Harry's GPS Suite Forum

Harry,

I'm requesting a configurable "Max Throttle" option or a different implementation for detecting max throttle from OBD sensors. This option is needed for any Porsche with PDK, and I would suspect GT4's that will do an auto rev-match on downshift.


For vehicles with dual-clutch transmission like Porsche PDK or other vehicles that perform an automatic rev-match on downshifts, I've found that an abnormally high throttle value is sent, which causes Laptimer to use that abnormal value as new max throttle. After which the throttle graphic and values never reach 100% even when your foot is to the floor.

See this video here.. starting out, you get max throttle: https://youtu.be/bHGQdpTCbd4

Early you will see the Laptimer displays full throttle (all green), but after several rev matching downshifts, the throttle never get to 100%. i.e. like this: https://youtu.be/bHGQdpTCbd4?t=11m14s

I verified this quite a while ago, by running an OBD and sensors trace on Laptimer, and discovered very high throttle value coming in.... (it occurs on all Throttle PIDs.. I've tried all the over PIDs))

I see several solutions:

1. Create an expert setting to allow user to manually set the max throttle value. However, this is difficult to determine without running a trace and wading through a mountain of data to find the correct hex value.

2. Create an option like "Ignore Throttle Blips" and detect max throttle that is maintained for a minimum of 1 second or more (perhaps a configurable throttle detection period, where a default of zero, would result in current instantaneous detection). or this can be called a Throttle Detection Window.


Thoughts?

There is a request on vehicle specific min and max values delivered for throttle already. I assume this will be added at some time. The above behavior is quite complex (and I hear it the first time). So I think I will ask users like you to run a sensor data log (using Raw or OBD Sensor loggings - see the FAQ in how to send trace logs) and evaluate the max value individually.

- Harry

I do have some (older) raw throttle trace log files that I was analyzing. The ability to set a "fixed" max value would be helpful, although cumbersome to determine. One needs to run an OBD and Sensor trace, find a place to run at full throttle, then go thru the trace log to find the actual max value and then convert from hex/octal notation.

The issue is that PDK sends an abnormally high value for throttle during an auto rev-match. This spike may only be a single value, and certainly not more than 1/4 second in duration. Hence my suggestion to use a moving average approach over a period of 2 or so seconds to determine maximum value.

For a 2 second moving average, you would need an array of 20 values (given a 10Hz GPS scenario). Fill the array with 20 samples, and take an average as an initial baseline (startup). Then as each new throttle PID value arrives, you would take the oldest value from the array, subtract 1/20 of it from the moving average, then store the new value in the array and add 1/20 of it back into the moving average. Then test the moving average for a new maximum. An approach like this seems to be doable with little processing overhead (to me anyway).

The 2 second window and 20 value array size could (should) be derived from a "Throttle Detection Window" setting (perhaps 1-3 seconds)... and the array size determined by the GPS or PID rate, multiplied by the window size.

Have you tried looking at the alternate throttle PID's? What you want would be either the pedal position or the actual throttle position rather than the commanded opening.

gplracerx wrote:Have you tried looking at the alternate throttle PID's? What you want would be either the pedal position or the actual throttle position rather than the commanded opening.

Thanks GPL.... I finally found a PID that works for Porsche PDK transmission vehicles. Position D is "Accelerator Pedal Position". The default PID seems to map to a PID which is "Relative Throttle Position"... During a downshift, the PDK logic automatically blips the throttle open and produces a high value. With Position D, I do not see this behavior (at least in neighborhood testing). I have a track event soon, where I'll be able to determine if this position works and shows full throttle continuously when my foot is on the floor

Just a follow-up to confirm that using Position D solves the issue I was having, per my post above. Using Position D, which is "Accelerator Pedal Position", my OBD data now shows what my foot is doing, versus what the ECU is doing to the throttle for PDK rev matching.