Observing weight drift with my setup

Hi everybody, thanks for your great job!

I’m building a prototype scale with arduino mini, HX711, battery, solar power, temp and humidity sensors. but cannot make it work! the weight drifts over night and day, it is not temperature related, it’s humidity related. I double checked.

Now I’m keeping the scale inside, room temperature is stable +/1 one degree. Humidity goes up and down today is 35% and the weight is down 70 grams, over a 20kg class M test weight.

I experienced that outside, at the beginning I thought could be temperature related, but soon I realized that the only thing driving the weight up and down was the Humidity! Is it possible? how can i solve?

Im’ using a single point 100kg load cell Class 3 IP54.

Anyone ? please help

thanks a lot

Bye

Dear @edilsite,

thanks for joining us - welcome! Take care about the flux just posted recently by @thias comes to my mind when reading about your observations:

Good luck with getting rid of this problem, we hope this helps.

With kind regards,
Andreas.

P.S.: Please use the globe button below each post for translating German texts into the language your browser is running on:
image

P.P.S.: We will be happy if you want to join our collaborative data collection effort as soon as you finished your hardware setup and will support you with the onboarding process if you like that.

1 Like

I made some changes in my design, still using same load cell and same Arduino, but I changed the load cell interface, now is connected with Load Cell nano shield from Circuitar ADS1230 microcontroller, but I’m still getting this bad result. this is NORMAL?

The scale is outside as you see from the temperature graph. Is protected by wind. The load is a fixed 20.000 grams Iron weight. It’s so strange no temperature related. just going up and down, it seems at fixed time interval

1 Like

Interesting, is it this board:

https://www.circuitar.com/nanoshields/modules/load-cell/index.html

I have used the ADS1231 as load cell interface. Is the 20 bit vs. 24 bit ADC the only difference between this chips?

Dear Gino,

thanks for letting us know about your observations. May we ask if you store the data locally or if you forward it to our platform at swarm.hiveeyes.org already that we could have a more in-depth look at the measurement data?

If temperature, humidity and weight would be accessible to us via Grafana, some of the other smart guys here are sometimes having fun to find correlations or explanations for the effects you are experiencing. Also, the geolocation where this setup might be running would also be helpful for further analysis as this can tell us more background information about sun position or other weather conditions acquired from other environmental sensors nearby. A 5-6 character geohash which you can compute at http://geohash.co/ will probably be okay for that, see also Verunschärfung von Geopositionen.

Regardless of this, I would like to list some resources with information about possible effects scattered throughout the forum here for your reference:

As some of the resources are in German, fell free to use the translation button below each post to get an idea at least about the content here.

With kind regards,
Andreas.

Probably is the only difference, but I need to compare the datasheet. The reason I tried that shield is the poor result I had with the small HX711 boards.

We have really a lot of stuff for making our projects but I think we have to keep the cost low.

There are other evaluation board like

or other products like the 32bit

Just for example

for transmitting data I do like the SIGFOX based product called mkrfox1200.

But my project has to be lower than 50E and I think I’m near that budget.

I will upload some more graph now that I’m testing under an hive

Thanks bye

1 Like

@edilsite may you talk a bit more about you project?! Is this a community project or commercial? Where comes the 50 EUR limitation from? Set by yourself? And for the electronics only. I would like to have a scale (scale hardware and electronic parts) under 100 EUR. ;-)

What load cell are you using? We found big differences e.g. between the “branded” Bosche load cells and no name cells from ebay.

1 Like

load cell used 150Kg class c3. it’s giving good results to me.

https://it.aliexpress.com/item/1PCS-X-100KG-150KG-electronic-platform-scale-load-cell-pressure-balanced-cantilever-load-weight-sensor/32344819630.html

the cost limitation is set by myself, no reason in building something expensive to monitor an hive. It must be reasonably cheap in my mind, but we are going off topic. I’m sorry

1 Like

As I promise I’m uploading a graph after some days of testing under an Hive. I’m satisfied but I’m also tired of this project, it was frustrating for the most part of it, and it took too much time. Now I’m monitoring temperature, battery charge, and weight. Hope the scale will work for years… I don’t want to work on this anymore. I will upload the list of materials as soon as I collect all the information and also the price for each piece. I’ think the drift was caused by a bad power connection.

1 Like

Dear Gino,

Thanks for letting us know you have been able to make it through and good luck that you won’t have to touch the setup anymore.

With kind regards,
Andreas.

It’s a bit difficult to make estimations about temperatur effects with your last graph because there is a living hive (and so changing in weight) on the scale. “k” ist “kg”? So it would be still 500 g change in one day.

In case you have stabilized the power connection you can measure a not changing weight and monitor temperatur in parallel. After some days you will be able to calculate a compensation factor and get better values even with “bad” load cells, see:

List of materials

load cell 12,00 (aliexpress)

load cell interface 15,00 (eletroshields.com.br)

generic 2.4 GHz transmitter receiver 1,50 (aliexpress)

generic charger voltage regulator 1,00 (aliexpress)

solar panels 2,00 (aliexpress)

battery 2,50 (aliexpress)

temperature sensor 1,00 (aliexpress)

total cost less than 40 euro

the metal frame for the scale was made of recycled pieces of metal, so no cost

Is the best and cheapest scale I can make, is working and this is enough. Drains too much power, needs more field test, needs a good designed circuit, needs more sensors (wind rain temperature pressure humidity etc.)

But I cannot make it better. Hope someone here can build it in a better way.

Bye