• Edward (unregistered) in reply to jgoewert

    NZ doesn't get them any bigger than 2cm or so. Even in rainy warm Northland.

    (If I discovered the ones of the size mentioned in the article, I'd probably have moved by now.)

  • (cs) in reply to Zylon
    Zylon:
    It's always fun when the picture for an article spoils the punchline.
    Wait, what? The picture? It gives away the punchline? Really? Damn.
  • Meee (unregistered) in reply to Kuba

    oh my god, i'd never pay you to put a weighbridge in for me and/or software.

    unless you are willing to loose a few thousand on your ideas, thats just going overboard :)

  • (cs) in reply to m0ffx
    m0ffx:
    Kuba:
    With the engine turned off, and the driver standing on the weighbridge, you should be able to tell the driver's heart *and* respiration rates, too. Heartbeat thrusts a good deal of blood out the aorta, and that gives a very noticeable kick to whatever you're standing on. Furthermore, heartrate is modulated by respiration, so those two can be separated out. Perhaps an el-cheapo scale won't be running a decent enough data acquisition system for that, but these days a 24 bit ADC runs ~$10/channel and that's all you need to measure what I'm talking about. You should get roughly 1/200,000 resolution at a 50Hz sampling rate.

    All it takes is some signal processing.

    And some bullshitting. It's not about the ADC, what matters is the precision on the strain gauge itself.

    You think that accuracy and resolution are the same thing. It's a common misconception, since they are not.

    In order to detect a heart rate from someone standing on a strain gage scale you really don't care if the scale has only 1% accuracy. Accuracy is due to various nonlinearities in the system. Nonlinearities mean that when you put 1kg mass on the unloaded scale it may register as say 1.1kg increment, but when you put it on a scale with 10 tons on it, it may register as a 0.9kg increment. The slope varies depending on how much load there is already on the transducer, and that's what affects precision. The more linear (less slope variation) your system is, the more accurate it is.

    Resolution almost always is at least an order of magnitude better than accuracy, if you have any sort of a "decent" A/D converter system.

    Which isn't going to be 6 decimal places - while that's available, it's needlessly expensive.

    A strain gage sensor with 6 decimal place precision (1ppm nonlinearity or better) is, to say it politely, untrivial. I don't even think they are "available" at all in off-the-shelf packages.

    Analogue noise in the system, prior to the ADC, will overwhelm the signals you're trying to detect. And then there's air motion.

    Air motion on a heavy (think ton+) weigh bridge will cause quite low frequency vibration that can be trivially removed. When you look for a heart rate, you look in a quite narrow part of the spectrum - perhaps 40 to 180Hz. A correctly designed strain gage conditioner will have 1 in 200,000 noise-free resolution with a couple Hz (or tens of Hz) of bandwidth. Heck, it's not even such a big deal, it's in Linear's application notes.

    Working on it a bit you can extend the bandwidth to say 200Hz, plenty enough to handle heartbeat. This is all on a system that has extremely good accuracy on the electronics side of things -- way more than needed for a weighbridge. Downgrading to an audio ADC, which is poor in DC, you can get even more resolution.

    The electronic noise thus isn't a problem, in general. Mechanical noise, faithfully transducted by the weighbridge, is of course a problem, but that's something that can be handled by properly characterizing the natural frequencies of the weighbridge itself, and applying some signal processing.

    I posit that with proper mechanical design you can have a truck weighbridge that can detect driver's heartrate, when said driver is standing directly on the bridge (not sitting in the truck). The only reason I'm fairly sure it could be done is that I've done it on a 16 bit system with a car (Volvo 940 wagon) sitting on the transducer (one wheel only), and me in the car (that's harder than standing directly on the tranducer) ;)

  • (cs) in reply to Meee
    Meee:
    oh my god, i'd never pay you to put a weighbridge in for me and/or software.

    unless you are willing to loose a few thousand on your ideas, thats just going overboard :)

    I don't think that you can get a weighbridge for much less than a few thousand anyway. The cost of electronics would be negligible compared to the cost of the mechanical structure. Electronics are so cheap these days that it doesn't really matter if you do the least that's needed, or you overdesign a bit. A 24 bit A/D converter costs $10 per channel, and that's something that attaches directly to the strain gage, without any signal conditioning. Many higher-end scales use those, and if anything, using such an "overblown" converter can reduce overall cost and increase reliability.

    Or, in other words: don't dismiss something just because you don't know jack $excrement what you're talking about.

  • Chris (unregistered) in reply to Anonymous
    Anonymous:
    Inhibeo:
    TRWTF is the laser sensor. Surely the weighbridge can detect the presence of a large truck...by its weight?
    They're not just for weighing trucks. If you're weighing a car, or an empty container, you can't rely on a weight limit to detect its presence.

    In the US, it's common to note the weight at every axle separately, as the total vehicle weight is one of many weight-related restrictions. The distribution per axle, whether that's a two-wheel or four-wheel axle, etc. is also specified.

Leave a comment on “It's the Only Way to be Sure”

Log In or post as a guest

Replying to comment #:

« Return to Article