Easy speed density? First test...
Mar 1, 2009, 01:50 PM
#1
Evolved Member
Thread Starter
Easy speed density? First test...
Whilst I am delighted with the way my car drives with our existing speed density patches, which I would say are complete, I think some are or will have difficulty with the RPM and MAP VE tables. Thinking about the strokers and big turbos where we are intending to rewrite the load patterns to be 1:1 with MAP once out of vacuum, I wondered how much it could be simplified.
So I flattened my RPM and MAP VE tables (rather than remove all their code) and so ran the load 1:1 with MAP (at 25C, temp comp is still active) everywhere - even in vacuum, and then edited the MAF smoothing table to sort the fuelling.
MAF Hz: VE%
25Hz: 80%
400Hz: 100%
1600Hz: 100%
1800Hz: 94%
I used only the above four bins, and it works great. Trims are tight, driving manners are good, WOT behaves as usual.
When I say I only used the four bins, I didn't edit the MAF Hz values in the smoothing table as they would also affect the scaling table which has a distinct shape that we would mess up. So I multiplied the existing MAF smoothing values at these points by the VE% and then linearly scaled between. We could combine the smoothing and scaling tables into one VE table that would start at about 60%, go up to 100%, and drop to 93%.
I saw no weird effects on knock control. Before the engine was full warm the medium rather than low LTFT adjusted, but at warm idle the low LTFT adjusted. I increased the closed loop load thresholds slightly. I took the 40-140 load axes in the fuel, timing, MIVEC maps and multiplied then up by 20% at 40, smoothing down to 0% at 140. At 200+kPa (or load - they are both the same now) I added a degree of timing at 6000-6500, and 2 degrees at 7000-7500 to get back to original timing since the load doesn't drop off unless boost drops off now.
Sounds a bit of work, not really, and for some people who are "starting again" with their maps for a stroker, big turbo etc, it might be quicker to dial it in.
You don't need to reduce it to four bins, but it can be made to work like this.
Now I don't want anyone saying VE is hard to tune after that
Discuss.
So I flattened my RPM and MAP VE tables (rather than remove all their code) and so ran the load 1:1 with MAP (at 25C, temp comp is still active) everywhere - even in vacuum, and then edited the MAF smoothing table to sort the fuelling.
MAF Hz: VE%
25Hz: 80%
400Hz: 100%
1600Hz: 100%
1800Hz: 94%
I used only the above four bins, and it works great. Trims are tight, driving manners are good, WOT behaves as usual.
When I say I only used the four bins, I didn't edit the MAF Hz values in the smoothing table as they would also affect the scaling table which has a distinct shape that we would mess up. So I multiplied the existing MAF smoothing values at these points by the VE% and then linearly scaled between. We could combine the smoothing and scaling tables into one VE table that would start at about 60%, go up to 100%, and drop to 93%.
I saw no weird effects on knock control. Before the engine was full warm the medium rather than low LTFT adjusted, but at warm idle the low LTFT adjusted. I increased the closed loop load thresholds slightly. I took the 40-140 load axes in the fuel, timing, MIVEC maps and multiplied then up by 20% at 40, smoothing down to 0% at 140. At 200+kPa (or load - they are both the same now) I added a degree of timing at 6000-6500, and 2 degrees at 7000-7500 to get back to original timing since the load doesn't drop off unless boost drops off now.
Sounds a bit of work, not really, and for some people who are "starting again" with their maps for a stroker, big turbo etc, it might be quicker to dial it in.
You don't need to reduce it to four bins, but it can be made to work like this.
Now I don't want anyone saying VE is hard to tune after that
Discuss.
Last edited by jcsbanks; Mar 1, 2009 at 02:31 PM.
Mar 1, 2009, 02:06 PM
#2
Evolved Member
iTrader: (153)
Join Date: Mar 2007
Location: Still in NC!! Loving retired life!!
Posts: 1,077
Likes: 0
Received 1 Like
on
1 Post
WOW!!! You ECU dissectors are making some amazing advancements. I'm still having a hard time grasping the whole MAF scaling stuff and finally finished reading the "other" Speed Density" thread. Way over my head.
