I have this weird knocksum 2 show up, (don't know if it's real), which I cannot get rid off it around 2500 and 2700 rpm, and it's only 2 knocksum, so the car shouldn't pull any timing. By looking at the logs, it seems like the car is pulling alot of timing when those occur.
I attached screenshot of the log.
Notice how timing jumps from 12 to 9, also from 11 to 9, also from 11 to 8 at pretty much the same rpm and load. Have you guys seen this before? Is this normal?
I will also post the timing map.

 

It's not pulling as much as you think because of knock. 2 counts of knock will be about 2/3 of a degree. Our logging won't show 2/3, you can round that up to 1.

The rest of the timing difference is because of how you are travelling through your timing map (interpolation between cells, etc) and what load you are logging. It's looks fine to me.

 

My High Octane timing map is identical to the low octane timing map, so there wouldn't be any interpolation between different maps.
As far as I know 2 counts of knock should not pull any timing, if 1* at most. But in my case it pulls 3*, which bothers me because I can't figure out why.
If you look in the map at 2500 and 140 it's 12*, there is no 9 around in adjacient cells, so I don't understand why it pulls it down to 9*.

 

There is a 9 in your 160 load column. And 2 counts of knock will show as 1 degree being pulled in your log.

IF you are logging the correct load for timing, even then the runs are so fast, that it depends on your logging speed, etc...you can easily be off by 10-20 load and over a cell or a half a cell than you think.

For example, if you are halfway to the 9, then your timing would really be 10.5. Take away 1 for knock, now you're at 9.5. Our logger won't show fractions of a degree, so that will probably show as 9 (again depending on exact timing and numbers, etc).

Logging with something like DMA and multiple load variables would show a clearer picture. But, as I said, it looks fine to me.

 

Talking about interpolation between cells.

The amount of timing pulled is not linear across the RPM range, as your logged data shows. It maybe worth trying to prove if this is phantom or real knock, octane booster, a touch of meths or race fuel.

 

The 9 is at 2250 with 160 column. The log shows the 9* at 2500 and 140. That's quite a bit away. I thought this was a more accurate process, but I guess interpolation can vary.

Thanks guys for explaining how it works. I think I understand now.

 

Guys, this logs is @36% TPS, it is not a good reference to work.

 

Have you logged Knock Voltage to see what that is reading?

 

I see strings of 2's and 1's in the logs, all the knock that I see is between 2500 and 3500 rpm at partial throttle, barely into boost. I get 0 knock at WOT however.

I've set my CEL knock flash on 2 counts, so I can see it directly while I drive, and it flashes slowly many times in the 2500-3000 rpm range.

I will log knock voltage and post the log.