Can you let me know how you did that so I can get rid of the Division zero error?

 

Unlikely. You were probably only _printing_ the time value to 2 decimal places. Try something like this instead:

print "%10.5lf" % time.time()



Yeah, that exception is just a place holder to remind me to check for weird rpm flux at the beginning of end of a run. That code is going to get yanked out of there anyway, since I have this new piece I cooked up last night with the help of my college physics book. It _should_ provide "accurate" torque/hp numbers. I did the whole thing and metric in my head was spinning trying to match it to data I see out on the internet

Anyway, you're famailiar with python, you're more than welcome to become the windows maintainer. Just say the word and I'll give you CVS access.

d

 

Thats good there is someone using windows that understands python... Python is quite a bit different then what I'm used too.

 

you'll get used to it. It really is a tremendous language to work with.

Anyway, divide by zero. When the intepreter throws that exception, it'll also give you a line number. Try to include that in your mesg to me so I know which division you're talking about.

I'll assume you're getting it somewhere around line 227 in pyxede.py where it calculcates the torque. the line looks something like this:

torque = (rpms[i] - rpms[i-1]) / (times[i] - times[i-1])


Obviously, if time[i] and time[i-1] are the same, then you're going to divide by zero. Just check and if so skip this iteration of the loop. voila!

....
if time[i] == time[i-1]:
continue
torque = (rpms[i].....

d

 

WINDOWS!!!! YES!!! The whole master plan is coming together!!

 

I got this error too.

 

There's a menu to set the gnuplot location:

File->Settings->Plotting

d

 

Nice I'm gonna go test it now...

 

Ok, but I'll release a new version in a few hours. I've made plotting a lot nicer.

d

 

Ahh you get me all excited about how to do dyno plots and now your coming out with nicer plots. Keep em coming donour. You still heading down here to sunny NM? I'll be going to our speedway (1.65mi course) next weekend to run my baby. i'm excited to use pyxede to do some logging.

 

It'll be more than a few hours. I thought I'd be able to get away without having the dyno data collection being threaded. I can't. That means pushing a whole bunch of thread stuff into the windowing code which is a headache (one of the main reasons I HATE writing graphical interfaces :-p). After I get off work, I'll straighten things out, but then I'll have to test...

EDIT: testing...While I'm at it, note to self: Mac OSX does _not_ like it when you try to force thread context switches 1000+ times/second

I'll be there on the 18th for a couple of days, and I'm moving for good at the end of july.

d

 

I'll try that out.

I'm somewhat familiar with python, so I'd definitely like to help out on the windows side. Let me know what I can do

 

New release -- 0.6 -- numerous improvements. It's the one to use from this point forward. A short description of the release changes are on the webpage, but there's much more than I care to rant on about.

d

EDIT: here's an example of the new, prettier plots.

 

The div0 error people were getting from samples with identical time stamps shouldn't happen anymore (version 0.6 and on). I've locked the samplerate at a max of 100hz. I didn't really like this solution, but I had to do it to avoid killing my slow *** laptop which runs OS X. The macs just don't like heavy thread context switching.

Test, test, test. Especially try the dyno code. If you can, do lots and lots of 3rd gear pulls from say 2000-6500 rpm on the highway. I live in downtown chicago, I can't get above 45 anywhere nearby.

To generate standalone windows programs, all you need to do is download py2exe. There's already script to build the executable: setup.py.

I generated basic API docs and put them here: http://www.donour.com/cars/pyxede/docs. That should give a running start to anybody who wants to modify the code or take the library and use it in a different application.

d

 

I have to admit I used my old (really old) notebook running Windows 2000 for testing. That said I didn't have much luck getting this running.

I was able to get the main window to come up and the a connection to the Xede made. On the other hand "Perform road dyno", followed by clicking on start caused pyXede to freeze.

In addition if I load road dyno data a window pop ups with what appears to be correct data (e.g. dustin's data). If I click on "plot" nothing occurs. The configuration correctly points to gnuplot and gnuplot runs correctly if started on its own.

I have to run out right now, but I will look into this a bit later. Yes I know this isn't much help unless you have a Win2000 box sitting there.