PDA

View Full Version : Laser Pool Measurement Device



Tspence
October 12th, 2005, 09:03 PM
Can anyone direct me to a company to markets a low cost laser measuring device that I could use to measure swim pools quickly and accurately?

Tom Spence
Chairman Lake Erie LMSC

Frank Thompson
October 14th, 2005, 11:24 AM
Hi Tom:

Our Michigan LMSC just bought one about 2 months ago for $350.00 from Home Depo. I assisted Milton French, our top ten chair and we measured the Dunworth pool in Dearborn which is a 50 meter pool and found that the measurement was easier and faster then using a steel tape. I will send you the information on what we purchased and hope this helps you.

michaelmoore
October 15th, 2005, 12:34 PM
Skip:

Can you send me information about it also.


michael

Frank Thompson
October 28th, 2005, 12:42 PM
To Tom Spence, Michael Moore, Dennis Tesch, and Mike Heather:

The laser that we purchased was the Leica Disto Lite 5. It was $375.00 when our LMSC bought it about 3 months ago. The price is now $329.00 and that does not include the target plate which is $20.00. There is another model called the Leica Disto Special Lite 5 and its now $399.00, down from $449.00 about 3 months ago. It has some more options. Extended Manufcturer's warranty for 5 years is available for $37.50

All of this information is available at www.lascolaser.com and hit the link products and you can search there and see where you can purchase one in your local area or just order online. We have used it once so far and its fast and efficient. It will become real handy if you have to take mutiple measurements before and after your days competion.

jroddin
May 18th, 2010, 11:38 AM
This thread has been inactive for five years and I'd like to "bump" it to current...

I would like to look into having my LMSC purchase a laser measurement device for the purpose of measuring/certifying pools (we currently use a tape measure and our officials have been getting on me to get something easier for them). However, I want to make sure we buy one that meets requirements for USMS. I found the following text in the LMSC Guide to Operations:


“… you need to be aware of the tolerance levels of your equipment over the distances required. It must be accurate to 1/100th of a meter (1 centimeter). To achieve this accuracy, the markings on the tape must be to the millimeter, thousandth of a foot, or eighth of an inch.”

To reiterate the last sentence, this tells me in order to be accurate enough the device must have markings to at least one millimeter, or 1/1000 foot, or 1/8 inch. To compare the three of these equally, this is what I convert them to be:
1 mm = 0.039 inches
1/1000 foot = 0.012 inches
1/8 inch = 0.125 inches

That is quite a spread, with the best one being 10 times more precise than the worst one. The guideline uses the word “or” so I take it this means as long as the measurement device meets the most lax standard (i.e. 1/8 inch) I will be ok? If that is correct, might I suggest to the powers that be the wording be changed to say the device must meet a minimum tolerance accuracy of 1/8 inch or 3mm (since 1/8 inch is equal to 3.175mm). That way if you buy one that measures all the way to 1mm you are still fine. But if the device can only measure to 2mm, it is unclear if we can use it or not.

Thanks,
Jeff Roddin
Potomac Valley Top Ten Recorder

chowmi
May 18th, 2010, 07:43 PM
North Texas LMSC has a Leica Disto A5 model. It measures in both metric and English. I used the instructions included with whatever the latest printout/updated version of the pool measurement form and instructions were available. You can find it under Administration/Guide to Operations/Top Ten/Pool Measurement. I recall also having a draft sheet with longhand calculations for the adjustment with touchpads.

The price was $449 free shipping no target plate but you did get safety glasses! Purchase date 01/18/08.

lascolaser.com

tjrpatt
May 19th, 2010, 10:08 PM
I just use a steel tape, stick some noodles for floating reasons, get a helper and measure the pool. It doesn't take it long to do it.

How would you measure the laser tape without it getting wet? Just curious. Someone should post a video on measuring the lasers correctly.

Rykno
May 20th, 2010, 02:20 AM
To reiterate the last sentence, this tells me in order to be accurate enough the device must have markings to at least one millimeter, or 1/1000 foot, or 1/8 inch. To compare the three of these equally, this is what I convert them to be:
1 mm = 0.039 inches
1/1000 foot = 0.012 inches
1/8 inch = 0.125 inches



I am not sure if someone has had time to update the measurement procedures since two days ago, but the text in the procedure shows a lower accuracy +/- 0.005m for the device.

jroddin
May 26th, 2010, 11:00 AM
I just use a steel tape, stick some noodles for floating reasons, get a helper and measure the pool. It doesn't take it long to do it.

How would you measure the laser tape without it getting wet? Just curious. Someone should post a video on measuring the lasers correctly.

I have a steel tape measure and it's simply not practical for 50m. Sure, we've used it to certify pools in PV but I feel like the uncertainty in our measurements really makes for a poor measurement. And because it's nearly impossible to make the tape taut over 50m, it's almost impossible for a measurement to come up short and you're really just doing a fire drill to cover your behind. I feel like a laser measurement is a much fairer method to use (assuming the device tolerance is acceptable) and that's why I'd like to see the LMSC purchase one.

