Stumbling After Cold Start

Recently, my car has started running really rough when it first starts in the morning. The throttle is very unresponsive, and under light load below about 1400 RPM (i.e. going up a hill, lightly accelerating, etc.) it stumbles pretty badly. It's weird, at around 1400 RPM it instantly picks up and feels fine. For testing purposes, I've pulled timing way back below a certain RPM, and that felt very similar to this - bog, bog, bog, then instant smooth acceleration. Also, it feels like the engine kind of jolts immediately after I add throttle (possible related to the acceleration enrichment function?). After the engine runs a few minutes (haven't been able to watch the datalog, but it seems like it's when the engine goes to closed loop), it seems to go away entirely.

I initially thought it was running lean, so I gradually increased startup enrichment to about 15 percentage points higher than stock, but that didn't seem to help or hurt. It doesn't smell abnormally rich, so I'm hesitant to blindly decrease enrichment.

I've made a few changes recently that could have caused this. I put long-tubes on it just over a month ago, but I don't recall seeing this behavior immediately after. A few days after adding the headers, I drove out from CA to school in VA, and I'm wondering if the humidity or cooler morning temps could have something to do with it. Lastly, a few weeks ago I discovered my PCV system was set up backward (the check valve was on the wrong side, causing the air to flow the wrong way and a strong vacuum in the crankcase), so I switched it to the proper orientation.

Looking back at my BIN dates, it looks like the stumbling popped up shortly after I fixed my PCV system.

Do you guys have any idea where I could start with diagnosis?

Can you record a datalog and post up?

Well, I took a couple datalogs where it was stumbling really badly, but ALDLDroid didn't save them for some reason (looks like there's a hidden 256MB limit for all saved datalogs?).

I took another this morning, but it didn't seem to be stumbling at all, which could have been for one of several reasons: it took a few minutes to get the datalog started, so it warmed up a little before I started driving; I left the house a bit later, so it was warmer; and it felt quite a bit less humid than previous mornings. It still sounded like it was idling rough though, so I'm attaching the datalog I took. (I attached a ZIP that has both the CSV and the XDL.)

I'll try to go for a drive early tomorrow morning to see if I can get a better log.

Here's a better datalog. Still wasn't doing it real bad for some reason, wonder if it's weather-related.

I viewed the first datalog this morn and didnt see anything that stood out. I need to find the more complete adx for 8D and will review both logs in more detail. I'll try to do it this eve but it may be late before I get a chance.

Brian

Thanks Brian, really appreciate it.
What do you mean by the "more complete" adx?

I have checked the log, you are running a little rich with BLMs around 120, and the moment you tap the throttle it gets momentarily lean from what I can see. Check your fuel pressure and have a look at the spark plugs.
If you still have stock injectors it's a good idea to check them or replace them anyway to D3 type injectors. Stock ones tend to finish their life after those 30 years.

EDIT: I read on other topics that your car is far from stock... I'd say to give it more AE on low temperatures in: "AE Delta TPS Pulse Width Factor Vs. Temp. ".
And work on the tune, log shows rich everywhere.

Thanks for looking at my logs. I tried checking the fuel pressure before I drove out to VA, but the gauge I got from O'Reilly was leaking like crazy. I don't have time (or tools) to pull a spark plug right now, but the last few times I've had them out they've looked pretty sooty.

I've got new 33.3 lb/hr injectors, so that's not a problem.

Tuning AE was on my to-do list, but man does it look complicated. Here's my current table; about how would you change it?
Attachment 13321

I have adjusted and updated my VE tables based on several long datalogs, but isn't it normal for it to shift rich and lean as temperature and humidity change?

Right, the fueling can change due to humidity and temp, but I suspect you are hitting a weird transition on the VE table when this stumble occurs. Are your VE tables smooth? Try increasing the TPS AE factor at the temperatures at which the stumble occurs.
"AE Delta TPS Pulse Width Factor Vs. Temp. "

Try using the XDF I attached, I tweaked it some to make it more intuitive, make sure tu use it in category view.

