Wildguzzi.com
General Category => General Discussion => Topic started by: Impulsive_Duc on December 30, 2014, 07:20:48 PM
-
This is a off-topic discussion forked from the "HowTo: GuzziDiag" thread ending in comment "msg1157177" (http://wildguzzi.com/forum/index.php?topic=69168.msg1157177#msg1157177).
Please feel free to join in and add your thoughts! :)
John, I've not found a wideband controller with dual differential narrowband emulated outputs. In your case, you would need two single channel units operating in parallel. It would get pricey!
Two LC-2's would be rather compact, but the wires are quite bulky. Then there is a LM-2 which has multiple differential analog outputs. The "advanced" version has 2 integrated lambda controllers, and also supports internal logging. (It is a bit pricey.) Finally, there is something like the WEGO-4 which has dual controllers but is rather "dumb".
If the 5AM ECU handles 0-5v from the sensors, that would suggest that a kind of "plug and play" wideband sensor kit is conceivable, where the rider could dial in the AFR required. A world of possibilities!:-).
(...)
I did actually approach a manufacturer here in Europe a couple of months back worth a suggestion that they think about producing such a wideband kit. Unfortunately they were not interested at this time. I think technically it shouldn't be such a big challenge, but there needs to be a certain ROI for the manufacturer, and I expect the potential Guzzi market is not big enough.
I have already achieved this. :D
(http://www.ducati.ms/forums/attachments/sport-classic/460242d1413727877-datalogging-cheap-off-stock-narrowband-lambda-using-ecu-elm327-harness_overview_800px.jpg) (http://www.ducati.ms/forums/attachments/sport-classic/460210d1413727664-datalogging-cheap-off-stock-narrowband-lambda-using-ecu-elm327-harness_overview.jpg)
I make this by hand. I also considered offering a such a premade kit using premium components. It occurs to me now it would also work for Guzzi riders. It is completely open-spec, so it is more getting the right parts assembled, tested, and wired correctly. As you say, there is little ROI; I am inquiring if CCW is interested in sponsoring a kit based off this design.
If I understand you correctly, I could use the "analog out" (<-digital out) data from the LM2 via BT, PLUS the data from a separate BT dongle connected to the ECU! Interesting. I didn't know it was possible to connect two BT devices simultaneously.
This was the code I submitted to Christian. ;D However, he deserves the credit for actually implementing the solution. There are still small bugs we are ironing out.
You can use either the Digital or the Analog out from the LM2. Although the analog is differential on the LM-2, I still recommend the digital out!
Innovate sensors typically output 3 ways: (2x) analog (1x) digital. My first design I tried routing analog 0-5v AFR into the ECU pin (22 signal, 32 signal gnd) through the stock plug. However, the delay from reading Lambda off the K-line using ScanM5X is caused by the ECU, and cannot be avoided. (More data = slower refresh Christian coded some clever workarounds to improve this. :) ) This is what prompted the second version of the design, which instead reads AFR from the digital output. There is no performance penalty since it is done from a different source! This increases logging rate from ~600ms to ~250ms.
One important note, the digital signal is not affected by grounding issues like ground loops or grounding offsets. Also, we are able to piggyback up to 32 other Innovate "channels" with no performance penalty (Temp, RPM, AFR, gyro, AUX, etc). For practical reasons, Christian capped it at 2 in ScanM5X: Lambda front/rear. However, there is no functional limitation! Christian is planning to make a universal Innovate logger for Android soon. ;)
I have been experimenting with using a wideband sensor to dial in the wanted AFR to the 5AM. There's a couple of issues.
1. The 5AM expects a narrowband Rich/Stoic/Lean signal only. You must program an analog output to emulate a narrowband signal, but you can still choose your target AFR as your 'Stoic' point.
2. You can't use an LC-2. The LC-2 has a single ended output. A differential output is required or you get a SERVICE error. :D. Fortunately this is a good thing, because it completely eliminates the offset voltage noise problem.