I don't understand your question about measuring a laser tape and measuring the lasers in your second paragraph. Can you please clarify?

jroddin
May 26th, 2010, 11:45 AM
North Texas LMSC has a Leica Disto A5 model. It measures in both metric and English. I used the instructions included with whatever the latest printout/updated version of the pool measurement form and instructions were available. You can find it under Administration/Guide to Operations/Top Ten/Pool Measurement. I recall also having a draft sheet with longhand calculations for the adjustment with touchpads.

The price was $449 free shipping no target plate but you did get safety glasses! Purchase date 01/18/08.

lascolaser.com

Thanks for the link. I called them and learned they now make a waterproof laser and it is only $219 (target plate is an additional $33). The accuracy is +/- 1.5mm and can measure up to 70m. The model number is Disto D210XT:
http://www.lascolaser.com/product/Leica-DISTO-D210XT-Laser-Distance-Meter-Now-available_Qitem776110-2.html

Battery power seems weak unless it's a typo on their website, but I suppose you should just plan to use rechargeable batteries and charge them up before each meet (post edit: seller confirmed it is a typo and should say 5,000 readings on a battery set and not 5).

The more expensive versions measure greater distances - but this one seems to be a perfect fit because we don't need more than 50m.

Jeff

MPohlmann
May 26th, 2010, 01:44 PM
I just received this message from one of our members who is going around the LMSC measuring pools:

I did end up buying a Bosch DLR165K Laser Distance Measuring Device for $149 plus tax at Lowe's on 4-30-10. I also bought red glasses online, which help a little in bright sunlight to see the red dot the laser makes on the target. I made a target to fit snuggly over the end of a pool without losing or gaining any distance. It is two vertical 2" by 4"s that extend below a piece of attached plywood (painted white) so that when the 2" by 4"s rest against the end of the pool in the water, the plywood is even with the end of the pool, but just above the gutter/water surface. Hopefully it can work with most gutters, otherwise another target will have to be used....maybe a touchpad.

I have successfully tested the laser to just over 50 meters here at home...something like 50.7 meters, which is this unit's max. If you are over that, your pool is too long! The greatest challenge may be to find the window of time when you can use this laser without the risk of anyone looking into it, especially if it is used at a USA meet where children are present. It might be a three person operation, one to position the target at one end (don't look at the laser), one to chase away curious onlookers, and one to position the laser at the other end. The laser device can accept the threads on the end of a small adjustable tri-pod that I already owned for my camera, so it can sit in a very stable position. The device is very easy to use and measures in English or metric within the tolerances prescribed by USMS. It measures to 165 ft and its accuracy is +/- 1/16". It includes 4 AAA batteries.

John Rempert

Rykno
May 27th, 2010, 12:44 AM
one tip that might already have been mentioned.

make sure that your laser is not reflecting off the water. when placing the device on the wall, lower in a way so that you can see the dot over the target coming down. if you just place it out there you can't really tell if the laser is going in a straight line or if it is reflecting off the water.

this could easily happen if the wall is not perfect, you measure too close to the waters surface.

Conniekat8
June 22nd, 2010, 03:07 AM
Oh nooooo!
*kitty peeks into the thread*
says hi to everyone and runs away screaming!
:bolt:

Meegan Wilson
July 28th, 2010, 03:22 PM
To my chagrin, perhaps because I'm an engineer and I'm supposed to know these things, I have been been assigned the responsibility of recommending/designing a pool measurement gizmo for the Florida LMSC. I was handed a 2010 rule book, the construction details of a laser-based system and pointed to this forum as a further source of information.

I've spent a day trying to figure this stuff out, and I am now profoundly confused. I hope that someone here can lead me to the light.

The only semi-sane description I can find, from page 40 of the rule book:

"""
107.21 Length
A. {A 50 meter pool must be 50 meters long unless you measure it in feet and inches where it is allowed to be 0.1mm shorter.]
B. [A 25 meter pool must be 25 meters long unless you measure it in feet and inches where it is allowed to be 0.05mm shorter.]
C. [A 25 yard pool must be 25 yards long regardless of the units used to measure it.]
D. [The pool can't be shorter,]
E. [even with timing pads in place.]
"""

To be fair, I probably should interpret the above rules as a simple accommodation for old folks (and sometimes even NASA scientists) who have a hard time thinking in these new-fangled metric units. The alternative is that it's a brazen attempt by USMS to abrogate the Meter Convention treaty of 1875. Under either assumption, the rules begin to make some sense: "A pool can't be shorter than it's supposed to be."

Seeking more detail, I find on page 37-38:

"""
105.1.7 ...
A. [You must use a steel tape or equivalent ... and record the length in]
1. [feet, inches and fractions thereof] or
2. meters and centimeters.
B. [Follow the stipulations of] USMS Pool Length Certification Form (Appendix B).
"""

Progress! The referenced form on page 134 promises an answer:

"""
Measurement procedures ... are found on the reverse side [of this form].
"""

Perhaps it's only a scrivener's error. Or perhaps the measurement procedure has been abandoned by executive fiat. If not, I am too stupid to grasp how adjusting qualification times for altitude outlined on page 133 relates to measurement of pool length.

