fugiwara

iTrader: (124)

I didn't assume, I base on the facts you provided in the posts and go from there!
If you don't give enough info, don't expect people to know whats on your little mind.

l2r99gst

iTrader: (2)

redlanevo, I know you have already done a million logs, but can you do one more just like your last one, with everything you had logged, plus iscsteps and load?

It's just that I have a log of not too long ago where I was testing the iscsteps and I have a spot in it where I was blipping the throttle and I am comparing that log of mine to yours.

Your RPM definitely drop faster than mine, but I want to see more data right around the time when it is dropping and trying to catch back up. I want to see if the iscsteps are doing anything out of the ordinary.

Also, let us know if you try the maf smoothing suggestion. I don't know if this table was ever disassembled or discussed properly, so I'm not positive on it's use. I am curious how that affect this as well. If it works the same way that the maf scaling does, then I think your problem is exactly opposite of what RoadSpike is talking about, but it's worth a shot nonetheless.

Upon a brief look at your log compared to mine, your airflow reading stays higher much longer than mine. Mine returns to normal after a second or so and yours stays high for a good 3 or more seconds. It's during this time that the RPM drop and stumble.


Thanks

RoadSpike

iTrader: (5)

There is a thread on it somewhere in the ecuflash subforum i just don't have a link to it right now. I'm pretty sure the problem is he's going lean as you said yourself the STFT is shooting up over compensating then dropping like a stone. I think if we can nab it at the start with more fuel it wont try to do the whole jump then dive scenario

take everything i say though with a grain of salt i'm not a god at this

RedLanEVO

iTrader: (16)

Ok. For load, I have to log ecuload, injpuslewidth, and afrmap in order to log load...so there will be some data you don't need.

Here is it...

l2r99gst

iTrader: (2)

Yeah, but that's the confusing part. In his log during the stumble, the wideband goes very rich, like 10:1 and the Hz readings stay somewhat high, like from 80-100. So, if the smoothing table is scaled the same way, that would be the 4th and 5th cells, and he would need to take out even more fuel, not add. But, if he adjusts those, then it will be affecting the cruise trim in that area. His cruise trims are getting better, but they are still negative, so maybe further adjustment will help more. Maybe he just needs even more correction in the cells starting aroudn 75 Hz and up.

There is definitely something very weird with his setup, though. I am trying to wrap my head around his latest log, but some of the values just contradict each other.

For one, he is idling anywhere from 12.5:1 - 14.5:1, yet his LTFT Lo and STFTs seem OK at idle. That makes 0 sense. Either his wideband isn't correct or the ECU isn't using the LTFT Lo at idle. There is evidence of the latter in his latest log. At the end of the log, the LTFT Mid starts changing at idle, which shouldn't be happening. In my experience with DSMs, the switchover point for LTFT Lo to LTFT Mid is around 70 Hz or so. His idle is very close to this sometimes, but below, so I don't know why his LTFT mid was changing. Actually, that could be part or most of his issue. If his idle and cruise are using the LTFT mid, each one could be changing it and screwing things up.


Eric

l2r99gst

iTrader: (2)

redlanevo, was your car showing a rich idle on the wideband with the stock pipe as well?

Also, don't worry about another log just yet with the load and iscsteps, etc. We have to figure out a couple of other things first.

RedLanEVO

iTrader: (16)

I posted the log you requested 2 threads above.

RedLanEVO

iTrader: (16)

The wideband with the stock pipe w/cone filter showed 13.8-14.6. It fluctuated between those numbers.

I do have a universal Bosch front o2 sensor. Maybe that's why?

l2r99gst

iTrader: (2)

I was going to ask about that next. When did you get this?

Also, that doesn't sound right even with the stock intake. Either the wideband is reading wrong or the front O2 is reading wrong. Maybe the addition of the Burschur pipe just compounded the problem and that's why you didn't have any issues with the stock pipe.

Something definitely doesn't seem right. I'll take a look at your latest log.

RedLanEVO

iTrader: (16)

I got the universal o2 a while back...but I'd say no more than one year. I would suspect the o2 since it's a universal and you have to make the connector which extends the wires, adding more resistance. I've heard that those are really sensitive to resistance since they have no more than 1v going through at any given time.

I got it before I actually went to a tech school for Automotive service and learned that universal o2's are not good.

l2r99gst

iTrader: (2)

OK, I have studied your log enough to know exactly what is happening. The question now is can we fix it with the tables that we have defined. That's what I have to think about for a while.

Basically, what is happening is that the maf is overreading after releasing the throttle. Instead of the Hz dropping back to normal quickly, it hovers higher than it should, and when the RPM drops, it causes your load to spike as a result. This overreading Hz value is causing way too much fuel to be injected during this stumbling phase. Only when the STFT has enough time to compensate and drop rapidly does the idle come back up to normal.

Another odd thing is that your iscsteps do not respond at all to the throttle blip. The ISC steps should rise with the throttle blip and ramp back down, wihch will help controlling the RPM drop below your target. I can post up some pictures of my log compared to yours.

Did you ever alter any of your ISC tables?

Edit: Check your PM..it would be easier if we jumped on IRC or somewhere where we could type in realtime instead of making a million posts here.


Eric

l2r99gst

iTrader: (2)

OK, here is a comparison of me blipping my throttle to you blipping your throttle.

A few differences to note:

1. In my log, ISC steps go from 38 to 73, and takes 1.5 seconds to come back to 38 again. The idle then slowly comes back to the target again.

In your log, the ISC doesn't even budge with the throttle blip. It starts at 58 steps and stays there. There is a very brief rise to 69 steps then right back down. But, there is no response to the blip at all, whereas mine follows the throttle blip.

2. My airflow Hz takes about 1.5 seconds to return to normal, which in my log was 31 Hz. It went from 288 Hz to 31 Hz.

Your log takes about 3.6 seconds to return to a normal Hz reading, from 220 Hz back to 57 Hz. During this whole period it is reading way too high. This overrun is causing the load to spike as the RPM dips.

3. After the throttle blip my load reading drops well below the starting point, then works it way back up slowly through a range of about 5 seconds.

Your load drops initially, but the overreading Hz values causing the load to spike back up to where it was when you were blipping the throttle. This load stays there for around 5 seconds before returning to normal.

I won't consider AFR at this point, since I'm not sure if your reading is accurate or not, but it is pretty evident that your MAF is overreading after the throttle blip and causing way too much fuel to be added, as evident by your injector pulsewidth rising, your Hz staying too high, your ISC not increasing, and your load spiking.

My log:

Your log:



I think if we could get your ISC to respond, we may be able to fix this, even with the MAF overreading.


Eric

RedLanEVO

iTrader: (16)

Check your PM

RedLanEVO

iTrader: (16)

Ok, nothing works. Time for a new intake. This one sucks.

RazorLab

iTrader: (8)

One thing to check is if your ISC is even working. Plug in a idle rpm like 1100 and see if it responds.

I have seen a few 03 evos with the little plastic teeth on the ISC broken and it obviously it just moves free and doesn't respond to any changes.

It's a good thing to check.