SSpdDmon
August 23rd, 2007, 12:42 PM
So, I think I've figured some stuff out. I've always asked people for hints when it comes to tuning for injectors. But, never got any solid answers. Well....here's what I did to help my situation out.
First, the symptoms:
The problem I was having with my setup was no matter what, it seemed I was having a rich decel issue. AFR's always seemed to be in the mid to low 13's. I would drop the MAF/VE or increase the IFR and still....rich decel.
Some background on the tune I'm working with:
My IFR table was reverse-engineered using a stock MAF curve. By setting my B0120 to 1rpm, I found the best line that worked so that my desired AFR matched my actual AFR on the WB. If you decide to use a spreadsheet to calculate your IFR table like most do, the following should still help.
B3701 (IPW Voltage Adjustment) - I have read in several locations the fact that the SVO30's have about a 30% greater offset than our stockers. So, I increased the stock table 30%. I then flattened the table out so that the values for the 80kPa row and the 0kPa row were equal by using a value somewhere in the middle.
B4001 (IFR) - See above and picture below.
B4003 / B4004 (Min. IPW) - I wanted to make sure the PCM would allow a small injector pulse width. So, I dropped both of these two tables to 1.0 across the board.
B4005 (Small Pulse Adjust) - I zero'd this table out for starters. This table and the B3701 table are added on to the desired pulse width. One too many roosters in the hen house for me. This table seems to be useful if you need to make pulse width adjustments based on injector on-time. I haven't got that far to notice a need. If you leave DFCO active later on, you probably shouldn't zero out the samller cells. The injectors don't cut all the way off during DFCO. I believe their pulse width during DFCO is determined by this table. Still have to look into this further.
Once I made the above changes, I logged my GM.IBPW1, GM.MANVAC, GM.VOLTS and BEN pids. The GM.IBPW1 pid shows the final commanded injector on time. During decel, it would bottom out in the low 1.80~1.82ms range. AFR's were still in the 13's. Based on the GM.MANVAC and GM.VOLTS pids, I could see where that put me in B3701. So, now I have a table to change, I know how long the final pulse width is, and I know what my AFR is. Since my AFR was low 13's and commanded was stoich, I knew my error was about 10%. With an IPW of 1.8, I found a 10% decrease would mean a new IPW of 1.62 for that MANVAC vs. VOLT cell (1.8*0.9=1.62). In my case, that's the 13.5V cell and the 75kPa row. The value in my B3701 table was .577ms after the changes from above. I needed to reduce it 0.18ms (1.8-1.62=0.18ms). That essentially is a 31% reduction to the value already in the cell (0.397/0.577=0.69). So, I highlighted the entire row and reduced it by 31% effectively shortening the pulse width. :) I had to do the same procedure to the 80kPa row as well.
The picture below is what is currently in my tune. Decel AFRs are no longer rich. :D It'll take a little more fine tweaking. But, it's close for now.
First, the symptoms:
The problem I was having with my setup was no matter what, it seemed I was having a rich decel issue. AFR's always seemed to be in the mid to low 13's. I would drop the MAF/VE or increase the IFR and still....rich decel.
Some background on the tune I'm working with:
My IFR table was reverse-engineered using a stock MAF curve. By setting my B0120 to 1rpm, I found the best line that worked so that my desired AFR matched my actual AFR on the WB. If you decide to use a spreadsheet to calculate your IFR table like most do, the following should still help.
B3701 (IPW Voltage Adjustment) - I have read in several locations the fact that the SVO30's have about a 30% greater offset than our stockers. So, I increased the stock table 30%. I then flattened the table out so that the values for the 80kPa row and the 0kPa row were equal by using a value somewhere in the middle.
B4001 (IFR) - See above and picture below.
B4003 / B4004 (Min. IPW) - I wanted to make sure the PCM would allow a small injector pulse width. So, I dropped both of these two tables to 1.0 across the board.
B4005 (Small Pulse Adjust) - I zero'd this table out for starters. This table and the B3701 table are added on to the desired pulse width. One too many roosters in the hen house for me. This table seems to be useful if you need to make pulse width adjustments based on injector on-time. I haven't got that far to notice a need. If you leave DFCO active later on, you probably shouldn't zero out the samller cells. The injectors don't cut all the way off during DFCO. I believe their pulse width during DFCO is determined by this table. Still have to look into this further.
Once I made the above changes, I logged my GM.IBPW1, GM.MANVAC, GM.VOLTS and BEN pids. The GM.IBPW1 pid shows the final commanded injector on time. During decel, it would bottom out in the low 1.80~1.82ms range. AFR's were still in the 13's. Based on the GM.MANVAC and GM.VOLTS pids, I could see where that put me in B3701. So, now I have a table to change, I know how long the final pulse width is, and I know what my AFR is. Since my AFR was low 13's and commanded was stoich, I knew my error was about 10%. With an IPW of 1.8, I found a 10% decrease would mean a new IPW of 1.62 for that MANVAC vs. VOLT cell (1.8*0.9=1.62). In my case, that's the 13.5V cell and the 75kPa row. The value in my B3701 table was .577ms after the changes from above. I needed to reduce it 0.18ms (1.8-1.62=0.18ms). That essentially is a 31% reduction to the value already in the cell (0.397/0.577=0.69). So, I highlighted the entire row and reduced it by 31% effectively shortening the pulse width. :) I had to do the same procedure to the 80kPa row as well.
The picture below is what is currently in my tune. Decel AFRs are no longer rich. :D It'll take a little more fine tweaking. But, it's close for now.