My visit to the UK was quite awesome and it was really good to catch up with all my old friends and work colleagues again and see how everything has changed, hopefully even convinced a couple to make the journey down under for a visit.
The downside is I’m still catching up on the very large backlog of work from the holidays, which includes some re-work and changes to one of my projects I created in 2007 for my customer Wine Technologies which I have to convert 700 lines of code from MikroPascal to standard MPIDE compatible C.
MikroPascal was a very nice language and is still written and supported but it does cost about $200 for the IDE and also my client (and I suspect most clients if you really ask them) prefer to be able to have the code in C so they can read/modify it if required.
It’s a shame in one way that I have to do this, Pascals a much more forgiving language which makes it real hard to make mistakes in syntax and structure, and it is so easy to pick up what you intended the code to do, even years later. I’ve sometimes written whole modules in one go, compiled them and they have worked, yes you can still put the wrong value in the wrong register and make a pigs-ear of the logic, but generally it was a nice way to go.
C was a language I was very happy with back in the DOS Turbo C days of 1986, Ironically I had to move from it to Turbo-Pascal when I worked at the government and I remember having the same feelings! when I first encountered C++ in windows in 1994 I shuddered at the complex monster of libraries and headers it had grown into and immediately started using Delphi (I always take the easy way out!)
Anyway, as part of the re-work I had to also re do the PCB. The new PCB was larger and customer decided to move from surface mount to through hole and to build them themselves again another change from my normal as most people prefer surface mount which are easier to mass produce offshore. The PIC18F2520 which the project was based on is also still manufactured but is still $4 so I reworked it with the pin compatible PIC18F23K22 as the C code compiled to about half the size of the Pascal code. The new chip is under $2 which should offset some of the costs as he will be making hundreds of these.
The catch2 production system is also still on the back-burner due to work pressure from my major client, the Mussel factory are sweet with this delay (they don’t pay a penny until its all in tested and working) and it gives them a bit of time to figure out a more efficient way of production s a few changes are on the cards when I pick it back up in a week or two.
I ordered and checked out the ESP8266 boards, mixed results with the first modules (with older firmware and chip) booting but failing to go ready, the second batch I ordered (after I almost gave up on the chips) are the newer ESP8266EX ones and seem to be playing ball, but seem a bit flakey in the firmware and power hungry at switch on. A few people in the maker community are doing a lot of work testing these and improving the firmware, with this in mind I have re-spun the mobile terminal board to use one and test properly.
I signed the NDA to get more technical info on the device but they do not give you any more data than is already leaked to the comunity, a strange way of doing business is not giving out technical documentation on what you are trying to sell! Anyway, Check out the progress on ESP8266.com and hackaday.io
I have also been designing the byte-code and comms protocols for the mobile terminal to ensure all traffic to and from the server is short and concise. The 2.4Ghz wifi band is very over-crowded and having everything in a single packet has a lot of advantages. Currently this is slow going as I really need to put some quality time aside where there are no distractions.