Bottom Belay, an exercise in Futility?

[skianddive]

Fellow Canyoneers,

Any mathematicians or physicists out there?

An article on page 5 of this SAR publication discusses the "fireman's belay" technique and some of its limitations.

http://tracker.sbsar.org/Issues/2004/tracker_04_04_low_res.pdf

Comments/opinions/whines????

[rcwild]

Hmmm ...

I understand the premise of the article and the cautions expressed should certainly be heeded, but ...

Using a bottom belay, I've stopped dozens of rappellers who lost control. I don't know how long it took to react. I don't know how much force I applied. I pulled, they stopped.

Perhaps the most important lesson to learn from the article is that belaying should never be taken lightly. When you are belaying, pay attention. Don't just grip the rope with one hand. Wrap the rope once or twice around your forearm so it can't get away from you.

I've heard people yell, "Belay's on!" before they even got hold of the rope. Wrong! When you say, "Belay's on!", your partner believes you are saying, "I have a firm grip on and am in control of the rope, 100% of my attention is focused on you, I understand I may have your life in my hands, you can trust me."

[rcwild]

Couldn't help noticing the cover story was written by ACA member, Sonny Lawrence. Nice article, Sonny.

[sonnylawrence]

The article on a bottom belay was written by Paul Stovall and Jeff Lehman. Jeff teaches chemistry at a college in San Diego. Paul runs the mechanical engineering department at the University of California, Riverside. Both are long-standing members of the San Bernardino Sheriff Cave Rescue Team. We have Paul analyze systems as to breaking strengths, etc. We use his lab for static and dynamic testing of hardware and software. In fact, if anyone is interested in breaking stuff, let me know! For example, last year we did dynamic testing of an old Petzl energica. It held up very well. We destroyed a rope with Fall Factor 2 dynamic loading of a Tibloc. The Tibloc looked brand new after the test. The rope was shreaded.

[jwurst]

I had a look at the spreadsheet and calculations and I must note that this was done for the absolute worst case scenario. The out of control rappel is being defined as a free fall in a vacuum (not even any air resistance).

Air resistence is negligible at these relatively low speeds but the friction from the descender on the rope would be substantial. This means that the pre-reaction time fall would not be nearly as far as what is calculated. This would result in a much reduced fall speed at the time the belay is applied, resulting in a much quicker stop or less force necessary from the belayer.

Experience is rarely wrong when working in the realm of things that can be seen and touched. If the study had included friction imposed by the descender during the pre-belay fall then I think the results would have been far less surprising and much less dangerous.

Jeff

[paulsailer]

Free fall with a belay device is absolutely much slower. If you wanted a a real world test, you should rig some weights to a belay device with a trip wire to let them fall while attached to the belay device. This would be real world information.


Are there any real world reports of people being either injured or saved by belay device in action.

[miro]

A rather long comment on the "bottom belay theory":

It is easy to note that whatever model is behind the calculator, it should NOT be trusted. To convince yourself about this, just go to that webpage
and enter in the calculator the following:
0 for time to notice the fall, and 0 for force from belayer
This represents no action from the belayer and we know what result we
expect in the bottom line, namely it should be the height of the rapeller
i.e. 100ft (default value if you din't change it). Yet the calculator gives
zero as answer.
Moreover, if you keep the force zero (so still no action from belayer)
the calculator's answer depends on the reaction time! That is of course
unphysical - this theoretical belayer can stop the fall just by noticing
early that the rapeller is falling, without even touching the rope!

Looking at the equations, the source of the problem is the following
inconsistency: After the rapeller looses control, his rack doesn't
provide any friction - (as Jeff noticed before) he is in free fall.
The moment the belayer notices (even without taking any action),
the rack starts to exhibit friction even for zero applied force.
This is not "worst case scenario", it is plain inconsistent.

To summarize: If the theoretical model that underlies the calculator
can't give correct results for trivial scenarios, why would you believe
its answers at all?

Cheers!
Miro

[bj]

My thoughtfully contemplated response

If the authors of the study are experienced SAR team members how do they reconcile their practical experience with the results of this study? Our practical experience tells us that the fireman's belay is a safe practice. However, based on the results of this study, someone who had not used the fireman's belay might consider the practice is downright reckless.

It seems to be that the authors of the study have missed some crucial factor between the theoretical and actual application. It troubles me that this article was published without further discussion of whether this is been a safe practice in the field.

