Need Bees Removed?
International
Beekeeping Forums
November 27, 2014, 05:57:44 AM *
Welcome, Guest. Please login or register.
Did you miss your activation email?

Login with username, password and session length
News: 24/7 Ventrilo Voice chat -click for instructions and free software here
 
   Home   Help Search Calendar bee removal Login Register Chat  

Pages: [1] 2  All   Go Down
  Print  
Author Topic: Hive Data Recorder  (Read 4199 times)
The Bix
House Bee
**
Offline Offline

Gender: Male
Posts: 427


Location: Arapahoe County, Colorado


« on: April 04, 2011, 11:51:57 PM »

Hi,

Below is a link to a chart that displays data recorded from one of my hives regarding weight and temperature.  A friend of mine with a real engineering background built this contraption which records temperature (inside and out) as well as the weight.  The microprocessor polls the information every ten seconds.  The "trendline" function was used in Excel to remove the "fuzzy caterpillar" effect with so many samples every minute.  Let me know if you want the raw data.  Would love to hear what you think.

--John

PS, the original document is a .pdf, trying to link it here: http://img842.imageshack.us/i/beedata.pdf/
« Last Edit: April 05, 2011, 06:39:38 PM by buzzbee » Logged
BlueBee
Galactic Bee
******
Offline Offline

Gender: Male
Posts: 4398

Location: Mid Michigan


« Reply #1 on: April 05, 2011, 01:25:04 AM »

Candy, candy, candy!

Very cool John, please tell us more.

What are you using for a weight sensor?  How are you getting the data from the micro into your computer?  If you happen to know what micro your friend used, that would be cool to know as well. 

Where is your inside temperature being measured?

Thanks for sharing.  Iím working on a similar system, but Iíve been bogged down the entire month of March building wooden ware are getting ready for springÖ.if weíre going to have one this year.  Iíve got my hardware designed, but I still have a lot of code to write for my system.  One design is using a Atmel XMega16 due to its precision ADC whereas another design is using a lower cost NXP 32bit ARM micro.
Logged
The Bix
House Bee
**
Offline Offline

Gender: Male
Posts: 427


Location: Arapahoe County, Colorado


« Reply #2 on: April 05, 2011, 08:54:15 AM »

Candy, candy, candy!

Very cool John, please tell us more.

What are you using for a weight sensor?  How are you getting the data from the micro into your computer?  If you happen to know what micro your friend used, that would be cool to know as well. 

Where is your inside temperature being measured?

The weight sensor is a load cell which translates the force of the weight into voltage...it's the same type of technology we used to measure rocket motor thrust Smiley

The microcontroller saves the polled data in a .csv format onto an SD memory card, like the one for your digital camera.

The outside temp sensor is on the north side of the hive so it never sees the sun.  The inside is right in the middle of of the hive, we laid the sensor atop the top bar of the #5 frame in the bottom box.  I run double deeps over the winter and haven't added any supers to this hive yet.

--John
Logged
deknow
Field Bee
***
Offline Offline

Gender: Male
Posts: 752


Location: Massachusetts


WWW
« Reply #3 on: April 05, 2011, 10:41:17 AM »

...I've looked into doing something similar.  the roadblock (for me) was finding a sensor I could use for weight that would stay calibrated under a continuous load and changing temperatures.  ie....if you lift the hive off the sensor (may not be practical to do depending on the setup) at any time of the year, are you getting a consistent weight of "zero"?  If so, I would love to know more details about what is being used (manufacturer, specific part numbers).  Thanks!

deknow
Logged
BlueBee
Galactic Bee
******
Offline Offline

Gender: Male
Posts: 4398

Location: Mid Michigan


« Reply #4 on: April 05, 2011, 12:10:13 PM »

Iím with deknow, I was also under the impression that load cells would have to be recalibrated.  Iíve never used one though, so I could be blowing hot air.

I was pondering a homemade (electronic) balance beam setup instead, but Iím interested to learn more about the sensor youíre using.

