Go Down

Topic: Measuring the weight of a beehive (Read 64559 times) previous topic - next topic

Trottel100

Hi,
i want only get the information over the Internet side, i have not much time, here is summer and the bees Need time and i have a normal Job, i would look to force the side.

best regards.

GoForSmoke

If you don't really need instant information and have someone visit the hives on a regular basis, it would be far easier and cheaper to log your data to an easily swapped SD card. You could probably train a monkey to do it so a student would get it right about half the time.
1) http://gammon.com.au/blink  <-- tasking Arduino 1-2-3
2) http://gammon.com.au/serial <-- techniques howto
3) http://gammon.com.au/interrupts
Your sketch can sense ongoing process events in time.
Your sketch can make events to control it over time.

Clemens


This may not be the right place to ask - but what is the best board I could use to achieve 1 input (weight reading from the hive from Shear Beam Load Cell), data send via GSM shield and ability to be solar charged? I have looked at the Arduino mini on the understanding that they use much less power however they do not seem to be compatible with standard shields and may not support GSM? All power savings I have seen appear to be for barebones custom boards/breadboards, I am not technical enough to solder a custom board and would rather buy a ready-made but basic board that will support the aforementioned functions.

Are there Arduino clones that support the same form factor for shields, but are configured for low power consumption? If so, what are they?


I think you do not need a board with a standard form factor for shields because there are no shields you will use. ;-)

The best hardware solution--without soldering / making the PCB by your own--could be the

  • Seeeduino Stalker
    http://www.seeedstudio.com/wiki/Seeeduino_Stalker_v2.3 and the

  • GPRSbee
    http://www.gprsbee.com


and you can use the

  • HX711 Weigh Sensor breakout
    http://imall.iteadstudio.com/hx711-dual-channel-weighing-sensor-module.html
    (similar board from DfRobot http://www.dfrobot.com/wiki/index.php/Weight_Sensor_Module_V1)



for reading the load cell.

NZbeek

Thanks a lot for the reply Clemens, what is the other benefit of the Seeduino over Arduino Uno for example apart from being 'plug and play'?

I had been thinking of the following.

Arduino Uno with Power Saving library by Rocket Scream (found this at comoyo.github m2m adventures measuring temperature of water and air, this seems to be in line with my project except I will record weight) Nick Gammon appears to offer other techniques to save power but it is too complex for me, so unless I need greater power reductions I won't go down that path. Hopefully a solar trickle charged battery that powers the Arduino, suggestions?

I will use this shear beam load cell http://www.ebay.com/itm/371036128442?var=640213319397&ssPageName=STRK:MEWAX:IT&_trksid=p3984.m1423.l2649
I know people talk about creep, temperature compensation and distortion due to long term loading on load cells. Can you offer advice on this particular cell? It has specs on the page but I don't know how to interpret, it talks about creep and temperature, are those figures ones I will need to use for calibrating or does it mean within those ranges, temperature for example, the load cell will be accurate?

I had intended to use one of the following MCP3551, ADS1231 (I think you used this Clemens) or INA125P - can you help tell me the difference between these, if any?! So far I am leaning towards 1 of the MCP or INA only because I have seen others use it so should be able to wire it up OK. Any advice here?

Finally, the most expensive part is a GSM shield, I haven't decided on what one yet, but hopefully can adapt the code M2M used for sending the data and I will use something like xively.com to store and present data (unless there is another suggestion here too?). Haven't looked too closely yet but there seems to be good information at cerulean.com and liubo.us on load cells and arduinos.

The way I would like it to work is for the weight to be recorded every hour between 5am and 7pm but to only send the data once at say 8pm, if it is too difficult to do an hourly reading and to send  then one reading at 8pm is fine.

I have lots of good references, can I just put libraries/code all together? Or will it be far more complicated. Are there any obvious problems with what I have outlined? What is the best way to start?

Thanks all, sorry for the long winded post!

NZbeek

Also (just making the most of the time difference) - instead of using watchdog which only allows an 8s sleep, would it be better using a real time clock to sleep it for the 1 hour? - assume this would be more energy efficient.

I thought I would use a 12v sealed lead/acid battery - would it be likely that I would even need to trickle charge by solar if the scales were only being used for maximum 4 months? - more likely 2 - 3 months. I realise that the on board voltage regulator on the UNO will 'discharge' any additional energy supplied, can anyone suggest the best external voltage regulation? - presume I would only need to use a step down regulator, but perhaps there is a switching regulator available that could switch up or down as needed?

If i can find answers to even half of my questions I will be doing well!

Thanks all

GoForSmoke

Big enough battery, you could run the rig for a year as long as it didn't have to keep itself warm too.

