Another 205 8v thread
- Thread starter flock
- Start date
Happy to report those changes made the tuning-by-feedback process much better. I went back to Peter's base map shape and adjusted for idle. Did a run up the street and back, turned off the car and quicktuned the areas where load and rpm were constant then went about linearising and smoothing out the table up to about 4000rpm. After a few runs it was getting much closer to target AFR and , with error on the rich side rather than lean.
Keeping in mind where I live has elevations between 0-60m with few flat stretches so I was trying to be conservative and efficient when making changes as possible. But at least there is varying load for which to map to. A trap I fell into at first was seeing long spikes where the wideband was going out of range. But going through the logs this only happen when I lift off the throttle going up and down all the hills. I haven't deactivated decel cut, I just avoid quicktuning when I lift off the throttle.
You can already feel the difference in that ignition map vs the old distributor, just going through the revs up to 4000rpm. After regraphing the distributor I could tell it lost some oomph going up hills, which was the sacrifice to stop it pinging. It has definitely gained some low-mid torque and feels a lot more eager now which is where the magic of mappable ignition is starting to shine on this engine. Or maybe it's all in my head because I haven't driven the car for a while...
I am confident now that after a few more runs with some main road cruising I can get this thing very driveable until I book into a dyno. Sometimes on decel, I swear I could hear/feel a backfire every now and then. Possibly the sudden transition from 30deg advance to 12degrees at zero throttle?
Keeping in mind where I live has elevations between 0-60m with few flat stretches so I was trying to be conservative and efficient when making changes as possible. But at least there is varying load for which to map to. A trap I fell into at first was seeing long spikes where the wideband was going out of range. But going through the logs this only happen when I lift off the throttle going up and down all the hills. I haven't deactivated decel cut, I just avoid quicktuning when I lift off the throttle.
You can already feel the difference in that ignition map vs the old distributor, just going through the revs up to 4000rpm. After regraphing the distributor I could tell it lost some oomph going up hills, which was the sacrifice to stop it pinging. It has definitely gained some low-mid torque and feels a lot more eager now which is where the magic of mappable ignition is starting to shine on this engine. Or maybe it's all in my head because I haven't driven the car for a while...
I am confident now that after a few more runs with some main road cruising I can get this thing very driveable until I book into a dyno. Sometimes on decel, I swear I could hear/feel a backfire every now and then. Possibly the sudden transition from 30deg advance to 12degrees at zero throttle?
Idle VE only went from 65 to 60. Compared to the mi16 map, between 1000-2000 its about 15points higher, and about 10 points higher 2000-4000. Overall, compared to the spot-tuning I did in the driveway it's about 5-10% less. Still curious why they are higher than the mi16, assuming the initial settings are accurate on both maps.
The difference it made was more how accurately the quicktune hit the mark, so to speak. That's quite an impressive feature for the age of the ECU. It can correct itself without feedback with the software in offline mode using the datalogs. If injector values are correct its pretty bloody good at getting the target AFR in one hit. Just have to keep a mental note of where you hit Q and made changes in case you go back over the same cell and hit Q again.
I'm sure the Elite series save a lot of time with autotune. But it is more satisfying sitting on the side of the road, interpreting the logs and fiddling with the values, then going for another run and looking over at the laptop seeing it hit the mark.
It's still early days and I've only changed up to 4000rpm. To be fair, when I found it ran decent at 2500rpm I more or less copied that entire row upwards and fluffed about making it resemble the 'shape' of your map with the big step changes at high vacuum. I haven't been able to hold it at enough load cells in those transient zones to define its own shape yet. But it feels right and that's what matters for this exercise.
The difference it made was more how accurately the quicktune hit the mark, so to speak. That's quite an impressive feature for the age of the ECU. It can correct itself without feedback with the software in offline mode using the datalogs. If injector values are correct its pretty bloody good at getting the target AFR in one hit. Just have to keep a mental note of where you hit Q and made changes in case you go back over the same cell and hit Q again.
I'm sure the Elite series save a lot of time with autotune. But it is more satisfying sitting on the side of the road, interpreting the logs and fiddling with the values, then going for another run and looking over at the laptop seeing it hit the mark.
It's still early days and I've only changed up to 4000rpm. To be fair, when I found it ran decent at 2500rpm I more or less copied that entire row upwards and fluffed about making it resemble the 'shape' of your map with the big step changes at high vacuum. I haven't been able to hold it at enough load cells in those transient zones to define its own shape yet. But it feels right and that's what matters for this exercise.
Last edited:
Dead time is definitely not accurate, as I only have a value at 13.8V. The Haltech measures 14-14.1V when running.
Tell me if I interpreted the values wrong:
From that I gathered;
Flow at 100% duty = 236cc/min
Dead time = 0.459ms @ 13.8V.
I estimated dead time as 0.430ms @ 14V and 0.400 @ 14.3V
If you think it’s worth going back and using their new machine to graph the dead time properly I’ll do it. Your injector’s dead time were double mine.
