Recording audio frequency signals
- [audacity] seem nice, but it is a cross platform tool and they are often farily unstable. Looks nice though.It has a VERY nice feature, when you press the large record button it will instantly start to record, if you stop and press rec again it will make a new recording in a new window. Perfect for us.
Signals to measure
- knock (see KnockListener)
- injector channels
- WideBand heater output
- wideband nernst signal
- HV: do not make contact, try to wrap a few turns around the HV insulation. Take extreme care.WhatEverYouDoYouDoItAtYourOwnRisk
- ignition transformer primary: this can kill as well. much bigger voltage division is needed than for 16V signal (eg. 5*470k in series insted of 100k) - be very careful
- logic level ignition signal
- supply noise
- ground noise
Best HW depends on notebook and the signal to measure.
- voltage divider
- R1=100k from signal
- R2=1k to GND
- optional: 1..10uF non-polarity (not electrolitic! a cap from GenBoard/VerThree/RescueKit is fine) cap to the soundcard input
- cable shield connected on one end only
This was sufficient for upto 16V signal (but soundcard mixer settings can tune this).
Note that there is a high-pass filter involved (the soundcard cuts freq below 20Hz), so the squarewaves are changed in a predictable way: possible to evaluate waveforms, but takes some getting used to. For us it immediately made temporary loss of injector signal (due to fully pressed TPS or processor reset) visible.
Possible to test and tune on the table with mdh commands. Eg. recording output from i259 chip at the IGBT gates and mdh02..72 for signal=low and mdh82..f2 for signal=high - make sure not to fry anything that's accidentally left on output, eg. a coil.