Data Logging, Combining 123+ data and AFR for tuning analysis Options

[Ishley]

Background: I rebuilt my ‘72 1.7 a few years back. I completely stripped the car back to bare metal… did a bunch of metal rust repairs… painted the car and eventually moved onto the engine. This Forum was a great resource as were some of the Youtube videos. For the engine rebuild, I basically kept the engine case and replaced everything else. All AA Performance…2056 with a webcam 73, 42X36 heads (no vents) and a 123 distributor… running DJET. The engine is actually a ‘73 replaced by some previous owner.

The DJET has also been completely rebuilt, MPS, throttle body and position sensor, new/old ECU (old one wouldn’t run the fuel pump), head temp sensor… wiring harness (Injector and Engine), 2.0 injectors along with about everything you can think of. I also have a Bursch exhaust. The car runs really good… and with the exception of a few failed parts, it’s perfect. I’ve put about 4K miles on it over the last 2 years.

Science Project: During the tuning process I discovered that the 123 + with bluetooth builds a log file every time you connect with your phone. I’m pretty sure this is the only 123 version that does that. The file resides in the File area of your phone (makes sense) and is a pretty rich set of data parameters. It's easy to dump the file into a spreadsheet.

I also installed an AFR gauge and a bung for a Wideband O2 sensor… and noticed that it had the ability to also log AFR readings. This got me thinking… Joining the two data sets together would be a great tool/ resource to help with the tuning of the car. So this past week I spent some time trying to get the AFR to communicate with the Windows laptop. I ended up upgrading/buying a real AEM gauge (mine was a Chinese knockoff) and some cables… and was finally able to collect data. I use Realterm software (free) and it can add a timestamp to each data point.

Next problem…. The AFR puts out about ten times the number of data points then the 123 distributor does. Also…Realterm date stamping is very limited compared to the AFR gauge. I haven't quite figured out how to marry the data completely together yet. I’m doing this in Google sheets… and find it slow and clumsy compared to using Excel (which I don’t have any longer). But I was able to chart the data sets individually and stack the charts on top of each other which got me most of the way there.

This was a cold engine (135F) and a quick drive( 3 Minutes) in suburbia …. And is just meant to show what is possible.

I also sent an email over to the 123 guys in Europe asking if they have ever considered building a tool and system like this (they said they are always evaluating ideas) There are now Blue Tooth enabled O2 Kits which could be a nice pair with a 123+ reporting tool. The idea works for FI or Carbs (and other cars)… as long as you have a 123+ and an O2 sensor/gauge that can log data.


Thoughts??

[GregAmy]

My answer involves Excel and X/Vlookups and marrying the data together...but that won't work for you.

My secondary idea is to pot both into data analysis collection, like a cheap-ass Race Technology DL1 Mk2 where you could collect wideband, advance, RPM, even CHT and MAP if you put sensors on it (I've done that with my street car's Microsquirt project). But unless you're willing to do some job box builds that may not work either.

I will offer that my fav wideband with data output for the race car is 14Point7's Spartan line, which has 0-5V data output.

https://www.14point7.com/pages/products

Data is enticing, even when you're not racing it. - GA, race data guy

[Ishley]

The problem is that the AFR data comes in in Hundreds of a second. You can had 5-8-15 reads per sec. The 123 data is one every 2-3 seconds and the time stamp only goes to the second. So when you look at the data... now you would have to choose which of the AFR data you were going to use. I'm think that you could average the AFR reads into second buckets...and even then you would have mismatches on a vlookup... where you had AFR data at a time stamp and now 123+ data.

I'm probably going to buy the new Excel so I have a more robust tool I can write code to handle the post processing better.

[emerygt350]

The problem is that the AFR data comes in in Hundreds of a second. You can had 5-8-15 reads per sec. The 123 data is one every 2-3 seconds and the time stamp only goes to the second. So when you look at the data... now you would have to choose which of the AFR data you were going to use. I'm think that you could average the AFR reads into second buckets...and even then you would have mismatches on a vlookup... where you had AFR data at a time stamp and now 123+ data.

I'm probably going to buy the new Excel so I have a more robust tool I can write code to handle the post processing better.

Libreoffice might be enough for this. It is free.

R and Rstudio is also free. That might make the bucketing easier, or just use perl or python perhaps.

[Ishley]

Libreoffice looks Like a good alternative. I Downloaded it and started tuning up my skills this afternoon. The sample 3 drive I did in my original assessment generated 19k data elements. So test drive will have to be short. Another option is to build a Script that samples the AFR gauge on every 2-3 seconds to cut down on the massive data. More progress coming.