Tell me if I interpreted the values wrong:
From that I gathered;
Flow at 100% duty = 236cc/min
Dead time = 0.459ms @ 13.8V.
I estimated dead time as 0.430ms @ 14V and 0.400 @ 14.3V
If you think it’s worth going back and using their new machine to graph the dead time properly I’ll do it. Your injector’s dead time were double mine.
If the deadtime table is 3D (fuel pressure vs volts) then revert to PeterT deadtime values and adjust injector size to get target lambda at the VE he suggests.
Deadtime also compensates for closing time dribble, you can empirically work out a percentage correction to the PeterT deadtime table as per this Emtron Tuning Tip:
Injector Dead times can be validated using the Emtune Software by using the Wideband Lambda control. If you add 10% to your VE table (@ 3000rpm / 80kpa load for example), you should see a corresponding negative 10% trim applied via the Wideband lambda control. If you don't, then you know your dead times need some attention.
By utilizing a dead time table available in your Emtune software that is close. The correct dead time for your injector can be quickly arrived at by simply globally moving the table up & down. The voltage slop of the dead times can be further validated by removing the alternator charge and allowing the supply voltage to drop away. Correct dead times allow the engine to operate correctly over a wide range of variable conditions
Last edited:
Found another tutorial link more applicable to your ECU posted by some dude called Pitcrew
https://www.205gtidrivers.com/forums/topic/176044-injector-dead-time/?do=findComment&comment=1544034
Thank you Alex and Peter.
I made a large adjustment to the flow rate which brought VEs into check with Peter’s base table but quick tune was still off the mark. So I took my injectors back for testing dead times and flow at 2.5bar on the injectors shop’s new machine. Shit in = shit out I figure.
There was a huge difference in flow rate.
186cc/min vs my calculated 215cc/min. And the dead time curve was way off. Hopefully I don’t stray into high duty cycles, else I might need to upgrade to yellow tops or 3bar regulator.
I just gave it another warmup and VEs are more like 48 at idle. Back to square one with on-road tuning but at least once the VE table is sorted, any changes should be tolerated with more accurate injector data.
More to come!
I made a large adjustment to the flow rate which brought VEs into check with Peter’s base table but quick tune was still off the mark. So I took my injectors back for testing dead times and flow at 2.5bar on the injectors shop’s new machine. Shit in = shit out I figure.
There was a huge difference in flow rate.
186cc/min vs my calculated 215cc/min. And the dead time curve was way off. Hopefully I don’t stray into high duty cycles, else I might need to upgrade to yellow tops or 3bar regulator.
I just gave it another warmup and VEs are more like 48 at idle. Back to square one with on-road tuning but at least once the VE table is sorted, any changes should be tolerated with more accurate injector data.
More to come!
What part number are they?
0280150734 standard DFZ ones. I won't be posting the data, but if anyone wants it I will be happy to share privately.
I just gave them a call to clarify. Their Bosch datasheet says they are 215cc/min @ 3bar with N-heptane. Converted to 2.5bar puts them down to ~200cc/min.
The fluid they use is closer in viscosity to E85 which is thicker than 98. The conversation got a bit nerdy, but the tech said I should be adding about 7% onto the flow rate but the dead time's won't change. Since his customer base is mostly boosted race engines with 1000cc injectors he uses that fluid to be on the conservative side.
If you can, upgrading now to 3 bar Mi16 fuel rail is well worth doing to get the best from engine mods.
Because the L-Jetronic is 'tuned' by (mostly) changing resistors, I believe the reason for the drop to 2.5 bar was to get the injector pulse width to match an existing 8v ECU.
I had considered that before dropping them into the shop to get them tested at 2.5bar. But hey, I'm happy to see how far I can push it using what I have, for the benefit of others. The next (possibly never) logical upgrade would be ITBs which would need a fuel system rework anyway.
What's the point of an Mi16 rail. Can't I just swap the regulator over from an XU10J2 or XU9JAZ which run mi16 injectors at 3bar?
Took it for a spin today to fill up the tyres. Apart from idle, I didn't touch Peter's Mi16 base fuel table. Not too dissimilar VE's, but hard to comment without a tune. The main thing is, it is very driveable! Already better than Jetronic.
Volume. I suspect it's bigger, hopefully @PeterT can confirm. The rail acts like a plenum, as pulsewidths increase (especially if 2 injectors fire at same time) if the rail volume is too small you can end up with standing waves that I've seen starve an injector of fuel.
I don't think the Mi16 fuel rail is any bigger than the GTi rail. If fact it differs significantly in mounting style and integration of the regulator. Additionally, the Mi16 has a damper in the system. I'd just keep using it at 2.5 Bar, if that's what you have the data for. The difference in volume between 2.5 Bar and 3 Bar is 9%, which is about what you've confirmed.
I said that, as being past tense, in that the map I gave Lachlan was from an Mi16 using the 734 injectors.