International Guild of Knot Tyers Forum

General => Chit Chat => Topic started by: jglohr on June 11, 2006, 09:57:28 AM

Title: Looking for a knot
Post by: jglohr on June 11, 2006, 09:57:28 AM
Hi. I am new to the forum, and am looking for a knot for a very specific application. I have been using the adjustable hitch grip with moderate results, but am looking for something better.

The application is a homemade drum, which uses a cloth drumhead tightened over a cylindrical drum shell. A hem is sewn all the way around the inside edge of the cloth to hold a drawstring with an opening for the two ends. An adjustable grip hitch is tied at the end.

(http://www.geocities.com/jglohr/drum1.JPG)

Using a piece of wood with a slot to hold the knot, the loose end is pulled upwards to tighten head.


(http://www.geocities.com/jglohr/tightening.JPG)

The result is a high tension rope, and the hitch works great for holding the tension, but the drawback is that as it tensions the hitch slides to the right, causing the material to stretch, and eventually tear if enough tension is added.

My question is, is there a knot which would stay self centered in the opening in the cloth? It seems to me that for this to occur, one must pull both loose ends of the rope, with a knot that can slide on both ends. Does such a knot exist?

(http://www.geocities.com/jglohr/tightened.jpg)

My other question is regarding rope material. Nylon works, although the stretch and high friction result in a very high force necessary to achieve proper tension, and it is very difficult to untie. Polypropylene works good because it slides easier and doesn't stretch as much, but the stiffer rope is harder to tie and dress the knot, resulting in less holding strength if not properly done. I haven't tried polyester yet. Cost is also an issue, which is why polypropylene is desirable.

Thanks for any input on this question.

Regards,

Josh
Title: Re: Looking for a knot
Post by: KnotNow! on June 11, 2006, 10:27:51 AM
Hi Jglohr (aka Josh),
 The darndest thing comes to my mind:  Drummers have drum solutions and mandolin pluckers have mandolin solutions.
 I am happy that you came to IGKT for a "knotting "solution to a "drumming" problem.
 As I understand "drum" there is a mega change in tension as the drum head heats and cools.  My buds (buds = good friends who love to sit by a fire and pound a drum all night) hold the drum over the fire and the tone changes and modifies?  Me, I can't stay up all night but does that make me stupid?
 You come to forum and want a rigid drum head connection?  I have it.  Use steel.   What I have found is that "drummers" want the variation from heat and humidity... the tone they love.. flexible and on the curve.
Title: Re: Looking for a knot
Post by: KnotNow! on June 11, 2006, 10:35:10 AM
Hi,
 I contacted some "drummers" in my turf.  The "tie off" is nothing to them.  You could tie off with any good holding knot.  The heat over the fire on the drum head and the subsequent tension is the thing.  Knots have almost nothing to do with the final tone.  Of course they are all a bit out of it.  What was I saying?
Title: Re: Looking for a knot
Post by: jglohr on June 11, 2006, 08:50:54 PM
Thanks for your replies. I have to clarify my drum a bit, because it is not a typical application, and heating it over a fire is not an option. I am trying to design a drum that would be merely used as a practice pad (IE it makes very little noise when hit.) The head is a cloth mesh, as opposed to the leather heads that are typically tied off with rope. So all the tension must come from the rope, and it should be easily adjustable. There is not much tone on a practice pad, but the tension must be matched to a real drum so it feels authentic. Comercial practice pads and the drums I am trying to replicate use lugs and a steel rim to achieve the same results, but I am going for the low-cost alternative.

Thanks,

Josh
Title: Re: Looking for a knot
Post by: KnotNow! on June 12, 2006, 07:19:37 AM
Hi Josh,
 I will sign out.  My local drummers play cedar hoop, hand made, leather skinned drums and have no info on "practice drums".  Good luck with your project.  As a knotter I am intrested but can not see what is needed for you as my skill as a drummer is limited to pounding on the bottem of my pasta pot with a "wooden kitchen scoop" (edited by me to keep the profanity filter from making nonsense out of my post.  This is fun because the filter knows "bad" words I have not learned yet).  I did bother to learn how the leather and cedar drums are made... just as a craft experience.  Some fun to be had there.  Good luck with the drumming and the drum you will make.
Title: Re: Looking for a knot
Post by: jglohr on June 12, 2006, 07:59:17 AM
Thanks for your help Roy. I'm always looking for something that probably doesn't exist. Really, the drum is irrevalent to my problem, I am mainly asking if there is a non-sliding version of an adjustable hitch (the loop being what moves). The only idea I had was wrapping several passes around both ends of the rope with a different one, and pulling evenly on both ends, versus just one end. This didn't really work because the wrappings didn't maintain their tension. I figured if there was such a knot, you guys would know. I can't really think of any other applications for such a knot.

