[propertyco]

Currently building out a PNP ecu for ford eec-iv using v0.4.4
My "shop" car currently runs on a PNP based on microsquirt.
Engines a Turbo 2.3
1000cc high-z injectors, 340lph pump, E85
Runs flawlessly on my MS ecu.


I have full sync for RPM using WTMTronic's optoisolator, seems reliable no issues.
I'm using STD95N2L power mosfets for injectors and fuel pump.

First issue i ran into was the CPU not properly powering the mosfets on, i put a IXDM602sia between the CPU and mosfet for the injectors, got them to fire without an issue this way, not sure what the issue is for this as gate threshold for these mosfets should be 1v, measuring gate prior to adding the gate controller i'd see 2.2v

I'll sort that issue out i'm sure, but the main issue i have right now is the car leaning out hard at 3k.
It'll act like it's spark cutting, though all engine protection is shut off, I wonder if my fuel pump fet is turning off and then back on.

Any one see anything obviously wrong with this? I'll attach a tune and a log, the tune is VERY ROUGH, i know, but it's just to drive around on 3psi or so, as to not waste my tuners time with a proper tune.

BTW, xls file is the short log i took, rename it for megalog to see it, couldnt upload it with the standard extention.

[jonbill]

What's a "shop" car, and is that the car we're talking about here?
your dwell voltage correction table is garbage - you want 100 rather than 0 for dwell % as a starting point.
I don't see much else terribly wrong. Have you tried larger VE numbers in the cells where it goes lean?

[propertyco]

Using TFI for ignition, i should've specified that prior, dwell time is controlled by TFI and not the ECU.

Car's a Merkur XR4TI.

I'll try larger VE numbers tonight, comparing to my MS tune it's not too far off, ratio wise, from what that sees.

[PSIG]

You may be right, looking for an external factor combined with crappy tuning (at this level). I do see the roughness, but all the data is parallel (DC, PW, TPS, etc to MAP, RPM & AFR). So it's basically doing what you're telling it to do, and while data response is expectedly crappy, it is consistent with no sync losses. It is obvious you have some lean bumps from no AE. In an hour, you should have it tuned 90% below boost and working on timing and AE. By then either the issue will be gone, or what's happening will be clearer.

I will also note your MOSFET number has been flooding in from overseas with many having very poor performance or failures. Requiring a driver may be a sign of that. Verify they are originals or not. Even if original they have poor ratings for stuff like injectors anyway. I would verify you're not getting any noticeable heat after operating a while and your fingers on the FETs. I didn't do any calculations, but I imagine 25VDS is going to put a lot of injector EMF energy through them.

Final suggestion is that you have several compensations (not corrections) that you are guessing (not found by tuning) and will confuse things as you tune, such as Cold Advance and IAT Retard, which can fight each other or give weird results you're not expecting, because they happen invisibly in the background. Tune compensations in as you find you need them, and how much. AE Taper is another common example of that, and the data will tell you obviously if you need it and how much when you get there, but not if it's already on and doing something. Have fun!

[propertyco]

FETs were from digikey, i'd certainly hope they're genuine but who knows now-a-days with how china pumps out fakes of everything.

Idling i could comfortably rest my fingers on the FETs, i should note the ground plane is rather large on this board.

I have some ST VND5Ns here i'll swap in for the injectors for the time being, maybe that'll help rule something out, just odd, looking at the logs, it leans out right at 100kpa (i live at sea level)

[PSIG]

Keep in-mind, the data is not simultaneous, so one action must be correlated with another by time shift in data. A good way to do this is simply find an obvious change (throttle quick open or close), and note the TPS should be before the fuel change, and by how much. Get that in your head, so you know the other changes follow, and by how much. I hope that makes sense.

An example is O2. Not only will it see a change in AFR with a throttle change (TPS obviously came first, followed by MAP and then PW some time this cycle, etc), but then O2 waits to see the exhaust from that event to sense down the pipe (hundreds of ms). Then, O2 lags in processing, reporting to Speeduino, then the data output called in-series. Some data channels report more often than others. So, we know the change in AFR is coming, and that it will lag far behind the action, even though it may appear in data to happen either nearly simultaneously, slightly ahead or far behind. For AFR, we have to calculate the different O2 lag at different speeds and exhaust flows, and use that to filter our data in MLV.

It sounds more involved and more work than it is, but the first step is to get the relationship in your head to better see what the data is telling you in a relative sense. OK, that really does sound much worse than it is... LOL