Real-time hive scale [graphs]

[tmk]

Hi everybody,

I'm fairly new to beekeeping (started last Aug) and am one of those guys who loves data. I would read a lot of posts and books that would talk about doing this or that 'when a flow is on', but never any advice on how to detect nectar flows other than opening and inspecting.

Some folks had rigs which would let them weigh their hives, and this seemed like a good non-intrusive method, but it has a low sample rate.

I know a bit about electronics, so i decided to have a go at building a scale that would permanently live underneath the hive, and could be polled at any time. I built one, and thought I'd share some of the early results with the forum here.

It's been running for a week, and over that time it looks like the hive has gained about 8 lbs. Not saying it's all honey, just that it weighs that much more now than it did a week ago.

I'll do a full writeup on how i made it eventually, but it was fairly cheap for the scale itself (< $50), if you don't count all the tools involved, computer to poll it, etc.

Some pictures:

The hive. Scale is the yellow thing at the bottom:


A graph of the last 24+24 hours (day over day):


There's a bit of a summary over at my bee blog, and some more posts discussing it if you're interested.

http://hivetalkin.blogspot.com/2011/...-they-eat.html

I'll be keeping the raw data basically forever and if there's a lot of demand i'll find a way to post it somewhere that anyone can get to. It is all 1 minute samples.

I live in the silicon valley area if it makes any difference :)

-tmk

[jip]

Awesome. Please post detail plans on how to build one. I'm interested and I'm sure others are as well.

Also, perhaps you should keep a page on your blog just for the data. I've already bookmarked your blog to check regularly.

I'm not far from you, here's my blog http://losaltoshoneybees.wordpress.com/

[Oldtimer]

Hey that's pretty interesting TMK, I'll be keeping an eye on it.

One thing that could add to the value of it, would be if you put a marker on the graph when some major event happened. So, for example, if the hive swarmed, you'd put a marker with the word swarmed. Or similar.

Would be interesting to see the real effect of some of those kind of things.

[jrbbees]

I look forward to getting your plans on how you built it.
Can we also get the software you used to take the measurement.
It would be nice to work out a way to go to a yard slip the scale under one end and take a reading. I know it would be the entire weight but it would note change.

If you sampled 10% of the hive in a yard it would give you some kind of projection.

[DarbCal]

I love this and data. Please post plans.

[buhbee]

I have tossed around ideas about how to monitor the weight of my hives, but have not yet come up with a good solution. I like what you have done and would love to do this on at least one hive here.

Greg

[tmk]

One thing that could add to the value of it, would be if you put a marker on the graph when some major event happened. So, for example, if the hive swarmed, you'd put a marker with the word swarmed. Or similar.

If you go to the blog link, i have done that for the weekly graphs. I'm recording weather since that seems to have the most impact on the hive weight.

e.g. The day after rain it has gained the most so far.

week-graph with annotations

-tmk

[tmk]

Also, perhaps you should keep a page on your blog just for the data. I've already bookmarked your blog to check regularly.

I have been tagging the posts with graphs with the 'graphs' tag, and those that talk about the scale with the 'scale' tag.

you can filter posts by tag, eg:

http://hivetalkin.blogspot.com/search/label/graphs

-tmk

[AstroBee]

Very cool!! You should build another and place a known weight (100+ lbs) so that you can get a handle on temperature variance. Can these strain gauges handle a nearly constant (DC) load and still produce an accurate reading? Wish I had the same skills with electronics....

Keep us posted.

[tmk]

so that you can get a handle on temperature variance.

I have another method of temperature compensation, and it's actually working pretty well. The line for lbs is not even close to the raw scale readings :)

Can these strain gauges handle a nearly constant (DC) load and still produce an accurate reading?

I'm not sure, but this one isn't 'on' constantly. I have an electronic switch (FET) which turns it on when a reading is to be taken, and turns it off after, so the scale is only 'on' for 4 sec or so every minute.

-tmk

[Bear Creek Steve]

tmk,

Thanks for your work. Like you I think that data when appropriately gathered and displayed can tell a very interesting story. I'm a tinkerer, but unfortunantly not real strong on electronics/computers. I have an un-branded made in China 330 lb. X 0.2 lb. battery digital platform scale which might be useful.

Please continue with your work, ideas, and graphing.