Josh
Title: Re: Looking for a knot
Post by: KnotNow! on June 12, 2006, 09:50:33 AM
Hi Josh,
 Now that we are not making a musical instrument there may be many knots which will slide to an adjustment and not slip back when released.  Every day I use a very old "slip and hold" knot.  It is the blackwall hitch.  The hook is fixed and the line is  moving.  This needs "hardware".
 But the rolling hitch, the midshipmen's hitch and just a whole "boat load" of knots would seem to meet your needs.  These knots will slide in one direction and when released will still hold.  Most will put a kink in the original main line.  As you tension your line these may hold.  However the "icecicle hitch" will hold without putting a kink in the main line.   I am sorry that icecicle or icecycle or other is not in my "spelling checker" at IGKT Forum.   The hitch has so many turns and locks that it will hold on the tapered frozen ice that forms on the eves of your home in midWinter.  Since the frozen water will not yeild or bend the hitch must apply all the stress in the structure of the knot.  The fun part of this is that many readers of this post may not know of ice on the roof.
 Ice or not (knot) if you need a knot that can be pulled tight on a standing line, to to draw tight a drum head or other object you may want to look at the series of knots that start with the midshipman and progress from there.
Title: Re: Looking for a knot
Post by: roo on June 12, 2006, 10:00:28 AM
Quote
Hi. I am new to the forum, and am looking for a knot for a very specific application. I have been using the adjustable hitch grip with moderate results, but am looking for something better.

Perhaps you could modify your design to involve grommeted eyelets in a smaller skin... lacing in a zig-zag pattern though holes in the bottom of your cylinder.  

I'm not certain, but it seems like it might make motion of the cord less important and it might even leave enough room for a Versatackle, which can easily produce a large amount of tension, and can also be easily released:

http://notableknotindex.webs.com/Versatackle.html

Grommeted eyelets may also be able to be used with your current design if the lacing is done in a "star"-like pattern.
Title: Re: Looking for a knot
Post by: KC on June 12, 2006, 05:49:58 PM
Very Nice!

A Versatackle to me is like the tightening of a Trucker's Hitch or Z-Rig; only with another loop eye to reeve bitters thru, allowing trailing the bitters thru both loops a number of times.  This gives much more power; by giving more distance to pull for same work/ distance of drawing ends of systems closer together.  The more turns taken on the moving load end, the more power output.

Continued (http://www.igkt.net/cgi-bin/yabb/YaBB.cgi?board=news;action=display;num=1150120313;start=0)
Title: Re: Looking for a knot
Post by: Willeke on June 12, 2006, 07:32:27 PM
Josh,
do you have, or can you make, a fixed point to use as anchor?

Willeke
Title: Re: Looking for a knot
Post by: jglohr on June 12, 2006, 07:54:40 PM
I tried several sliding hitches, such as the rolling, midshipman's, tautline, etc. All work good for tightening, but the problem is the same with all of them - because of the friction in the drawstring seam, instead of the entire loop moving when tightened, the hitch moves, causing the it to slide to the right and stretch the material. So I need to either fix the hitch from moving, or have a self centering hitch. I like the grommet idea, but the idea is to make a dozen of these drums as quickly and cheaply as possible. I may be able to have an anchor, but the forces may be to high, as the material is only  5mm MDF.

Thanks,

Josh
Title: Re: Looking for a knot
Post by: Dan_Lehman on June 12, 2006, 11:03:32 PM
Quote
... because of the friction in the drawstring seam, instead of the entire loop moving when tightened, the hitch moves,
But this is given your particular tightening method--of pulling the end UPWARDS
through the wood--; can you pull more in line with the circumference,
holding the knot in place and pulling the end to/through it?  --that could have
a lesser problem, perhaps, of the end's movement drawing the hem towards
the knot with it; but can this disturbance be countered by some iterative
fiddling with the stuff?