Josh
Josh
Mar 1, 2009, 02:28 PM
#4
Evolved Member
Thread Starter
This is why I started a new thread, I did not want to confuse, more to discuss with those interested. This is not a change of direction I don't think, just another option, there are loads of ways of doing this conversion, and different methods may be useful on different engines.
The method in this thread is a deliberate simplification to the point of absurdity, yet it actually works well, and the end result is a very neat mapping arrangement where 0 PSI gives you 100 load, 14.5 PSI gives you 200 load, 29 PSI gives you 300 load. More like a standalone and it tidies up some of the odd design decisions Mitsubishi have made in their calibration of load (which as I've said before isn't always strictly load in the sense of it being always proportional to air mass per cylinder).
The method in this thread is a deliberate simplification to the point of absurdity, yet it actually works well, and the end result is a very neat mapping arrangement where 0 PSI gives you 100 load, 14.5 PSI gives you 200 load, 29 PSI gives you 300 load. More like a standalone and it tidies up some of the odd design decisions Mitsubishi have made in their calibration of load (which as I've said before isn't always strictly load in the sense of it being always proportional to air mass per cylinder).
Trending Topics
Mar 2, 2009, 04:20 AM
#9
Sweet
So, in summery
Flat line the RPM VE table to 100% (adjust 500 - 1500rpm to suit, and 7000 - 8000 rpm)
Change 1600hz to 1800hz in the smoothing table only ?
Then change the 4 values in the smoothing table as per above ?
So, in summery
Flat line the RPM VE table to 100% (adjust 500 - 1500rpm to suit, and 7000 - 8000 rpm)
Change 1600hz to 1800hz in the smoothing table only ?
Then change the 4 values in the smoothing table as per above ?
Mar 2, 2009, 04:40 AM
#10
Evolved Member
Thread Starter
Also make the MAP VE table 1:1.
If you change the 1600Hz to 1800Hz it affects both the scaling and smoothing tables.
As well as changing the 4 values you need to smooth the values in between - for me this was 1.5% per 25Hz - note the non-equal spacing of the Hz values mean you can't just use Ecuflash's interpolation function.
If you change the 1600Hz to 1800Hz it affects both the scaling and smoothing tables.
As well as changing the 4 values you need to smooth the values in between - for me this was 1.5% per 25Hz - note the non-equal spacing of the Hz values mean you can't just use Ecuflash's interpolation function.
Mar 2, 2009, 04:45 AM
#11
Also make the MAP VE table 1:1.
If you change the 1600Hz to 1800Hz it affects both the scaling and smoothing tables.
As well as changing the 4 values you need to smooth the values in between - for me this was 1.5% per 25Hz - note the non-equal spacing of the Hz values mean you can't just use Ecuflash's interpolation function.
If you change the 1600Hz to 1800Hz it affects both the scaling and smoothing tables.
As well as changing the 4 values you need to smooth the values in between - for me this was 1.5% per 25Hz - note the non-equal spacing of the Hz values mean you can't just use Ecuflash's interpolation function.
Also with the MAF smoothing table, there seems to be some different settings in the XML's for different roms.
Mine is setup (luckily) already to Read in MAF HZ so goes from 19 - 1600, and the units are from 128 - 133 (unit8) (however I am using a E9 maf sensor, and I noticed your rom the maf sensor units are slightly different, so maybe I should change mine to the E9 settings whilst it's running on MAF).
Mar 2, 2009, 07:04 AM
#13
So Im looking at my maf smoothing table and it is scaled in Hz. and ranges from 192-1600. As for the scaling on the right it ranges from 144 to 160 back down to 150's.
Im trying to read this and make sure I got it right. So I multiply the values that I can edit that range from 144-160-150 by the 4 values above in the 25hz, 400hz,1600hz,1800hz ?
I think Im lost. I notice that I dont have a 25Hz on my maf smoothing tables so maybe its not rescaled. Ill have to search that one up.
Im trying to read this and make sure I got it right. So I multiply the values that I can edit that range from 144-160-150 by the 4 values above in the 25hz, 400hz,1600hz,1800hz ?
I think Im lost. I notice that I dont have a 25Hz on my maf smoothing tables so maybe its not rescaled. Ill have to search that one up.
