Dr.Jeff
True Classic
Going back to my prior post about knock sensing technologies.
As mentioned there are basically two general approaches; to determine knock levels as part of the initial tuning/setup process, and to sense knock real time with onboard engine management to control it. The technology for either approach is nothing new, but perhaps more refined in recent years. It's my understanding one major reason for this is faster processors, able to sample much more frequently in much smaller increments. As a result more aftermarket products are incorporating this enhanced technology (especially for the second application onboard the ECU), at a notably lower cost than before - although that's a relative thing.
After review of the information I collected at SEMA, looking at what's new on the market, and reading a bit more on the subject, I've come to a decision (at least for now). For my particular turbo project, the interest is more for the first application noted above - to determine at what point knock exists when setting up the initial tuning, so the tune can be kept conservatively below that threshold. Therefore I will stick with the advice I've been getting all along and not attempt to incorporate knock sensing into the onboard engine management system. Furthermore I do not intend to invest in one of the advanced detection devices to use with my initial tuning. At some point I plan to get the car on a dyno and let an experienced professional do some fine tuning with more elaborate equipment. But until then I'd like to see if I can get the baseline stuff sorted out enough to have things working reasonably well. That way the tuner can focus the dyno time on the fine tuning.
To do this I like the idea of using a laptop connected to a Bosch knock sensor, with a oscilloscope type sound analysis software. I've been reading about use of FFT and bandpass filters to help isolate the knock signal from engine noise. Granted it does not utilize crank position "windows", RPM, load, or other parameters to assist with detailed analysis. But that really isn't critical for this intended use. The goal here is to simply identity the existence of knock, so this method is sufficient. An even simpler method is to use any of the "listening" devices and just hear the knock. Unfortunately for me than won't work, my hearing is severely compromised. Too many years of heavy metal music, machinery operation, and hanging around racing engines. So I need to rely on other senses, and a visual display seems best for this.
The setup is quite simple. Bosch "donut" sensors are very broad scope (<5 to >20 kHz) and generate a low voltage signal (100mV). Therefore it can literally be plugged directly into the microphone jack on the computer using a shielded cable with a "minitimer" connector on the sensor end and a 3.5mm mono plug on the other. The sound card is capable of processing the signal and the analysis software is available as a free download and not difficult to use. If you wish to listen with headphones you can plug them into the headphone jack (sound card adjusts volume). Doesn't get any easier or less costly (total of about $15).
The (over simplified) principal is to play with the tune until you induce some knock, as detected on the laptop. Then back off the tune until the knock disappears, and finally back off a bit more for a safety cushion. Some road testing while a passenger monitors the laptop under various driving conditions will help with the process. Add some notes on RPM and relative load and you should have enough to know where the ignition maps need to be. Record AFR's during the process and the fuel maps can also be set (autotune will also help a lot with that). As stated, it isn't intended to be highly scientific and certainly not definitive. It is only to help get some sort of baseline until dyno testing fine tunes it.
As mentioned there are basically two general approaches; to determine knock levels as part of the initial tuning/setup process, and to sense knock real time with onboard engine management to control it. The technology for either approach is nothing new, but perhaps more refined in recent years. It's my understanding one major reason for this is faster processors, able to sample much more frequently in much smaller increments. As a result more aftermarket products are incorporating this enhanced technology (especially for the second application onboard the ECU), at a notably lower cost than before - although that's a relative thing.
After review of the information I collected at SEMA, looking at what's new on the market, and reading a bit more on the subject, I've come to a decision (at least for now). For my particular turbo project, the interest is more for the first application noted above - to determine at what point knock exists when setting up the initial tuning, so the tune can be kept conservatively below that threshold. Therefore I will stick with the advice I've been getting all along and not attempt to incorporate knock sensing into the onboard engine management system. Furthermore I do not intend to invest in one of the advanced detection devices to use with my initial tuning. At some point I plan to get the car on a dyno and let an experienced professional do some fine tuning with more elaborate equipment. But until then I'd like to see if I can get the baseline stuff sorted out enough to have things working reasonably well. That way the tuner can focus the dyno time on the fine tuning.
To do this I like the idea of using a laptop connected to a Bosch knock sensor, with a oscilloscope type sound analysis software. I've been reading about use of FFT and bandpass filters to help isolate the knock signal from engine noise. Granted it does not utilize crank position "windows", RPM, load, or other parameters to assist with detailed analysis. But that really isn't critical for this intended use. The goal here is to simply identity the existence of knock, so this method is sufficient. An even simpler method is to use any of the "listening" devices and just hear the knock. Unfortunately for me than won't work, my hearing is severely compromised. Too many years of heavy metal music, machinery operation, and hanging around racing engines. So I need to rely on other senses, and a visual display seems best for this.
The setup is quite simple. Bosch "donut" sensors are very broad scope (<5 to >20 kHz) and generate a low voltage signal (100mV). Therefore it can literally be plugged directly into the microphone jack on the computer using a shielded cable with a "minitimer" connector on the sensor end and a 3.5mm mono plug on the other. The sound card is capable of processing the signal and the analysis software is available as a free download and not difficult to use. If you wish to listen with headphones you can plug them into the headphone jack (sound card adjusts volume). Doesn't get any easier or less costly (total of about $15).
The (over simplified) principal is to play with the tune until you induce some knock, as detected on the laptop. Then back off the tune until the knock disappears, and finally back off a bit more for a safety cushion. Some road testing while a passenger monitors the laptop under various driving conditions will help with the process. Add some notes on RPM and relative load and you should have enough to know where the ignition maps need to be. Record AFR's during the process and the fuel maps can also be set (autotune will also help a lot with that). As stated, it isn't intended to be highly scientific and certainly not definitive. It is only to help get some sort of baseline until dyno testing fine tunes it.