- In the Ducati firmware for the 5AM at least, it is possible to disable narrowband lambda regulation by turning off 2 flags (or using a "DP" open loop map). This makes the sensor input not affect anything. Then you can feed a 0-5v signal to the pin, and it is read correctly without affecting fueling! It is just another byte of data to the ECU at that point. I have experimented numerous times and have datalogs to demonstrate.
The analog outputs can be programmed from (0-5v) wideband to (1-0v) narrowband @ 14.7. Although it is highly shunned by the engineers, as you suggest, it can also be changed to (1-0v) narrowband.... @ 13.8. ;) Essentially, turning the stock ECU into a "PowerCommander-5 with AT-300 autotune" for just the price of the sensor and harness. It is then possible to edit which regions the Narrowband input affects by editing the "Lambda correction map" in the ECU. - The LC-2 analog AFR lines are differential to the common ground (shared with the heater). After consulting technical documents, I solved the ground and calibration issues and was able to read analog AFR through the ECU to .1 AFR accuracy (with a LC-2!). The old LC-1 did have an additional "signal ground" wire, which in my conversations with Innovate engineer Adam, was eliminated for "simplicity". However, these units have numerous other problems I won't discuss here.
However, I am an engineer so did all my designs properly. ;-T While my harness is capable of the above, I have it set so instead it honestly reports analog narrowband (14.7) to the ECU, while reporting true wideband to the datalogging app ScanM5X. I get the best of both worlds: my bike is street legal, and I am able to log wideband data!
If I ever change my mind, I can reprogram the ECU to a "DP performance" map, disable lambda, and set the wideband sensor to output a 0-5v signal again. (Or, continue to read it digitally.)
I'll think about it and maybe order a couple of new BT devices.
The biggest problem for me as a die hard Apple user will be to get the beta version of the App to work! :-).
I recommend you purchase the device pictured in the video posted. I have sent configuration instruction for Christian to share. Most Bluetooth-RS232 adapters are $60 or more, the one I created is maybe $10? You merely need to supply it with 3.3-5v power, I accomplished this with 3x "AAA" batteries. It is called a "HC-06 RS232 DB9" adapter board. I am working on making a proper housing and other tricks to make it "better". :)
I have also uploaded the newest version of ScanM5X (1.10e). Please test it and let us know of any bugs! ;D
It should work with any Guzzi using 59M and 5AM units.
-
Thank you for this. Beetle will now have the arduous task of trying to explain this in a way my ageing analog brain can comprehend. Thing is if I want to do this stuff I have to learn the new skills required. I feel lucky to have people so willing and eager to share their knowledge and skills.
;-T
Pete
-
Thanks Impulsive Duc,
I have installed the updated app 1.10e. Unfortunately I live in a cold and wet country so I won't be testing it any time soon! :-(.
I'm looking at BT RS232 adapters with built in power and I suppose I should be able to connect one up to my LM2. I still think for data logging purposes, the LM2 alone is all I need. But it certainly would be interesting to be able to see all the other ECU data in the log too. (That was what I hoped to get by hooking up the OBD2 connector directly to the LM2, but that unfortunately didn't work via the adapter cable that we need on the Guzzis).
The dynamic tuning aspect is what interests me the most, and I do think there must be a market for that kind of kit (basically just wideband sensors and a dial, going straight into the standard ECU) without having to piggy back with other devices.
The possibility to adjust the AFR via a "Lambda correction map" sounds fascinating too, but I have never seen that table in the map I have. Maybe it's hidden, or just a Ducati thing?
Thanks again
John
Edit:
However, I am an engineer....
To my lasting regret, I am not.. :-). So I have to read all these kind of posts at least twice and very slowly! :-).
Sent from my iPad using Tapatalk
-
The LC-2 analog AFR lines are differential to the common ground (shared with the heater). After consulting technical documents, I solved the ground and calibration issues and was able to read analog AFR through the ECU to .1 AFR accuracy (with a LC-2!).
ID, how did you solve the ground issues? I've tried various combinations to power the LC-2 and feed a programmed NB signal to the ECU. The result is always a
heater current failure DTC. I have a kludge solution (dummy load) on the heater supply from the ECU, but it's not very elegant. :D
The answer eludes me. I can't see it for looking. ;D