Draft buffers

[Andy Carson]

Tim,
Were the average peaks taken from the graph shown or from a different set of data? On the graph shown, I see a max peak force of ~775 with the chain and ~725 max with the rope… Also, I calculated an average peak force of ~670 with the chain (775+750+725+525+575+675/6) and ~580 with the nylon rope (475, 350, 575, 650, 700, 725/6). There are some rounding errors here, I’m sure, because I am estimating numbers off the graph, but these numbers give about a 13% difference in peak force… Maybe this matters, maybe not… More importantly, I notice a difference in the duration of the forces needed to start a load with a chain versus a nylon rope before returning to the low forces needed to maintain forward momentum. In the data set shown, the oxen with the chain had to maintain a pull of over 400 lbs for an average of 0.6 seconds ((4+4+4+2+2+2/6)*0.2) while the oxen with a rope had to maintain a pull of over 400 lbs for only 0.3 seconds ((2+0+2+2+1+2/6)*0.2). As power expenditure is best estimated by the area under the curve, this is a pretty big differance! Even if the peak force was similar, maintaining that heavy starting pull for only half the time cuts the power required by half. Although I can’t make heads or tails of the maintainence draft, I think this data does show a beneficial effect of the nylon rope on starting draft.

[Tim Harrigan]

The peaks were from this data. You will notice that I changed the symbols now in the graph, the circles were obscuring some of the diamonds in at least one instance. I did report incorrectly though, average starting peak was 614 for the nylon and 664 for the chain. I will email you the data file so you can look it over a little more closely. The average draft was from a 1500 ft pull with each towing device, not from this graph. Give the variability in the numbers, only six starts, and the inability to start the team exactly the same in each case I might need a little more convincing than what I see here.

[Andy Carson]

Thanks for the raw data Tim,
I focused in on the power expenditude over the first 0.8 seconds for each start (most very high draft points are in the first 0.6 seconds), so four data points for each start. To calculate total energy expenditure during the start, I added up all the draft forces for these data points and divided by the time interval (0.8 sec). The nylon rope was very consistant in the force required to start the load at 1645, 1605, 1570, 1845, 1795, and 1805 lbf/sec (average = 1711 +/- 118). The chain was a lot more variable with 3170, 3095, 3150, 1555, 1645, and 1930 lbf/sec (average = 2424 +/- 792). This represents about a 30% decrease in power required to start the sled with the rope. The p value (unpaired two tailed t test) between these groups is 0.05, which is statistacally significant with 95% confidence (although just barely). An unpaired two tailed T-test is a pretty strong test in biological systems (at least in my hands), so I am convinced this reduction in power required to start is a real phenomenon.

[Tim Harrigan]

I enjoy your enthusiasm for the analysis, and there probably is a small buffering effect. I would be more comfortable with the statistical component with more control over the dependant variables. I would like to see more contribution to the 1500 ft pull to help get over the ‘practical benefit and worth investing in’ test.

[Andy Carson]

Yes, controlling this system is tough. One might be tempted to simply repeat the starts many times, but I would speculate the animals learn quickly how to most efficiently start and move this particular test load and starting and stopping over and over again might not represent a “real world” situation. If you look closely at the data, you can see the first 3 starts with the chain were horribly innefficient using almost twice the power of any other starts. The last 3 starts with the chain were much more efficient, and I would speculate the oxen learned exactly how much force was required to start this load and exactly how fast to apply it. This lesson may or may not apply to the next load they start and it is highly possible that thier timing or force will be off if the load is increased or decrease by 25%, leading to several more inefficient starts. With the nylon rope, there was much less variability and they seemed to be learning over time that they could start with a heavy push and the energy wasn’t wasted. Every start with the rope had a higher peak force, although they were all of very short duration. This is similar to what my horse learned with the buffers. It is amazing the physical techniques animals can learn without really being taught, but they complicate data analysis… Tim, in the data set you sent me, I can see hints of harmonic oscillations in the nylon rope (especially between starts 1 and 2) with a time interval of between 0.8 and 1.2 seconds. Did you see harmonic oscillations in the 1500 ft pull? That seemed to be a limitation in some of the buffer set-ups…

[Tim Harrigan]

Yes, I agree that the animals will adapt to the load and I was concerned about that when I was doing the tests and that was why I only did six starts. I think I mentioned this ability to adapt in regards to the front-to-back load shifts that I posted earlier where the draft increased at different hitch angles when the load shifted from back to front when pulled with a tractor, but there was no change in draft when Will and Abe pulled the load and the load shifted back to front.

There are trends in the starting forces, I checked and I started with the chain and then went to the nylon rope so there could have been some adaptive behavior that makes me suspicious of jumping to conclusions in interpreting what is really going on.

I do not doubt that the draft buffer concept is valid and I appreciate your close inspection of the data but I have been trained to be skeptical. I have 4 nylon traces but I do not use them. If you want to try a set I can send them to you for testing.