[Robarabian]

Throw it into Chat GPT and see what it does.... "Analyze this for (add what you want)".... might be interesting to say the least.

Background: I rebuilt my ‘72 1.7 a few years back. I completely stripped the car back to bare metal… did a bunch of metal rust repairs… painted the car and eventually moved onto the engine. This Forum was a great resource as were some of the Youtube videos. For the engine rebuild, I basically kept the engine case and replaced everything else. All AA Performance…2056 with a webcam 73, 42X36 heads (no vents) and a 123 distributor… running DJET. The engine is actually a ‘73 replaced by some previous owner.

The DJET has also been completely rebuilt, MPS, throttle body and position sensor, new/old ECU (old one wouldn’t run the fuel pump), head temp sensor… wiring harness (Injector and Engine), 2.0 injectors along with about everything you can think of. I also have a Bursch exhaust. The car runs really good… and with the exception of a few failed parts, it’s perfect. I’ve put about 4K miles on it over the last 2 years.

Science Project: During the tuning process I discovered that the 123 + with bluetooth builds a log file every time you connect with your phone. I’m pretty sure this is the only 123 version that does that. The file resides in the File area of your phone (makes sense) and is a pretty rich set of data parameters. It's easy to dump the file into a spreadsheet.

I also installed an AFR gauge and a bung for a Wideband O2 sensor… and noticed that it had the ability to also log AFR readings. This got me thinking… Joining the two data sets together would be a great tool/ resource to help with the tuning of the car. So this past week I spent some time trying to get the AFR to communicate with the Windows laptop. I ended up upgrading/buying a real AEM gauge (mine was a Chinese knockoff) and some cables… and was finally able to collect data. I use Realterm software (free) and it can add a timestamp to each data point.

Next problem…. The AFR puts out about ten times the number of data points then the 123 distributor does. Also…Realterm date stamping is very limited compared to the AFR gauge. I haven't quite figured out how to marry the data completely together yet. I’m doing this in Google sheets… and find it slow and clumsy compared to using Excel (which I don’t have any longer). But I was able to chart the data sets individually and stack the charts on top of each other which got me most of the way there.

This was a cold engine (135F) and a quick drive( 3 Minutes) in suburbia …. And is just meant to show what is possible.

I also sent an email over to the 123 guys in Europe asking if they have ever considered building a tool and system like this (they said they are always evaluating ideas) There are now Blue Tooth enabled O2 Kits which could be a nice pair with a 123+ reporting tool. The idea works for FI or Carbs (and other cars)… as long as you have a 123+ and an O2 sensor/gauge that can log data.


Thoughts??

[Front yard mechanic]

I stared at your afr graft and it looks like your not in sinus rhythm

[Ishley]

I think I hit at least 4 stop signs. Takes a bit for me to get out where there is any rhythm.

[FlacaProductions]

Interesting - my wideband is bluetooth and also has a hardwire 5v output that I'm not currently using.

https://pe-ltd.com/product/wideband/

[Ishley]

The Bluetooth log from the 123 is collected on your phone. I don’t know of any other way to get the data. I used the serial connection from the AEM AFR gauge which a 2 wire connection TD and GRND. The afr just streams everything that the gauge sees… which is quite a bit of data. Most AFR gauges have a +5v wire which can be integrated into a modern ECU… but that’s not what a 914 has unless you are running MegaSquirt or Haltech.

The 123+ data log has the following data;

TimeStamp;CurveNr;RPM;Advance;Vacuum;Temperature;CoilCurrent;Voltage;ShiftLightA
ctive;GPSms;raw

There is an extra field which is always -30. I assume it is the max vacuum range. The temp aligns well with oil temp. I don’t use vacuum on my 123+ but I have it connected and I get the data. It’s a good indicator for throttle position. So with this data I can really see what happens when I accelerate, cruise, gain speed, my advance etc. I linked it with the AfFR data so I can see if my tune is in line with what I want. I really don’t want to go lean for extended periods.

The AFR data is just;
Timestamp, AFR

I find just watching the AFR gauge flash numbers helpful… but this datasets really helps confirm where my tune is and where I might want to make adjustments.

I’m running a non-stock engine with a mix of injection parts,
Cam etc. There is no book on what to do… just trial and error.

This same setup is useful for DJET and carb tuning.

[worn]

Thanks for starting this thread! It seems like this ought to be easier than it is. You have time vs voltage for everything going on, so making graphs should be possible without buying a lot of hard and soft products. At any rate, as someone who loves data from A/F.gauges and the like I appreciate your effort. I worked at a major university and always had excel for free. Even though I still teach some it has been hard to get some of the tools that I took for granted.