A few weeks ago I issued a questionnaire to try to establish why there seems 
to be such inconsistency in the results obtained by different people with the 
Rowlands Flowform. The questionnaire was sent to everyone I could find in 
the archives who had entered discussion on the kite, and also posted to 
rec.kites. I sent a supplementary questionnaire with a further 2 questions to 
all the respondents. I ended up with 18 completed questionnaires, including 
my own on my recently completed flowform, and one completed at Old 
Warden by a member of the Great Ouse Flyers. 
 
Postings prior to my survey had led me to believe there might be some simple 
explanation, such as the use of regular (crinkley) ripstop, but no such 
conclusions can be drawn from the results, and in fact, they point to no 
definite conclusions at all (except that analysing statistical results with
Excel isn't the push-over I thought it would be!) 
 
My method of analysis was as follows: 
 
Firstly, I entered all the results into an Excel spreadsheet, entering "n" to 
represent the nth reply to a question. 
 
I then formulated a composite metric of the degree of success of a 
respondent's kite as a combination of the replies to Q5 (performance), Q6 
(Needs a drogue?) and Q10 (performance in comparison to other soft kites), 
weighted in that order. On the basis of this metric, I sorted the rows 
representing the 18 respondents into decreasing success order by sorting 
them on the columns representing Q5, Q6 then Q10. 
 
Finally, I plotted other columns as scatter diagrams, hoping to see a rising
or falling trend in the points, representing a negative or positive
correlation of the column with performance. But I was largely disappointed. 
(Where people gave multiple answers, the spreadsheet contains things
like "1,2", which don't plot properly - but they're sufficiently few not
to invalidate the results - sort it yourself if you like!)
 
The conclusions I have come to on the basis of this analysis and otherwise 
are as follows: 
 
1. There is an apparent slight correlation between success and experience, 
based on a combination of the replies to the 3 sections of Q1, but the 
correlation is probably barely significant if at all. (Unfortunately, I can't 
remember enough statistics to determine the significance level - maybe 
someone else can?) 
 
2. At a similar level, there seems to be a slight negative correlation between 
success and the respondent's estimation of his neatness and accuracy, on 
the basis of a combination of the relies to Q2, Q3 and Q4. This may be 
spurious (in which case the correlation with experience may well be spurious 
as well) or it could be that the most skillful craftsmen are the ones most
aware of their limitations. I offer no opinion. No doubt others will. 
 
3. Material: most people used crinckley ripstop, including all but one of the
top 10. One used icarex, one tyvek and 2 balloon fabric, but these didn't
seem to have any significant effect. 
 
4. 13 out of 18 respondents used Rowland's profile (which doesn't match the 
length he gives for the back). Only 2 used Dan Weinreb's (Dan and myself) 
and one split the difference. One other openned the front vent by 1cm at Jim 
Rowlands suggestion (by private mail). (Incidentally, in the same private mail, 
Rowlands gave the thumbs-down to Dan Weinreb's profile.) The profile 
doesn't seem to be significant as far as I can tell from this survey. (I'm
sure I remember a posting or web page which I saw months ago, written by
someone who seemed to know, saying that the profile of a flowform isn't
that critical.)
 
5. Most people oriented the grain on the flares with the leading edge, but 
several people, all of them being amongst the most successful 9 respondents,
did differently. Since they used several different orientations between them,
the only conclusion I can come to is that it doesn't greatly matter (though I 
imagine it might if you didn't edge-bind very strongly). 
 
6. Q24 on whether you would start all over again, or can the project, if you 
trashed it, was semi-serious. I considered that those who would start again 
might be the more painstaking and so the more successful. But no such 
conclusion was evident. Out of 13 who gave a definite answer, 10 would have 
started again. One (who shall remain nameless) declared that he *never* 
made unrecoverable mistakes. Incidentally, he came second to bottom in the
success-stakes! 
 
7. The question about colours was also only semi-serious, but wouldn't it 
have been curious if we'd found that green kites never flew? The most 
popular colours were yellow and white, with blue and red next. Nobody used 
brown or violet. 
 
8. The supplementary questions were to do with wind speed and variability. 
Following a poor first test flight in gusty conditions with my own flowform. I 
wondered whether the less successful builders might be expecting too much 
in poor wind. This wasn't evident from the replies. Most people would want a 
wind speed of 2bft, and some 3bft before getting it out, though Andrew would 
get his out indoors! 
 
Miscellaneous Ramblings: 
 
Unfortunately, the mystery remains unsolved. If anyone wants to research
it further, then I think  the next step would be to get 2 or 3 flowform 
owners of varying success rates together and to examine the kites both in the 
air and on the ground. I still think much of the explanation may be varying 
expectations in varying conditions, though I remember a posting or web
page from many months ago (of course, I can't find it now I want to)
from someone and his friend who'd made 2 identical soft kites which
performed very differently. I'm not sure I'm convinced by Andrew's
hypothesis that this kite is simply extra-sensitive to tolerance - being
soft, it's variation in shape with changes in wind is likely to exceed
normal sewing tolerances, though I could believe variations in fabric
stretch, particulary 2nd quality with off-true weft, could be a factor,
even though this wasn't revealed by the survey.
 
Fom my own (very limited) experience, there are 2 distinct modes of failure 
leading to a ground impact: 
 
The first is directional instability. Sometimes, the kite drifts to one side
and falls, maintaining much the same angle of attack and causing the line to
follow the surface of a cone as it goes down. For the flairs to help this,
they need more area towards the rear - note that an aircraft invariably has
its tail fin on the tail - if it were on the front, a little yaw would cause
it to catch the wind on its side in such a way as to make the yaw worse. It
would be fairly simple to extend the flairs at the back (or just some of
them) to check out the effect. Actually, since air flows through the
kite, the ribs will act in the same way and the jet of air out of the
back may make it worse still. Thinking about it, if I'd just invented
this kite I'd probably dismiss it as a non-starter! Yet it flies great
for some people! How?
 
The second failure mode is luffing. In gusty conditions, particularly if the
kite is already yawing, a gust can hit it from the side and strike the upper
surface of the canopy, causing it to collapse. It doesn't normally recover.
The same may happen if it's flown beyond the top of the wind window,
particularly if the bridle point is on the high side, but I haven't actually
observed this. 
 
Enough from me. I am attaching a copy of the questionnaire with the 
supplementary questions merged in, and the survey results as raw text with 
fields separated by tabs, for easy import to your favourite spreadsheet. 
 
I've also put the raw text results, the Excel spreadsheet, and this posting,
on http://www.geocities.com/CapeCanaveral/Hangar/6767 
 
- Philip