I am trying to clear up a few things in my PV logs regarding knock:
1) how does the log report an event? Knockcnt? Spark knock? Spark Advance?

2) what exactly is short term and long term timing? Is adaptive knock considered long-term? Is spark knock short term —ECM’s reaction to a knock event? And is Advance showing long term? I tried zeroing out adaptive knock table , but not really seeing difference in my logs.

3) I take Advance F and Advance R and subtract it from the Spark Advance table. Is this short term or long term. ?

4) how does spark intensity correlate to knock event? Is there a threshold I should monitor? I really want to know if and where I am getting knocking.. example: I have 2, 20 min logs. One log has adaptive knock zero’ed. Both logs showing plenty of activity of timing removal and knock intensity, but only 1 or 2 knockcnt. Bike runs good and can’t hear any knocking whatsoever.


Sorry for the long winded post. I hope these questions make sense. It seems the deeper I dive into timing the more confused I get 🤷

 

I was told just a couple like that was nothing to worry about.

 

thanks. I heard that as well. Also, those few events happens after a hard acceleration—so I added more file in my PE table. Going do some logging today.

 

I've been doing a lot of logging and retuning lately since I moved to CO where the highest octane is 91. Had a bunch of knocking on my old tune and pretty much have it dialed in now. First off, in my experience with Harleys you will never get rid of all knocking. But that's ok because adaptive knock retard exists to protect the engine. On hot days or when you're beating the **** out of it, it's just going to knock occasionally. That's just the reality of air cooled engines. Those who say their engines don't knock just don't notice it because they don't have their PV hooked up. Plenty of knocking happens even if you can't hear it, but generally speaking if it's minor enough that you aren't hearing it, it's not gonna do damage.

The spark tables are "long term" in that every time you start the engine, that's what it's going to do. Adaptive retard happens during the ride once it detects knocks in a certain area of the MAP, and is only maintained for that ride. Once you key-off, it's gone. So every ride starts fresh going off the spark tables in your tune.

I just realized I don't know what bike you have. I'm speaking based on Dynas. No idea if the same is true on the newer bikes.

 

Sorry, I have a 2021 softail with SE 447 cam and Vance and Hines Pro Pipe 2-1 and a Power Vision tuner.
i agree with you and think my tune is coming along pretty good as the bike runs great. My VE seems good as my fuel trims are less than 3%, so I am now looking to timing. So before I start removing timing , I need to know how to identify real knock events in my logs.

For example, in my logs, I see timing removed by: Spark Knock F and Spark Knock R and Advance Spark F and Advance Spark R. But KnockcntF and KnockcntR counts do not correlate with either of the two.

I guess what am asking for are specific descriptions of the spark related fields in my log:
Knockint — Knock sensor output?
Spark advance — long term spark?
Spark knock — short term?
knockcnt — current events?

So what do you monitor for knock as you log? What tuner do you use?

also want to get my head around this so I can created a tune for Sturgis as it’s pretty much only 91 octane there 😀

thanks!

steve

 

No unfortunately the adaptive knock retard is only removed after, if I remember correctly, 50 key ons of no activity. It's not even reset when you flash a map, unlike adaptive fuel values. This is why when tuning you zero out adaptive knock retard tables.

 

This must be new behavior with the M8s cuz I assure you that's not true with my 2013 Dyna. Every run starts fresh. How do you even zero out the retard tables?

 

I also thought knock retard was reset on key on—and Spark Advance decayed over time, something like 500 miles? But the more o research this, the more confused I get. I wish there was a good Delphi document on how the ECM controls knock 😂

 

I use a PV on my 2013 Dyna, so there may be some differences, but it seems like it's very similar.

Knockint - no idea, I don't have this
Spark advance F/R (deg) - this is just the current degrees the spark is advanced at that particular part of the map, otherwise known as your ignition timing, and corresponds to your Spark Advance tables in your tune. This needs to be decreased if you start getting knocking in that area.
Spark knock F/R (deg) - this is how many degrees the spark is being retarded by the adaptive knock retard in that particular area of the map. So as knocks are detected you'll start seeing the numbers go up here, and there will be a corresponding drop in degrees in the Spark Advance column.
Knockcnt F/R - this is simply the running tally of knock events detected in each cylinder. On my Dyna, this number is reset at every key off cycle. Probably does on yours as well.

I keep my PV connected and mounted to my handlebar and set up live gauges to show me Set Lambda, MAP, Closed Loop Integrators for both cylinders, and Knock Count for both cylinders. If on a particular ride I start getting knocks, it hit the log recording button to save a log and then take it back to my computer to see what's up, and then adjust my spark tables and/or fuel as I see fit. It's a slow process, but I ride to work every day so it gives me ample opportunities to get good log runs. Check out my spark table deltas compared to the tune I was using at sea level. You can see how much timing I had to pull back in the center just to get the knocks under control using 91 octane vs 93. I still have a few here and there, but nothing like before. Keep in mind that at mile high, my MAP doesn't go above about 85kpa.

 

According to the WinPV description of these settings, the adaptive knock retard does decay over time and you can adjust how much and how fast in the settings. But again, there may be some differences between my 11 year old Dyna and the new software.