I like the SD and CSV for data xfer.   My XMega16 hardware is designed for a RS-485 wired link back to my hives (about 300í feet away) so I can view real time data on the computer when I want, however that does require a lot of wire.  The 32 bit ARM design uses a LPC1342 with a built in USB peripheral and I was planning to use that on my nucs; more for controls rather than instrumentation.
Logged
The Bix
House Bee
**
Offline Offline

Gender: Male
Posts: 427


Location: Arapahoe County, Colorado


« Reply #5 on: April 05, 2011, 02:57:12 PM »

deknow and BB,

All those questions are far too technical for me.  You guys are so much smarter.  My reason for posting is that I just thought that some might be interested to see the results and speculate/discuss what the data mean.  I like the data and I'm looking forward to watching the trends going forward.  I plan to use this tool to help identify the nectar flow and figure out when to drop the supers on.  Would love to hear what anyone thinks about that.

--John
Logged
deknow
Field Bee
***
Offline Offline

Gender: Male
Posts: 752


Location: Massachusetts


WWW
« Reply #6 on: April 05, 2011, 03:20:16 PM »

well
i asked the question about what sensor is being used for 2 reasons:
1.  i'd like to build such a device and was not able to find an apprpriate sensor for continuous use
2.  if an inapprpriate sensor is being used i would question the data
is it possible to ask the person who built it?
Logged
edwardw
New Bee
*
Offline Offline

Posts: 21

Location: Fort Collins, CO


« Reply #7 on: April 05, 2011, 03:54:23 PM »

The person who built the device is kind of a tin foil hat wearing recluse.  Doesn't get out much. And, like John, is much more interested in the meaning of the data rather than the method of getting the data.

Edward
Logged
BlueBee
Galactic Bee
******
Offline Offline

Gender: Male
Posts: 4398

Location: Mid Michigan


« Reply #8 on: April 05, 2011, 04:04:56 PM »

John, first of all we are NOT smarter than you.  You are a ROCKET SCIENTIST!  Seriously, Iím impressed  Smiley

Iím also looking forward to seeing your data.  A couple of observations of your first data set was the outside temp seemed to be reading 90F+ for a while.  Iím assuming that sensor got some sun?  I was also a little surprised to see it so warm in the hive at night.  Maybe that was due to the sensor being fairly close to the brood.  

Science is all about fitting data to theories, and vice versa, so itís great to have data.

Logged
The Bix
House Bee
**
Offline Offline

Gender: Male
Posts: 427


Location: Arapahoe County, Colorado


« Reply #9 on: April 05, 2011, 04:36:48 PM »

A couple of observations of your first data set was the outside temp seemed to be reading 90F+ for a while.  Iím assuming that sensor got some sun?  I was also a little surprised to see it so warm in the hive at night.  Maybe that was due to the sensor being fairly close to the brood.

I think you're right about the outside sensor...the late afternoon sun may be catching it as the side of the hive actually faces NW rather than N.  I'll check that out tonight when I get home and make sure it's shaded.

The brood nest is moving upward, but most of it is still below the sensor, at least it was on Saturday. Smiley
Logged
BlueBee
Galactic Bee
******
Offline Offline

Gender: Male
Posts: 4398

Location: Mid Michigan


« Reply #10 on: April 05, 2011, 06:01:43 PM »

Supposedly, you should see a solid 95F when the brood is over the sensorÖor so Iíve been told.  Maybe youíll be able to track the movement/expansion of your brood chamber by just looking at the instrumentation.

I think youíll learn a lot from your instrumented hive even if the weight sensor isnít perfect.  Something is usually better than nothing!

John, is the instrumentation on your super strong hive, or is it another hive?
Logged
deknow
Field Bee
***
Offline Offline

Gender: Male
Posts: 752


Location: Massachusetts


WWW
« Reply #11 on: April 05, 2011, 06:25:22 PM »

poor data is unlikely to lead one to (or support) valid conclusions.  it would be nice to know how accurate this data is or isn't likely to be over time.

deknow
Logged
edwardw
New Bee
*
Offline Offline

Posts: 21

Location: Fort Collins, CO


« Reply #12 on: April 05, 2011, 06:39:34 PM »

Due diligence in design, testing and implementation of the data recorder has been done.  

If you are looking for a reading that is accurate and precise to .01 gram, well, you aren't going to find it.  If you want a reading that is accurate and precise to .1lb, that is much easier to implement.


Edward
Logged
BlueBee
Galactic Bee
******
Offline Offline

Gender: Male
Posts: 4398

Location: Mid Michigan


« Reply #13 on: April 05, 2011, 07:02:25 PM »

I did take a look at load cells on Digikey and found this one that looks promising, but it's expensive. 
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=MSP6953-ND
You can find the data sheet on that link and review it. 

