Microsquirt staged injection problem

[TimoL]

I have encountered a following problem when trying to enable staged injection with microsquirt.
If I enable the staged injection the AFR goes really rich even when at 0% staging.
I'm running sequential fuel & spark using MAP phase detection.

I've attached tunes and logs for both cases (staged off and on). I forgot to wait for the 02 sensor to warm up when data logging the staged injection tune but I know the AFR would have been at the bottom of the scale the whole time (10.30).

The logs aren't very easily comparable because with the staged the engine ran so rich that it would not idle. But it looks like the PW1 doesn't differ that much between the tunes so why is it running so rich with staged injection? I don't have the secondary injector mounted yet.

HW: Microsquirt v3
FW: Extra 3.4.2.
Engine: 1 cyl, 4 stroke

Any help would be greatly appreciated.

-Timo

[TimoL]

Update:

I just switched to 4 squirts per engine cycle and to an untimed injection for superior throttle response.
And the problem went away, AFRs stay completely the same trough the VE table. I guess I will never know what the problem was but it doesn't matter anymore really.

By the way:
Do I really have to set the number of cylinders to 4 if I want to run 4 squirts per engine cycle?
This seems to mess up the MAP readings what ever settings I try for MAP sampling.
Not that it's a problem right now since I'm running pure Alpha-N because it seems to be the only way to get non-horrible throttle response.

[dontz125]

You need to tell the controller how many cyls you actually have. Lying to the uS is a workaround for certain problems in certain instances, but you really need to know what you're doing, and it can be fraught with unintended consequences.

Telling the controller your single is a 4-cyl will confuse the heck out of the MAP signal, because it will be reading the MAP signal every 180 degrees, which will give 3 garbage signals and one reasonably accurate reading.

You have 'Multiply MAP' turned off. Any particular reason why? Even with NA and Alpha-N, the MAP signal will often improve the tunability of the VE map.

Part of your issue may be the Fixed Injection Timing angle; you have it set to 270 when running, but 90 when cranking or staged. I note that the default timing map is set to all-90 as a starting point. That 270 may be doing strange things.

[Blown88GT]

There is an issue with staged injection using sequenced batch fire.
Sequenced Batch Fire - Feature is not working
http://www.msextra.com/forums/viewtopic ... 90#p477890
Injector driver?
http://www.msextra.com/forums/viewtopic ... 01&t=65775

The next firmware release will include Sequenced batch-fire tooltip.
http://www.msextra.com/forums/viewtopic ... 91&t=66021

This issue also occurs with MS3
http://www.msextra.com/forums/viewtopic ... 31&t=66098

[dontz125]

? Staged sequenced batch =/= staged sequential, unless I'm missing something ...

[TimoL]

You need to tell the controller how many cyls you actually have. Lying to the uS is a workaround for certain problems in certain instances, but you really need to know what you're doing, and it can be fraught with unintended consequences.

Telling the controller your single is a 4-cyl will confuse the heck out of the MAP signal, because it will be reading the MAP signal every 180 degrees, which will give 3 garbage signals and one reasonably accurate reading.

You have 'Multiply MAP' turned off. Any particular reason why? Even with NA and Alpha-N, the MAP signal will often improve the tunability of the VE map.

Part of your issue may be the Fixed Injection Timing angle; you have it set to 270 when running, but 90 when cranking or staged. I note that the default timing map is set to all-90 as a starting point. That 270 may be doing strange things.

Yes, I understand that settings the num of cyls to an incorrect value is not a good practice but the only issue to come up is the messed MAP signal, which is not used at the moment. The 4 squirts per cycle gives a lot better throttle response than sequential and good transient fueling is a must for my application.

I found that 270 deg angle to give the best throttle response when running sequential. I didn't bother with the cranking timing because it did not seem to affect anything. I don't even know if the RPM is fully synced until at idle because of the MAP phase detection?
And you're right, looks I completely missed that staging injection 1 timing. I will fix and see if it makes any difference.

And as for the Multiply MAP option. I have tried all kinds of a tunes during the past year and payed with the Time-Based Accel Enrich.