As for material, you cite chemical composition, but that's only part of the story
(and IIRC, nylon's pure coef. friction is lower than PP's), another being structure
of the cordage.  The cord in your photo looks to be a sort of diamond braid, which
has a relatively rough surface; "solid braid" nylon would be smoother (and not all
so stretchy, though I'd think that your braid wasn't so much either).

I've mused over having a 2nd rope in which the hitches would be formed, through
which one would equally draw the two ends of the tensioning rope, but in practice
I think one will find that things don't work out to their theoretical niceness
--one end would pull through whilst the other just thought about it.  :P
(I tried Rolling H. with a Prusik orientation, 1/2 turns; loaded on both ends,
when pulling down against the 2 turns, the one turn pushes the knot and prevents
it from gripping--this would be the tightening mode--; going the other way, it
grips.  (This idea arose from a mistake in the revision of On Rope for
its 2nd edition, where presumably a layout person inverted images and changed one
word of text to match, yielding sliding hitches that will do ONLY that--slide!))

--dl*
====
Title: Re: Looking for a knot
Post by: jglohr on June 13, 2006, 06:46:53 PM
Yes, I agree that pulling upward isn't helping, but I have yet to find a good way of gripping the hitch and pulling tangentially. I think in the end it would still slide, because the tension is so high that there is a huge frictional force all along the rope and hem.

I also experimented with the second rope method. What I saw as the problem was that on a normal hitch increased tension causes the hitch to self tighten, whereas with a second rope nothing keeps it tensioned. So I was wondering if there is another knot that I don't know about that achieves the same purpose without a second rope.

The rope in the picture is popypropylene. My main reason for using pp is cost at this point. It is obvious that cost is a big factor in the design of this drum.
Title: Re: Looking for a knot
Post by: DerekSmith on June 14, 2006, 01:22:27 PM
Hi Josh,

I don't think you need a different knot, just a method of using this one to best effect.  Dan cracked it by identifying that you need a tangential tensioning force, so how about this as a means to obtain just that.

(http://igkt.pbwiki.com/f/drumring.gif)

First tie the slip/grip knot inline with the ring rather than sticking out of the ring as you had been doing.  Leave a good end on this knot, (I will call it the S/G end).  Tie a loop in this end in order to apply tensioning pull.  The end you previously used to haul on (coloured red in the picture) I will call the Red end.

Having made and dressed up the knot nice and tight, slide the knot along the cord to gently tighten the ring into position and make sure the knot is central in the gap you want it to stay in.

Now for the tensioning.

Throw a cable over a beam and tie it to the loop on the S/G end so that the drum is suspended about shoulder height.  Then tie a loop in the Red end about a foot from the ground so you can put your boot into it.  Hold the drum up with your hands so the line falls straight from the beam to your boot with the drum sort of fixed on one side.

Gently press with your foot to tension the knot.  As you drag the Red end through the knot, the force will be transmitted arround the loop and up to the beam (which is effectively pulling in the other direction).  This is the same effect as if you had grabbed hold of the S/G end in one hand and the Red end in the other and pulled the ends apart.  The knot stays still as the loop tightens because there is no unballanced force dragging it in just one direction.

However, because it is your foot in the loop, you can add a lot more tension than if your were just pulling by hand.

Let the tension off with your foot and allow the tension in the loop to tighten up the knot.  This will remove slack from within the knot and lets you check the drumskin for correct tension.  If you need more tension, press down with the foot again, but harder, to pull more cord through the knot.  Again, release the foot pressure and recheck tension in the skin.

If you need to (and your stitching can take it) you can even jump up and down to pump in more tension and each time the S/G hitch moves forward over the red part, it locks in that tension like a one way ratchet.

Technically, you should be able to release the S/G by pushing the nose of the knot back towards the Red end.  In practice, (with a few hundred pounds put onto it) the knot locks rock solid and cannot be released without cutting it off.

Thank you for posting this challenge. I have often wanted a good knot to apply tension progressively, but generally have finished up with the parcel hitch which is very weak and seriously limited.  Your use of the S/G hitch has been the all important catalyst to solving the problem.

Thanks - and good luck with the practice drums.
Title: Re: Looking for a knot
Post by: Dan_Lehman on June 14, 2006, 05:11:27 PM
Following up on Derek's idea, another hitch to try in the same orientation
is what I call the ProhGrip (aka Blake's) hitch.  Various sites show this, incl.
http://storrick.cnchost.com/VerticalDevicesPage/Ascender/KnotPages/KnotBlakes.html
which has several other friction hitches.  Loading the end should defeat the grip
of this hitch to enable the tightening as Derek shows for the other hitch.
You'll still have to deal with the biased draw of hem on the side being drawn
through the hitch, though.

