Data Analysis Software - Motec, Gems, AEM

[JJLudemann]

Hey guys,

Having some issues with GEMS itself.
When you have the data loaded and try to switch from time to distance mode it's goes nuts.

As viewing data by time is nearly useless for comparison.

Any help appreciated,

Mat Walker
DREAM Motorsport (F4, FFord)
Pedders Racing ( GT86 Australian Production Touring Championship)

I know this is an old post, but I think it's worth it to document my experience.

All the basic steps worked perfectly: Download & install Perl, download GEMS & install application, download & install DLOG99 application, download & install Perl Text::CSV module, download the Perl conversion script, run it on a RaceCapturePro2 log file, import the resulting file into DLOG99, save it as .stf file, open the .stf file in GEMS. Yay!

When I clicked the GEMS View > Toggle Time/Distance to distance, I could no longer scroll the trace window left/right, and the cursor would only take huge jumps. Similarly, the cursor on the track map would only make a few jumps around the track. My "Distance" channel was good, incrementing smoothly to the correct lap length and resetting to zero at the start/finish line. But I noticed that the "Lap Distance (m)" channel would only go from 0 to 3 over a lap. As in a lap length of 3 meters. This also meant I couldn't color the track map with a channel, because apparently it only does that once per meter of Lap Distance.

So, I went to Data > Distance Channel Setup... unclicked the Distance channel in the first box, selected the second box "Calculated Distance" based on the Speed channel, (click OK, of course), and everything's suddenly wonderful. Lap Distance (m) is now correct, I can use both Time and Distance trace overlays, and the track map is all kinds of pretty colors.

I don't know where the Lap Distance channel is getting messed up, because I checked my .csv file in Excel, and the Distance channel there is correct, and the Distance channel in GEMS is also correct.

I hope this is helpful.

Regards,

-JJ

[DigiPot]

Hi All!
First of, thank you very much for spending time making this toolkit and publishing your hard work, its really appreciated!

I have been able to use it somewhat, having found a log file somewhere here on site and putting it through the converter to get a file open in GEMS.
No warnings were given,and I can make graphs appear. So far so good.

However, when opening the .stf file in GEMS, I cannot seem to make it understand laps and the channels that are in the file. It does not understand what is Engine speed, Vehicle Spped, Damper movement etc.
All the channels are in the "channel values" list, but not sorted correctly.

Don't know if anyone has experienced the same ?
The file I have used as starting point is the Laguna Seca file:
https://www.dropbox.com/sh/o6u4kil4xdv4 ... 0logs?dl=0

But I think it is formatting correctly, when opening the csv file I see the following headers:

Code: Select all

Utc,Interval,Battery,AccelX,AccelY,AccelZ,Yaw,Pitch,Roll,RPM,TPSA,TPSB,PPS1,PPS2,OilLEV,P_APS,RPMDash,S_Veh,P_FRP,P_FP,FuelUs1,EOP,ECP,MAP,P_FBrake,P_RBrake,T_Br_FL,AAT,AirT,ECT,GBT,EOT,EGT1,EGT2,P_FLT,T_FLT,P_FRT,T_FRT,P_RLT,T_RLT,Latitude,Longitude,Speed,Distance,GPSSats,LapCount,LapTime,Sector,SectorTime,PredTime

I am confused as to why dlog99 would want to capture LapCount and LapTime, I thought that was the software's job, polishing the GPS signal etc ?

[Bryan@UnderbiteRacing]

Hi -

I'm still coming up the learning curve with my RaceCapture and GEMS, so maybe you've already tried these things.

• - Have you added a GPS Beacon on your track map near the start / finish line and adjusted the range (I've been using the time window) until laps appear?
  - Does your channel list have the default channels, or channels with the same heading as your CSV file?

To your last comment, I wasn't expecting the software to polish the data; I was actually hoping it WOULD use the lap count and lap time data to segregate the laps so I would have to use the GPS beacon mentioned above.

The reply is very late, but hoping to generate a little more traffic in this data analysis section.

Cheers,
Bryan

Hi All!
First of, thank you very much for spending time making this toolkit and publishing your hard work, its really appreciated!

I have been able to use it somewhat, having found a log file somewhere here on site and putting it through the converter to get a file open in GEMS.
No warnings were given,and I can make graphs appear. So far so good.

However, when opening the .stf file in GEMS, I cannot seem to make it understand laps and the channels that are in the file. It does not understand what is Engine speed, Vehicle Spped, Damper movement etc.
All the channels are in the "channel values" list, but not sorted correctly.

