Coffee that you can trust.
 

 





















 

Running Green Bus and the roaster seems like the culmination of a lifetime of tinkering.

To me, fixing stuff is a metaphor for the more important, non-stuff, aspects of life.

Hence this blog.  It concerns the technical aspects of building and running the Green Bus and its associated support mechanisms.  Buying, renovating and maintaining a bus will either drive you mad or strengthen your philosophy.  This particular example is a 1996 International diesel with a Bluebird body that was transformed into a mobile kitchen in Ontario, Canada.  As with all such adventures, it is a work in progress. 

Throw a coffee roaster into the mix, and things get even stranger.

If you have any comments, you can reach me at:   felix@greenbus.ca .

 

 

technical


December 10, 2014
On being grounded.

The Green Bus roaster is a TJ-067.

The good folks at Mill City Roasters in Minneapolis sold it to us, and we are quite happy with it.  It is solid, quiet, easy to control, and works equally well with propane or natural gas (after changing the jets).

Up until now, I controlled it manually, and logged roasts manually using pen and paper.  The roaster, however, came with a USB interface and an extra temperature probe to allow roast "profiling" with a computer.

It proved a vexing technical problem.  I was getting inconsistent temperature readings from the bean probe.  Often they were several hundred degrees off, and other times they just got stuck in one place, or they fluctuated  in a seemingly random manner.

I was using a linux box, and I put the problems down to my limited ability to diagnose driver issues in trying to get the profiling program (Artisan) to talk sensibly to the USB interface (a Phidget 1048).  After spending many hours searching the internet for information and trying new approaches, I conceded defeat and hauled out my old Windows XP laptop.

Within 10 minutes I had Artisan loaded onto Windows, the Phidget drivers were installed, and the temperature readings seemed to work miraculously.

Pleased, I rearranged my workspace to incorporate my old laptop, and did a final test run to make sure everything worked.

Alas, the temperature readings resumed their wild and inconsistent behaviour, just as they had under the linux box.

So I made myself a cup of coffee while I pondered this latest attack by the forces of Entropy.  Maybe it was the coffee that fixed it.   Maybe it cleared my mind, focused my attention, and allowed me to see that the program worked properly whenever the laptop ran on its battery, and messed up whenver I had the power supply plugged in.

It was a ground loop problem.  There was a ground voltage difference between the roaster and the computer, which can be fixed by using a USB isolator, by insulating the probe, using an ungrounded thermocouple, or by keeping the laptop ungrounded. 

How simple problems can seem once a clear perspective is found.




Archives