:)
Title: Re: Looking for a knot
Post by: jglohr on June 14, 2006, 07:11:25 PM
Thanks Derick,

That looks like the exact solution I need to my problem. I'll give that a try, and post the results.

Josh
Title: Re: Looking for a knot
Post by: jglohr on June 14, 2006, 09:50:13 PM
I just got done testing the inline method of tying the knot. At first it I thought it would work perfectly. But it actually did the same thing as before - the knot slid to one side. After thinking about it more I understand why.

Here is a diagram showing the forces acting on the rope:

(http://www.geocities.com/jglohr/drumfbd.gif)

There are two possibilities for the knot to move when tightened:
1) the red end slides through hitch clockwise, subsequently tightening the loop, while the hitch remains fixed (desired case)
2) the hitch slides counterclockwise along the red end. The red end and entire loop remains completetly fixed.

In order for case 1 to occur, F_friction hem must be less than F_friction knot. Because F_friction hem is a huge force when any tension is applied, the entire loop essentially is locked in place. Thus, the force applied can only overcome F_friction knot, which makes the hitch slide instead of the loop (case 2).

In most other applications, this adaptation of the knot would work as desired, because there wouldn't be a frictional force tangential to the loop. That is why it works with a free rope.

All was not lost, however, as this method is still better than the method I was using. If I pull a little bit, and pull only on the hitch end, I can work the knot back over to where it should be. Using this method of applying a load, adjusting, and applying a load again, I can achieve the results I am looking for.

This has been a good learning experience. Thanks for all the interesting posts.

Josh
Title: Re: Looking for a knot
Post by: jglohr on June 14, 2006, 11:19:17 PM
Now, my question is, which of the multitude of slip/grip knots is the easiest to untie/loosen when in tension? This needs to be adjustable, which means easy to loosen as well. The adjustable hitch grip was pretty easy to untie in the non-tangent setup, but is rock solid in this setup.
Title: Re: Looking for a knot
Post by: jglohr on June 14, 2006, 11:46:05 PM
After looking at me previous post I started thinking more on the forces. Now I think I was wrong - for both cases to occur the frictional forces in the hem would act against each end.

Now when I look at it I think you are right Derick. It should stay centered. Yet it doesn't.

I need to think about this one some more.
Title: Re: Looking for a knot
Post by: jglohr on June 14, 2006, 11:54:44 PM
Ok, let's try this again.

If I am thinking about it right now, F_friction hem is not acting as shown in my diagram, but in fact acting in the opposite direction in the bottom half of the circle (assuming when tensioning, there will be a fixed point exactly 180 degress from the knot, and on the bottom half, the rope will be stretching counterclockwise from this point, and on the top half clockwise. Thus, both case 1 and 2 must be happening equally, as the forces are balanced and equal. Thus, while the ideal case 1 is happening, case 2 is also equally happening, resulting in some movement of the knot.

Let me know if this makes sense, or if my logic is wrong.
Title: Re: Looking for a knot
Post by: Willeke on June 15, 2006, 12:47:45 AM
Josh,
as a member you can modify your old posts. That way you do not have to have 2 or more posts telling what you have missed or done wrong in a previous post.

(You are welcome to post as often as you feel needed but when reading I feld you were not happy with the posts.)


Willeke
Title: Re: Looking for a knot
Post by: DerekSmith on June 15, 2006, 12:51:27 AM
Of course Josh, you are right.

If you fix the opposite point of the cable to the hem (effectively achieved by the friction in the hem), then as you pull up the loop, the circle gets smaller.  So BOTH pieces of cord become too long and move away from the desired point.  Thats not a problem for the red end, as it is all the same, but the knot end effectively moves forward because the piece of cord from the half way point to the knot is now too long for the reduced size half circle.

Back to the drawing board - sorry.
Title: Re: Looking for a knot
Post by: Dan_Lehman on June 15, 2006, 01:13:20 AM
With Derek's general set-up--eye in end of hitch, oriented to pull tangentially--,
one should be able to control where the hitch is, by holding in place this eye
as the hitched part is drawn through the knot, or pulling/releasing the eye to
whichever direction seems needed.  However, so long as predominantly the line
is drawn through a stationary knot, that line is moving ONE way through the
entire and increasingly tight-upon-it hem, and will drawn the hem with it.
Maybe there is a strong force, thus, for the hem to arrest this drawn-end movement
and coerce compensating movement of the knot, though I'd think that one could
position one's hold on this so as to at least become aware when this force occurred.