When I asked someone who should know, I was handed an older, unredacted version of this form. It may or may not reflect the intention of the current rules. However, I find that:

"""
* [25 and 50 meter pools can be a little short.]
* [The measurement quality must be 5mm]
:
* [Use the shortest horizontal distance between ends from -0.8m to +0.3m]
* Measurements may be conducted [with or without touchpads].
"""

By "quality", I'm sure "accuracy" is intended, not "resolution". If the pool measurement is made without touchpads, I presume the pads are measured separately to determine their maximum thickness.

All this seems to contradict the earlier rule, at least for the metric sizes. A rational person, having measured a pool to be 49,999.9 mm (perhaps by using cesium interferometry or placing a series of platinum bars end to end) would report this pool to be 50m 0cm even though it was in violation of 107.21(D). An alternative to lying is to report the length in Olde Units, but only because this example was contrived to allow it. If the pool measures 49,996 mm, one's only recourse is to abide by the rules, fib, and report the pool as conforming[1][2].

Increasingly confused, I turned to the available literature (including this forum).

Fred Piggot provides a compelling argument against the use of a steel tape by describing a complex procedure for correcting such a measurement for temperature, tension, catenary sag, manufacturing tolerance, etc.

Jeff Rodin quotes an anonymous source who insists any measurement must be "accurate to [the centimeter]" and argues that any measuring tape must be ticked in "[mm , 0.001ft, or eighth inches]". I can find no justification for either assertion in the rules.

Jeff points out that these requirements are wildly inconsistent, but I suspect this is simply a nod to commonly-available commercial products instead of an actual requirement.

Jeff questions whether a tape ticked at 2mm intervals can even be used to measure pool length, but I can find no basis in the rule book for his apprehension. Obviously, eliminating all but three ticks would be desirable: 25y, 25m and 50m (corrected for the various errors that Fred mentioned).

In his steel-tapes-are-the-work-of-Satan piece, Fred justifiably advocates using a contractor-grade time-of-flight laser "tape" and describes a jig that will accommodate various pool gutter geometries.

Bob Griffiths suggests an improvement to Fred's laser rig: incorporating 4-foot 0.0005"/in-quality levels in the laser-holding jig and target to reduce the mechanical error to 1.2mm.

If you trust the marketing hype of the manufacturers, the best contractor-grade time-of-flight units I've been able to locate are accurate to 1.5mm. Ignoring other problems, this places a lower bound on the possible measurement error of circa 3mm. The water-resistant Leica DISTO DXT, recommended by Jeff Rodin, seems to fit the bill nicely. I have one on order.

In a message to MPohlmann, John Rempert advocates a cheaper, less rugged Bosch alternative, reports it works out to 50.7 m, but then drops a bombshell, "If you are over that, your pool is too long!" While I understand it one is unlikely to construct a pool to disadvantage one's swimmers, I am unable to find any such limitation in the rules. If such a rule exists and I have overlooked it, I would deeply appreciate it pointed out to me.

MY ACTUAL QUESTION:

In my current state of confusion, I suffer the delusion that to adhere to 107.21(D) using this rig, a pool must measure no shorter than 22.863, 25.003 or 50.003 meters before one can deem it conforming.

However for metric pools, one can bend the rules by adhering to Appendix B (unredacted) instead. Since my apparatus will conform to the accuracy (quality?) requirement, pools that measure 24.995 or 49.995 meters will be blessed because they can reasonably be reported to be 25m 0cm and 50m 0cm, respectively.

Does anyone share either hallucination? Or do you have alternate interpretations, statistical methods, calibration ideas, non-Euclidean time-space properties or whatnot that would allow shorter pools to be deemed conforming?

Jim Wilson
Gainesville, FL

[1] True story: At a sanctioned Senior Games meet at the University of Florida movable-bulkhead pool the post-meet lane measurement came up well short of 107.21(D). The meet director protested, and in support of the protest presented the results of a pre-meet measurement done by a "professional engineer". The form attested all six lanes measured "75 ft, 0.000 in".

Apparently one way of telling the truth is to hire a professional, with a cesium interferometer in his truck, to lie for you.

[2] A mathematician, an engineer and and an accountant are interviewing for an administrative position. As part of the interview, each is asked, "What is two plus two?"

The mathematician assures the interviewer that the set of integers is closed under addition, and the result is surely an integer.

The engineer uses a slide rule and reports, "3.97; at least that's close enough for government work."

The accountant declines to answer the question but instead asks, "What do you want it to be?"

The UF consulting engineer must have minored in accounting.

Chris Stevenson
July 29th, 2010, 10:30 AM
Okay, I refuse to print out that post and use a highlighter to ensure I find all the salient points. So I'll restrict myself to saying that the Rule Book is messed up and didn't print the reverse side of the form. The current measurement form, with procedures, answers (I think) many of the questions posed. It is in the Guide to Operations (http://www.usms.org/admin/lmschb/), a place well worth knowing to look up operational details.

Here is a direct link to the form:

http://www.usms.org/admin/lmschb/gto_rectab_pool_measurement.pdf

No, the pool cannot be shorter than the nominal length. The tolerance of the measurement device must be +/- 5 mm or better. If the measurement is made without the touchpads in place, a correction of -1 cm per touchpad is applied.

Let me know if you have more questions. Keep it short for those of us with ADHD. (Well, I don't have it, but I know people who do and I'm using that as my excuse.)

