I thought I would start this diary exclusively for a watermelon scale project that I have started to work on. This will be an ongoing endeavor because I’m sort of making things up as I go. Last March, I watched Frank Mudd give a presentation to a group of giant watermelon growers down south. I found it very fascinating to hear that Frank (and other growers) use scales under their melons to monitor the growth. It seemed like an important tool, so that you can see what affects the growth of the melons throughout the season. I knew I couldn’t justify buying scales so I got thinking of ideas to make my own. After a lot of research I had an idea to build a scale using load cells. I don’t know if this has ever been done DIY before with watermelons. I do know that some beekeepers build these types of scales for monitoring the weight of beehives and home brewers use these scales to monitor keg weight. My thought was to modify the scale to fit under the frame of our watermelon hammocks. The first thing I wanted to do was build a prototype scale just to prove to myself that this might be feasible. If it worked, I would expand the design to fit under the watermelon hammocks. I was able to purchase all the components that I need form Amazon and eBay.
Pictured below are all of the electrical components that I will need to make this scale.
This is a close up of the loads cells. They are each rated for 50kg’s and I will be using 4 to make the scale, so the total rating of the scale will be 200 kg’s or 440 lbs. A load cell is a piece of spring metal that has a strain gauge firmly attached. As weight is pushed down on the metal (bump in the center) the resulting change to the resistance in the strain gauge can be measured as voltage. This voltage change is directly proportional to the amount of weight applied to the load cell. So the amount of weight can be calculated from the load cells output. The load cells have to be wired in a very specific way for them to work correctly. I wasn’t going to make this into a “how to do this” type of thing but if someone wanted more specific details of what I am doing I could provide that. To tell you the truth, I don’t really know yet if this will work out for me, but I’m going to keep on with it hopefully until I get a functioning scale. Or hit some sort of roadblock that I cannot pass.
The load cells have to be connected to an HX711 amplifier, because the change in resistance in the strain gauge is such a small value it must be amplified so that the microcontroller can read it. These are surprisingly inexpensive. I bought 4 load cells and it came with an HX711 amplifier for $10 on Amazon. The pins have to be soldered on to the board.
I need some sort of display to readout the weight on the scale. I have decided to use this OLED display. It is very small but with the right sized font it should be just what I need. I paid a little less than $10 for this one on eBay.
The microcontroller is next on the list. This is basically a small computer and is the brains of the whole thing. This is a genuine Arduino Uno and I paid $42 for this one on Amazon. You can buy a Chinese clone board for $8 that is supposed to work just as well. I did buy one at first but couldn’t get it to work for me. This is the part of the project that has caused me the most trouble. This board uses C++ computer language to program it and until a few weeks ago I didn’t have a clue what that was. The code must be written with all the information in it that you will need to operate the scale as designed. It will only accept one sketch (a fancy name for lines of code) at a time so everything must be compiled into one sketch in order for it to work properly. For instance, most scales will re tare to zero every time the scale is powered off and then back on even if it has weight sitting on it. So I had to write the code so that I could tare out the weight of the watermelon hammock at the start of the season and then have the weight of the melon saved in the memory when it was powered off. I’m going to power it with a 9v battery and just turn it on once a day. The scale also needed to be calibrated so that it will be fairly accurate.
I made a small proto type scale (pictured below) and have it connected to my computer with a USB cable for programing. I joined an online Arduino forum where I had to ask a lot of questions about programing and microcontrollers. I was helped by some amazing people from around the world and now after a lot of hours of trial and error I have a working sketch that works exactly like I want it to. I should be able to save this sketch for future scales. I plan to make two scales if this one works. I have all the electronics just sitting on a chair for testing purposes. I have a plywood board over the top of the loads cells.
This next picture shows what is under the plywood. The load cell are recessed into the wooden frame. I soldered new wire to lengthen the wires on the load cells so that they would be long enough to work on the real scale.
This is a close up of the area that the load cells fit into. I used a router to make a recess in the wood so that when the metal tab bends slightly it has room below.
This next image shows the OLED display connected temporarily to a solderless breadboard just for testing. The scale seems to work well. It retains the last weight in the memory when I power it off and the scale is fairly accurate. My next task is to design a weatherproof case that I will be able to house the electronics shown below. Not the breadboard, but everything else. I need to be able to see the OLED so I need a design with a window for viewing the display.
I thought of a lot of different designs I could have possibly used but eventually decided I would make a wooden box to house all of the components. To me it made the most sense. I enjoy woodworking and I could make a wooden box very easily. I will seal it up very well and apply some spar varnish.
This is the inside of the front panel. I have a recess around the OLED window big enough to place a glass microscope slide to act as a window. I will use silicone to attach this glass to the wood. The OLED will be held up to the window for viewing with wingnuts.
This is the dry fit.
This is the finished box after 3 coats of spar varnish. I made a silicon gasket for under the lid to seal out the weather. I also did some wood burning on the box.
Now I need to install all of the components into the box and use a 9 volt battery to supply power.
Here I have it connected to the prototype scale. The OLED display is held up behind the glass window.
It seems to work great. Now, I need to build a larger frame for the actual scale and get some wire loom to put the wires into. That will be several weeks down the road. I need to concentrate my efforts right now in the garden to hopefully get a melon worthy of putting on a scale.
The next step in this project was to build a stand to keep the control box off the ground.
Then I made a box out of aluminum that will fit over the control box. This will hopefully keep rainwater from getting inside the box.
For the last few weeks I have been working on the next steps to complete the watermelon scales.
First, I made a frame out of 2” angle iron.
Added a hardwood top to the frame. The frame was made exactly the same size as our hammocks.
Routed a recess for each load cell and painted everything black.
Sealed the load cells in place using silicone.
Connected all the wiring to the control box and put all the wire into split loom.
To protect the load cells from moisture I put a small piece of plastic over each one and then covered that plastic with aluminum tape. I want to keep rain water out as much as possible. Once I had everything wired up I had to connect my computer to the Arduino board and upload the program. I then tared off the weight of each hammock and saved that weight in the memory of the board. I also had to tweak the software a bit to calibrate the scale slightly to get a more accurate weight. So each time the scale is powered up it will subtract the weight of the hammock and only display the melon weight.
Here I have one of the frames sitting in the grass just before I was ready to put it under the melon.
I had to test it out one more time before using it so I put the hammock on top and weighed myself. It was very close to my weight so I am confident that the scale is fairly accurate.
Here is the first melon on the scale. I’m not happy with the shape of the melons this year. I can’t get a decent shaped melon for some reason. I have removed a lot of pear shaped and crooked ones and I had to make a decision to keep one because we are getting late in the season. I have had this melon on the scale for 2 days now and found out that this one loses 2 lbs. during the day for some reason and gains 5-6 lbs. during the night. Not sure if this will continue but that is what has happened in the last 2 days. It weighs 31 lbs. this morning.
Here is our other melon sitting on the scale. This one is not growing as fast so far and I don’t like the shape of it as well. This one loses very little weight during the day. It puts on a little weight at night but so far only a couple of lbs. So each melon seems to have its own growing cycle. It will be interesting to track this as the season progresses. My next project is to make a stand to set the control box on. I also want to build a cover that will go over the stand to shed rain. I made the control box as watertight as I could but I would feel better if it had an additional cover over it.