Following on from the above suggestion of graphing e-motor speed
instead of engine rpm in our NVH signal history, please see the information below:
Having knowledge of the e-motor construction is paramount and so using our Mk VII e-Golf (which has been reverse engineered here
topic22767.html) we can see our rotor speed and position sensor generates 5 electrical cycles per rotation of the rotor
- 1
Using the NVH “RPM Signal” option “Channel D Square wave/Tach” we connect channel D (via a x1 test lead) to the e-motor speed and position sensor
(G713) signal (See configuration below)
- 2
Using the above RPM Signal option and populating the Advanced Vehicle Settings dialog with the
total gear ratio (9.756:1) we can also obtain road speed (See configuration below)
- 3
With gear 1 selected, the vehicle was driven and the graphing of e-motor and road speed was achieved. However, note the spikes present on the RPM/Road speed signal due to electrical noise present on the G713 signal. These spikes occurred with the vehicle stationary in “Drive”
- 4
Below we have repeated the test above (5.0 Pulses per RPM) using “current” within the “U” phase via the TA167 BNC 200A/2000 A clamp
https://www.picoauto.com/products/curre ... rent-probe connected to channel D
TA167 is required as it can handle the typical e-motor current (10 mV/A in 200 A mode) and can be connected without “probe detection” by the NVH software if using a PicoScope 4425A. The NVH software does not permit the use of the BNC+ 2000 A clamp to Channel D
at this stage
- 5
Above we can see how progressive acceleration (approx. 50 seconds into the signal history) returns sensible e-motor and road speed values (approx. 10,000 rpm @ 75 mph). However, under other driving conditions the phase current is subjected to electrical noise, resulting in either spikes or crazy e-motor rpm values (> 20,000 rpm)
To view the noise present when using either the current clamp around an e-motor phase or a x1 test lead connected to the rotor speed and position sensor signal, we have used PicoScope
- 6
Note above the electrical noise present within phase “U” during over-run (freewheeling) along with the dramatic rise in current during braking regeneration and selected regeneration using phase shift
Below we have added a frequency math channel to highlight the effects on the graphed e-motor speed during the various driving transitions above
- 7
Applying a 10 kHz low pass filter
improves the graphed e-motor speed dramatically but does not fully cure! (at present we
cannot filter the RPM input into channel D of NVH)
- 8
Below we have the noise level captured with the vehicle in “Drive” but stationary; when selecting “Park”, the noise level diminishes considerably. (Hence the spikes present in the signal history when motor speed and position sensor and potentially phase current are used for graphing e motor RPM)
- Phase current
- G713 Signal
To conclude at this stage:
• Where possible, use an e-motor speed and position sensor signal as your “Square wave/Tach RPM” input (knowledge of rotor pole count will be required)
• Using phase current is possible but challenging due to electrical noise during driving transitions
NVH software recommendations:
• NVH to permit connection of BNC+ 2000 A clamp to channel D
• Filtering to be permitted when channel D used for “Square wave/Tach RPM” input
I hope this helps and rest assured the software team are working on a solution via CAN
Take care……Steve