My First response

What a theoretical hunk of crap. Have these guys ever been in the field or are they just wacky engineers that sit behind a computer crunching numbers all day. Now every numb nut, newbie is going question the safety of a firemans belay. How does crap like this get published?

[paulsailer]

The article suggests through its mathematical model-- what we already know--and what common sense and a little experience with falling objects would tell us, that the longer the bottom belayer waits to begin to pull down on the rope, the faster the person on rappel will fall, the more force there will be the belayer and rapeller, and the greater distance the fall.

What would be useful to know would be the real figures for speed, distances and forces in real situations, and at what point it will be too late to stop the fall at various distances.

A real world tests with real ropes, different belay devices, (and weihts instead of a person falling) would provide real world answers.

A mathematical model with some coeficient of friction on the rope for the belay device would be more accurate than one which assumes 0 friction.

We know there is substantial friction from a belay device

A good example of a real world type of test is the bucket drop conducted by the SCMA to introduce novices to forces when belaying a falling climber. With a 100 lb weight releasted via a trip wire falling 20 feet before the belay takes up the slack there is quite a jolt on the belayer, but much much less than there would be without a belay device.

[jwurst]

I wouldn't be that hard on the authors of the article. The experiment was surely done on their own time and just for fun and published just in case anyone was interested. From the sound of the first paragraph, Paul Lehman did the calculation just for fun. If someone was paying it would surely have been much more thorough. He showed it to his friend, Jeff Lehman, who thought it was interesting and asked if he could put it in their search and rescue group's monthly newsletter.

To me the intent appears to be, like Rich said, a reminder to stay vigilant while belaying. The article says: "Does this mean that a bottom belay is not effective, and shouldn’t be used? Of course not."

Its all just food for thought. And it has gotten some people thinking. Anyway, its almost a year old now so not worth much fretting.

Jeff

[jeffbell]

I would suggest that one of the most crucial issues is for the fireman belayer to position himself so that he is able to quickly take up both the slack and stretch of the rope. The rappeller needs slack to maintain control himself. A rope normally stretches – I've heard numbers like 5% or so. This can add up to several feet very quickly – especially on a long rap. I remember Bo Beck (Zion SAR) telling me of a situation of a rappeller losing control on a 150' or so free hanging rap. Bo did not have time to take up the rope slack and stretch by pulling hand over hand. He was luckily standing where he could jumb off a 5' or 6' perch to a landing below. He spun around and wrapped the rope around himself as he jumped. He said had he not done so, the rapper would have been history.

Another crucial issue is making sure that the rappeller does not slide down the rope onto the fireman belayer or others. Self safety first; safety of others second; safety of the rappeller third.

Jeff

[sonnylawrence]

Okay, I am the messenger. Please don't shoot me! Here's food for thought.

I am told there was video that was presented by Tim Kovach at ITRS (International Technical Rescue Symposium) two years ago. A guy took a 150lb weight and dropped it and had somebody at the bottom try to stop it. I think these were drops of over 200 ft. The belayer (big burly firefighter) could not stop the weight unless he tensioned the rope prior to the subject rappelling. It was VERY eye opening to watch that weight tank every time.

If there is enough interest yet in this subject, I can try to contact Tim Kovach on the search and rescue egroup and get more info.

[sonnylawrence]

Paul Stovall asked me to post a response to the bottom belay article. He was the primary author.
Sonny


I just wanted to clarify a few things about the article on bottom belay that was published in the April 2004 Tracker and brought to this forum by Paul Sailer. This article was originally attended for fellow search and rescue personnel and fellow cavers that wanted to instill the importance of being completely observant of the person they are belaying. The 4-Bar BMS Micro-Rack and 7/16� PMI easy-bend ropes are the most common among these groups and are the only ones considered.

Let’s start with the assumptions to the statement problem.
1. Type of rope: 7/16� PMI easy-bend ropes, less than two years old with minimal usage and a clean sheaf.
2. Descender: 4-Bar BMS Micro-Rack with out the use of the hyperbar.
3. Minimum response time of the belay person: ½ second.
4. Belay force applied directly below the person rappelling.
5. Person is unconscious.
6. Rope is homogeneous throughout its length.
7. Force from air drag was eliminated do to its small relative force, after a 22’ free fall of the average sized adult it only contributes to about ten pounds of resistive force.
8. Length of rope 150’+: This is important because on short rappels the impulse that the belay person imposes on the rope contributes to most of the rappels deceleration and on long rappels it is adsorbed by the ropes elasticity and therefore your decent is more gradual.

