I have worked a bit with a load cell, building my own pressure platform (see pictures here: http://www.flickr.com/photos/25468334@N08/sets/72157637376790655/)

However as I wanted something with a much lower height, I bought a kitchen scale and opened it (pictures here: http://www.flickr.com/photos/25468334@N08/sets/72157637373028495/).

There are four load cells in there, but as each of them has only three connectors, so firts it made me think they are of either half-bridge or quarter-bridge type. You can see on one picture that the four cell inputs on the PCB are labelled E+, S-, S+, E-. Usually, S means “sense” and E means “excite”. My conclusion: the cells are of quarter-bridge type and make a full bridge together.

Next step: connection of the sensors to an Arduino.

All the scales, and the load cell from Ebay, have been able to scale negative as well as positive. If you can find a scale locally, and it’s operational, you can just try pulling the scale platform from the base to check for negative readings.

You can order the load cells from Ebay, but I would try for scales at local shops which you might be able to test, bargain shops and markets are best places to start.

A 10bit A/D converter will give you between 1 and 2 grammes accuracy while a 12bit A/D converter will give you less than a gramme accuracy for a 1kg load cell. I’ve only worked with the INA125 and this chip probably wouldn’t work for you if you want to read both positive and negative from the load cell.

Mark

Thanks.

The load cell is used to measure the stepper motor effort during extrusion. The data from the sensor can then be formulated to provide a complete set of information without the need to collaborate with other sensors or data.

Errors like filament slips, stepper motor stalls, filament jamming, filament quality and bowden failure can all be detected just from the load sensor alone. A lot more analysis is possible and waiting to be explored, and will be demonstrated with new software being released shortly.

Thank you for your interest,

Mark

Thanks.

Harry

The project turned out much bigger than realized – with the different elements involved. However, a lot of progress has been made, the model files are done and the write-up almost complete.
The software for calibrating the load cell was re-written from scratch to better suit the project. The software and the load cell electronics documentation will be drafted this week.

I’ve also drafted code to work with the small 1.8″ TFT mobile coloured screens. This is like the next stage of the project that allows real-time graphing without the need of a computer. It’s cheap and compact and currently on test on the 3d printer.

The project notes will be well worth the wait. You’ll need both Sli3er and Skeinforge to print the parts in case you need to get familiar.

Mark

The beauty of the force sensor is that you can use it alone and still get good information without the need for verification. However, a high resolution stepper motor encoder is planned because the information it will provide, with the force sensor, will be greatly expanded.

I think any sensor system for the extruder needs to report more than just errors because a good extruder hot end combination could make a potentially expensive sensor add-on mostly redundant.

That ADNS-9800 Optical Laser Sensor looks awesome and might provide a nice compact stepper encoder. It would be interesting to know if any type of filament has enough texture contrast for the laser sensor to work consistently. I’ve had experiences where the optical mouse has not worked properly on some surfaces.
https://www.tindie.com/products/jkicklighter/adns-9800-optical-laser-sensor/