Meegan Wilson
July 29th, 2010, 02:25 PM
Thanks for the link to the current procedure. It is much less ambiguous than the older copy that was handed to me. Also, I suspect my copy may have been dumbed down a little so as to be understood by civilians. Let's stick to the 50m case, and make sure I've got it right:

1. A 50 meter pool is nominally 49.9999 meters long. (Sure, it's weird, but 'dem's the rules!)

2. If I can somehow delude myself into believing my measurements are accurate to 2.5mm, I cannot in good faith certify a pool measuring less than 49.9999+0.0025 = 50.0024 m.

(The Leica's still in transit, but I'm guessing it will resolve only 1mm. If so, my rule must be modified: )

2'. I cannot in good faith certify a pool measuring less than 50.003 m.

3. An offending pool, measuring 50.002 m will be recorded as "50m, 0cm", but I must indicate somehow (with an asterisk like in baseball?) that the pool is not known to conform.

4. Oh, A SCY meet can be held in a 25m pool.

This is all your fault, anyway. If you still lived here, you'd be doing this.

Jim

Meegan Wilson
July 30th, 2010, 02:34 PM
Drat! The Leica DXT just arrived, and I've been had!

1.5 mm accuracy is only promised up to 10m. At 50m the accuracy is only 7.5 mm, and that's 2.5 mm to the bad, even when mechanical errors are ignored. If one is borderline delusional about mechanical errors, the DXT is barely adequate to measure short pools (25m or 25y), and only pools measuring 5mm longer could ever be certified. That's nearly a quarter inch for those keeping score.

In retrospect, I should have read past the marketing dreck on the Leica web page and located the actual user manual where they weasel out of their accuracy promise.

(I do find it bizarre they derate their accuracy past 10 m at 0.15mm/m. In the absence of nearby black holes, I sense only two error sources: quantization error {0.5mm} and timebase slop {the rest}. The former wouldn't change with distance. And, unless the pulse detector is made out of cheese, the latter would be proportional. {I wonder if they use Blauschimmel or Räucherkäse; I can't distinguish the two by smell, and if I open it, I don't feel right about returning it.})

After downloading the manuals for their entire Disto product line, I see that even the most expensive model D5 only promises 4.5mm accuracy at 50m, and there is no practical way I can see to reduce mechanical error below 0.5mm.

If Leica can't do it, I'm not sure anyone else can either. Time to study those cantenary equations, I guess.

Does anyone have a contact at NIST where I can get a quote for the manufacture and delivery of a 3-notched steel tape?

Jim Wilson
Gainesville, FL

Conniekat8
July 31st, 2010, 09:52 AM
Drat! The Leica DXT just arrived, and I've been had!

1.5 mm accuracy is only promised up to 10m. At 50m the accuracy is only 7.5 mm, and that's 2.5 mm to the bad, even when mechanical errors are ignored. If one is borderline delusional about mechanical errors, the DXT is barely adequate to measure short pools (25m or 25y), and only pools measuring 5mm longer could ever be certified. That's nearly a quarter inch for those keeping score.

In retrospect, I should have read past the marketing dreck on the Leica web page and located the actual user manual where they weasel out of their accuracy promise.

(I do find it bizarre they derate their accuracy past 10 m at 0.15mm/m. In the absence of nearby black holes, I sense only two error sources: quantization error {0.5mm} and timebase slop {the rest}. The former wouldn't change with distance. And, unless the pulse detector is made out of cheese, the latter would be proportional. {I wonder if they use Blauschimmel or Räucherkäse; I can't distinguish the two by smell, and if I open it, I don't feel right about returning it.})

After downloading the manuals for their entire Disto product line, I see that even the most expensive model D5 only promises 4.5mm accuracy at 50m, and there is no practical way I can see to reduce mechanical error below 0.5mm.

If Leica can't do it, I'm not sure anyone else can either. Time to study those cantenary equations, I guess.

Does anyone have a contact at NIST where I can get a quote for the manufacture and delivery of a 3-notched steel tape?

Jim Wilson
Gainesville, FL

You woud need high precision (3 PPM or better) total station survey instrument, and several sets of repeating measurements to even get close to the kind of precision the rule book is asking for... and even if you can reach that precision, the accuracy may still be questinable - especially if the instrument is handled by an amateur, or hasn't been recently calibrated.

With regards to the .9999999's, counds like someone was trying to account for not using 2.54 cm in an inch as the conversion base from feet and inches to meters and the other way around. If you use one of the more commonly used conversion factor, which says 1 meter = 39.3700... inches, you don't get and exact conversion, because 100/2.54 has an infinite number of decimal places, and a rounding error gets introduced.

It also neglects to consider that such precise measurements are vitrually unobtainable in common practice. They could, but not in last 10 minutes before or after the meet, and not with a teeny handheld laser instrument commonly used to measure home improvement material quantities.

Like in the case of the Leica DXT, people don't even know to check for it's precision limits. Not to mention taht when you use laser beams near water and on surfaces of different reflectivity, other errors can be introduced. Like laser beam echo from water. In case of Leica Disto, all you have to do is look at the picture , and know it won't do. If you're *really* interested why, I can get into the long explanation...... It's about as obvious to a surveyor taht it won't do as it may be obvious to a pilot that an ultralite may not be able to cross the atlantic carrying 200 passengers.

