The AFR error is huge at the lean end. Gauge will read 22ish while MS is reading 17ish. I do believe I will try connecting via CAN, which people say will work between AEM and MS/Extra.
First tank of fuel: 15mpg. Yuck. Going to say "berk you incorporate AFR table" and go back to tuning the old fashioned way: Unified VE table and pull fuel at various points in cruise until it stops feeling good, then add a little. It's iterative, sure, but less iterative than getting the VE table perfect, then tweaking the AFR table, then realizing it is undershooting or overshooting because the injector characterization is imperfect, etc. But if I get it running well, I can maybe datalog the final tune, get a histogram of the AFRs it likes, and use that as an AFR table for running closed loop. Maybe.
In reply to Pete. (l33t FS) :
All fair points, and I haven't tried it myself, just thought I'd pass it along in case it gave you a path to usable parts.
I took a look at the the pair of rotors that I have. Not sure how you measure them, but the gaps look like about 0.08 or 0.09 mm at all the corners which is just inside factory spec. Much tighter along the length the groove. If I'm measuring wrong, let me know.
These came out of a running street engine.
One bearing needs replaced for sure, the other might be ok. None of the side seals were stuck, so I think these are usable. Let me know if you want them or need more info.
In reply to Pete. (l33t FS) :
A while back you had talked about using pb blaster and a rag to clean parts up.
I tried it on a plastic clutch cable last night. Works PERFECTLY. Thanks for the tip! Owe you a Pepsi for that one.
Thinking out loud.
When I started the car after a good hot soak, it was running extremely lean.
The fuel calculation takes air temp into account. Higher air temp means less dense air means less fuel.
So the computer was taking out too much fuel for the less dense air. It was overshooting the calculation. This means my injector dead time is off. But which way?
Because I like big fat numbers, let's say when running correctly the dead time was 2ms and the pulse width was 10ms for a total of 12ms of on time. If the temp correction wants to cut 25%, then it will cut the pulse width to 7.5ms for a total of 9.5ms of on time. If it wants to add 25%, then it bumps the pulse width to 12.5ms, for a total of 14.5ms.
Now let's say I have the dead time wrong and have it set to 3ms, and have tuned around that error so the pulse width is 9ms, to give the 12ms of on time. Temp correction cuts 25%, cutting pulse width to 6.75, for a total of 9.75ms, or adding 25% makes the pulse width 11.25 for a total of 14.25. That's an undershoot error, it is not leaning enough and it is not richening enough.
But lets say I have the dead time wrong the other way and have it set to 1ms, and I am at 11ms pulsewidth to get the 12ms. Subtract 25%, that makes 8.25 for total of 9.25, add 25% makes 13.75 for a total of 14.75. An overshoot error.
I am going to add some deadtime. I'm currently on .95ms because that was as good a place as any to start. I found my old MegaTune files and see that when I started using these injectors, I was using 1.2ms. Sounds as good a place as any to start...
Was checking under the car to find a squeaking noise.
ARGH
Pete, I assume you've confirmed you have healthy battery voltage? If your battery or charging system is tired, it would could be aggravated in a hot start condition further throwing your dead time off. Seems like a stretch to me though. How do you know you just haven't got the fuel maps correct for those IATs?
what about just running no mufflers? or even better a rotary muffler
In reply to Mezzanine :
Air density is a calculated value and not something to tune for.
Another clue is that when running with "incorporate AFR", it undershoots or overshoots.
All fuel calculations assume that the injector characteristics are correct. If they are not correct, then you are going to have a bad time.
EvanB (Forum Supporter) said:
MrChaos said:
what about just running no mufflers?
What is interesting is that you cannot hear the Nova doing a BURNOUT over the Starlet idling.
In reply to Pete. (l33t FS) :
That's some old school d42
Meanwhile, in another thread...
In reply to Pete. (l33t FS) :
I cruise by here and a few other threads looking for rotary content, and to see what other mischief is up.
Then you post those vids. The brap is calling.
And I just got rid of my check engine light and ordered a/c stuff. E36 M40 IT!
Videos TONIGHT. I promise.
Backtracking.
Yeah, that happened. I would like to thank the Miata's ability to spin out on short notice...
I uad been posting slower raw times than John or Evan. At the lunch break, John said something like "I'm only 17 seconds behind you!" Huh? John had a spin and was hitting all the cones, while I had been driving uncharacteristically clean, even though I was way, way, WAY overdriving until my tenth and final run. Turns out your times improve if you don't have to re-collect the car after every corner.
You saw the muffler carnage pics, but what you did not see are what I found after runs were over. I think I found the cause of the binding problem.
Tires are GRINDING into the upper chassis rail. I had been suspecting this, which is most of why I painted it all white.
So, wat do? I want to go with smaller tires, even though that would hurt straight line traction. I would also have to re-engineer the rearend, as I cannot fit 14" wheels over the brakes. (I tried. It was laughable) I could cut and section the body, something else I had been meaning to do, but there is just no time. So, on with plan "compromise".
I raised the front suspension five turns. The spring seats used to be resting on the strut ears, for an idea of how much I raised it. Had to realign it, of course, because bumpsteer.
Fortunately it does not seem to be running hot on the highway like it did with the old struts. I re-painted the fender, we will see how it works tomorrow...
how bout some bump stops in the rear
oh i thought the tire grinding was on the rear
such a short course. the shortest course the local region runs is 70ish seconds in the fast MA cars. the usual courses seem to be 100 seconds plus.
They canceled everything until mid July for us. So I am living vicariously.
Thank you for the Rallycross fix!
In reply to MrChaos :
Maybe drive faster? 
Was faster than all the MA cars