Don't know if anyone has experienced the same ?
The file I have used as starting point is the Laguna Seca file:
https://www.dropbox.com/sh/o6u4kil4xdv4 ... 0logs?dl=0

But I think it is formatting correctly, when opening the csv file I see the following headers:

Code: Select all

Utc,Interval,Battery,AccelX,AccelY,AccelZ,Yaw,Pitch,Roll,RPM,TPSA,TPSB,PPS1,PPS2,OilLEV,P_APS,RPMDash,S_Veh,P_FRP,P_FP,FuelUs1,EOP,ECP,MAP,P_FBrake,P_RBrake,T_Br_FL,AAT,AirT,ECT,GBT,EOT,EGT1,EGT2,P_FLT,T_FLT,P_FRT,T_FRT,P_RLT,T_RLT,Latitude,Longitude,Speed,Distance,GPSSats,LapCount,LapTime,Sector,SectorTime,PredTime

I am confused as to why dlog99 would want to capture LapCount and LapTime, I thought that was the software's job, polishing the GPS signal etc ?

[Bryan@UnderbiteRacing]

I ran into the same issue when trying math channels in AEMData/GEMS.

My work around has been to create the math channels in excel after converting with the perl script and before importing into Dlog99.

Yeah, that's easy for certain things, harder for others... like calculating time slip between arbitrary laps which is what I'm most interested in.

Did you ever figure out a way to determine time gaps between arbitrary laps and use that information for focus analysis of data in GEMS?

Thanks,
Bryan

[wizrd54]

I have created my own Python script to convert the sparse CSVs into a tab delimited text file that can be read by Bosch Motorsports WinDarab. I doubt many people are using this program because a full license is expensive, but it's top notch. If anyone wants it, let me know.

[MCScott]

@wizrd54, I'd be interested in getting your Python script - WinDarab looks pretty interesting, even the free version.

Thanks,

Scott

[wizrd54]

@wizrd54, I'd be interested in getting your Python script - WinDarab looks pretty interesting, even the free version.

Thanks,

Scott

Here you go. The variable names and code is a bit of a mess because this evolved over time and was mostly me coding at track days. One note is that every once in a while I'd end up with a file without GPS lock at the beginning. This seemed to mess up the time stamps in the very beginning causing a few negative values. My script doesn't like this so I just deleted the lines in the beginning that had negative times and re-ran it through my script.

I also interpolate everything so that they are at a fast rate. My engine speed is usually sampled the fastest (10ms maybe) so I interpolate to make all of the channels this fast as well. For example, if RPM is sampled at 10ms and vehicle speed at 100ms then I will have to fill in the vehicle speed every 10ms. I do this by holding the previous value at each time stamp until the new value is sent at 100ms.

My script doesn't make a native WinDarab file either. You need to use its import text file function. Open the main menu (orb in top left) then click Import/Export -> Import from text file -> select ASCII file. You can leave the channel definitions box blank.

You can convert more than one RaceCapture at a time, but can only import one at a time to WinDarab. Very important note is that the script was made using Python 3.6 NOT 2.7. Let me know if you have any questions.

It sounds like a lot, but isn't bad...but maybe that's just because I made it.

1. Run script
2. Select RaceCapture files to convert
3. Converted files will be placed in the same folder as the script file
4. Open WinDarab
5. Import each file into WinDarab
6. Analyze files like the pros!

https://drive.google.com/open?id=1-Laf0 ... bxVlET0Nxl

[imstimpy]

I cannot find DLOG99 for download from GEMS.

Any updates on how to proceed with utilizing the GEMS DataAnalysis without paying $400 for the Pro and the native csv import?

[jpf11]

imstimpy, pm'd you.

If anyone else needs files I can share them via google drive. I just need an email address.

[imstimpy]

jpf11,

I sent you a reply via PM. I didn't get a notification that you'd sent me the original PM. Did you get mine?

[imstimpy]

Much to my dismay, the free versions of Motec i2, AEMData, and GEMS DataAnalysis all require the data be in a proprietary format other than CSV. As such, the ability to use alternative data analysis software is contingent upon a conversion, either the no longer available GEMS DLOG99 or a bespoke converter that nobody has yet written.

I found some files from before and after a DLOG99 conversion from www.flyingratms.com/downloads . With some time I may be able to recreate the STF but I don't have much time for projects like that at the present.

[AM150]

Much to my dismay, the free versions of Motec i2, AEMData, and GEMS DataAnalysis all require the data be in a proprietary format other than CSV. As such, the ability to use alternative data analysis software is contingent upon a conversion, either the no longer available GEMS DLOG99 or a bespoke converter that nobody has yet written.

I found some files from before and after a DLOG99 conversion from www.flyingratms.com/downloads . With some time I may be able to recreate the STF but I don't have much time for projects like that at the present.

Getting it into the free version of GEMs is a hassle and the free version lacks math channels which are required to create the incredibly useful timegap/timedrift channel. I have been trying to buy the paid version of GEMs but they make it extremely difficult here in the US. The only dealer for them wants me to email my CC number and then text him the 3 digit code on the back... No paypal?

The AEM version of GEMs is free and they have math channels, but there is something in there that can identify an AEM file vs. an external file and blocks the use of math channels on external files (I can't tell what it is about them that triggers the blocks).

I have even contacted the fine folks at the now defunct traqmate to see if I could purchase the source code for their analysis software but unsurprisingly they were not interested.

It's frustrating that there is not an easy option for good analysis software for RCP.