OTOH, I've seen pools get certified, with 20 years old crinkled up steel tape, with a catenary sag likely causing an inch or two error, and have them called 'good' *shudder.

Even though I am a civil engineer and a land surveyor, and deal with precise measurements of this kind all day long, I just stay away from the whole thing. It makes so little sense taht it gives me one massive pounding headache. The whole rule set needsto be rewritten so it makes sense.... But, since I'm *just* an engineer, and not a high lama swimmer or a championship committee chair etc, what do I know. Armchair quaterbacks are the experts. Even worse, when I try to explain it, everyone glazes over, or tries to convince me it's not so... so I gave up. it's that whole, professionals vs. people who think they know how to do it... Professionals are seriously outnumbered in this case.

Meegan Wilson
July 31st, 2010, 07:53 PM
I'm an electrical engineer. I prefer sitting in front of a CAD station in my air conditioned office to slinging a theodolite through mosquito-infested swamps. I'll happily defer to you on the practical aspects of making measurements to the accuracy required by the rules.

But as it turns out, my understanding of these pocket-sized laser pulse time-of-flight devices was way too naive. My self doubt started when I realized that to resolve 1 mm required counting a clock ticking at 150Ghz or so. Now I'm sure there's a three-letter agency that can pull such counters out of their parts bin, but I doubt such a device could be powered very long by two AA cells. Google Scholar found me "Laser ranging: a critical review of usual techniques for distance measurement", Amann, et.al., Optical Engineering, Jan 2001, pp 10-19. For the time being, this paper is leaking out of some university library database: http://bit.ly/9wEq8j

These little doodads may be made from cheese, but it's pretty darn sophisticated cheese.

First, I had envisioned a sharp pulse, but the pulse is quite smeared out (into a Gaussian shape) with no distinct features to use as a reference. But by comparing the outgoing and returning pulses one can normalize the two and trigger on any point on both normalized versions. Furthermore, this whole process can be magically automated.

Second, a clock rate of a few tens or hundreds of Mhz will only allow the instrument to resolve to the tens of centimeters for one round trip. But the pulses are generated asynchronously from the timer ticks, and this introduces a kind of dithering that (statistically at least) allows a much higher resolution when the experiment is repeated a few thousand times (automatically, of course).

That's probably why the phrase "2-sigma" appears in most specs; the measurement is accurate within 1.5mm with very high probability, but not with certainty.

I'm still suspicious of the time base; instability there would introduce a systematic error increasing with the distance measured, just like the specs describe. If one could tighten up the time base, I can believe a sufficiently accurate measurement could be made, but I don't think these cheapo $300-500 units will cut the mustard unless modified.

Have a look at Fred Piggot's contraption. The beam is high enough above the water that I think your worries of multi-path might be assuaged.

But since the underlying accuracy of the rangefinder is in the tens of cm, do they meet the better-than-5mm accuracy requirement? Is "highly probable" sufficient? Or is "certainly" required? "Probably" would suit me, but I'm an engineer, and 3.97 is close enough for government work.

Jim

swimmieAvsFan
July 31st, 2010, 08:04 PM
Even though I am a civil engineer and a land surveyor, and deal with precise measurements of this kind all day long, I just stay away from the whole thing. It makes so little sense taht it gives me one massive pounding headache. The whole rule set needsto be rewritten so it makes sense.... But, since I'm *just* an engineer, and not a high lama swimmer or a championship committee chair etc, what do I know. Armchair quaterbacks are the experts. Even worse, when I try to explain it, everyone glazes over, or tries to convince me it's not so... so I gave up. it's that whole, professionals vs. people who think they know how to do it... Professionals are seriously outnumbered in this case.

just so you know, the current championship committee chair you're being snarky about is himself an engineer, and probably deals with more precise numbers on a daily basis than you do...

Chris Stevenson
July 31st, 2010, 09:22 PM
Let's stick to the 50m case, and make sure I've got it right:

1. A 50 meter pool is nominally 49.9999 meters long. (Sure, it's weird, but 'dem's the rules!)

2. If I can somehow delude myself into believing my measurements are accurate to 2.5mm, I cannot in good faith certify a pool measuring less than 49.9999+0.0025 = 50.0024 m.

(The Leica's still in transit, but I'm guessing it will resolve only 1mm. If so, my rule must be modified: )

2'. I cannot in good faith certify a pool measuring less than 50.003 m.

3. An offending pool, measuring 50.002 m will be recorded as "50m, 0cm", but I must indicate somehow (with an asterisk like in baseball?) that the pool is not known to conform.

You're making this way too complicated. The nominal length of a 50m pool is exactly 50 meters; you don't get to make it shorter by double-converting your units. Since "nominal" is a theoretical value, it has an infinite number of significant figures, so "exactly 50" means what it says.

In any event, there is no point in using more decimal places than the precision of your measuring device, so ditch the 4th decimal place if your device is only precise to (say) 1mm.

I believe you are also confusing measurement precision with "minimum distance." First secure a device that can measure the pool length with the required precision (5 mm). Once you have it, measure the length in every lane. If the touchpads are in place, then the measurement can't be less than 50.000m. If the touchpads are not in place, then the measurement can't be less than 50.010m if you plan to use one touchpad in competition, or 50.020m if you plan to use two.