But one cannot have the two ends (i.e., knotted end & hitched end) make
equal movements in opposite directions (so as to equalize the line's draw through
the halves of the hem), because the knot is around the other line and would
have to move into the hem whatever distance was needed to balance.

--------
There is a (theoretical  ;) ) way to, er, *centralize* things, so try this tactic:
1) size your cord so that it wraps around the desired circumference TWICE and more;
2) form a Prusik Hitch around air in the center, and bring each end, in opposite
direcitons through the hem, around through this hitch;
3) work the hitch snug/tight;
4) pull ends in opposite ways to tighten evenly (in theory--that's all I do here!)

In this set-up, having the doubled cord through the hem might help in some
way diminish the friction against it?  In any case, the cord will be moving equally
in both directions, so the hem should be *unruffled*.

As for ease of untying, well, c'mon, ain't 2 outta 3 okay?  Is this g'mnt work?
One could get extra clever (pain on the tying end) and arrange to make this set-up
in reverse, beginning with a loop to cross the ends (which loop will be enlarged
in pulling it wide to tighten...), run ends through hem and each emerges to tie
one half of the Prusik, being themselves tied off (Reef Bend) to form the Prusik's
cross-part (I've seen it recommended to orient the typical Grapevine bend
in a climber's/caver's "prusik sling" positioned here so as to keep the knot from
jamming so much (YMMV))!  That's maybe too clever by half, but, hey, it's only
keystrokes for me!   ;D

The above, yes, entails more cord.
But one expects to suffer the usage of this, so if it's a better mousetrap in that,
it might be worth it.  (Btw, the idea of cost of these fiddly bits being relevant
surprises me--it just isn't all that much cordage!)

--dl*
====
Title: Re: Looking for a knot
Post by: jglohr on June 15, 2006, 02:53:26 AM
I believe it is in fact possible for both ends to move opposite. Both ends are being pulled with equal force, which means that, assuming frictional forces to be equal around the entire hem, deflection should be equal on both sides of the "equator" (a line drawn from the knot to a point 180 degrees). On the non hitch end, this deflection is achieved by sliding through the hitch. On the hitch end, because the force is directly pulling the hitch, the deflection is taken up by the hitch moving, and it is moving directly into the hem. Which, of course, is the whole problem, because it stretches the material eventually to the point of tearing.

Another way of thinking about this is, that the hitch is not in fact stationary as previously described - the force is directly pulling it into the hem.

All this is assuming the drum isn't fixed. If I hold the drum in my hand and vary the angles of the pull force I can effectivly make the force not equal and oposite (IE, hold the drum tightly and pull only the hitch end, but in the opposite direction to pull it back after it has moved. I think this might be what you were saying Dan. Another option may be to put a nail or some hardware as a stop for the hitch, although given the low strength of mdf I'm not sure this would work.

I am intrigued by the Prusik idea, my one question is, what keeps tension on the hitch? Being independent of the tension in the main line, there is nothing keeping it from coming undone. This line of thinking is what I was originally looking at.

Yes, cost really isn't that big of an issue, other than that I eventually plan to make 12 of these things for an entire set. I just like to design my projects as cheap as possible, because DIY projects not planned out correctly almost always end up costing me more than buying the thing new. Now I notice that PP isn't really much cheaper than Nylon (at least not at the store I was at).
Title: Re: Looking for a knot
Post by: Dan_Lehman on June 15, 2006, 09:23:23 AM
Quote
Another way of thinking about this is, that the hitch is not in fact stationary as previously described - the force is directly pulling it into the hem.

Well, again, one should be able to simply hold the end-eye of Derek's knot so
that the hitch doesn't move with the end and thus the end is drawn through
the hitch, whose only movement should be a bit inward along the radian--you
again are holding it against moving with the draw of the end through it,
but not pulling away from that draw!

Quote
I am intrigued by the Prusik idea, my one question is, what keeps tension on the hitch?