Do you leave the hives unvisited for 4 months? Could you top up a smaller battery from the electrical system of a small vehicle once in a while? Solar panels might need occasional cleaning from pollen or bird droppings or who knows until it happens over months of time, but hey they are used and do work to power a LOT more than you need.

That's why I say, if you have someone checking the hives regularly, you could save out the whole GSM thing for less than $10 a unit.
It's not just the money either. GSM raises power requirements and adds a lot to the software. What happens if ET calls home and no one answers? ET has to be able to call again, with the data, until that job is done. Just being able to make the call means a library or more to fit in. That's already been done elsewhere, it's possible, but not a complete no-brainer if you're doing your own custom project -- just some non-trivial thing more.

Is there some golden moment that if you catch in a day or week to be on the scene will pay for the extras?

BTW, from news I see we have Japanese wasps killing hives here. Are the wasps enough bigger that a screen could lets bees through but not wasps and would the bees navigate such a screen? People here make bird houses with small holes just to keep bigger birds from robbing bluebird nests as bluebirds eat mosquitos. A fine wire mesh might do the same for bees.
1) http://gammon.com.au/blink  <-- tasking Arduino 1-2-3
2) http://gammon.com.au/serial <-- techniques howto
3) http://gammon.com.au/interrupts
Your sketch can sense ongoing process events in time.
Your sketch can make events to control it over time.

Chagrin

Grumpy Mike wrote the RadioHead library which would let you set up a low cost mesh. Aside from your "manager" nodes I would imagine that you would be able to power most of your hives with an inexpensive solar panel and small battery -- using the Arduino itself to do simple charge management.

With respect to the load cell you select make sure you have a credible way to mount it. A shear beam cell with a single point of contact wouldn't be practical unless you had the hive mounted on a pivot. Also, shear beam cells need rock-steady mount points (steel/concrete); if you try to screw the fixed end into wood I would expect it to start pulling its screws out and thus require constant recalibration. As an alternative you can get three-wire load cells (cheaper on eBay) like those that are used in bathroom scales, mount them in the corners of the hive, and would be able to wire the four of them together into a full bridge for a more practical mounting solution.

NZbeek

I had thought that the load cells as used in bathroom scales suffered more from the effect of creep and distortion?

The frame for mounting will be stainless steel. Again, to visit some of these hives means a 5 hr return journey, the hives at times will be left for a fortnight, but if the honey starts to flow heavily in that period then a revisit to super sooner will mean more honey collected in a box that we can more easily harvest (rather than the bees packing it down in their brood boxes).

I am not familiar with the Japanese wasp, in New Zealand we have German wasps, which can be a problem. Generally entrance guards, which reduce the size of the hive entrances, are sufficient but still requires a strong hive for defence. If there is a size diffence then certainly this could work - size is the principle for keeping queens out of honey supers, it is also used in pollen trapping where the bee can fit through but not the pollen grains on its legs.

Can i expect issues using the HX711 premounted? - it was suggested to me that it might be more temperature sensitive than other chips like those I mentioned??

Clemens

#68
Jun 26, 2014, 12:08 am Last Edit: Jun 26, 2014, 12:16 am by Clemens Reason: 1

Thanks a lot for the reply Clemens, what is the other benefit of the Seeduino over Arduino Uno for example apart from being 'plug and play'?


34.5 mA.(see "Baseline-Arduino Uno" http://www.gammon.com.au/forum/?id=11497 ) vs.  600 uA (http://www.seeedstudio.com/forum/viewtopic.php?f=16&t=3911) plus you have RTC, SD card, solar charger circuit, bee socket

Quote
Arduino Uno with Power Saving library by Rocket Scream (found this at comoyo.github m2m adventures measuring temperature of water and air, this seems to be in line with my project except I will record weight) Nick Gammon appears to offer other techniques to save power but it is too complex for me, so unless I need greater power reductions I won't go down that path. Hopefully a solar trickle charged battery that powers the Arduino, suggestions?


LowPower lib makes more or less the same as described in Nick Gammon article. Consider to use Narcoleptic lib if you will use time information (e.g. without RTC) it recalculates millis for more accuracy but use the right version, see http://forum.arduino.cc/index.php?topic=121096.msg1768493#msg1768493 . About the charging approach have a look at the Seeeduino Stalker  spec http://www.seeedstudio.com/wiki/Seeeduino_Stalker_v2.3 be aware that it is the "Stalker" not an other Arduino clone from Seeeduino.

The most load cells are temperature compensated, I don't know how problematic creep is. For the accuracy we need to measure hive weight is seems to be not problematic. But please read the whole thread there are useful information and a hint to use not 3 wire load cells.  