The calculations from the spread sheet ONLY factor in the prior assumptions. Any other rappel device or rope will give different results.

Now let’s clarify Miro’s response on the ACA forum to using zero time in the “Time before noticing the fall:� cell. The only time that you would use zero for the time variable is in a statics application and this is clearly a dynamics application and requires some unit of time and therefore zero is an invalid parameter. Besides the fastest response time the average human is around ¾ of a second.

As for modeling for the worst case scenario that’s generally what you want to do when you are trying to emphasize the importance of being on your toes at all times. Research shows (Kovach et. al. ITRS 2003) that on long rappels (200+ft) that the bottom belayer has a difficult time determining when a rappeller goes out of control and how far up he/she is. Being extra vigilant helps mitigate this to a degree. I know I feel a lot better if the person belaying me is a little paranoid about staying alert.

As for the friction between the mini-rack and the rope; it is incorporated in the calculations both during the fall and during the entire time the belay person is applying a load to the rope. These numbers were obtained on an Instron tensile testing machine at a speed of 10 inches a minute. During a fall without a belay the testing showed approximately 55_lb of resistive force. It would be wonderful if I had the facilities to check this in its dynamic conditions so I could look at things like the force accrued from friction impulse, the amount of energy transferred into heat do to the rapid bending of the rope and relation between kinetic friction and velocity “this may change depending on the ropes dynamic behavior around a bend at high speeds�, but unfortunately I don’t. Anyone interested in donating a drop tower?


The assumptions in a typical cayoneering scenario will be different, to a degree. Either way the work presented in this spreadsheet is a good starting point for a discussion.

A couple of years ago Tim Kovach at the International Technical Rescue Symposium presented work where he performed some tests using a break bar rack on long rappels. The rack was loaded with the number of bars that experienced rope technicians would use on a drop of approximately 300 ft. Off the top of my head I can’t remember the progression of the number of bars, but it was a standard rack without the “hyper bar�. I think he may have started with 4 bars. He used a 150 lb weight with the end of the rope on a redirect at the bottom of the drop. The redirect was there to keep the bottom belayer out of the fall zone. The weight attached to the rack was released with a solenoid so that the loading was consistent between drops. This entire exercise was video taped. The bottom belayer was a big, burly firefighter, who would begin tensioning the rope when the signal that the load was released was given. It was amazing to watch that load smash into the ground time after time. The bottom belayer spent most of the time removing stretch in the rope, so that he couldn’t get enough tension in time to stop the load. The rope was low-stretch (static) 11.1 mm rope. Not until all bars were in the rack could he get the load to stop. I know I weigh more than 150 lb (more with my pack), and I can’t start a 300 ft drop with 6 bars engaged, so this test was rather eye opening. That belayer was primed and ready on solid ground with good footing. He wasn’t awash in a pool or stemming over some obstacle; he could bring his entire weight and strength to pulling on the rope.

As Jeff Wurst quoted in his response on the ACA forum: “the intent appears to be, like Rich said, a reminder to stay vigilant while belaying. The article says: "Does this mean that a bottom belay is not effective, and shouldn’t be used? Of course not."� THANKS for your input Jeff!

Cheers, Paul

PS. Please feel free to pick this apart. It may give us all something to think about.
.

[jeffbell]

I quote from Sonny and Paul in the previous posting:

"The bottom belayer was a big, burly firefighter, who would begin tensioning the rope when the signal that the load was released was given. It was amazing to watch that load smash into the ground time after time. The bottom belayer spent most of the time removing stretch in the rope, so that he couldn't get enough tension in time to stop the load. The rope was low-stretch (static) 11.1 mm rope."

This does not even take into account the slack in rope needed by a rappeller to properly manage his own decent. Furthermore, the firefighter was given the signal when the load was released. He wasn't required to determine it on his own while watching with the sun in his eyes or bullshitting with someone.

This just re-emphasizes the point I made three postings back that all the theory in the world doesn't mean squat if the bottom belayer is not in a position to quickly take up the rope slack and stretch. This may well mean that he needs to be in a position to run like hell or jump off a perch of some kind. But what ever the bottom belayer does, he needs to stay out of the fall zone.

Jeff

[jwurst]

Paul,



As for the friction between the mini-rack and the rope; it is incorporated in the calculations both during the fall and during the entire time the belay person is applying a load to the rope.

