JPI EDM log file converter

Wize said:
Howdy xander75,

Thank you for doing all of this! While I have a fair amount of software work experience I'm just running a standard x86 based Windows PC right now without doing much development work and don't care to set up any servers on this PC. I actually was looking to get the CSV files so I can combine them using Alteryx script into a flat file that I use as a database for a Tableau workbook of all my flights and engine data. One of the big limitations with the Savvy platform is the inability to easily compare historical data or get data sets beyond a single flight visualized easily. I.e. I like comparing my carb temp vs. OAT and setting like this picture below using a scatter plot. Tableau lets me create all sorts of interesting graphics and, perhaps most importantly, allow me to filter out bad or otherwise erroneous data.

I'll hang tight until another solution comes along. It takes me about an hour to extract a year's worth of CSV files out of the JPI files. And then I have to filter them for all the instances when the JPI instrument was only on for a ground run or something that I don't want to consider as a true "flight". So that's probably a quicker process than setting up a virtual server to run Grafana, mysql, et al.


View attachment 134885
Click to expand...
ah thats very cool @Wize !! yeah the mining that one can do from the engine data is amazing!!
if you like to give it a try, install docker desktop and I'll share you the descriptor file to build a container from that descriptor which is then a fully configured and baked linux distro (about 1Gb), you'll point your browser to localhost:8080 (or whichever port you chose for the apache webserver in the container and you're good to go).

If you don't wanna go through that, you can also share your jpi file(s) with me and I'll convert them to csv's if you like no problem!

cheers!
xander
 
Really great work, thank you for this. It works great for me on MacOS.

Flysto is another free tool that does flight analysis. (I'd link to it but my account is too new to post links)

They accept JPI data but it must be converted to CSV using the JPI Windows tool. Would be amazing if this tool and Flysto worked together!
 
So, newb question. My 830 rx 232 gps data for fuel calculations. When I download my JPI from the thumb drive to EZ trends, all I see is graphs of cht and egt data. When I cat the file, I don't really see anything that even looks like gps data. My JPI sw ver 1.4. Which sw version includes GPS data in the download, or am I missing something?
 
dave5k said:
Really great work, thank you for this. It works great for me on MacOS.

Flysto is another free tool that does flight analysis. (I'd link to it but my account is too new to post links)

They accept JPI data but it must be converted to CSV using the JPI Windows tool. Would be amazing if this tool and Flysto worked together!
Click to expand...
hey dave5k, awesome to hear!! glad it worked out! I never heard of that tool you mentioned, but if you can share a PM or however so a detail on that one I can reach out to them to consolidate efforts!
cheers!
xander
 
Chip Sylverne said:
So, newb question. My 830 rx 232 gps data for fuel calculations. When I download my JPI from the thumb drive to EZ trends, all I see is graphs of cht and egt data. When I cat the file, I don't really see anything that even looks like gps data. My JPI sw ver 1.4. Which sw version includes GPS data in the download, or am I missing something?
Click to expand...
hey chip, few things that I know: even in the latest fw that I have, the gps data reads zero and is not easy to decode based on what the implementation says it should. for sure you need a gps input on your EDM to have ground speed, LAT/LONG data. I'll review from teh info that I have what fw would include the control blocks data that would include the gps info, but you may be out of luck there. for sure, it shoudl export more then just cht and egt. will get back to you. xander
 
xander75 said:
hey chip, few things that I know: even in the latest fw that I have, the gps data reads zero and is not easy to decode based on what the implementation says it should. for sure you need a gps input on your EDM to have ground speed, LAT/LONG data. I'll review from teh info that I have what fw would include the control blocks data that would include the gps info, but you may be out of luck there. for sure, it shoudl export more then just cht and egt. will get back to you. xander
Click to expand...
Thanks. I know the 830 is getting the 232 data, because I get fuel required to waypoint etc. I do get a red x in fuel to destination should I shoot an missed approach and subsequently go direct to a waypoint that was in the flightplan prior to the approach, rather than add a new destination after the missed. I'm trying to figure out if this is bug in the JPI, or if the 232 signal is actually dropping out.
 
