6

Last few months I am working on personal project -- I want to measure weight of beehive during the year and then save the results. Beehives has usually from 100 to 200 kg and during one day, the weight can change from -5 to +5 kg. The ideal precision should be 100 g.

I have working prototype with MCU, amplifier and one 20 kg load cell. I put beehive on mechanical platform and weight is scaled down in 1:10 ratio -- thus I can measure 200 kg weight. The precision has to be 10 g.

The results from my prototype are:

  • My cheap load cell changes its output with temperature (I expected this, I can correct the results when I save temperature together with weight).
  • The output of load cell drifts with time. I put 3 kg weight directly on load cell and after few days, the output was much lower than first day. I think this is called creep?

The second issue is problem, because load cell has to measure weight during whole year so it will be under permanent pressure.

Should I buy more expensive load cell? Are load cells designed for permanent load? Or should I look for another method how to measure weight?

I was thinking about precision multi turn potentiometer. I am not mechanical engineer, but if I can easily change weight into movement, I may be able to use potentiometer as weight sensor.

Or maybe use spring instead of load cell. But how can I convert spring deformation into electrical signal? Spring compress with more weight, so maybe some kind of distance sensor?

vasco
  • 957
  • 2
  • 15
  • 25
  • Don't know how they vary with temperature, but maybe a [Strain Gauge](https://en.wikipedia.org/wiki/Strain_gauge) in bridge configuration for your spring option. – jippie Feb 02 '14 at 16:40
  • 6
    Just a small nitpick: kg is mass, N is weight. – Ignacio Vazquez-Abrams Feb 02 '14 at 16:46
  • why not get a better loadcell. Alternatively, if you know how much the measured weight changes with time, use compensation. – Andy aka Feb 02 '14 at 17:05
  • If your load-cell isn't 4-wire, you need a better load cell. [wheatstone bridge](http://en.wikipedia.org/wiki/Wheatstone_bridge) is really the only properly reliable topology. I know spark-fun and other sell little three-wire load cells, but they're almost universally horrible. – Connor Wolf Feb 03 '14 at 13:20
  • 1
    @IgnacioVazquez-Abrams: Thank you, I was not aware there is difference in mass-weight. English is not my native language. – vasco Feb 04 '14 at 08:55
  • Right now, I am using 4-wire load cell, so my topology is Wheatstone bridge. I tried to use four 3-wire load cells from personal scale, but after few months they lost precision. – vasco Feb 04 '14 at 09:00
  • 5kg!? What kind of flow do you have these on, anyway? – Erik Friesen May 04 '14 at 01:53

2 Answers2

3

Here is one way to avoid load cells, bridges and all that with something possibly easier. Since you have a way to mechanically support your hive (Which sounds huge! Mine are not near that heavy even when three boxes full).

How about a simple balance and a water tank on the other arm? Pump water in or out to get balance. There are a lot of nice ways to measure depth of water in a container. You can also continuously fill till balanced then measure contents as it empties. A dump bucket as used in rain gauges could work for both in and out.

Compression of a spring with or without a lever is also good. Have you looked at the load cells that you hang something from?

Personally I like measuring ml of water to get grams - sealed of course or your bees will keep moving it into the hive :-)

C. Towne Springer
  • 2,141
  • 11
  • 14
  • Thank you. Mine hives are made from solid 2cm thick wood, size is 485x485 mm. I'm using 4-6 boxes. The idea with water container is interesting, I'll think about it. What do you mean by 'load cells that you hang something from'? – vasco Feb 04 '14 at 08:51
  • I was thinking of using a cell under tension of this type http://www.transducertechniques.com/tll-load-cell.aspx and hanging the hive from an overhead support. – C. Towne Springer Feb 04 '14 at 16:01
0

Like a apothecary or laboratory balance you would unload the cell from the pan between measurements.

Having no load eliminates creep.

Other errors can be detected by alternately loading a reference mass or your hive.

A tiny gear motor that turns small cams that select scaled hive lever or reference mass alternately or neither in rest phase. Run motor one turn to take a reading.

With cam in rest position rough handling of the hive will not stress the load-cell.

A motorised balance arm might also work as you do not need to take fast readings. Move the pivot point or counter mass and calculate the hive mass like a doctors balance, or a trade weighing platform.

A proper mass balance will usually have less drift than a scale (load cell or springs) over time.

EDIT:
Some cool ideas to be seen on Google image search.

An alternative might be to use the simplest bathroom type scale interfaced to your readout circuitry and then have a gear motor load and unload a 1 to 10kg reference mass before or after your reading in the false bottom with the bathroom scale. This will likely be much cheaper than other more elaborate mechanical arrangements.

KalleMP
  • 4,039
  • 1
  • 14
  • 32