Since weight is the reason for this site, I assume everybody has a decent set of scales. I don't and I WANT ONE. (After all, weightweenies need to weight. Obviously, weenie is no problem.) Currently, I use a double beam balance for small items; accurate but limited. I have better than average bathroom scales but I don't don't trust it to better than +- 1/2 lbs.
Weighed my bike at local bike shop on their spring scale. It worked pretty well but won't get down to ozs. accurately.
I need a digital hanging scale with a capacity of 50 lbs., metric mode, decent accuracy, resolution, and repeatability. The best I have found is from Mc Masters-Carr, http://www.mcmaster.com/ (this is an INCREDIBLE CATALOG. CHECK IT OUT FOR GENERAL HARDWARE) for $100. Can anyone suggest better? I don't want to spend a fortune.
ALERT!! Avoid Exacta Digital
They are out of business but their WEB site is still up. They are alive enough to cash my $30 charge but no scales made it to my door.
BUCKY BALLS HAVE CAJONES!! :wink: [/b]
You are right. I don't need 50 lbs capacity for bikes (at least I HOPE not!). But if I spend the money, I want as much flexibility as I can get. There are more things to weigh than bikes.
I have found fishing scales that handle around 22 lbs. at attrctive prices. I may compromise and settle for these. But, if you check their specs, accuracy and resolution aren't so good. Makes me not trust their overall quality. Since we are spending hundreds of dollars for grams, it would be nice to spend with knowledge.
I'll continue to check eBay but previous searches yielded nothing. Thanks
BUCKY BALLS HAVE CAJONES!!
Since I trust Mc Masters-Carr, I used some of their less expensive hanging digital scales as examples. To be consistant I tried to use scales with similar maximum capacities. Accuracy is rated as a percentage of their full scale(maximum) capacities. For their 55 lb, $98.28, " Triple Scale" model, I came up with an accuracy of +- 4.4oz and +-125g. For their 60 lb, $297.98, "Dual Scale" model, the accuracy comes out at +-.77oz and +-21.6g.
Neither of these scales are good enough for weight weenies buying carbon bottle cages, like me. Pulling some weights out of the Colorado Cyclist catalog, the Douglas Carbon cage (33g) could weigh out from 0g to 158g (!) on the less expensive model and 11.4g to 54.6 on the more expensive scale. And there is no guarantee of consistancy from one session to another.
There is no doubt that the $297 scale is worth the extra money. But that is almost the cost of a set of CAT Claws. Do we care that much? If yes then I cheer our dedication to the cause. If no, then what the hell are we doing here?, I ask.
Given these hard to swallow facts, I must ask how everybody is coming up with their numbers (including the site examples)? Hate to burst your bubble. Does it matter?
BUCKY BALLS HAVE CAJONES!!
Joel, I have a good Ohaus double beam balance which will accurately weigh up to 2000g with the supplied aux weights and with resolution of 1/10g. This will handle most bike things. I suggest this type of scale for all weenies that care about accurate weights. They are not too expensive. Check Ohaus link below.
Joel, how in the world do you get your whole bike on the scales? I would like to see that and learn how to fold my bike. A 5kg scale is of interest. Who is the manufacturer, what model, & where did you get them (ebay?). Can you provide a link? Thanks.
Problem with small scales is that the pans don't handle physically akward items like wheels and complete bikes. Unless you want dissemble the item and weigh everything individually and then add everything up, as Wally suggested, double beams or any small scale may not get you there.
Wally96, your suggestion is good. Thanks. But unless you have very good scales, there is a problem. The full scale error is present on each measurement. Each item weighed will be wrong by an amount equal to the error. As each item is added, so is the error. If we assume that the error is not statistically distributed, the errors will add in one direction (high or low but not both) and will have a more or less constant value. For example, if we weigh 10 items, each with an error of 10g, the total error after all is added will be 100g. Thus, weighing a COMPLETE bicycle will give only a one time scale error which may be acceptable, as Joel said, even with less expensive scales. Weighing individual items and adding maximizes the error. Not good.
Also, from a practical point of view I want to weigh my whole bike without taking it apart first. Since I bought a fully assembled, custom specified bike, I did not have the opportunity to weigh everthing first. Oh well....
I really like where this discussion is going; from scales to proper weighing technique. Any input from folks who are knowledgable about this scientific subject would be appreciated. This something all weight weenies should be interested in. Thanks for your thoughts. Lets keep it up.
BUCKY BALLS HAVE CAJONES
I'm going to add my 2 cents in on this scale subject.
Karma, you seem to be revolving around two concepts in error. Accuracy and precision. You can precisely measure small parts, but the accuracy of the entire bike comes into question, error is propagated, and this concerns you?
As you have noted, the cost of a balance scales with precision. So for example, I have solenoid based balance in my lab which has a useful range of 200g. And is precise to 1/100 mg. So theoretically I could weigh 199.99999g accurately and precisely. This balance is too costly for any normal usage. The cost the service contract on this balance is more then my bike.
