MS1 parallel install: idle and enrichment issues?
 

1997 OBD2 Miata
MS1 installed in parallel with dedicated GM IAT and CLT sensors, and LC-1 O2 sensor
*MSnSE HiRes 11d firmware
Kraftwerks Rotrex C30-74 supercharger kit (w/ DeatschWorks 22S-00-0350-4 injectors, DeatschWerks 255 lph fuel pump, vortech style FPR, and intercooler), which peaks at ~10psi
*IAT sensor mounted in cold-side end-tank of intercooler
*CLT sensor mounted in water outlet at rear of engine (w/ coolant reroute)
Stock coils and wires, 1-step colder NGK copper plugs gapped to .030

I'm trying to get the car running reliably on the MegaSquirt (in part because it seems to handle the larger injectors far better, but also in preparation for the eventuality of removing the Rotrex and installing a turbo). I'm using Al Hounos's basic spark map and I think I've got a fair fuel map put together with VEAL, but I have some issues.

I've tried to turn off nearly all enrichment tweaks (EGO, Baro, Accel, OverRun, etc and CLT/IAT-related correction is set to "normal correction") while fleshing out the VE table. I've also set AFR targets to 14.7 in idle, cruise, and deceleration cells.

Generally speaking, with IATs around 60F-70F degrees and CLT around 190F-199F, the car drives nicely. Most of the time (though not always) it idles pretty well too.

The car starts fine but idles roughly until CLT is up to temp and the car has been driven around for a few minutes. Often it is a bit lean, so I manually bump up VE in the map idle cells, but once up to temp it is rich and I have to readjust. I assume this means that I need to work on warmup enrichments?

Even when it is up to temperature, sometimes the idle is a bit rough. I've attached a log of it running rough at idle after the car was warmed up and driven for a bit. I've also attached my current tune file, if someone would like to look it over and share some advice. If I can get it to idle and run at colder engine temps, that would be a great hurdle to clear.

One would assume, yes.


what size are the injectors you are running? 8.4 Req fuel suggests a 360cc injector? a 1.48ms deadtime seems a bit larger for an injector that small, coupled with a fuel pump that's goign to raise your base fuel pressure a whole bunch, you might have trouble getting small enough pulse widths at idle to maintain 14.7 and nede to target closer to 13.5-14.



One thing I see that would be hurting you is a 10% fueling difference between 500 and 800 RPM. Those first 6 cells of 40 | 30 should all pretty much be the same value. The smoother you get you spark and fuel at idle, the smoother idle will be.

Once you have your VE table setup at full temp, THEN you can alter your enrichment tables.

The cells around 30-40kpa @ 500rpm were made higher back when the car was still NA and I hadn't really touched them even though I'd dropped the values at 800rpm and higher. I had made them richer to help catch the idle if/when it would sometimes droop below 800. I will drop them closer to the values at 800 and see if it helps. Thanks for catching that!

The documentation that I got with the injectors follows:
I chose 360cc/min because the regulator on the rail should average around ~42-43psi (I think?), and the 1.48ms because injector voltage generally hovers just above 13 volts. Did I misinterpret the data?

Sometimes the car idles perfectly well at 14.7, but there are other times when it definitely seems smoother on the slightly rich side (13-14).

I made some very minor tweaks to the WUE this afternoon and the car was much smoother starting up and idling up to operating temperature. Of course, the ambient temps weren't very low, but it was definitely an improvement.

I think that you are onto something with the injector dead time. I messed around with it a bit and the car idles MUCH better at both rich and lean AFRs with smaller dead time numbers. I'm still not certain what time is ideal though. I tried using the log+spreadsheet method, but I think I've got too much data noise. While setting various limits for the data, I'm seeing a likely range of between 1.16 to 1.22, if I'm understanding it properly. I'd still like to refine it a bit more though. I might have to try the dead time logging a few more times to get more precision.

I've attached my working spreadsheet, if anyone would care to take a look and share some feedback? K1 is the lower limit and L1 is the upper limit. Limiting the upper range to ~2.5 and slightly lower seems to clean up the trend line a little bit but it is still messy though.

EDIT: I fixed some errors in the spreadsheet and now the trendline offset is looking more like 1.4 - 1.5 again. :/ But it looks like it's being skewed by noisy data and the car definitely seems to idle better with a lower value. Suggestions?

After futzing with the method in the "How to find your injector dead time" thread, I decided that 1.13ms seemed to fit the data plot fairly accurately, so I set the dead time to 1.13 and re-tuned the fuel map. So far, with what little testing I've had a chance to do, it seems to be much more stable.

That said, I'm still not confident that I've got things right and I'm still a bit concerned with the deviance from the published table that was included with the injectors. Was I just interpreting the data incorrectly? When I go to the DeatschWerks website, it lists a minimum PW of 1.1ms. That value corresponds to 17-volts on the chart that they included with the injectors, but I know that I'm not getting 17 volts to the injectors (generally between 13.1-13.9 volts).

Has anyone else used DeatschWerks injectors, and if so can you share your experience/expertise?

They have really shitty deadtimes.

^ I think you're right. After continued issues, I switched things back to the documented values and decided to just let it idle where it was happiest (at 12:1) for now. After another few VEAL sessions things seem to be shaping up finally and behaving a bit more consistently. Some changes in weather to really alter the conditions/variables might be helpful to verify though.