That may have been the intent, but the spreadsheet isn't doing that. The formula for cell C13, the Distance of fall during elapsed time, is using 32.2 for the acceleration. This is value for acceleration due to gravity alone when measured in ft. To include friction you would need to use acceleration = 32.2 - (1.465*Fb + 91.783)/(rappeller mass). In this case the belayer has not begun to pull so Fb is simply the weight of the rope below. Trying this out I get the pre-belay fall distance of just over half of your value for a 200 lb rappeller, substantially less than half for a lighter load and getting closer to yours as the weight goes up towards infinity.

I did my own calculations using your force of resistance. The results are in a second sheet of a copy of the same spreadsheet now located at http://myweb.cableone.net/wuorst/bottom_belay.xls. My free fall times are substantially less even though I didn't account for rope weight during free fall. My stop times can be substantially more! I think the reason is due to an error: Your resistance force is 1.465*Fb + 91.783. The total force on the rappeller is Ft = Fg - Fr (Ft = total force, Fg = force due to gravity). In order for the rappeller to slow down Fr must be larger than Fg. The amount of pull force (Fb) necessary for that depends on the rappeller's weight but for a 200 lb person that would be about 74 lbs. Anything less will allow the rappeller to continue to fall faster. 74 lbs would keep the rappeller at a constant speed. Anything more would begin to slow them down. I believe this is a major inaccuracy. I use this same BMS micro rack for my caving endevours on a variety of ropes, including BWII, PMI Pitrope, and PMI ez-bend. I weight about 195 and normally use the hyperbar but even without I find a have plenty of friction. If I pull the bars all the way apart I have much less friction, but I still don't think it takes 74 lbs of force to simply hold my weight. Anyway, the bars would have to be held apart because they naturally ride up.

If anybody wants to see my calculations I can post them. I won't bother with it now because I suspect most people are already rolling their eyes at this.

Jeff

P.S. I think we all agree that a belayer stand in a good place and maintain vigilance regardless of how effective the belay may be. Having established that, this is a physics thread to statisfy a curiosity rather than to establish correct safety technique.

[sonnylawrence]

Here is the link to the bottom belay article referenced earlier in these posts.

http://tracker.sbsar.org/Kovach_Bottom_belay.pdf

[sonnylawrence]

Here is the summary from the bottom belay article in the previous post's link:

"In summary this testing suggests some points to consider for an out of control
rappeller.
1. The length of the drop matters. The further the rappeller is from the bottom
belayer, the less likely the bottom belayer is to notice an out of control
rappeller.
2. The greater the rope length between the bottom belayer and the rappeller,
the less effective the belay effort may be.
3. The slower the reaction time of the bottom belayer, the less likely he or she
is to be successful in his or her belay effort.
4. The gripping ability of the belayer may be more important than the overall
strength of the belayer.
5. The size of the belayer may be misleading."

I am curious if anyone disagrees with the above points. Are these conclusions supported by the way Jim Kovach did his research?

Are there messages here for new canyoneers as to the diameter, amount of stretch and length of rope they should be using while doing bottom belays? So for example, perhaps a new canyoneer should not do a bottom belay on anything less than a low stretch, 200 foot (or shorter), 11 mm rope until he/she is competent at stopping a fall. But then, how does such a canyoneer get the experience needed to feel that competence?

[koentje]

The article on a bottom belay was written by Paul Stovall and Jeff Lehman. Jeff teaches chemistry at a college in San Diego. Paul runs the mechanical engineering department at the University of California, Riverside. Both are long-standing members of the San Bernardino Sheriff Cave Rescue Team. We have Paul analyze systems as to breaking strengths, etc. We use his lab for static and dynamic testing of hardware and software. In fact, if anyone is interested in breaking stuff, let me know! For example, last year we did dynamic testing of an old Petzl energica. It held up very well. We destroyed a rope with Fall Factor 2 dynamic loading of a Tibloc. The Tibloc looked brand new after the test. The rope was shreaded.

I think the biggest "problem" was created by using a rack as the descender of choice for these tests. Pulling down on a rope doesn't create a whole lot of extra drag on a rack, pulling up does.
I bet that with a descender which brakes downwards the tests have a somewhat different conclusion.

On static ropes and with a figure of 8 bottom belays do work. A anecdote: I was guiding a group of beginners and at the start of the canyon I gave them a full briefing. I also mentioned that they could be confident on rappel because they were always belayed - either from above or from below.