Chip Sylverne said:
Thanks. I know the 830 is getting the 232 data, because I get fuel required to waypoint etc. I do get a red x in fuel to destination should I shoot an missed approach and subsequently go direct to a waypoint that was in the flightplan prior to the approach, rather than add a new destination after the missed. I'm trying to figure out if this is bug in the JPI, or if the 232 signal is actually dropping out.
Click to expand...
do you happen to have other devices connecting to the gps info (like an gpss enabled auto pilot or a electronic flight display like an aspen pfd?) that should confirm if the flight plan is automatically pointing to the next (missed approach) waypoint and back to the hold.
If that is the case then it has to be something in the jpi.

I did some more checking, while unrelated to your situation, the engine data logging for LAT/LONG is problematic, but groundspeed recording should work. also hobbs time is not functioning (unrelated also to gps input obviously)
you may want to open a bug report with jpi, although I havent had much luck getting traction on some of the bugs uncovered due to what sounds like prioritization and resourcing.

x
 
xander75 said:
do you happen to have other devices connecting to the gps info (like an gpss enabled auto pilot or a electronic flight display like an aspen pfd?) that should confirm if the flight plan is automatically pointing to the next (missed approach) waypoint and back to the hold.
If that is the case then it has to be something in the jpi.

I did some more checking, while unrelated to your situation, the engine data logging for LAT/LONG is problematic, but groundspeed recording should work. also hobbs time is not functioning (unrelated also to gps input obviously)
you may want to open a bug report with jpi, although I havent had much luck getting traction on some of the bugs uncovered due to what sounds like prioritization and resourcing.

x
Click to expand...

I do have a gpss as well as an mfd that rx RS232 and those did work normally. I haven't had a chance to get out and fly since, but it may be that I just went "direct to" an earlier flightplan waypoint rather than clearing the "suspend" message after the missed first.I plan on doing some more testing next week to see if I can nail it down.

Thanks for looking into it!
 
I downloaded the osx version and I can't for the life of me figure out how to get it to work. I open "Terminal" and change to the folder I have jpi_osx saved in (the same folder with my JPI files) and type jpi_osx -f U241026.jpi and I get a "command not found jpi_osx" error. What am I doing wrong?
 
try
Code: 
./jpi_osx -f U241026.jpi
 
GeorgeC said:
try
Code: 
./jpi_osx -f U241026.jpi
Click to expand...
That worked thank you! Now I have a couple of other problems...the .csv doesn't show all of the sensor data, I have a 6 cylinder engine but I'm only getting data for cylinders 1-4 plus OILT, OILPR, FF, MAP, and RPMLO. Also the -d command returns "Invalid Args!"
 
I can also help out porting your code to python to get around the platform/M$ issues....
 
Hey everyone, I just saw this thread. I wrote a JPI logfile parser awhile back, but just now threw it up on github in case it's useful. It's on github at hoche/libjpiedm.

It's really just the C++ library itself, although I did provide a sample app to indicate how it can be used. It only parses the later logfiles, although I may add in the older logfile parser too as I've been asked to help with parsing stuff from older EDM700s. I mostly test with data from the 930 in my Six. I'd really like to test with more variety of airplanes and EDM models though.

I've tested it fairly extensively on several platforms so that should all work fine. I'll probably be providing a code-signed binary for the Mac at some point (as well as binaries for Windows and Linux) but right now it's here on a compile-it-yourself basis.

I've also included a doc on how the JPI file is structured. There's (at least) one major bug in it that I'm still working out, which is how to determine the flight times. It seems that the sample rate can shift based on something (LOP/ROP find maybe?) and I haven't quite figured out what it is, although the values in the MARK field appear to give an indication of when it's happening.
 
Ok, fixed a bunch of things in my lib:
* figured out the flight times and when it shifts from 6sec to 1sec and back
* fixed an issue with date/time stamps on Linux vs Windows and OSX.
* redid the data record checksums calculations.
* added github's continuous-build action and some rudimentary CTest targets to make sure changes don't break it.
 
You must log in or register to reply here.
Back
Top