I'm not hearing any knock but the knock sensor hit 104 (units?) on a boosted run. How does the sensor output correlate to damaging knock? Should the timing be retarded above a certain level?

 

I've seen some 100+ readings as well and have never heard any knock. Not sure what to make of it.

I tried to load a couple of my runs into Megalog Viewer and have been unsuccessful. What format do they have to be saved in? .csv?

 

It should be in the .csv format but you have to remove the quotation marks from the header - See Turbo Roaches post below. I opened the file up with note pad and cut and paste TR's header and resaved.

https://www.miataturbo.net/forum/t33439/

 

The NB factory sensor output needs to be filtered to help the signal to noise ratio. I posted a schematic for such a filter in the ECu section.

 

Here is a good thread on the Adaptronic forum:
Knock Logging...
Based on this thread the level you are seeing is about right. The Adaptronic has a programmable filter built-in which is what the thread is about.

Based on some use of electronic det cans, you can still have some light knocking that you will not be able to hear listening normally. So be careful.

 

I glanced at that thread, and they dont seem to be talking about frequency filtering that I'm talking about.

 

Does it give you a line level output that you'd plug into an amp of some sorts?

 

No it is more like a microphone than a line output; it is a passive device. But it has enough power though to plug it into an amp.

Jason you are right as far as i can tell the Adaptronic setup is just a gain versus RPM filter, not a frequency filter.

 

I found this in the manual - appears there is filtering.

 

FWIW the circuit I designed is more specific for the miata - it has a narrow bandpass filter tuned for the miata block's knock frequency. It also provides the power necessary to power the factory knock sensor. Its output voltage will be compatible with knock inputs designed for piezo sensors.