Yup! Scales again. My assumption that the small transistor was causing my display issues was correct – I rebuilt the soft on/off part of the circuit using an N-Channel power MOSFET and all was well.
Here’s the “finished” base unit before I reassembled it:
From left to right we have:
- CISECO Xino Mini
- INA125 / MCP3202 – Instrumentation amp (& voltage reference) + SPI 12-bit ADC Converter
- Soft on/off switch based on a 555 timer
- Button de-bouncing board (incl. inverting Hex Schmitt Trigger IC)
You can also see a small piece of strip board that was used to attach the wires from the base unit of the scales. When I hot-glued that in place, something went wrong and I lost on of the ground connections. I then threw that piece of board out and soldered the wires directly to where they needed to go on the boards in the above picture. Hot glue for strain relief and it’s all good now.
Right. The proof of the pudding and all that…
Have been titting around with the firmware and I have to say that the results are extremely disappointing. I’m getting a lot of fluctuations and the “reference” values I use to calculate weights seem to vary depending on what weight was placed on the scales last. Maybe I’m not supposed to use the analogue reading with no weight on the scales as zero but if these things are linear then I can’t see why not.
This could well be the usual story of these prototyping boards suffering from lack of ground plane and all the wires being too ready to pick up stray signals. I think I’ll put this on the back burner for now and may return to it when I figure out how to layout circuit boards in Eagle or DesignSpark.
If I can’t get good accuracy then they’re not much use as scales. It’s been a fun project to tackle though and I got to learn about a few things. There are just so many other things I want to get on with at the moment that I can’t justify spending any more time on this right now.