Measuring the weight of a beehive

LEGO-lars:
2. upload the measurements to Cosm.com using a gprs-module. (I just smoked my gprs-shield so this can take some time :))

Thanks for this update! May be this combined board is usable: http://imall.iteadstudio.com/im120411004.html It is an Arduino, GSM, XBee, SD-Card on one board. It is cheap, but I do not know how power saving can be done.

Perhaps the new Arduino GSM shield is more power saving and battery compatible but there are some inconsistent information atm:

When <400 mA peak current is correct this would be an interesting board. But you have to pay nearly 100 Eur for the board only!

Lars, what resolution do you have with your scale? We had some ideas and came to the solution not to use an amp but instead an external ADC with a higher resolution than the 10bit Arduino, do you use an other board (e.g. the DUE has a 12 bits resolution)?, see http://hackerbee.com/2013/03/18/100-hive-losses-this-year/ or did we forget something? and a resolution of max load / 1024 is with the INA333 possible? I think it amplifies only the signal or does it do any digitalization?

Hi Clemens :slight_smile: Thanks for the link. A lot useful information on that page.
You are right in that the arduino 10 bits of resolution is a bit low.
However, there is a way to increse the resolution of the ADC called oversampling. http://www.atmel.com/images/doc8003.pdf
In short, 16 10-bits sampeles are equal to one 12-bit sample.
In my scale, I have no problem to get a stable resolution of 0.1kg.

http://dl.dropboxusercontent.com/u/128833299/bees.htm

Uses 50kg Load Sensor PPSEN-10245 for weight monitoring.

Photo Interrupter GP1A57HRJ00F to count the bees going in and out of the hive.

What has happen to this project?
Is there any farther progress?

Hi toadstul.
I have been quite busy with some other projects lately (and got a bit bored :blush:), so nothing has been done on this project in the last 6 months.
However, before I took a break I was doing great progress.
I've got the PCB for the wireless scales manufactured. I've got the the power consumption low enough for the scale to work several seasons before battery change is necessary. I have made the base station with the GPRS-module, a wether station, and a solar panel.
I've also got the whole setup working on my workbench, and uploaded data to Cosm.com.

What remains is to make the actual scale that fits under the beehive.
I also have some problems with the wireless trancievers (probably due to a designflaw on the PCB :blush:).
Then I have to make everything watertight.
On the computer side, I would like to download the data into a spread-sheet, and use some algorithms to calculate which beehive needs attention.

I plan to start up again this winter when I will have some more spare time.

I could probobly post some pictures and explanation of the current setup if there is any interest.

LEGO-lars

Measuring bee-hives? :fearful:

NOPE. Nope nope nope. nope.... nope...

shudder nope!

Is there any particular reason you bother yourself with power consumption as you can put a car battery with a solar panel charger to rid yourself of power consumption problems.? It's not like you'll leave your bees to themselves for two seasons? :slight_smile:

Or am I missing something?

The two main reasons for the fiddling with the power consumption, is cost and flexibility. If this system turns out to work, I would like measure all our 200+ hives (my father works full time as a beekeeper). Therefor the cost of each weightscale is critical. Powering the each scale with 2 AAA batteries would be a lot cheaper then using a car battery and solar panel, even if I could power several scales from the same battery.

The second reason is flexibility. When the scale is completely wireless, everything can be hided under the hive with no wires lying everywhere to be tripped over. The hives are also moved a couple of times during the season (to get more honey), so rigging the system up and down have to be easy and fast.

A last benefit with my wireless modular system is that it is easy to add or remove modules from the network. The plan is to let the base station in every bee-farm auto detect the wireless scales without intervention from the user.

Lego-lars

Too bad you can't just use one of these:

Imgur

(Or can you?)

Hi LEGO-lars
Sorry about the delay of my answer(i was changing isp)
What method did you follow to lower the power consumption?
What chip did you use to amplify the signal from the loadcell?
You say ''I plan to start up again this winter when I will have some more spare time. I notes that you are from Norway, I think you already have winter up there :stuck_out_tongue: !! So please start again your project asap.
Pictures and explanation...? of course there is interest !

On my side I have buy already tow arduino uno, tow loadcell 160kg, tow ina125, tow DS3231 rtc, tow DHT22/AM2302 Digital Temperature And Humidity Sensor, and i am waiting an Arduino GSM Shield (antenna connector).
I plan to connect all these together and get daily reports of weight to my mobile. That will help me to know when the nectar flow stars or ends.

Power consumption can be done mainly by changing to an barebone Arduino setup, think the most cited article about this is http://gammon.com.au/power .

The initial outcome is that, to save power, forget about using a development board. Further savings (like reducing clock speed) would be overshadowed by the huge overhead of the voltage regulator and USB interface chip.

I think Lars used as OpAmp an INA333 as linked in an older posting above: http://bildr.no/view/1409684 and no additional ADC so there is really an option for better reading than 0.1 kg resolution.