Speed Density:
good steady state driving, does not obviously work above 10% to 20% throttle because of the MAP signal going static at that point, absolutely horrible throttle response (even at small throttle openings)

Alpha-N with Multiply MAP:
better throttle response than SD, ok steady state operation, some issues at low loads

ITB(multiple tries):
throttle response bad again, but idles well and works well at low load conditions

This all lead to the conclusion that whenever MAP is used as a fueling factor, throttle response goes down, being worst with SD and ITB since the MAP affects the fueling through VE table and Multiply MAP.
So I pretty much decided to give up on MAP signal and run pure AN with EGO correction enabled.

[TimoL]

There is an issue with staged injection using sequenced batch fire.
Sequenced Batch Fire - Feature is not working
http://www.msextra.com/forums/viewtopic ... 90#p477890
Injector driver?
http://www.msextra.com/forums/viewtopic ... 01&t=65775

The next firmware release will include Sequenced batch-fire tooltip.
http://www.msextra.com/forums/viewtopic ... 91&t=66021

This issue also occurs with MS3
http://www.msextra.com/forums/viewtopic ... 31&t=66098

I see. I will do reading on those when I've got the time.
May do some scoping if I decide to fiddle with the settings.

But right now I don't see any benefits going back to sequential

[rickb794]

You might benefit from timed injection.

Remember with staged injection if the injector is not spraying at the back of the valve, spraying in the airstream will change the tuning of the intake tract by changing the weight of the air.

[TimoL]

Yes and I don't mean that seq does not have benefits it's just that the gains do not outweigh the losses in my case.
The transient response just happens to be the top priority. I have a quick action throttle and TPSDot values above 2000 are not uncommon.
I would love to have ITB mode with seq AND good throttle response, I'm just not able to get it working.

If someone has a similar engine running ITB I would gladly listen to their advice...

[TimoL]

Update:

I got the secondary injector mounted and I changed back to staged sequential injection without the previous problem occurring. I verified that it was not the injection angle param mentioned earlier. Must have been some other setting...

The setup works quite smoothly despite the VE table needing a lot of work and the inj2 dead times. The current dead time of 1.55 ms was determined visually. I had the injector out of the bike with running a very small PW and I increased the dead time until I saw the injector spraying.

There is a one strange thing happening sometimes when the staging comes on. I get these sudden changes in staging percentage even when the RPM and TPS stay pretty constant. This causes a momentarily loss in engine power which is quite noticeable when riding the bike.

Isn't the table based staging supposed to be relatively smooth because of the interpolation between the cells? For example if a have the staging percent set to 10% @ 3000 RPM and 20% @ 4000 RPM I would get 15% @ 3500 RPM? Currently I have effectively used only half of the table and I could double the resolution possibly alleviating the spikes.

But I would like to know if this behavior intentional or not? You can see the spikes occurring roughly at 6163s, 6400s, 6428s and 6456s. I also verified that they match the exact times when I felt the pause in engine torque by viewing the logs in msDroid with the map view.

[jsmcortina]

The current dead time of 1.55 ms was determined visually. I had the injector out of the bike with running a very small PW and I increased the dead time until I saw the injector spraying.

Sorry, but that is not the correct way to determine dead time and will give you a very wrong answer.

Flow your injector, it is easier and quicker than it looks. https://www.youtube.com/watch?v=lJgi8l-2XG4

James

[TimoL]

The current dead time of 1.55 ms was determined visually. I had the injector out of the bike with running a very small PW and I increased the dead time until I saw the injector spraying.

Sorry, but that is not the correct way to determine dead time and will give you a very wrong answer.

Flow your injector, it is easier and quicker than it looks. https://www.youtube.com/watch?v=lJgi8l-2XG4

James

Thank you for link, the vid explains the procedure quite well. I was planning to make an Arduino based test bench for injectors but since MS2 has testing capabilities why bother.
But the last time I played with the injectors test mode I had to re-flash the firmware.

So the sudden jumps in staging are normal and the incorrect dead time is most likely the culprit? I will try to get the injector(s) test flowed as soon as I can and see if it makes any difference.

[jsmcortina]

But the last time I played with the injectors test mode I had to re-flash the firmware.

MS2/BG firmware perhaps?
I've never heard of that with MS2/Extra.

James

[TimoL]

But the last time I played with the injectors test mode I had to re-flash the firmware.

MS2/BG firmware perhaps?
I've never heard of that with MS2/Extra.

James

Yeah, that happened last year. It must have been running the pre-flashed FW back then, which was indeed MS2/BG firmware.