Here is a picture of the nylon rope. The dense rubber insert is compressed when the rope is under tension.

[Andy Carson]

Taking into account the animals ability to adapt, I am not sure I believe the buffering effect either… I wonder if the best way to test starts would be to vary the load between starts enough that the animals really don’t know what to expect. Tim, I wonder if you might remember how long it would have taken to transfer between the chain and the tow rope. I wonder if the oxen would remember and expect the same load if the changeover took several minutes. I am pretty sure if I walked my horse away from the sled to change over and then brought her back, she wouldn’t expect the same load. On the other hand, if I left her to stand in place while I switched the traces, she would likely expect the same load (especially if I was quick about it). If the changeover took a while and the oxen really didn’t know what to expect, the most fair comparisons would be the first few pulls. This is where the differance in power is huge, with the rope reducing the total power required to start by nearly half.

[Tim Harrigan]

In my test I started with the chain, made 6 starts with short pulls and then made a steady pull for about 1600 ft. Then I swapped the chain for the nylon rope and did the same thing. So I would say the team knew what to expect. There are a lot of variables associated with starts and because they are of such short duration I am not sure we will sort it out, particularly when the differences are small such as I think we see here. It is not easy to measure physical systems that are under biological control. I still think one would need additional sensors to measure pressure points etc. and probably still be left with more questions than answers.

For our purposes, at some point it has to pass the test of practicality. Does a buffer clearly make a difference, and where and when? A start is a sort of shock load just like a rock or other stationary object, but the team has some control over a start. Even if you vary the load they still assess the load the instant they begin to step into it. I have seen my team very quickly make a distinction between a heavy load and a fixed object. And they did not have to slam into the load to do it. It is actually quite impressive what they know about a load.

This is a reason I liked the frequency of pulling force charts that I showed earlier for the wagons. The rubber-tired wagon passed both the test of statistical significance and the test of a pretty clear practical difference. You can have a little more confidence in the assessment when you are looking at 5000 measurements over 1/4 mile or so rather than 5 measurements over 1 second. The starts could be done, I just think it might be beyond the scope of what we can do these limited measurements.

[Andy Carson]

Tim,
Where did you get that rope from and do you have an additional information about it? I have been reading a little about rope and have found it fascinating how many difference ways rope can be put together and how many different things can be accomplished by this. Particularly important here is that some types of rope are designed to absorb shock and not return that energy. This is particularly important for climbing rope when if the energy was returned a fallen climber would bounce up and down the mountain side as if on a bungee cord. Ouch! Apparently, this property is sometimes designed into tow ropes as well, so that a jerky pull is smoothed without slingshotting the towed car into the tow vehichle. It seems that pure rubber has nearly the same resilience as spring steel and would likely return most of the energy. Pure synthetic ropes are more likely to absorb and dissipate the energy rather that return it. They are also subject to a phenomon called “creep” where the rope under constant tension slowly elongates without energy storage or return. The rope you showed is somewhat of a hybrid between the two and I am not sure if it will act more as a “rope,” a “bungee,” or something inbetween. Do you have a better idea?

[Tim Harrigan]

http://www.fanac.ch/zep.htm

http://www.fanac.ch/zep-f-.htm

Here is and English translation,
http://translate.googleusercontent.com/translate_c?hl=en&ie=UTF-8&sl=de&tl=en&u=http://www.fanac.ch/zep.htm&prev=_t&rurl=translate.google.com&usg=ALkJrhhrCF–Akti8qqNHTVbnr_J-HWjog

[Andy Carson]

Awesome site. It is great to see all the things we have been learning laid out in roughly translated german. Amazing! Apparently, both my spring buffer and the rope buffer are not new in concept, but are simply modern versions of the 1880 “horse-saver” that improved draft efficiency by about 14%. One of the pictures looks, uh, pretty familiar… Very interesting read.

[Roscoe]

@Tim Harrigan 18251 wrote:

Here is a picture of the nylon rope. The dense rubber insert is compressed when the rope is under tension.

Where did you get that rope from? I exactly used this ones in Switzerland and i was very satisfied (working & pulling). Sadly, I sold most af the tack before I moved to Canada. Now I’m wondering where I can get them on this side of the pond.

[Tim Harrigan]

I am not sure if there is a source in Canada or the US.

Claude Fanac
Schlosserei/Schmiede & Apparatebau
8588 Zihlschlacht
Switzerland

Tel 071/422 4745

[Andy Carson]

Carl,
I’m curious if you have had a chance to try out the buffer yet. Do you have any thoughts you might share? Don’t worry about hurting my feelings if you don’t see an effect. I feel like I do, but I wanted to hand it off to someone else to see if they see the same thing…

[jac]

Andy the front springs on my wagon we use in parades are made of… I dont know what ?? looks like glass fibre.. and are a lot lighter. Would something like that be any use for an alternative to steel ??
John

[Author]

Posts