Quote
I had intended to use one of the following MCP3551, ADS1231 (I think you used this Clemens) or INA125P - can you help tell me the difference between these, if any?! So far I am leaning towards 1 of the MCP or INA only because I have seen others use it so should be able to wire it up OK. Any advice here?


Be aware that these are different ICs the INA is an OpAmp while the ADS and the HX711 has a gain amplifier but also a ADC with a higher resolution than the Arduino's internal ADC (10 vs. 24 bit). But read the post from Lars in this thread http://forum.arduino.cc/index.php?topic=113534.msg1173511#msg1173511 about using oversampling to reach 12 bit in case you will an OpAmp only.

Quote
Also (just making the most of the time difference) - instead of using watchdog which only allows an 8s sleep, would it be better using a real time clock to sleep it for the 1 hour? - assume this would be more energy efficient.


The 8 vs. 60 x 60 sec seems not to be that problem if you have the right board, see http://lowpowerlab.com/blog/2012/12/24/moteino-coin-cell-battery-tests/ how about 3 years with a 2 x AAA pack? ;-)

Clemens

The frame for mounting will be stainless steel.


Seems to be a good decision if you have the ability to weld it. You can have a look at the professional hive scales like this one: http://www.beewatch.biz/index.php/stockwaage-41.html It is also possible to use two alloy plates like this http://www.emsystech.de/wp-content/gallery/penso/penso-flex.jpg and mount a case on it http://www.emsystech.de/wp-content/gallery/penso/penso-flex-beispiel.jpg. But I think the best mechanical and easiest way is to use the load cell only on one side as Lars did http://forum.arduino.cc/index.php?topic=113534.msg1473176#msg1473176 and take the load twice.

Chagrin

The most load cells are temperature compensated, I don't know how problematic creep is. For the accuracy we need to measure hive weight is seems to be not problematic. But please read the whole thread there are useful information and a hint to use not 3 wire load cells.


If you're talking about "creep" due to changes in temperature, yes, a three-wire load cell used singly will not be accurate. When you use them in pairs, however, the temperature effects are equal within each three-wire load cell, the change in resistance in the strain gauges will be cancelled out, and the "creep" isn't an issue -- or at least no different than a four-wire load cell.

NZbeek

I will research shortly, but I need a bit of a break from all the reading!- are there further power savings on the seeduino via library function? - or is it just so low out of the box that it isn't worth it? Am repairing other honey extraction now for a change.

For load cells I will probably build 2 identical side by side but use the 4 x cheap load cells as used in body weight scales, etc.

I dont think solar charging is going to be a requirement but will deal with it if and when I do, no point going down that path just yet and it seems as though the sealed lead acid will be surplus - just going to need to figure out a good voltage regulator now.

I have bought the HX711 for now, it gives me more flexibility for resolution and will be more straightforward I hope.

I dont see the advantage of using a one side mounted load beam opposed to the double end mounted? The beewatch configuration is exactly how I imagined building mine - with the spread of the "H" platform I would have thought the weight would register fairly accurately on the load beam? - exact weight is less import than weight trend, though an accuracy of +/- say 3kg would be tolerable.

NZbeek

Also, in this article
http://alanbmitchell.wordpress.com/2011/10/09/design-of-the-extended-battery-life-power-controller-for-the-arduino/

he designs a custom 'power controller' - is this essentially unnecessary if using an RTC module to control power on/off?

Gopster

NZBeek I am building a very similar set up to yours. However I will be using 2x 50Kg load cells in a full bridge and a HX711 to measure 1 side of my hive (1/2 of the weight).

I have purchased a DS3231 RTC that has an alarm to wake the arduino, I am assume this will use less power than the 8s sleep cycle. I am hoping that a 3600mAh Lithium battery will be enough to upload twice a day for a couple of months.

I am still waiting on the components to come from China. I think my code will be based on this http://arduino.cc/en/Tutorial/GSMExamplesXivelyClient or possibly the water temp sensor you mentioned earlier.

If i was to use the GSM example code could I just add the hx711 library, set the pin and configure the scale in the setup, then add this in the loop?
Code: [Select]
scale.power_up();
  int sensorReading =(scale.read_average(5));   
scale.power_down();


As an int will it be rounded to the nearest kg?

Sorry so many questions its just taking so long for things to arrive from China

NZbeek

#74
Jul 11, 2014, 12:10 pm Last Edit: Jul 11, 2014, 11:54 pm by NZbeek Reason: 1
Great questions gopster, unfortunately I am too much of a newb to answer. I have similar questions though, maybe some one else  on this forum could come up with some suggestions? It might be useful to post your Proposed code?

Bump

Go Up