Mar 2, 2009, 08:15 AM
#14
Whilst I am delighted with the way my car drives with our existing speed density patches, which I would say are complete, I think some are or will have difficulty with the RPM and MAP VE tables. Thinking about the strokers and big turbos where we are intending to rewrite the load patterns to be 1:1 with MAP once out of vacuum, I wondered how much it could be simplified.
So I flattened my RPM and MAP VE tables (rather than remove all their code) and so ran the load 1:1 with MAP (at 25C, temp comp is still active) everywhere - even in vacuum, and then edited the MAF smoothing table to sort the fuelling.
MAF Hz: VE%
25Hz: 80%
400Hz: 100%
1600Hz: 100%
1800Hz: 94%
I used only the above four bins, and it works great. Trims are tight, driving manners are good, WOT behaves as usual.
When I say I only used the four bins, I didn't edit the MAF Hz values in the smoothing table as they would also affect the scaling table which has a distinct shape that we would mess up. So I multiplied the existing MAF smoothing values at these points by the VE% and then linearly scaled between. We could combine the smoothing and scaling tables into one VE table that would start at about 60%, go up to 100%, and drop to 93%.
I saw no weird effects on knock control. Before the engine was full warm the medium rather than low LTFT adjusted, but at warm idle the low LTFT adjusted. I increased the closed loop load thresholds slightly. I took the 40-140 load axes in the fuel, timing, MIVEC maps and multiplied then up by 20% at 40, smoothing down to 0% at 140. At 200+kPa (or load - they are both the same now) I added a degree of timing at 6000-6500, and 2 degrees at 7000-7500 to get back to original timing since the load doesn't drop off unless boost drops off now.
Sounds a bit of work, not really, and for some people who are "starting again" with their maps for a stroker, big turbo etc, it might be quicker to dial it in.
You don't need to reduce it to four bins, but it can be made to work like this.
Now I don't want anyone saying VE is hard to tune after that
Discuss.
So I flattened my RPM and MAP VE tables (rather than remove all their code) and so ran the load 1:1 with MAP (at 25C, temp comp is still active) everywhere - even in vacuum, and then edited the MAF smoothing table to sort the fuelling.
MAF Hz: VE%
25Hz: 80%
400Hz: 100%
1600Hz: 100%
1800Hz: 94%
I used only the above four bins, and it works great. Trims are tight, driving manners are good, WOT behaves as usual.
When I say I only used the four bins, I didn't edit the MAF Hz values in the smoothing table as they would also affect the scaling table which has a distinct shape that we would mess up. So I multiplied the existing MAF smoothing values at these points by the VE% and then linearly scaled between. We could combine the smoothing and scaling tables into one VE table that would start at about 60%, go up to 100%, and drop to 93%.
I saw no weird effects on knock control. Before the engine was full warm the medium rather than low LTFT adjusted, but at warm idle the low LTFT adjusted. I increased the closed loop load thresholds slightly. I took the 40-140 load axes in the fuel, timing, MIVEC maps and multiplied then up by 20% at 40, smoothing down to 0% at 140. At 200+kPa (or load - they are both the same now) I added a degree of timing at 6000-6500, and 2 degrees at 7000-7500 to get back to original timing since the load doesn't drop off unless boost drops off now.
Sounds a bit of work, not really, and for some people who are "starting again" with their maps for a stroker, big turbo etc, it might be quicker to dial it in.
You don't need to reduce it to four bins, but it can be made to work like this.
Now I don't want anyone saying VE is hard to tune after that
Discuss.
John, can you explain how you came up with the percentages`applied to the MAF Hz table?
Mar 2, 2009, 08:53 AM
#15
Evolved Member
Thread Starter
Emperor, I think you are looking at the MAF scaling rather than smoothing table.
Cossie1, I multiplied the cell contents by my new percentage. e.g. 131*80%=105. I scale then as percent128 so 128 shows as 100%.
wreckleford, I took them from the VE I get at those airflows. You can either do this from logs, or use the calc I posted in the other thread to work out airflow from load and RPM and work out your VE from there.
Cossie1, I multiplied the cell contents by my new percentage. e.g. 131*80%=105. I scale then as percent128 so 128 shows as 100%.
wreckleford, I took them from the VE I get at those airflows. You can either do this from logs, or use the calc I posted in the other thread to work out airflow from load and RPM and work out your VE from there.