[brentp]

Hi,

On the heels of our latest release, (https://www.autosportlabs.com/racecaptu ... available/) we are in the position to dramatically improve analysis features for RaceCapture/Podium so we can offer a suite of useful feature that makes RaceCapture and Podium a fully stand alone system.

We have a list of high priority features, focusing on analysis and video integration capabilities we'll share shortly.

What can help our effort is to understand how you want to do analysis, and to know what your 5-10 items you need to see to make analysis complete for what you need.

Please feel free to reply here with your list; we'll follow up with news and official announcements as well. Thanks!

[AM150]

Hi,

On the heels of our latest release, (https://www.autosportlabs.com/racecaptu ... available/) we are in the position to dramatically improve analysis features for RaceCapture/Podium so we can offer a suite of useful feature that makes RaceCapture and Podium a fully stand alone system.

We have a list of high priority features, focusing on analysis and video integration capabilities we'll share shortly.

What can help our effort is to understand how you want to do analysis, and to know what your 5-10 items you need to see to make analysis complete for what you need.

Please feel free to reply here with your list; we'll follow up with news and official announcements as well. Thanks!

I think the features are going to be based on use case, an HPDE guy will be different to an enduro guy who will be different to a time attack guy. I'll give you my perspective as someone responsible for managing the data systems for a national level time attack team. When the car returns to the pits here's what I'm doing:

1 - I go through tables that we have setup that show me min-max values of various channels that we have setup logging what's going on with the car and engine. Any values of concern (like a low oil pressure reading) I then look at the graphs to get more detail. So being able to see summary data and quickly jump to the point in the graphs that that data occurred is important. If I have an issue I can then relay that to the crew chief to decide what he wants to do about that.

2 - I then set about reviewing the performance channels. I'll start comparing lap time against distance travelled for the current session and previous sessions or laps within the same session (either on that day or previous days). It's really important to be able to visualize the graphs based on distance and time, although we use distance most often.

3 - The first graph I am going to look at is the time drift graph, this shows me where the two laps are gaining or losing time in comparison to each other and try to figure out why. The figuring out why is the tough part, sometimes it's a different line, sometimes braking at a different point, carrying more or less speed, etc. etc. So here I am adding plots to graphs on the fly. I need to be able to add seemingly unrelated channels to the same graph and scale the data to where it makes sense.

4 - I then need to mark these areas and review with the driver to see what's going on and where we think we can make improvements. He will often ask me to look into specific areas or turns where he did something different and wants to know if it worked well. Video integration here would be AWESOME.

5 - Being able to see a summary of segment analysis. Turn by turn, straight by straight, see distance travelled min speed, max speed, time taken, etc. and be able to build up a theoretical best lap time. This gives us a moving target of what we're shooting for, but it is a target.

Please spend some time with GEMs, the configuration, being able to set up different views, tabs, etc. is great and that the level that we need with our time attack car but there is a learning curve.

Being able to add notes to a session would be a very nice feature to have, so many times we've got towards the end of day 2 and comparing to the first session of day 1 and we can't remember what the setup was.

Honestly, I don't know what licensing of GEMs looks like, but they've done it with AEM. I would reach out to them to find out costs so you can do a build vs. buy analysis.

[imstimpy]

Here is my list of features requests and bugs from the last time I was really thinking about what was needed (about a year and a half ago)

Top Features

• Automatic segment generation at a minimum, configurable segments saved to the track database would be ideal.
  Time delta between selected laps at current location (see SoloStorm Tutorial #2 3:34)
  Per segment summary providing entrance and exit speed; average lat, long, speed; and distance traveled (see SoloStorm Tutorial #2 2:58)
  Traction circle through automatic selection of lat/long accel channels at a minimum, configurable X/Y plots would be ideal (see GEMS Understanding the Friction Circle 3:07)
  Multiple channel trace windows (see lower window of VBOX Circuit Tools)
  Theoretical best lap with indicators of which segment from which lap are best (see upper left window of VBOX Circuit Tools)
  Zoom on the channel traces
  Traditional red/blue crosshairs for cursors on multi-lap analysis (see SoloStorm Tutorial #2 1:34)
  Channel summary per lap (min, max, average, etc)
  A standard white background with light grey grid overlaid with crisp channel traces

Bugs

• The lap times should all be in base 12, base 60 not base 10, base 10 (e.g. 1m32s vs 1.53s); time deltas should be base 10, base 10.
  The color of the channel traces for the same channel of different laps should never be the same color.
  The data loaded when selecting the fastest time is the lap AFTER the fastest lap.
  The channel/data overlay should not obscure the latter portion of the channel traces
  The channel/data overlay clips channel name with longer log names (e.g. Jon_CotA_1... vs Jon_CotA_1_27_18 RPM)
  The inconsistent use of the check, trashcan, and x. Sometimes trash is delete, sometimes x is delete. Sometimes check is OK and x is cancel, sometimes check is select and x is OK.

SoloStorm Tutorial #2
VBOX Circuit Tools

GEMS Understanding the Friction Circle