This figure shows the effect of the muffler valve as installed in all XB12:
My mate and I decided to give it a try with our XB9 also, to find out if it might work similar there too.
Pash kindly pointed me to a free engine simulation software, which allows to compute one cylinder only, but this would be fine for just a proof of concept. I modeled one cylinder and an exhaust with the valve opened and closed and started the simulation runs. The figure below shows one of the results:
Attachment:
File comment: VE
ve.jpg [ 19.92 KiB | Viewed 2097 times ]
ve.jpg [ 19.92 KiB | Viewed 2097 times ]
Apparently the valve increases VE (a indicator of the air mass available for combustion) which in sequence would allow to add fuel and increase torque. With the XB12 muffler and the valve motor installed we made a few rides and ended with the switching points shown below, which are not too far away from the simulated results:
Attachment:
File comment: switching points
amvsetup_new.jpg [ 45.67 KiB | Viewed 2097 times ]
amvsetup_new.jpg [ 45.67 KiB | Viewed 2097 times ]
My mate has a Sinus module installed (an early predecessor of the Remus Powerizer, but mapped for the stock silencer), which will provide enough fuel, but I don't have and decided to adjust the fuel maps to cope with the new situation. All enrichments were switched off (WOT, OL, front cylinder temperature correction ...), but I left EGO correction and AFV active as I wanted to watch how these were changing. Next I logged some rides, to adjust the closed loop area first, because I wanted get a stable AFV before adjusting the open loop regions.
Attachment:
File comment: ego corr 1
corr1.jpg [ 68.77 KiB | Viewed 2097 times ]
corr1.jpg [ 68.77 KiB | Viewed 2097 times ]
That seemed to be a pretty large variation, and it took serveral rides to adjust the map until I ended with this:
Attachment:
File comment: ego corr 7
corr7.jpg [ 70.71 KiB | Viewed 2097 times ]
corr7.jpg [ 70.71 KiB | Viewed 2097 times ]
That's still not perfect, but more uniform than before and it didn't seem worth any more efforts. Additionally a wideband O2 sensor was installed in each exhaust header, and I tried to adjust the front map to make the front cylinder run slightly rich and the rear cylinder (which is monitord by the o2 sensor) about stoich:
Attachment:
File comment: lambda 7
labmda7.jpg [ 328.72 KiB | Viewed 2097 times ]
labmda7.jpg [ 328.72 KiB | Viewed 2097 times ]
There are apparently some lean spots left, but these numbers are averaged and weighted, so all values must be taken with a grain of salt. With WOT enrichment and open loop correction enabled again, these lean spots will disappear then.