During the first rappel (70 ft) one of the participants starts to go down, looking good, obviously confident. About halfway down the rappel the guy suddenly takes a jump off the wall and completely lets go with both hands, no warning whatsoever ?!?
The bottom belayer was completely taken by surprise but managed to stop him without any problems.
The guy's explanation was: " I was wondering if you were bullshitting about belaying us or not, I decided to check it out"...

Personnally I don't want anybody belaying me from below anymore, someone nearly broke my back "fooling around": I was going real fast down a vertical 90 ft waterfall when a guy down grabbed the rope to let me have an extra long shower about 2/3 down. It didn't slow me (I'm a big and heavy guy), it just stopped me dead in my tracks, nearly breaking my back from the shock.

Bottom belays do work: but it depends on the kind of rope (and single or double), the length of the rappel and the descender used.

Sonny, I might take you up on your kind offer to test some things to destruction :-).
I've been looking for tests on the use of "half-a-double-fishermans-knot" to attach/fix a carabiner at the end of a cowstail. To my knowledge such tests have never been done and I know a lot of people looking for answers on this one (including me). A lot of cavers actually.
Let me know if you're interested, we could work out a "test scheme" ?

Koen

[sonnylawrence]

Yes, speaking for Paul Stovall, we would be delighted to test knots, etc. Please formalize what you want to test and I will arrange it with Paul. Thanks for your thoughtful response to these posts. This looks to be fun and enlightening!

Jeff Lehman and Paul Stovall have been having discussions related to Koen's suggestion to test a cow's tail. Does it matter if it is made out of low or high stretch rope? Would 9mm high stretch be equal to 11 mm low stretch, etc.?

As to the half double fisherman's knot on a carabiner. That is often used in creating an "edge kit" for a rescuer to have a safety line while working around the edge of a big drop.

[jeffbell]

I tend to disagree with #2, but agree with #5 in Sonny's list. I'm 160 lbs without pack, rappelling gear, and additional shitola. I use a pirana and figure 8. It has been my experience – not that I've had an inordinate amount – that in the case of free hanging raps between around 150' and 300' the rope is heavy enough that I am constantly fighting and lifting it up to get going. This would suggest that if something happened to someone above 200' or so, he might actually get stuck there and present the opposite problem.

Somewhere in the range of 150' to 100' feet I can finally get going and need to begin applying some breaking. I tend to let myself reach a reasonable speed, and then break to near zero and start again. This way I'm not "riding the break" and getting my break hand (and glove) over heated. I haven't found it particularly difficult to stop myself. However, I try to be well prepared to do so with the proper set up, including a leg loop.

As for Sonny's #3, I would suggest that "success" includes not just stopping the rapperller, but doing it in a controlled way so as to not cause what Keon experienced: "…it just stopped me dead in my tracks, nearly breaking my back from the shock."

As for Sonny's #4, I would suggest that gripping is likely to be less of a problem than being able to take up the slack and stretch in the rope fast enough.

Jeff

[sonnylawrence]

Jeff Bell, do you think there are significant real life concerns as to the diameter of the rope and low vs. high stretch or are these just theoretical concerns with little real world impact?

[mrklusman]

I've enjoyed reading these posts 'cause there is some good critical thinking taking place.

I agree with the comments about descender types (or more importantly - how direction of pull impacts braking). The effectiveness of a bottom belay seems to be impacted quite a bit based on my unscientific observations.

Those of us performing in a leadership role need to be sure we understand the limitations of the bottom belay given a particular descender type.

Finally - it reinforces the notion that belaying is serious business and, because you are holding someone's life (potentially) in your hands, it is NOT the time to be BS'ing with the pretty girl next to you, digging your lunch out of your pack, staring of into space, etc. I can't tell you how many times I see belayers at various climbing spots not paying attention.

Rob

[koentje]

Yes, speaking for Paul Stovall, we would be delighted to test knots, etc. Please formalize what you want to test and I will arrange it with Paul. Thanks for your thoughtful response to these posts. This looks to be fun and enlightening!

Jeff Lehman and Paul Stovall have been having discussions related to Koen's suggestion to test a cow's tail. Does it matter if it is made out of low or high stretch rope? Would 9mm high stretch be equal to 11 mm low stretch, etc.?

As to the half double fisherman's knot on a carabiner. That is often used in creating an "edge kit" for a rescuer to have a safety line while working around the edge of a big drop.

Thanks a LOT ! I'll give it some thought, ask around and let you know what might be a good set of tests - we can take it from there.