It looks like it would provide data accurate to within 3% over temp and 1% error in linearity.  However I canít decipher from the data sheet how much drift there is from the zero point.  I also canít figure out why a Wheatstone bridge is so expensive.

Edward are you the guy in the tin foil hat?  The man behind the curtain?  Good job!  Can you tell us what load cell/sensor you are using?  Does the zero point in these load cells drift over time if they are under constant load/mass?
Logged
edwardw
New Bee
*
Offline Offline

Posts: 21

Location: Fort Collins, CO


« Reply #14 on: April 05, 2011, 07:20:54 PM »

A Wheatstone bridge is so expensive because it is in a value added product.  Someone had to produce very accurate film resistors that could be bonded with metal without de-laminating. Then they had to use repeatable steel/aluminum so it would be repeatable.  Finally they tested it over a wide range and gave you a calibration curve and told you how it reacts when you loaded it.  It isn't simply 4 99 cent Radio Shack resistors wire wrapped together.

If you move away from Digi-key to the industrial world where they weigh things a lot (Concrete batch plants, grain hoppers, etc.) you'll find load cells that will work for this application.  If you load the load cell up to full scale, let it sit and then take the load off you'll find how much it creep there is.  In this case, there is less than 2 pounds creep in 500lbs.  Acceptable for weighing bee hives.

Edward
Logged
BlueBee
Galactic Bee
******
Offline Offline

Gender: Male
Posts: 4398

Location: Mid Michigan


« Reply #15 on: April 05, 2011, 08:49:56 PM »

Sounds good Edward, can you tell us where to find industrial load cells?

I acknowledged early on I donít know squat about load cells.  So let me ask you this.  What is the difference between the load cells used in the industrial world vs the ones found in my $19 scales? 
Logged
deknow
Field Bee
***
Offline Offline

Gender: Male
Posts: 752


Location: Massachusetts


WWW
« Reply #16 on: April 05, 2011, 09:15:24 PM »

Hi Edward,

I can understand if you want to keep your parts list proprietary.

2lb creep over 500lbs sounds just fine for this application.

without some test data (or at least a data sheet for the part) however, i remain skeptical.  the applications you cited don't generally require that a sensor be under load for long periods of time (months) and remain accurate without re-zeroing (or recalibrating).

deknow
Logged
edwardw
New Bee
*
Offline Offline

Posts: 21

Location: Fort Collins, CO


« Reply #17 on: April 05, 2011, 09:23:36 PM »

A Google search reveals many suppliers.  Omega and Honeywell are some of the bigger suppliers.  There are many smaller ones that do a good job as well.

Thee difference is that your $19 bathroom scale load cell probably only has a maximum capacity of 25 pounds.  And it isn't rated for a wide temperature range or very accurate. It also depends on the type you are looking to implement.  Compression (pancake type), bending beam or shear beam.  All have pro's and con's.  It isn't about keeping the parts list proprietary, it is more about people wanting to be handed a silver spoon and doing no work.  I did my research, found the part and tested it. 

As for how they perform, these are from the same supplier that a lab at my local college uses.  They built a single story house on a platform using these load cells and kept continuous weight measurements and then after it was built they saw how wind loading was transferred through the frame.  You can't build a home in a day and these were under load for over 6 months.  I think the creep was inconsequential.

Edward



Edward
Logged
BlueBee
Galactic Bee
******
Offline Offline

Gender: Male
Posts: 4398

Location: Mid Michigan


« Reply #18 on: April 05, 2011, 09:49:21 PM »

Fair enough Edward, but can you disclose how much a load cell appropriate for a hive costs?  Are we talking $20, $50, $100?
Logged
edwardw
New Bee
*
Offline Offline

Posts: 21

Location: Fort Collins, CO


« Reply #19 on: April 05, 2011, 09:52:30 PM »

Under $50 should get a very capable load cell.  You won't be getting a 500# one for that range, but if you use some mechanics you can get it to work very nicely.

Edward

Logged
Pages: [1] 2  All   Go Up
  Print  
 
Jump to:  

Beemaster's Beekeeping Ring
Previous | Home | Join | Random | Next
Powered by MySQL Powered by PHP Powered by SMF 1.1.20 | SMF © 2013, Simple Machines | Sitemap Valid XHTML 1.0! Valid CSS!
Page created in 0.775 seconds with 22 queries.

Google visited last this page November 21, 2014, 06:01:52 PM
anything