But as it turns out, my understanding of these pocket-sized laser pulse time-of-flight devices was way too naive. My self doubt started when I realized that to resolve 1 mm required counting a clock ticking at 150Ghz or so. Now I'm sure there's a three-letter agency that can pull such counters out of their parts bin, but I doubt such a device could be powered very long by two AA cells. Google Scolar found me "Laser ranging: a critical review of usual techniques for distance measurement", Amann, et.al., Optical Engineering, Jan 2001, pp 10-19. For the time being, this paper is leaking out of some university library database: http://bit.ly/9wEq8j

These little doodads may be made from cheese, but it's pretty darn sophisticated cheese.

First, I had envisioned a sharp pulse, but the the pulse is quite smeared out (into a Gaussian shape) with no distinct features to use as a reference. But by comparing the outgoing and returning pulses one can normalize the two and trigger on any point on both normalized versions. Furthermore, this whole process can be magically automated.

Second, a clock rate of a few tens or hundreds of Mhz will only allow the instrument to resolve to the tens of centimeters for one round trip. But the pulses are generated asynchronously from the timer ticks, and this introduces a kind of dithering that (statistically at least) allows a much higher resolution when the experiment is repeated a few thousand times (automatically, of course).

That's probably why the phrase "2-sigma" appears in most specs; the measurement is accurate within 1.5mm with very high probability, but not with certainty.

I'm still suspicious of the time base; instability there would introduce a systematic error increasing with the distance measured, just like the specs describe. If one could tighten up the time base, I can believe a sufficiently accurate measurement could be made, but I don't think these cheapo $300-500 units will cut the mustard unless modified.

Have a look at Fred Piggot's contraption. The beam is high enough above the water that I think your worries of multi-path might be assuaged.

But since the underlying accuracy of the rangefinder is in the tens of cm, do they meet the better-than-5mm accuracy requirement? Is "highly probable" sufficient? Or is "certainly" required? "Probably" would suit me, but I'm an engineer, and 3.97 is close enough for government work.

I don't have the device so I can't answer your questions with certainty (no pun intended). But from your description, here's what I would say:

-- don't confuse the uncertainty of the measurement of a single pulse with the overall uncertainty of the device. There is the effect of averaging, after all; for completely random error the precision scales with the inverse of root(N).

-- don't sweat the time jitter either. While I am not familiar with this device, I have done a decent bit of work with pulsed lasers (nsec widths). If you split off a portion of the light and trigger your detection electronics off it then the time resolution is much better than the pulse-to-pulse time jitter.

-- as far as "measurement certainty," there is no such thing. It is all about probability. So 2-sigma is fine.

Again, you are making this too complicated. Look up the manufacturer's specified measurement accuracy and accept that it is correct. If it is 5mm or better, then you're ready to do your measurements. Hopefully they aren't too short.

Note added in proof: I just scanned the Opt Eng paper. I assume that most laser devices for pool measurement are of the pulsed types. From what I can tell (although they don't quite state as much), optical triggering is indeed used. (I really see no other way to do it, and it is very easy to implement...a small glass slide, for example, reflects 5% of the pulse and transmits the rest, so that would work.) And while it is true that the statement "to obtain 1 mm accuracy, the accuracy of the time interval measurement should be 6.7 ps" is pretty daunting, the authors also state that "averaging enables millimeter or even submillimeter precision to be achieved." I assume a laser diode is used for typical pool measurements (YAGs would be overkill) and, with their high frequencies, the time required to achieve 1mm accuracy would be pretty short.

Meegan Wilson
July 31st, 2010, 11:55 PM
...The nominal length of a 50m pool is exactly 50 meters; you don't get to make it shorter by double-converting your units.

I didn't want to make it shorter. 50m makes my life much simpler. I thought USMS wanted it shorter. I inferred this from the pool length cert form you referred me to. Indeed, the error is down in the noise, but why confuse me with all this rounding-to-hundredth-inch smokescreen?


-- as far as "measurement certainty," there is no such thing. It is all about probability. So 2-sigma is fine.

You know that, and I know that, but most civilians tend to see the marks on the tape as somehow "certain", and may feel uneasy about statistical methods used to reduce the error interval from a much larger value with Monte Carlo techniques.


I believe you are also confusing measurement precision with "minimum distance." First secure a device that can measure the pool length with the required precision (5 mm). Once you have it, measure the length in every lane. If the touchpads are in place, then the measurement can't be less than 50.000m. If the touchpads are not in place, then the measurement can't be less than 50.010m if you plan to use one touchpad in competition, or 50.020m if you plan to use two.

English is my first language, and I was sure the newer form demanded my interpretation, but I am certainly willing to go along with yours if that is the accepted practice. Let me try one more time to see if I understand this practice which is certainly simpler to implement:

1. By whatever (unbiased) method, half the pools measuring 50.000 meters will be in reality shorter; the other half will be longer, but that's OK because as long as we're within 5 mm it's good enough for government work. Regardless of reality, all such pools may be certified for competition.

2. Practical advice: If your pool lasers out just a little short, you should try again next week because your luck may be better.

3. Practical advice: Try when it's hotter or colder to see which makes the timebase in your laser gadget run faster and your pool measure a teeny bit longer.