The two ends (well, two LOADED parts--no ends are "loose" here) run in opposite
directions around the drum, passing head-on at 180deg point, continue over
each other's outset to re-meet at ground zero (the knot) and dive through it
still in opp.directions.  Pulling on these ends pulls around to tighten the knot.
(Though one will want to set the knot with some pre-tensioning, to draw down
the wraps.)  The tensioning of the ends loads the hitch.
In the "too-clever-by-half" way, presented as one way to have things easily
untied, you tie this set-up in reverse, essentially having the first set-up
with ends welded together after passing through the hitch, then cutting the
Prusik's *back*, and then tying that cut back (which spans the wraps).
Otherwise, untying will entail some bit of gripping each end and pushing on
the Prusik to draw out the end.  Maybe pliers.  You'll be best to see what sort
of tension & tightness arises from the materials and need.

(-;
Title: Re: Looking for a knot
Post by: jglohr on June 15, 2006, 10:59:39 PM
That's the whole problem, it is impossible to just hold the hitch without drawing it away. By holding it, and applying a force on the other end, you are applying an equal and opposite force on the end you're holding. Because there always must be a constraint with an equal and opposite reaction to remain statically balanced, and in this case it is the hitch end, the only way to avoid this is to hold the drum, so the drum essentially becomes a contraint. I don't like this option because the drum isn't that strong.

I understand the Prusik idea now, and tried it out on a rope outside the drum. I think it is the closest to what I am looking for in terms of equall draw on both ends. I will sew up a drumhead and try it out.
Title: Re: Looking for a knot
Post by: DerekSmith on June 28, 2006, 09:39:52 AM
Hi Josh,

I don't know if this issue is dead yet, but it has been bugging me since the realisation that both ends of the drum loop move out of the seam equally when you apply tension.  To achieve the goal you set, both ends would have to move through the knot as tension is applied.  Also, you wanted a means to slacken off the tension, so the knot really had to be a good slip/grip hitch.

The nearest I have come up with so far is to tie the working knot(s) using a second cord so that both of the tensioning ends can flow in or out of the seam without the knot moving.

In the following picture, the drum tensioning cord is the white loop with ends C and D.  You can use the overhead support and tensioning footloop as in the first proposal to apply tension to the ring.

In the red cord, I have tied two simple four turn slip/grip hitches in a back to back configuration.  In use you will have to fix the ends A and B in order to stop the two knots moving towards oneanother as tension is applied.  Alternately, have a helping hand hold the two ends in tension while you put tension onto C and D to tension the drum.  I found a simplesolution was to tie A and B respectively to the ends of a short length of cane to keep them in place.

Slackening the knot is easy.  Simply retension C and D to take the tension off of the hitches, pinch the two hitch 'noses' together and then relax the tension in C / D.  The white cord will then slip through the hitches until you stop pinching them together.

(http://igkt.pbwiki.com/f/drumloop02-sml.gif)

Here is a closeup of the knot 'assembly'

(http://igkt.pbwiki.com/f/drumloop03-sml.gif)

In practice you will need to tie the two knots as close to oneanother as possible.  I have left it rather long in this example in order to be able to show the layout.

Can anyone think how we could incorporate both these SG hitches into one knot?

Derek
Title: Re: Looking for a knot
Post by: Dan_Lehman on June 28, 2006, 07:21:14 PM
Can anyone think how we could incorporate both these SG hitches into one knot?
Derek
Put a Half-hitch between the knots, to hold the opposed draw-string ends together
when tensioning!?

--dl*
====
Title: Re: Looking for a knot
Post by: jhonny3 on July 10, 2006, 12:10:16 AM
Not sure if this will help...though the hem might be a bit of a problem, but you could tie two inline adjustable hitches, essentially making a water knot but with adjustable hitches.  I don't have either reference or digital camera to make heads or tales of this, but I'll try to better describe it as follows.  Same as the Water knot, but instead of overhand knots, tie adjustable hitches.  The hitches allow them to be moved from the initial point and still hold. I've found it holds very well, indeed, I've never seen it slip, the biggest problems with this solution lie in working the hitches through the hem and the rope doubling when you move the hitches.  Hope it helps, and if you still need clarification feel free to ask :D

EDIT: OK, so I missed the second page of the forums the first time, so this is veerrryyy close to DerekSmith's idea, only instead of using a second piece of rope, just use the same rope.  Now that I look at it, Derek's might be better...try em both and see how they work! ;)  Also, now looking at the original post of mine, I realised that in my tired state I said water knot...I meant fisherman's bend...um...really! ;)