Koen

[jeffbell]

Response to Sonny's question: " …do you think there are significant real life concerns as to the diameter of the rope and low vs. high stretch or are these just theoretical concerns with little real world impact?"

These are ABSOLUTELY practical issues, not necessarily theoretical ones.

(1) The diameter of the rope makes a significant difference. Try rappelling on an 11mm rope and then on a 9mm rope and you will notice the difference. The larger the diameter, the greater the breaking capacity. The difference is not as large as the difference between a double strand and a single strand. But the principle is the same. Beginners should be made aware of this.

Furthermore, it is also much easier for the bottom belayer to get a good grip on a larger rope.

(2) Next time you do a long rap, try to mark or watch the spot at eye level and see how much it bounces up when you get off rappel. I've heard that some ropes can stretch as much as 7%. If that were true – and I'm not stating that as a fact – then a 200' rap could have as much as 14 feet of stretch. But that is only part of the story. There is the extra slack that needs to be taken up. The rappeller needs slack in the rope to properly manage his descent. It the rope is too tight, he can't maneuver properly. I can't put a figure on that, but my feeling is that it can be relatively significant, especially when combined with the stretch.

Ask yourself how long it would take to retrieve 15' to 20' feet of slack and stretch before you could start breaking a rappeller's fall. Once you did it, would you be in such a panic that you would be likely to slam the breaks on so hard you might do serious injury to the rappeller?

We know that bottom belaying can and does work. Incidents have been reported. It has also been reported that taking up the slack and stretch has caused major problems. I would therefore suggest that tests, measurements, and/or studies be done to get a handle on how much rope a bottom belayer most likely must retrieve, and how to best achieve that. In other words, how much stretch and slack is typical, and how should the bottom belayer retrieve it in order to manage a controlled descent.

Jeff

[paulsailer]

It occurs to me that people doing bottom belays should practice to learn from experience what they need to do to make it work.

One person could be rapelling and the other doing a bottom belay. Periodically the person doing the bottom belay would try to stop the rapell. Granted this would not simulate a fall. But it would teach the belayer how much tension
would be needed to stop the rapell. And how much time it would take the belayer to apply that tension from various position, for example at nearly right under the rapeller or as is sometimes necessary when the rapeller is rapelling over a pool --at a signifciant distance from the rapeller, and at different angles etc.

This way a person doing the belay would not be actually doing it for the first time if he or she needed to actually stop a rapeller from falling.

[paulsailer]

I suppose another practice could be doing the bottom belay with a weight instead of a person on the rapell device, in which case the weight could be set up from a tree limb or an overhang and the bottom belayer could see what would be needed after falls of different durations.

[sonnylawrence]

I have been on the giving end and receiving end many times with a bottom belays. The last experience was when I was on the guided rappel through the waterfall at Cebeque Canyon. Our guide decided I needed to stop mid-waterfall and experience the rappel to the fullest. I am suspicious my wife put him up to it.

It takes very little force to stop someone who is rappeling at normal speed. That's easy. It would be safe to do scenarios so everyone could experience that.

What I have not done is catch someone who is totally out of control, screaming down the rope. Discussions like this thread have helped keep me conscious of the importance of technique and physical limitations, such as being able to determine a person 300 feet above me is in fact moving fast, out of control. It would be difficult to safely try this scenario in a way that is relevant to canyoneering.

In the past I did other belay practices that were quite instructive. A group of us used a firefighter training tower to learn the standard belay for a lead rock climber. I was anchored to a wall on the fourth floor. A hundred pound weight was tied to the rope and dropped off the 5th story. I tried body belay, Stitch link and Figure 8. Sometimes I knew when the "person" was falling. Sometimes I had no warning. I did some of the scenarios with eyes closed so I could not see the weight flying past the window. When it hit, I was lifted off the ground. Needless to say, the body belay was my least favorite. I learned many lessons, one of which is to have the belay come off the anchor instead of me. The exception is if the anchor is not bomb proof. In that case the entire system needs to be as dynamic as possible to absorb the shock.

[paulsailer]

I also did a practice belays using a falling weight, and the first time, I wasn't in line with the rope and was jerked into a rock wall, Its a great way to learn how to do it right

I think practice with weights would be the best way to practice stopping a falling rapeller with a firmans belay

No matter what speed the weight fell, it would not be the speed or force of a person or weight in a free fall situation, so training using weights would make sense, it would also make sense to have another rope to stop the weight
so it did not hit the ground