bathroom scale hysteresis

John, My guess is that the hysteresis is caused by friction in the mechanical workings.

Tam

Occam would agree...

--
John Miller
email domain: n4vu.com; username: jsm(@)

On Sun, 29 May 2005 19:20:02 GMT, John Miller wroth:

Occam might agree, but only if he never read the whole thread.

Jim

Occam should have grown a beard - much simpler.

Ken

Our bathroom scale has exactly the same problem - the same reading is repeated when you weigh yourself a number of times.

If the first reading were, say 160.3, further readings would be either

160.3 exactly, or something outside the range of 160.3 plus/minus 1.

These results, I believe, is a result of clever firmware and not mechanical or sensor characteristics.

About ten lines of code in C or C++; not at all hard to add.

-JS2

I have a hard time believing that they would do that. Why would they go to the trouble of "lying" to users of their product?

If I noticed my scale doing that, I would not be happy. Fortunately, my

4-year old Tanita does not seem affected by this problem (I've tested it a number of different ways, and never once noticed the problem described by the OP).

Perhaps it's limited only to low-end scales.

Oddly enough, nobody complaining about this "problem" has mentioned brand names either...

--
GG
http://www.WeightWare.com
Computer-Assisted Weight Management

Having spent some time in the weiging industry, i can assure you that the cheap portable bathroom scales that you buy are neither: a. repeatable b. accurate c. linear

In fact, most will become more innacurate in the most common weight ranges.

In AU, a cheap set is $20, a good repeatable set is $400. The expensive scale can be recalibrate and remains linear. Should it not be linear a multi-point cal can be done. The cheap set can only be thrown.

Most cheap sets are innacurate of the shelf as they are made in china and fail to take into account local gravitational constants

Do you have any experience with the newer digital models? Not the low-end ones, but ones like the Tanita with ostensible 0.2 lb resolution, priced around $50-90 (US).

For instance,

or the more expensive German-made Soehnle brand

GG

Hi, Thanks to all who took time to reply.

Yesterday I went to a Bed, Bath, & Beyond store where there were several models on display.

I brought with me several 0.5 liter water bottles (approx 1 lb.) and half-consumed one (approx 0.5 lb.) so that I could adjust my weight in

0.5 lb. increments. I tried 3 models and found two that I think exhibited this programmed hysteresis, and one that did not.

The $40 Thinner TH300 uses load cells and reports to 0.5 lb. Holding an empty shopping basket and full pockets, I measured at 181.0 six times in a row. Then, holding ~0.5# in my basket, I still measured

181.0. Again, holding ~1#, I still measured 181.0. Then, when holding ~1.5#, apparently past the programmed hysteresis threshold, scale reported 182.5.

The $50 Tanita BF679 reports to 0.2 lb. It did not have hysteresis. Each measurement was independent, reporting like 178.2, 178.4, 178.6,

178.4, ... (I wasn't holding an empty shopping basket for this one).

The $60 WeightWatcher WW60 (Scales by Conair) uses load cells and reports to 0.1 lb. I (plus basket) measured 180.7 six times in a row. Then holding ~0.5 lb., I measured 180.7 again. Then, holding ~1#, I measured 181.9 .

So, I see this programmed hysteresis in some models, with a breakaway delta of ~1#.

I had thought I originally noticed this two months ago in several Taylor models and a Tanita, but I wasn't able to test any Taylors yesterday and my one Tanita test didn't show it.

Best regards, John Ruckstuhl

Outstanding report...thanks!

My experience with Tanita has been the same - no hysteresis noted.

Consumer Reports magazine tested several models of body fat scales in the last year or so. As I recall, they liked the Tanita the best - they warned that body fat was quite variable, and probably not very accurate, but that the body weight reported by Tanita was better than most other units (not sure if they tested for hysteresis).

GG

The high end Soehnle scales can be calibrated, but the ones show look like you standard cheap crap. I could well be wrong here. Some Soehnle stuff is quiet good.

Tanita, never seen anything but cheap crap.

The good scales I am familiar with are manufactured by a company called AND. Try do a google for AND though!!

I suspect that many modern, microprocessor -controlled bathroom scales>

There's no such thing you jerkoff.

I presume that he was referring to scales with an LED display. I imagine that these *do* have a microprocessor of sorts inside, to perform analogue-to-digital conversion on the output of a strain-gauge attached to the part you stand on, and then to convert this reading to signals which control the segments of the display.

Now who's the jerkoff? ;-)

"Martin Underwood" wrote in news:42a1b925$0$21780$ snipped-for-privacy@ptn-nntp-reader01.plus.net:

The "digital" bathroom scales I've disemboweled only digitized the rotation of a disk. The disk would have had numbers on it, if it were a "mechanical" scale. The disks had black and transparent radial bands, that were read by a quadrature decoder/counter, and displayed on 3 LED digits. The mechanics were exactly the same as in the old bathroom scales we've had for a nearly a century.

That's true. But the point in question was the "no such thing" put forth by a previous poster. That's clearly not true.

--
Noah

snipped-for-privacy@gmail.com wrote: [snip]

[snip]

Do you think that this behaviour would result if the ADC resolution is coarser than the display resolution, for example the LSB represents 1.5#? Or maybe the ADC has missing codes. This wouldn't require tricky programming, just a bad ADC.

Occam like that, too.

--
Noah