[KLRthur]

It lives! '97 0.993L 3 cylinder Geo, stripped interior, building as a Lemons entry. There's an issue with the EFI though, in the attached log right around 420s the engine starts bogging and stuttering. The WUE bit is a red herring. Cutting ignition and restarting returns to normal operation. I drove it after this for 30 minutes with no issue. Could AE put it into oscillation like this? Just checked and found my TPS sensor filter is at 10, recommended 50? O2 at 11, MAP at 20, Baro at 122.

Suggestions would be welcome!

PXL_20250105_065507362.jpg (488.75 KiB) Viewed 35577 times

The crank pulley was machined to provide 36-1 timing:

PXL_20250202_000011320.jpg (494.87 KiB) Viewed 35577 times

A printed housing for the NO2C and LM1516 trigger board:

PXL_20250212_050249760.jpg (491.54 KiB) Viewed 35577 times

[jonbill]

it's oscillating between rich misfire and lean misfire. the VE table is a little lean at high load and the AE table is .... mental, making it way too rich on acceleration.
I'd suggest copy the original from a base tune. you're adding 150% extra fuel at all TPSDot values.

[KLRthur]

After a parking lot AE recalibration the car is driving much better, thank you for pointing that out jonbill.

A 15 minute drive for dinner and back, ran well but stalled just sitting waiting for the parking garage door to open. I have included a truncated log since the 2MB log didnt want to upload. Stall right at 86s.

Maybe it's obvious, but the PW jumps up and yet the engine still leans out and dies? O2 sensor is right in the manifold. Maybe it's actually flooding? I could reduce the entire first column to match the rest of the table better? Those values would have been left over from the table generator. I have used the autotune extensively and the cell changes have settled down to +/- a percent.

Clipboard01.jpg (435.13 KiB) Viewed 33477 times

[PSIG]

Your sync seems to be borderline. Is that true? 7 sync losses in 1.6 seconds, although I imagine some of those are bumps as the engine quit. More sync loss on start. Why are you holding idle so high?

[KLRthur]

This is the best CAS configuration to date though it's obviously got some problems. Occasionally on startup it has what seems like a mild backfire, more of a 'chug' and then starts. I setup the VR load resistor with 10-turn pot + scope, and the plugs out while cranking the engine. It drives well but that startup issue is present.

sync-loss.jpg (74.26 KiB) Viewed 33169 times

I did not know about the sync loss counter, thank you, I will get back at it with the scope. Do you think this is the culprit RE the hot stall? Happening right before the stall, I'll check other logs.

hotstall.jpg (64.77 KiB) Viewed 33169 times

The idle is high because Suzuki implemented their idle control in a weird way. A DC motor with a travel stop switch advances a plunger against the throttle cam using a worm gear so it doesn't have to maintain a holding current. I attempted to replicate this with a stepper and plunger assembly but it was getting hung up. (below photo is not the final revision, I added more su The next attempt will either be ordering the exact size of linear stepper and better bearing consideration, or just drilling a port for idle bypass and using a more typical configuration. I am doubly confused as there's been some issues with the stepper driver /EN line getting held low by recent fw. Need to check in on that again but my last sanity check saw the driver never disabled and the stepper heating up quite a lot.

idle-motor.jpg (477.19 KiB) Viewed 33169 times

Behind the spring in the following photo is a throttle-stop screw. It's now set at the lowest I can get away with and still have the car start and idle but it does warm up to about 1200RPM.

idle-installed.jpg (450.78 KiB) Viewed 33169 times

[KLRthur]

More sync loss when cold, but at 1900RPM, should be solid then.

anothersyncloss.jpg (69.77 KiB) Viewed 33168 times

These should have been points instead of teeth since I am using a VR sensor? I should try a hall sensor here?

crankpulley.jpg (476.5 KiB) Viewed 33168 times