May have solved 100% E85 on LNF puzzle
03-07-2011 | 01:24 PM
#1
Thread Starter
Member
Joined: 04-09-08
Posts: 377
Likes: 0
From: Seattle, WA
May have solved 100% E85 on LNF puzzle
I am happy to report that we may have figured out (what should be) the last piece of the puzzle of running 100% E85 on the LNF. It's subject to further development and testing, but we seemed to have gotten past a major milestone.
I was actually reviewing a post on another lnf tech forum yesterday, from 2009 around the time when we sorted out the fuel rail pressure stuff. The post seemed to imply that the only way to increase fuel flow on the LNF was to raise the rail pressure because the injection window (amount of time we have to inject fuel on the LNF) was fixed. It also seemed to imply that the problem with injection window size was due to injection carrying over into the compression stroke and that going outside the injection window meant injecting fuel while the exhaust valve was still open causing soot build-up due to an improper combustion.
This implies that the ECM has a hard-set STOP injection point, and calculates the start of injection backwards from this point, meaning that any "time" that needs to be added goes onto the front of the injection pulse.
I don't believe this accurately describes what is happening, and maybe this has been debated here, already - if it has, I apologize for rehashing old discussions.
After some interesting discussions with the guys at ZZP who have more hands-on experience exercising the fuel supply limits of the LNF than I do, it started to sound like the problem wasn't the injection pulse starting too early, but rather it was ending too late - e.g. the injection is still happening when the ECM was trying to ignite the air-fuel charge! There is some math to explain this:
I've heard various numbers - some say injection window misfire sets in at "around 5.5ms" but I got some more detailed information from ZZP - they say that injection window misfire sets in at around 35% injector duty cycle, which is a more accurate way to describe the issue because it takes RPM into account. For example, at 7000 RPM, this would be 6ms.
We can use this figure of 35% IDC to estimate when the LNF starts its injection cycle. There are 720* of crankshaft rotation for all strokes, so 35% of this, or 720 * .35 = 252*. We also have to take into account ignition timing, which happens at, conservatively, 15* BTDC. So we add 15* to our 252*, and that puts the estimated start point of injection at around 267* BTDC.
It would seem this also dispels the theory that carrying over the injection into compression stroke is what causes injection window misfire. There is only 180* of crankshaft rotation from TDC (start of intake stroke) to BDC (start of compression stroke), but we can hit as much as 252* of injection window.
Our injection window size, could be, in theory, accounting for, say 20* of spark advance, 360 - 20 = 340* which would be 47% IDC, and 8ms (33% more fuel mass even at 7000 RPM!). The reality is, though, due to cam phasing events (e.g. exhaust valve still open - we don't want to spray fuel while the exhaust valve is still open), the window would likely be less than 340* maximum. But, then again, maybe we can get there - the LF1 engine (SIDI 3.0L V6) starts its injection almost that early.
With regard to running E85, raising the fuel rail pressure has helped us immensely, but that only shortens the required pulsewidth, it doesn't increase window size, and to run 100% E85 we need as much as 40% more fuel mass, and cranking the rail pressure up all the way doesn't *quite* get us there. I am Broke got somewhere around 67% E85 I believe with rail pressure at highest.
So, to summarize, by combining raised fuel rail pressure and changing the start of injection point on the LNF, we can increase the fuel flow potential, likely enough to run 100% E85 on the LNF. This is the part that we believe we've sorted out now. I'll be working on testing over the next few days, and a few people have offered to help test this out.
If it works out, we could be running 100% E85 in our LNFs within a matter of weeks.
I was actually reviewing a post on another lnf tech forum yesterday, from 2009 around the time when we sorted out the fuel rail pressure stuff. The post seemed to imply that the only way to increase fuel flow on the LNF was to raise the rail pressure because the injection window (amount of time we have to inject fuel on the LNF) was fixed. It also seemed to imply that the problem with injection window size was due to injection carrying over into the compression stroke and that going outside the injection window meant injecting fuel while the exhaust valve was still open causing soot build-up due to an improper combustion.
This implies that the ECM has a hard-set STOP injection point, and calculates the start of injection backwards from this point, meaning that any "time" that needs to be added goes onto the front of the injection pulse.
I don't believe this accurately describes what is happening, and maybe this has been debated here, already - if it has, I apologize for rehashing old discussions.
After some interesting discussions with the guys at ZZP who have more hands-on experience exercising the fuel supply limits of the LNF than I do, it started to sound like the problem wasn't the injection pulse starting too early, but rather it was ending too late - e.g. the injection is still happening when the ECM was trying to ignite the air-fuel charge! There is some math to explain this:
I've heard various numbers - some say injection window misfire sets in at "around 5.5ms" but I got some more detailed information from ZZP - they say that injection window misfire sets in at around 35% injector duty cycle, which is a more accurate way to describe the issue because it takes RPM into account. For example, at 7000 RPM, this would be 6ms.