If everyone agrees on this methodology, I would advocate changing the "Measurement Procedures" blurb to:

1. Drop the red-herring conversion to English units.

2. Change "A pool must always be at least as long ..." to "A pool must always measure at least as long ...". To me, the word "be" connotes a much closer relation to reality.

Jim

Chris Stevenson
August 1st, 2010, 07:00 AM
You know that, and I know that, but most civilians tend to see the marks on the tape as somehow "certain", and may feel uneasy about statistical methods used to reduce the error interval from a much larger value with Monte Carlo techniques.

English is my first language, and I was sure the newer form demanded my interpretation, but I am certainly willing to go along with yours if that is the accepted practice. Let me try one more time to see if I understand this practice which is certainly simpler to implement:

1. By whatever (unbiased) method, half the pools measuring 50.000 meters will be in reality shorter; the other half will be longer, but that's OK because as long as we're within 5 mm it's good enough for government work. Regardless of reality, all such pools may be certified for competition.

2. Practical advice: If your pool lasers out just a little short, you should try again next week because your luck may be better.

3. Practical advice: Try when it's hotter or colder to see which makes the timebase in your laser gadget run faster and your pool measure a teeny bit longer.

If everyone agrees on this methodology, I would advocate changing the "Measurement Procedures" blurb to:

1. Drop the red-herring conversion to English units.

2. Change "A pool must always be at least as long ..." to "A pool must always measure at least as long ...". To me, the word "be" connotes a much closer relation to reality.

Jim

Most people are pretty comfortable with the idea of averaging and intuitively trust an average over a single measurement, whether or not they've heard of the Central Limit Theorem. I don't think it's a problem.

(In my mind, a slightly bigger quibble is the reference to a maximum tolerance of a measurement device. Tolerance -- maximum allowable deviation -- is a function both of the precision of the measuring device and how it is used. In the rule I think the maximum tolerance is more properly referring to the pool length (and its measurement), not the capability of the device. But it is just a small quibble and I'm not sure that anything is served in going into all that on the measurement form.)

To answer your questions:

1. You got it. Though I would point out that, because of how the rule is written, most certified pools have measured out at above 50.000 m. That means that -- assuming only random error -- the fraction of such pools that are truly shorter than 50 m is substantially less than half.

Actually, the situation is even better than that because all lanes of the pool are measured and if ANY ONE of them is too short, the pool is not certified. So I think it is reasonably unlikely that a certified pool is in reality too short.

When I first read the rule, the measurement scientist in me was slightly offended that it didn't just say something like "the pool measurement must be 50 m or longer, within the precision of the measurement device." Assuming an unbiased measurement procedure, such a situation means that certified pool could not be proven (in a statistical sense) to be shorter than the nominal length. But I suppose I see the value in the way it is written.

2. I suppose, though you are biasing your measurement procedure by cherry-picking your measurements. For example, if you hired surveyors to do it for you (not uncommon) I am not sure they would cooperate.

3. I don't think temperature variations over the range likely to be experienced at the pool would make a significant difference is measurement accuracy, but feel free to try.

To respond to your suggestions:

1. Some people use devices that measure in English units so I think that needs to stay. The uncertainty in conversion should be less than that of the measuring device anyway.

2. I will bring it up. I do like your wording better, but I'm not sure the rest of the Records & Tabulation committee will agree. They may feel that your wording invites abuses (you know, like doing lots of measurements and choosing the one that gets the pool certified :)).

Meegan Wilson
August 1st, 2010, 01:44 PM
Or do you have ... ideas ... that would allow shorter pools to be deemed conforming?
I wish I hadn't phrased it quite that way. I've been fighting your impression that I'm trying to bend the rules in favor of too-short pools. I'm only trying to understand common practice and avoid inflicting an overly-strict procedure. To probe what's acceptable, I tend to let my cynical side show through. Any "Practical advice: ..." I offered was in that vein.

My real intention is/was to devise a system, packed in a foam-lined case that would fit in the trunk of a compact car. The system operation could be described in a half-page instruction manual so I can hand off the job to an amateur (or even a meet director) who would certify a pool. Other than initial testing and perhaps periodic checking of calibration, I hope to never touch the thing.

(Since I no longer swim, I was only mildly amused at the UF-pool-too-short fiasco. My feelings were similar to the one I had for the Billy Martin/George Brett pine-tar incident. But I can assure you some other principals had a substantially more violent emotional response.)


(In my mind, a slightly bigger quibble is the reference to a maximum tolerance of a measurement device. Tolerance -- maximum allowable deviation -- is a function both of the precision of the measuring device and how it is used...)

I wish you hadn't shared that thought. I was all set to report back that there was no way to make sufficiently accurate measurements with inexpensive pocket-sized devices, but the most-expensive Leica D5 barely squeaks in under the wire at 4.5mm.

I had interpreted "measuring device" to mean "measuring system", including the jigs used to assure conformance to the rule at various depths, but parsing the procedure closely seems to confirm your interpretation. Only the specific component of the system used to perform the actual measurement is required to be accurate.

Please understand, I'm not trying to rationalize sloppy construction of the other components of the system or abuse this error source to certify non-compliant pools, but only to point out that these (statistically-independent) error sources will degrade overall accuracy outside the 5mm limit.

