SparkFun OpenScale

As a load cell interface for a computer, this board works fairly well - when you finally figure out how to use it correctly! Spark Fun has put a number of different features on this board - not all of which you may need or use - but which makes it flexible and useful for a number of applications. The firmware has 13 different options or modes which can be set using a terminal program on Windows. That is both the good news and the bad news. For example, if you happen to mis-type the number when changing the serial port baud rate (yes, I did this), the board will use that rate even if it is not one of the standard rates. Once the terminal screen is gone, and if you don't know what that mis-typed number happens to be, you've just "bricked" communications. Fortunately there is a hardware reset feature, but you have to manually short two contacts and then reconnect the board - the "reset" button installed on the board doesn't do this (not really sure what it does, but it has a great tactile feedback click!). Anyway, for my application I was finally able to get the board reading my load cell and it is working well now. I cut the jumper to increase the sampling rate to 80 Hz and I get nice stable readings or reports (with one exception - see below!) at 40 Hz (note: the REPORT rate, or message rate to the computer will always be lower than the sampling rate due to internal overhead in processing and serializing analog samples). The "exception" that I mentioned (Spark Fun people, NOTE THIS PLEASE!), is that about once in every five hundred (or so) reports, there will be an erroneous force value that appears as a "spike" in your data. This is NOT a single character error (dropout or bad character), but an actual incorrect load cell value that is significantly different than the "true" values in the data stream. I did see a mention of this in a Spark Fun forum from several years ago, and it was supposedly going to be fixed in an ensuing firmware update, but as of the current version 1.2 it is still there. Because these spikes are easily detected and infrequent, I found a simple way to remove them in my software, but if you are using this without custom software then this may be a problem for you. For my use, I would give the board 4 stars, but because of the glitch and the learning curve to use the board well, it only gets three stars for general purpose use. If Spark Fun would update the firmware to remove the data spikes, and provide a bit better documentation, this could be a five star product.

I'm on my second board, second set of load cells, multiple wiring configurations. I'm giving up.