le Roux]

By change in velocity I assume you are talking about acceleration and not the derivative of the acceleration which is jerk.

Hi Gordy

I used "change in velocity" based on your "the timing for how the velocity of the fly leg is affected by a change in the rod leg velocity". However, the two legs' velocities in this case change virtually all the time, so it is somewhat pointless to merely talk "change in velocity" without going to slope changes, trends, patterns, increases and decreases in acceleration (which is jerk).

Can you point to some of the time points in those two curves to explain where you see that changes in the rod leg's acceleration is accompanied by a simultaneous change in the fly leg's acceleration?

If you don't mind, I'd rather use your data from the file you sent me some days ago, as the huge fly leg velocity toward the end of your external model curve effectively dwarfs out the readability of earlier events. This data is smoothed somewhat differently to your later version, but I can zoom in a bit and therefore work with it.

Here is your data, with the pink areas marking where negative direction rod leg jerk and positive direction fly leg jerk occurred:

The dot and time instance is where you jumped from one fly leg marker to another on the same run. Hence the squiggle.

The lighter coloured area is a longish gradual jerk according to your smoothed data.

The relationships should be obvious.

To the casual reader: On a velocity chart, slope indicates acceleration, curves indicate a change in acceleration (jerk). The concavity of such a curve indicates the direction of the jerk. On an acceleration chart, slope indicates jerk.

Here is the same exercise on my measured data of the same cast:

All the best,
Dirk