J'ai donc envoyé ces résultats à Maurice Hilarius (le concepteur du boîtier additionnel) pour connaître son avis; voici ça réponse :
"The stock CANSCAN unit from Harrison is not calibrated for the O2 sensor, and the data is essentially useless.
Also, the communications chip on the standard CANSCAN is too slow and gets overloaded when used on an RX-8. CANBUS makes a lot of data. If you want some meaningful data from it, you need to get the upgrade of chip and software to support RX-8 and logging.
Also, you might want to try a newer map, and/or calibrating to your car."
Et dans son message suivant (je pense que Rasp appréciera !) :
"As for the CANSCAN, the A/F (O2) calibration is not useful in the standard version.
We calibrated the sensor output for use with tuning the RX-8 by logging simultaneously with an NGK sensor which is known to be accurate.
The reason it is not calibrated in the stock ECU maps is because Mazda do not use the sensor for "open loop" mode.
They only use it when close to 14.7:1 to maintain lean burn in "closed loop"
When one goes over 70% throttle position and above a certain rpm ( which varies depending on what gear you are in) they then transition to "open loop" and use a lookup table rather than the O2 sensor.
After we developed a calibration curve for the stock Bosch LM42 sensor they used, Dan incorporated the calibration in the CANSCAN version we sell for the RX-8.
This version also has a faster processor and UART chip to handle the greater sample rates needed to accurately log more than two factors. For example, one wants to log: throttle position, rpm, O2S1, and ignition advance.
That is a set of 4 data sets, and one needs the faster chip on the CANSCAN to manage that.
The new USB version of the CANSCAN that is now just shipping is even faster and will accurately log up to 10 data sets.
BTW, if you want to upgrade an existing CANSCAN to CANSCAN RX-8, the cost is $45 and includes a new chip and different software."
Je vais donc upgader mon CANScan avant de poursuivre le tuning du CZ ...