We can use this figure of 35% IDC to estimate when the LNF starts its injection cycle. There are 720* of crankshaft rotation for all strokes, so 35% of this, or 720 * .35 = 252*. We also have to take into account ignition timing, which happens at, conservatively, 15* BTDC. So we add 15* to our 252*, and that puts the estimated start point of injection at around 267* BTDC.
It would seem this also dispels the theory that carrying over the injection into compression stroke is what causes injection window misfire. There is only 180* of crankshaft rotation from TDC (start of intake stroke) to BDC (start of compression stroke), but we can hit as much as 252* of injection window.
Our injection window size, could be, in theory, accounting for, say 20* of spark advance, 360 - 20 = 340* which would be 47% IDC, and 8ms (33% more fuel mass even at 7000 RPM!). The reality is, though, due to cam phasing events (e.g. exhaust valve still open - we don't want to spray fuel while the exhaust valve is still open), the window would likely be less than 340* maximum. But, then again, maybe we can get there - the LF1 engine (SIDI 3.0L V6) starts its injection almost that early.
With regard to running E85, raising the fuel rail pressure has helped us immensely, but that only shortens the required pulsewidth, it doesn't increase window size, and to run 100% E85 we need as much as 40% more fuel mass, and cranking the rail pressure up all the way doesn't *quite* get us there. I am Broke got somewhere around 67% E85 I believe with rail pressure at highest.
So, to summarize, by combining raised fuel rail pressure and changing the start of injection point on the LNF, we can increase the fuel flow potential, likely enough to run 100% E85 on the LNF. This is the part that we believe we've sorted out now. I'll be working on testing over the next few days, and a few people have offered to help test this out.
If it works out, we could be running 100% E85 in our LNFs within a matter of weeks.
03-07-2011 | 01:30 PM
#3
I recently got to ~70% E @ 2800 psi but it was iffy. Sometimes it was smooth, other times it wasn't depending on how hard the midrange hit. I was seeing ~45% IDC when it started popping popcorn. Then the Low Fuel Rail pressure CEL popped. Hopefully you won't have the low pressure issue.
Thank you for the continuing effort.
Thank you for the continuing effort.
Last edited by Iam Broke; 03-07-2011 at 01:34 PM. Reason: sp
03-07-2011 | 01:48 PM
#7
Thread Starter
Member
Joined: 04-09-08
Posts: 377
Likes: 0
From: Seattle, WA
I recently got to ~70% E @ 2800 psi but it was iffy. Sometimes it was smooth, other times it wasn't depending on how hard the midrange hit. I was seeing ~45% IDC when it started popping popcorn. Then the Low Fuel Rail pressure CEL popped. Hopefully you won't have the low pressure issue.
Thank you for the continuing effort.
Thank you for the continuing effort.
03-07-2011 | 01:51 PM
#8
Senior Member
Joined: 09-23-10
Posts: 770
Likes: 0
From: VA
03-07-2011 | 01:55 PM
#9
Senior Member
Joined: 07-23-10
Posts: 3,451
Likes: 0
From: Carlsbad, CA
So not enough fuel is getting pushed to the high pressure pump? or the pump cannot sustain the pressure? Heck if it isn't getting enough from the stock fuel lift pump I'm sure a walbro diesel lift pump could flow more than enough haha.
03-07-2011 | 01:59 PM
#10
Thread Starter
Member
Joined: 04-09-08
Posts: 377
Likes: 0
From: Seattle, WA
We ran into this quite extensively on the 2010 Camaro SIDI V6 with turbo kits on them, but on those cars there's an FPCM that controls the fuel pump speed so we were able to raise the pressure on the low pressure pump as well to help.
But on our cars it is a fixed-rate fuel pump so what we get is what we get. If, after raising fuel rail pressure there are persistent P0087 codes, then the low pressure pump needs to be upgraded.
03-07-2011 | 02:09 PM
#12
Senior Member
Joined: 07-23-10
Posts: 3,451
Likes: 0
From: Carlsbad, CA
In some cases, the stock low pressure pump isn't sufficient to sustain fuel supply to the high pressure pump on the engine. That's why some people with raised fuel rail pressures can get P0087 codes and limp mode.
We ran into this quite extensively on the 2010 Camaro SIDI V6 with turbo kits on them, but on those cars there's an FPCM that controls the fuel pump speed so we were able to raise the pressure on the low pressure pump as well to help.
But on our cars it is a fixed-rate fuel pump so what we get is what we get. If, after raising fuel rail pressure there are persistent P0087 codes, then the low pressure pump needs to be upgraded.
We ran into this quite extensively on the 2010 Camaro SIDI V6 with turbo kits on them, but on those cars there's an FPCM that controls the fuel pump speed so we were able to raise the pressure on the low pressure pump as well to help.
But on our cars it is a fixed-rate fuel pump so what we get is what we get. If, after raising fuel rail pressure there are persistent P0087 codes, then the low pressure pump needs to be upgraded.
lol
03-07-2011 | 03:05 PM
#23
Thread Starter
Member
Joined: 04-09-08
Posts: 377
Likes: 0
From: Seattle, WA