Bear Creek Steve

[Nabber86]

If you go to the blog link, i have done that for the weekly graphs. I'm recording weather since that seems to have the most impact on the hive weight.

e.g. The day after rain it has gained the most so far.

week-graph with annotations

-tmk

Interestng diurnal fluctation on the weekly graph. I assume that is worker bees going out for forage and returning home in the evening.

You must be an engineer?

[tmk]

Interestng diurnal fluctation on the weekly graph. I assume that is worker bees going out for forage and returning home in the evening.

I was initially expecting to see a dip in weight in the morning as the workers leave, and a bump in the evening when they all call it quits, but i have seen neither. It may be that the outflux of workers is gradual - eg they scout and recruit, so no big dip. Also in the evening it does seem to be that they come in gradually and leave less and less frequently... no 'rush' hour to get home

There are 2 main patterns i've noticed:
* my fix-the-crappy-scale compensations lend a 'wave' to the graphs which move them by about .5lb. This is much better than the alternative (REALLY wavy)
* The hive gains wait up to about 6 or 7pm then loses weight over the next 12 hours. I am assuming this is due to evaporating the nectar at night. It is definately NOT temperature induced fluctuation.

There is a very real possibility that i am not giving enough credit to the accuracy of the compensation and that the wavy patterns are real weight fluctuations. I need to spend some time and see if i can verify one way or the other.

A local BK and I have been looking at this data for a while now and he initially thought the bees were foraging a long way away, which would also lend itself to the patterns you see.. a dip follwed by a spike.. bees find a far off nectar source, recruit tons of workers, they all leave together, and return full of nectar 30 mins later or so. Hard to say.

You must be an engineer?

Something like that, yea. Electronics design is not my main line of work though. just a hobby.

-tmk

[Nabber86]

I love graphs. Does the program that you are using let you develop equations for the data? I am wondering if the growth would be linear or expontential (I am leaning towards linear, since there is only 1 queen laying at a fairly constant rate). Of course the daily fluctuations would mean a polynomal solution and that can get really messy really fast.

my fix-the-crappy-scale compensations lend a 'wave' to the graphs which move them by about .5lb.

Oh no. That is beginning to sould like data manipulations performed by the Climatic Research Unit (CRU). Next thing you know is the growth will appear exponential and you will predict that the hive will soon reach the mass of the known universe. RUN!

[tmk]

I love graphs. Does the program that you are using let you develop equations for the data?

I can graph based on any manipulation of the data i want.. the lines you see are heavily modified

When i was developing the compensation stuff, i used gnuplot and my brain.

If you're interested on how the compensation works, here are a couple posts i wrote regarding that:

http://makingthingswork.wordpress.co...curve-fitting/

http://makingthingswork.wordpress.co...voltage-sucks/

-tmk

[Fuzzy]

TMK,

looking good. The data is quite interesting. Glad you posted the pics. It's the little things that make unexplained differences.
a. That is a nice "bowl" you placed on top of the hive. You might consider inverting it, such that the moisture runs off.
b. This is the middle of the spring "build up". So some of the weight gain is new brood and bees. The hive population (if it doesn't swarm) should maximize about the end of March to the middle of April. That is when we get our first major flow.

But, nice start -- Fuzzy

[tmk]

a. That is a nice "bowl" you placed on top of the hive. You might consider inverting it, such that the moisture runs off.

It's just a piece of plastic with a brick on it. Water doesn't seem to pool up on top, though it does get stuck in the 'flutes' (it's corrugated plastic, like cardboard). Doesn't really hurt anything and keeps the water from dripping on the bees through the roof :)

-tmk

[tmk]

Graphs for the last week. It's been rainy in california, and the bees haven't brought in much.

I'm going to try and update the blog weekly. This will probably be the last time I post an announcement or graph to this thread. Add the blog to your RSS or bookmarks or whatever if you want to stay in touch :)



The above link will show only the blog posts with graphs in them, in theory.

-tmk

[Humanbeeing]

Hey data guy, Goddard Space Flight Center wants your data. Go to this web site and click the Scale Hives and Climate tab. Good job on the scale.


http://honeybeenet.gsfc.nasa.gov/Honeybees/Forage.htm

[tmk]

Thanks, I sent the folks over there a note.

Anyone else on here participate in this study?

-tmk