a scale being linear in its use range is an indication of its usefulness.
depends on how much effort you want to make.
i would use the short time comparison to a good balance at a doctor’s office, a person gets there every year or three. good enough for my body concerns.
if you had three sets of free weights you could use those or haul the scale to a gym. you could load the scale (with a platform on blocks in the human feet position) and check for linearity and accuracy.
As mentioned, this doesn’t calibrate the scale at all. Assuming the scale shows half-pound digital increments, all this says is that the scale is linear to better than 12% in the vicinity of 150 lbs over a range of about 4 lbs. Even assuming the same quality of linearity from 0 to 150 lbs (and there’s nothing showing that that’s a good assumption!), this translates into a possible 18 lb error at 150 lbs.
The only way to calibrate this is to put a known weight on the scale in the vicinity of 120 lbs.
Hazle: How about this… Place your empty kitchen waste bin on the scale. Also, get a gallon jug. Now, fill the gallon jug with 1 gallon of water using a precise measuring cup (since you’ll want to mark off a precise 1 gallon mark.) Dump that gallon into the bin. Continue filling to the gallon mark on the jug and dumping the water in. Do this until there are 14 gallons in the waste bin. That will weigh 116.8 lbs plus the weight of the waste bin.
Now empty the water and dry the waste bin (or, better, do this part first). Stand on the scale and weight yourself, and then pick up the empty waste bin. Take the difference and add that to 116.8 lbs. The result (say, 118.5 lbs) should match the scale’s reading when the bin is full of water.
Note that even a 10% calibration error in the calculated waste bin mass via the picking-it-up method translates into only a 0.2% error on the final bin+water mass.
What else do you have to do tonight? Nothing! Get to filling!
Ok. Started from scratch. Scale set to zero. I put a four-lb bag of sugar on the scale and it did, in fact, show a weight of four pounds. I weighed in at 110. I then held the bag of sugar, hopped on the scale and it showed a weight of 114. So what does this mean? Not trying to be dumb. But doesn’t this mean that my scale is accurate?
Anecdotal evidence:
I weighed myself on four different scales in the past few weeks, and the results I got ranged between something like 74.X kilograms and 85 kilograms (some of these had a supposed precision of tenths of kilograms). I find an inaccuracy of 10 kilograms (more than 22 lbs!) rather startling - or maybe the inaccuracy is just 5 kg, with the actual weight being right in the middle between the two extremes.
Precision does not equate to accuracy. I might have scales that tell me my weight is 84.002 kg but my weight may actually be 92 kg. The first figure is very precise but inaccurate. Digital readouts often have displays with high precision because it is easy to do and it gives the impression of accuracy.
depends on how much effort you want to make. knowing the bag of sugar adds 4 pounds to the total is an indication of some good function that takes seconds with the bag which was already at hand.
i’ve created gallon measures to approximate weight. when you get to that 120 pound total it isn’t that accurate.
if you had three sets of free weights you could use those or haul the scale to a gym. you could load the scale (with a platform on blocks in the human feet position) and check for linearity and accuracy.
Oh heck, that’s true with lots of instruments. We have a lab full of DVMs, and it’s apparent there’s a marketing war amongst instrument manufacturers with the number of digits displayed. Both Keithley and Agilent sell “8.5 digit” DVMs. Truth is, those last couple of digits have no meaning whatsoever when it comes to absolute uncertainty. They *might *be of value when looking at a *change *in voltage, but only if you take great care in ensuring noise is not corrupting your signal (or you do a ton of averaging). For 95% of measurements, those last couple of digits are random number generators, and can be safely ignored.
Back to the OP… the only way to measure the actual uncertainty of a scale or balance is to calibrate it with NIST-traceable weights over the range of the device. No one is going to do this for a cheap home scale, so the next best thing is to simply put a known weight on the scale that is between 100 and 200 lbs. The easiest method is to weigh yourself using a much better scale (like one at a gym or doctor’s office), and then weigh yourself using your home scale within a few hours. This will give you an idea of the absolute uncertainty of your scale in your weight range. At this point, you can either adjust the zero knob on your scale, or simply compute an offset error. Once this is done, you may want to measure the gain error of your balance by weighing yourself while holding a known weight.
Also, my room mate thinks it’s insane that I stick my drinking straws in the dishwasher and reuse them. But, that’s a different thread, I suppose. Anyway, thanks!
You didn’t mention your roommate is a psychiatrist. Even so, drinking straws are seldom a measure of sanity.
But, but…what if I weigh them first?
Most spring-based home scales are easily off by 3-5% when new, and become more inaccurate through use. Cheap home bathroom scales are essentially measure the deflection of a spring per Hooke’s Law (F = kx), with k being the spring constant. If you weighed 120 lbf but the scale only indicated 110 lbf, that would be about an 8% precision error. For a nominal 4 lbf bag of sugar, that would be 0.33 lbf difference, probably not enough to easily detect. In addition, at loads that are significantly higher or lower than the measurement range, the spring may not behave in a very linear fashion, so the precision error may not be that distinct.
Cheap polystyrene disposal straws? You realize that because of the narrow aperture little of the detergent or hot water really enters the inside of the straw, so putting it through the dishwasher does almost nothing to sanitize it. The correct way to clean something of that nature is to submerge it in soapy water with a small amount of bleach or antiseptic cleaner, and then thoroughly rinse it out just as you would a hydration tube or piping on food processing equipment.