I see this two options: Use an additional ADC plus OpAmp / PGA or a chip who has both integrated, see my posting under http://forum.arduino.cc/index.php?topic=198139.0 - so you see, you are not alone. :slight_smile: you may also have a look at the Seeeduino Stalker who has a lot of build in (charger circuit, RTC, SD) http://www.seeedstudio.com/wiki/Seeeduino_Stalker_v2.3

Hi again :slight_smile:
Here are some pictures of my project so far.

This is the wireless sensor node. The Atmega328 is running at 1 Mhz internal ocilator. I am using a 32 khz clock-crystal connected to the xtal pins and timer 2, to make a the AVR work as a RTC.
To amplify the loadcell I am using the ina333 op-amp and no additional ADC, but with some oversampling and black magic, I get something that looks like 12 bits of resolution :slight_smile: As clements has pointed out, it is possible to use an external ADC, and probably get a better resolution. In my case I could not find anyone that works over the whole voltage span of two AAA batteries.
A P-channel mosfet cuts the power to the loadcell and op-amp when they are not needed.


This is the prototype for the base station. It is using the sim900BE GPRS-module, and have interface for a DHT22 temperature/humidity sensor, and a rain gauge. It also have a solar charge controller.


This is a prototype of the actual scale that fits under the bee-hive. The plan here is to only measure one side of the hive and expect the other side to be equal. This would make the scale much easier to make, and I could use a smaller and cheaper loadcell. If this method is accurate enough, remains to be seen.

Just wondering how much the bees weigh and if there's times when they're all inside like at night.

A single bee has a weight of about 0.1 g. So a big hive has 3-4 kg bees. But most changing factor relating to weight is nectar flow and evaporation. Bees bring nectar to the hive with 80 % water. Honey has in the end under 20 % water. The difference must be reduced by the bees.

Hi Lego-Lars
Have you got at suitable loadcell for you bi-hive weight ?
I am very interested to here more about your project. Maybe we could help each other. I have acces to very cheap 120 kg full bridge single point loadcells In a month or so.
Is I possible to here more about you have build your hardware, and maybe a little about your software.
I am a be-keeper too, here in DK.
Sorry for my English, it's a little rusty!

Hi Lars, there is a recommendation for an ADS1231 as ADC: 24bit and 128x PGA sounds good to improve accuracy a lot:
http://forum.arduino.cc/index.php?topic=198139.0

I've seen a 12 bit ADC home game controller unit that's already got instructions to use short shielded wires and needs care to not get garbage readings. What care must be taken past 16 bits? They aren't snap-together Lego blocks.

Hi,
how Long stays the battery?
more than one month?

Best regards

Hello,

I am working on the same project! I started in July 2013 and at first, I used 4 load cells from cheap digital scale. I connect them together to form Wheatstone bridge, excitation voltage and amplification was done by INA125, analog-digital conversion with internal ADC in ATMega328P. For wireless communication, I used ZigBee at 2.4 GHz. My results were:

  • Do not use cheap 3-wire load cells. They drift with temperature and time. They lost precision. Creep is problem.
  • Do not use 2.4GHz if you need to communicate over 100 m. This frequency if full of Wifi/Bluetooth/noise. Everything from grass to trees will reflect your signal.
  • INA125 is good, but expensive.
  • 10bit ADC in ATMega is good, but not sufficient.

In my second version of electronics, I used 868MHz (yes, I'm from Europe) mesh modules from RadioCrafts (they call it TinyMesh). Maximum transmit power is 9dBm. I am testing them last few months and it looks very promising. I am able to communicate for distances over 150 m. Frequencies 868 or 433 MHz are much better in those projects.

The electronics is powered from 2AA batteries via step-up converter. With this, I'm getting quite stable 3 V. Also, I removed RTC and external crystal and for waking up MCU from power-down I'm using watchdog in interrupt mode. Thus, MCU will wake-up every 8s.

Instead of 4 small load cells, I used one 4-wire load cell (LEGO-lars used similar). This is my last problem with this project. The reading from ADC is not stable and it drifts with time. I put 2 kg on my 20kg load cell and after 2 weeks, the readings from ADC were lower then first day. Do you have similar experience with load cell?

Now, last month I am working on 3rd version of electronics. The digital part remains the same, but I completely redesigned analog part. I removed INA125 and used precision opamp MCP6V07. They have very low temperature drift and low noise. Instead of internal ADC inside of ATMega, I pick MCP3551 -- 22bit sigma-delta, low-speed 13sps. This new version of electronics will contain 2 opamp and 2 ADC, so I'll be able to measure 2 beehives with one electronics.


I am not fan of load cells. I would like to use multi-turn potentiometer. Unfortunately, I'm not mechanical engineer and don't know how to change weight into movement. It would be necessary to use some kind of lever and spring and gears to change small changes in spring into several turn of potentiometer shaft. Not sure if this is doable.

This is my second version of electronics.