Unfortunately I have more work in front if me, if that is the intention of the rules and not just the ravings of some metrically-libertine tow-headed nutjob that swims fly fast. There's only room for one nutjob in this discussion, and I claim that position.

Incidentally, I am told that neither Fred Piggot's contraption, which apparently hails from Michigan Masters, nor the improvements suggested by Bob Griffiths are widely known. I suggest these secret documents somehow be declassified and made available to this forum.


...because of how the rule is written, most certified pools have measured out at above 50.000 m.
Because of how the rule is written and because the danger of striving too hard for the ideal may subsequently entail corrective action involving jack hammers, dump trucks, cranes and multimillion-dollar contracts.

I do like your wording better, but ... the ... Records & Tabulation committee ... may feel that your wording invites abuses
In "A pool must always be (measure?) at least as long ...", I'm sure they could wield a particular definition of "always" to smack down a cherry picker.

Jim

Chris Stevenson
August 2nd, 2010, 10:28 AM
I've been fighting your impression that I'm trying to bend the rules in favor of too-short pools

I don't have that impression at all. Fire away with your questions. Hopefully I can answer them eventually.


Incidentally, I am told that neither Fred Piggot's contraption, which apparently hails from Michigan Masters, nor the improvements suggested by Bob Griffiths are widely known. I suggest these secret documents somehow be declassified and made available to this forum.

I don't know what these are. Can you elaborate?


In "A pool must always be (measure?) at least as long ...", I'm sure they could wield a particular definition of "always" to smack down a cherry picker.

I'll put it on the agenda (which is pretty full right now but we'll get to it eventually :)).

I assume you are working with Meegan's daughter Marly on all this? She is Top Ten recorder of your LMSC and would be the one who is in charge of keeping records of certifications of all the pools where you hold meets. When you certify a new pool, please be sure to let Walt Reid know so he can add it to the national database:

http://www.usms.org/~rectabs/poollengthdb.xls

jroddin
August 3rd, 2010, 07:29 AM
just so you know, the current championship committee chair you're being snarky about is himself an engineer, and probably deals with more precise numbers on a daily basis than you do...



Actually, Connie is a civil engineer and land surveyor which means she knows about 100 times more than I do about taking precise measurements. I design parts so I know how things are "supposed to be" but she works on the practical side of things.

PS. I didn't interpret her comments as being snarky. She pointed out a naive mistake I made in interpreting the tolerance of the Leica Disto by taking what they said on their website (i.e. marketing) to be true instead of digging into the technical manual to find out the true limits of the device.

Meegan Wilson
August 4th, 2010, 10:45 PM
I submitted this stuff to Wikileaks, but they were too busy vetting other things so I got Meegan to more or less wrangle permission to release the hush-hush description of the laser/target holding apparatus I misattributed to Fred Pigott along with a critique and some suggested improvements from Bob Griffiths.

In a series of emails, Fred denied any responsibility for the document and instead credited Frank 'Skip' Thompson for the concept. I've done a bit of editing, and without doubt I've introduced an error or two to both Skip's original and Bob's improvements. Sorry, Skip. Sorry, Bob.

These older documents are mainly interesting from a historical perspective, because Bob's recommendations appear to have been accepted in their entirety by USA Swimming, and the resulting document is probably a better reference.

Here's the old stuff: http://floridalmsc.org/forms/PoolMeasurement.pdf

And here's the new: http://www.pacswim.org/certifyingpools.pdf -- See particularly Part 2 and page 9 where the contraption is professionally illustrated by someone with a good knowledge of Inkscape or Corel Draw. This is almost exactly what I had pictured in my mind after stealing Bob's ideas.

Jim, impersonating Meegan

Chris Stevenson
August 7th, 2010, 02:40 PM
Thanks, Jim. I skimmed both documents and they look useful, esp the USA-S one.

Conniekat8
October 1st, 2010, 01:02 AM
just so you know, the current championship committee chair you're being snarky about is himself an engineer, and probably deals with more precise numbers on a daily basis than you do...

Jeff is also is also a very dear personal friend of mine.

Conniekat8
October 1st, 2010, 02:32 AM
PS. I didn't interpret her comments as being snarky. She pointed out a naive mistake I made in interpreting the tolerance of the Leica Disto by taking what they said on their website (i.e. marketing) to be true instead of digging into the technical manual to find out the true limits of the device.

Well, I wouldn't exactly call it a naive mistake. It's one of those things that one is not likely to know unless they are a land surveyor who has education about it and deals with those things every day.

Best thing anyone outside of having a degree in Land surveying can do is fins a local Land Surveyor in a pnone book, tell them what they need to acomplis, what precision is needed (this is very important for the choice of methods and instruments) and ask for advice. Most Surveyors are likely to be quite happy to give advice on this. They're most likely not going to try and sell you anything, because coming out for an hour or two to mesure a pool is typically cost prohibitive for them.

Lot of times Conuties and Cities also have a Land Surveyor on staff as a part of their Public Works office. They may be even more willing to offer advice. If the pools are city owned and run, there may be even a way to get them out there to do it for you. For example, if a City Parks and Rec. department contact the City Engineer....