Also, color me disappointed that no one has yet introduced a Monty Python’s The Holy Grail reference into this discussion.
Stranger
Take your scale to a gym, put a range of barbell weights on it and record the accuracy over a range.
Damn it! Can’t I be right about ANYthing?
If you really are worried about the actual number of your weight, check with your doctor next time you’re in for a check-up. A 5-10% error doesn’t strike me as unusual for a scale. I have several scales around the house, electronic and spring-based, and they all float within a 5-10 pound range for my body (165 lbs). So that’s about up to 6% error.
What’s it really matter if you’re 120 when the scale reads 110? You say your clothes fit right. You seem to feel good about your body. Who cares if the actual number is 120 and not 110?
Fill the bath until there is enough water in it to immerse yourself until floating.
Mark the level with an oil pencil.
Get in the bath, float and allow the water to settle.
Mark the new level with the oil pencil.
Get out, get dry and top up the bath to the first level (some water will have been lost when you got out)
The number of litres it takes to increase the level to the second mark is your weight in kilograms.
Pretty much; the issue you are dealing with is linearity, where a scale weighs accurately at progressively heavier loads up to scale capacity.
We check scales using increasing and decreasing load tests, or with substitution or strain where applicable.
For example, there’s no practical way to check a 120,000 lb. truck scale to capacity; the best we can do with known, certified weights is put 21,000 pounds on it; 16,000 pounds in the form of 16, 1,000 lb. state certified block weights inside of a motorized weight cart weighing 5,000 lbs, also state certified.
We can however put a fully loaded truck on it that show an indicated weight, for sake of example, of 75,000 (rounding for sake of example). If we then put our 21,000 lb test cart on it and it goes up to 96,000 lbs, we know the scale is accurate and within tolerance.
This is essentially what you did with your test using both yourself and the sugar sack.
We check large capacity (30,000 to 100,000 lb.) tank scales in a similar fashion. We can typically get 2 to 4 1,000 lb block weight around the tank and lift them up with chain hoists. So we’ll check the empty tank at 4,000 lbs, and then fill the tank with material (usually water) and check it with progressively heavier loads.
As noted upthread, the foundation upon which your scale sits is critical; firm, flat surfaces work best. Try weighing yourself with your scale on the bathroom floor, and then go set it on carpet and try again, and see the difference.
Environment can play a factor, and any “bathroom scale” worth its name will have adequate moisture protection. But even industrial grade moisture proofing (IP69K) will break down over time.
Another possible factor affecting accuracy with an electronic bathroom scale would be the battery; these things are already using a pretty low excitation voltage to begin with, and an old, weak battery can cause inaccuracy, but I doubt you’d see linearity issues because of a battery; linearity is typically a design or mechanical issue, in my experience. If your bathroom is dusty, material build-up around your scale’s weighing element, or between the weighing platform and the scale chassis, can cause inaccuracy.
For those of you recommending using her weight at the doctor’s office: I always get weighed with my clothes on. Do they have some standard amount they subtract off to account for that? You’d have to know what that was for the doctor’s office measurement to be useful.
What you say is true if the scale is strain gauge based.
Some of the digital scales, however, are a mechanical scale with a rotary encoder on the mechanical readout. I happen to have one, and know this to be the case, because I took it apart and saw how it worked. I was actually interested in seeing the lever system that sums the forces on the four corners.
Speaking of which, if you measure accurately you could rig a lever to multiply the weight of a light but known weight. Of course you have to compensate for the weight of the lever, but you can balance that to zero without the sugar sack load, so it needn’t be a source of error.
Mechanicals are a little before my time, but there’s still some out there in service, and I’ve had some experience with them, just not in the bathroom.
And you’re right, I did assume an electronic scale, rather than consider the possibility of it being a mechanical.
Did it have two armatures interleaved, with one having a “nose iron” connected to the weight indicator? If it was connected to a rotary, I’m assuming some sort of a compensated spring element to provide counterpoise. There’s an actual technical term for the “nose iron,” (for all the weighing components of a mechanical scale) but it currently escapes me.
Calibrating mechanicals can be a PITA. I much prefer electronic “Press 1 to Zero, 2 to Span” digital indicators, but the old mechanicals do have a certain elegance to them.
When they aren’t rusted out pieces of junk.
It had one V shaped lever with the two ends of the V at the front corners, and the center resting on the spring element. The rear corners were supported by levers connected to the midpoints(via shackles) of the arms of the V. There was a sort of bell crank that pulled on a rack that ran the pinion connected to the needle and a slotted disk for a couple of photointerrupters in quadrature.
Here is a link that shows the lever arraignment, albeit with a strain gauge sensing element.
My former boss invested in a gravel pit as a side business. The guy that ran the scale house needed to be with his daughter at the hospital, so I had to fill in while they found a temp to sub for him. The pit was pretty easy to access, so I got to admire the ingenious way they brought all the forces to a single strain gauge.
The pit was in an area with no utilities, so they would fire up a genny to run the scale and AC in the scale house if they were feeling wimpy. Looked at me like I was nuts for suggesting that maybe a yardarm would have served them better than the electronic readout.