Page 1 of 1

New knock control tables

Posted: Tue Aug 30, 2011 1:45 pm
by Honda9ower
I'm interested in using the new knock control tables in version 1.5.1.0, just in case I get a bad tank of gas, but just want to check my understanding first. I have read the help file several times and understand what is known of the Honda setup, but with the Hondata 'knock ignition limit' table being relative to the main ignition tables its a little more complicated. My current kal was tuned without knock control, so I realise I can't just enable it without it changing the timing the engine will run.

The way I see it, to run the stock knock control setup I need to adjust any part of the main ignition tables in my current kal where there is a negative value in the 'knock ignition limit' table. To run the same ignition value under knock free conditions I need to add (advance) the main ignition value by this negative amount. So, for example if I have +27* in the main ignition table before I enable knock control, and the value in the 'knock ignition limit' table is -1.4*, when I enable knock control I need to change the main ignition value to +28.4*.

If the knock retard table at this point in the kal is 10*

under knock free conditions I will get:
Ignition advance = minimum(MBT Ignition, knock ignition limit - (knock retard x knock control)) = minimum(28.4, (28.4 - 1.4)-(10 x 0)) = 27*

With enough knock to give a knock control value of 50% I will get:
Ignition advance = minimum(28.4, (28.4 - 1.4)-(10 x 0.5)) = 22*

Am I correct?

It seems that in this calculation the magnitude of the value in the 'knock ignition limit' table is not important, it is only relevant whether it is positive or negative. Presumably the magnitude is important in other calculations we don't know about so I will go by the rule of thumb to "use low values where the engine is likely to knock".

Re: New knock control tables

Posted: Tue Aug 30, 2011 11:32 pm
by Honda9ower
Having thought about this, my last sentence is of course wrong. I was confused by the statement in the help file that "Note that the ECU will only retard timing if the (dynamic) knock ignition limit is less than the (MBT) ignition value". By dynamic knock ignition limit I assume this means knock "ignition limit - (knock retard x knock control)". So, in my example if the 'knock ignition limit' was not negative, but was say 4*, with a knock control value of 50% I would see the ignition retarded by 1*. If the 'knock ignition limit' is greater than 10*, knock control would not alter the ignition timing at all.

Re: New knock control tables

Posted: Wed Aug 31, 2011 3:08 am
by impc
In theory this should work as you described, but in reality the results are little different:

1. Main ignition 43 deg, Ignition limit 8 deg, ignition retard 12 deg, knock control 79%, from formula this is: 8 - (0.79 * 12) = -1.5 deg
Actual results (from datalog):
Knock retard: 0 deg.
Knock ignition limit: 23 deg <- this number from datalog is most confusing, and i can't match it to any table value.
Ignition: 42 deg.

2. Main ignition 35 deg, Ignition limit -1.5 deg, ignition retard 12 deg, knock control 80%, from formula this is -1.5 - (0.8 * 12) = -11 deg
Actual results (from datalog):
Knock retard: 10 deg.
Knock ignition limit: 0 deg.
Ignition: 22 deg.

As you can see that values are still still ~1-3 degrees off. Is there any other component in final ignition values, or there are some rounding errors in ecu during computation i don't know.
I also asked about knock control from another thread, but received no comments from hondata or community.

Re: New knock control tables

Posted: Wed Aug 31, 2011 3:36 am
by Honda9ower
Thanks, thats very useful info. It clearly works, just maybe not exactly as we would expect. I saw your earlier post, and noticed your datalog was at a cam angle of 24*. It may be that the difference is due to interpolation between the 20* and 30* cam tables.

Re: New knock control tables

Posted: Wed Aug 31, 2011 10:12 am
by impc
Honda9ower wrote:Thanks, thats very useful info. It clearly works, just maybe not exactly as we would expect. I saw your earlier post, and noticed your datalog was at a cam angle of 24*. It may be that the difference is due to interpolation between the 20* and 30* cam tables.
unfortunately this is not a case. In 30deg. map it's -1.9 and -0.5 deg. And it's not possible to get 4 deg. interpolating these values (only if you make sign errors during computation :) )

Re: New knock control tables

Posted: Sat Sep 03, 2011 3:02 am
by Honda9ower
I tried out knock control last night, and it was just as unpredictable as you found impc. I had no knock throughout my test, but K.Control was sometimes greater than zero (Hondata did warn that this might happen!).

On a WOT run the values in the kal file were: main ignition value at 23*, K.Limit at -2*, Knock retart of 12*. I had zero knock, and K.Control was recorded as zero throughout the run, I therefore expected K.Retard to be 0*. However, it was recorded as 6* and the actual IGN value logged was 17*. This resulted in a very noticable drop in power.

Looks like I need to study this a bit more if I intend to use it. For now it's switched off and I have a bottle of octane boost in the boot should I notice any knock occur after a new tank of gas.

Re: New knock control tables

Posted: Sun Sep 04, 2011 11:23 pm
by sass
Hondata.
I would like to offer you in the next release of K-manager, to activate
"Use stock knock control logic"
through the inclusion of a separate channel, I can choose from a list same as in the picture.
Image