The Tronbike E-Meter

Travis came through with the serial analyzer and I felt compelled to tear off the fairings and tank to get access to the nasty little serial cable for what is to be the Tronbike E-Meter.

Leveraging my collection of ancient RS-232 adapters and cables, the laptop and the analyzer, I found my software was actually fine, my problems were:

1. Laptop under Linux does not control the Keyspan USB adapter properly. It works fine under Win XP running on VMWARE on the same Linux laptop.
2. I need a null modem on the IPAQ cable. This might be actually reversing the null function IN the cable in order to manage the Alltrax. Either case, it works.

The analyzer was a bit tricky to get right, but I found myself confirming the hex dump content and null modem/non-null-modem orientation. Plugging the wires between IPAQ and Alltrax and voila!, I was suddenly reading data bytes on the IPAQ console.

This was great news, but it led to a series of late night programming binges that continue to tonight. The UI has remained the same (see screenshot sample, ignore the numbers), but the architecture has changed now - twice. I went from separate executables for reading and display to a single forking python executable. This presented all sorts of exciting lessons in signal handling and interprocess communication in Python. The goal was to start and stop a single app to do reading and display of the data.

Attempt 2 occurred last night where all forking and multiple processes were removed, with the polling of the Alltrax now being driven by the GUI directly. No more forks, one executable, 1 process. Less IO to the disk, so it runs faster.

I also added a data logger function to record every actual use of the meter, and did so in an object-oriented fashion. During the refactoring I added more object-orientation and found it useful to manage scope and state of the system, the first O-O success story I can remember having (old Fortran hack here!).

All this fresh software was just plugged into Tronbike tonight, on its stand, and below is the first
data set from the logger:

Columns are:

1. unix epoch time since 1970
2. throttle position %
3. controller temperature in C
   
4. voltage (notice what happens when I flip the solenoid, goes to 63
5. shutdown (not sure what this is)
6. amps (bike is on stand, no load, might want this as floating point value)
7. battery amps (calculated from amps as per Alltrax)
8. watt-hrs in pack (calculated)
9. projected time at current usage (calculated)
10. percentage left in pack (calculated)
11. time since start of run
    
12. watt-hrs pack capacity (5 * 12volts * 55a-hr)

Log Snippit:

1209855312.417228  21 19 62 0 0 0 0 3299 9999 99 171.479993 3300
1209855312.917160  21 19 62 0 0 0 0 3299 9999 99 171.980005 3300
1209855313.417234  21 19 62 0 0 0 0 3299 9999 99 172.480015 3300
1209855313.937162  21 19 62 0 0 0 0 3299 9999 99 172.999987 3300
1209855314.427109  21 19 62 0 0 0 0 3299 9999 99 173.489878 3300
1209855314.927228  19 19 62 0 0 0 0 3299 9999 99 173.990005 3300
1209855315.437221  17 19 62 0 0 0 0 3299 9999 99 174.500002 3300
1209855315.977232  16 19 62 0 0 0 0 3299 9999 99 175.040012 3300
1209855316.457197  15 19 62 0 0 0 0 3299 9999 99 175.520065 3300
1209855316.977281  13 19 62 0 0 0 0 3299 9999 99 176.040042 3300
1209855317.487231  12 19 62 0 0 0 0 3299 9999 99 176.550008 3300
1209855317.987198  10 20 62 0 0 0 0 3299 9999 99 177.049960 3300
1209855318.487206  8 20 62 0 0 0 0 3299 9999 99 177.550002 3300
1209855318.987278  6 19 62 0 0 0 0 3299 9999 99 178.050053 3300
1209855319.487112  0 20 62 0 0 0 0 3299 9999 99 178.549970 3300