Error propagation in addition, how does it affect accuracy? Lets say you had weights of two components with absolute uncertainties, 1000+/-5g and 10+/-1g. The error would be 1000+/-(0.5%)g and 10(10%) relative uncertainties. Added together they would be 1010(+/-x) were x is the square root of the added squares of the absolute uncertainties, i.e. 5.1g So the result is 1010+/-5.1g. As you can see from this example, the heavier weights propagate their error more (regardless of the relative uncertainty). So to be practical, you would want a balance that is capable of the most accuracy with the largest capacity for your object.
Your example of 10*10g error, is 100g error, your right. But you forgot to mention the fact the number you were measuring is now 10x more precise! In general, the more measurements you have, the more precise your number is, not the other way around. The accuracy of any weight is solely dependant of the quality of the balance, and if it’s properly calibrated. That’s why you don’t weight little tiny things on big balances, and big heavy things in tiny balances.
I hope you are comfortable accepting error. It sounds like you’re about to put down some cash, and I hope you find a good deal. Good luck. Sorry to hear about the $30. Let us know what you find.
BTW, cleaver with the Buckyballs, i.e. spherical graphite. There are a lot of people (like me) who are trying to use things like buckyballs as structural components in synthetic resins. Unfortunately buckyballs do not really reinforce in the same manor carbon/glass fibers do, and the cost of getting them into a process-able matrix is prohibitive. Exfoliated clay structures are showing some promise as nano-reinforcements, but their mechanism of reinforcement is again different then carbon/glass fibers, they restrict the chain movements of the polymers, this great for making strong materials, but really becomes a big pain in the ass to make parts from them. The future if synthetic resins will someday outperform carbon fiber composites, but for now, we will ride our bikes without Buckyballs.
Thanks for your 2 cents. No, the pic doesn't qualify but it's better than nothing especially if that's you in the middle. So, 1/2 point just for adding pizazz to the site. Your scales should satisfy anyone's hunger for accuracy. Bet the lab doesn't let your bike near it.
It's getting late and I don't have the time to persue your weight example. I do have some questions. If I understand a little, you are proposing that the error is distributed statistically and a least squares approch will give the actual error. Is this close? If so, my example was faulty in that I excluded the statistics and, by imference, I weighed the same item 10 times. In that case I agree, I think. But, the real life case of weighing a collection of bike components, not the same item 10 times, the errors would simply add up. Any measuring device will have a constant offset error, which is signed, and an statistical uncertainty which may add or subtract from the constant in a statistical manner. Do I have this right? So, my example was too simple. I think the sum of the individual errors will still be larger than the error when weighing the aggrigate even though the value of the agrigate is a larger value. Please correct me.
Whew! I was beginning to think that no one was going to make note of Bucky Balls. I'm not a chemist but have been following BB's, like I watch football (& the Cowboy's won!!), for about 10 years. You are the first that I have talked to that actually works with them. You gain my hall of fame. I have read about BB nanotubes which seems to hold promise in nano engineering and semiconductors. No, I really didn't think I would be riding a BB bike. But then, did the inventors of lasers foresee DVD's. Probably not. Once unleashed, technology can gain a life of its own.
Thanks for your thoughtful post.
What you call ‘offset error’ is accuracy. Accuracy is how close you are to the actual value. For example, you take a dog and measure its mass 5 times. And get 10.1, 10.2, 10.3, 10.2, and 10.1kg. You take the average, 10.2. And calculate the error (the previous method I had mentioned), +/- 0.17. So you determine the dog to weigh 10.2+/-0.17kg, or 10.2+/-1.6%. But the dog actually weighs 12.0kg. Your degree of accuracy is off by 47% because you have a shitty balance, even though you measured the same number 5 times. Looking at the errors, your measurement is very precise at 1.6% error, but not very accurate 47% error. Repetitive measurements help with precision, or statistics as you refer to it. And again, accuracy is dependant on the balance.
Getting an accurate measure is always more important then a precise one. That’s why I had suggested to you to get a balance that is appropriate for your total object of weight. So to a first approximation, it doesn’t matter if you weight things in parts or as a whole, it’s more about getting a quality balance that is properly calibrated, and thus accurate.
To be technical about this all, this isn’t really statistics. While it seems like it is, because we evoke roots of added squares, stats is usually expressed in sigmas (or deviations from a Gaussian distribution). Lets not get into that. Remember when your teacher curved the class grades, what he/she was doing was fitting the grades of the students to a Gaussian distribution, where the max of the curve was average of the class. Anyway, more information then you wanted…
You are a very smart and interesting person. I never thought there were ‘fans’ of science or scientists. Especially chemists, who get a bum wrap most of the time. Anyway, keep up the positive enthusiasm. I assure you, and the rest of the weight weenies out there, that new better materials technologies are coming. But instead of precipitating from the aerospace industry, technological improvements will come from the laboratory. It just might take a little longer, since the cycling market isn’t big to fund dedicated research programs.
BTW, I noticed you were from NM, I had a question for you, is the Gila national park open to cyclists? I’ve always dreamed about tooling around near Gila. I wonder if it’s as good as I imagine it to be?