I've smoothed my VE tables pretty extensively, although there are a few little areas that are a little jagged. Here's the graph of my lower VE table:
Attachment 13324

Whoops, found the right table. Here it is:
Attachment 13323
It's stumbling around the 20 °C cell, so I guess increase it there from 1.00 to 1.25 to match the above cells?

Hmm, that's a nice-looking XDF. I like how you added the patches section - maybe I'll finally try patching my BIN if it's that easy. The XDF I'm using (wish I could give credit, but I don't remember who on this forum I got it from) breaks parameters up into sections though, which I've really gotten used to. Check it out. Maybe if I have time I'll make a hybrid of the two.

TPS AE is just a short burst. If it still stumbles, it may need the MAP AE increased also in that area. I booted into windows this eve and replayed your logs. Its a little rich overall but not terrible. Only other thing that stood out was a slow oscillating integrator Do you feel a surge while cruising? I viewed the log using the monitors in Tunerpro and it looks like the integrator is driving the o2 crosscounts. Usually this means you do not have enough proportional gain. Green is O2 voltage, Red is TPS and Blue is the Integrator. If you don't feel a surge, I wouldn't fool with it.

Quote:

---

Originally Posted by 90Vette

The XDF I'm using (wish I could give credit, but I don't remember who on this forum I got it from) breaks parameters up into sections though, which I've really gotten used to. Check it out. Maybe if I have time I'll make a hybrid of the two.

---

Right, I removed all the breaks as I prefer category view.

Quote:

---

Originally Posted by 90Vette

It's stumbling around the 20 °C cell, so I guess increase it there from 1.00 to 1.25 to match the above cells?

---

Yes, increase that but also adjust the surrounding cells to smooth the table.

I increased the AE table like so:
Attachment 13329
and hooooly crap that was bad. Whenever I touched the gas, the engine just about died. After about 30s to a minute though, it ran great. Check out the datalog I took, and let me know what you think.
Since adding AE TPS fuel seemed to make it worse for the first little bit, I'm wondering if it's running too rich. The weird thing is, it would have been in a cell with 20% enrichment (20 °C), which doesn't decrease until 56 °C. I don't understand why it would run better so quickly after starting.

Quote:

---

Originally Posted by sturgillbd

TPS AE is just a short burst. If it still stumbles, it may need the MAP AE increased also in that area. I booted into windows this eve and replayed your logs. Its a little rich overall but not terrible. Only other thing that stood out was a slow oscillating integrator Do you feel a surge while cruising? I viewed the log using the monitors in Tunerpro and it looks like the integrator is driving the o2 crosscounts. Usually this means you do not have enough proportional gain. Green is O2 voltage, Red is TPS and Blue is the Integrator. If you don't feel a surge, I wouldn't fool with it.

---

How exactly does MAP AE work? Somehow I've only ever read about TPS AE.
Also, what are you looking at to determine whether TPS AE needs to be adjusted? I'd looked at O2 volts vs TPS %, but I couldn't find a clear pattern. Do you use the integrator for that?
Interesting about those oscillations. I haven't noticed a surge at cruise, but when I'm coasting in neutral I notice the idle speed oscillates up and down a few hundred RPM. Could that be related? Because that drives me crazy.

Quote:

---

Originally Posted by dzidaV8

Right, I removed all the breaks as I prefer category view.

Yes, increase that but also adjust the surrounding cells to smooth the table.

---

Oh wow, I didn't even realize category view was a thing. Don't know how I've never played around with that drop-down. That's a lot better, thanks!

Too much AE can cause a stumble too. Too little usually cause a pop back through intake. Too much causes the engine to bog. TPS AE is a short burst as throttle blade opens. MAP AE follows though when the throttle is held open during acceleration. If adding AE makes it worse, you may try to remove 10 % of the TPS AE from your original value at 20C then try 25%. View your log and watch for a TPS increase. If the O2 voltage goes lean, It usually needs more AE. If it goes rich, you need less AE.

I've driven it the past couple days with the tune with more AE, and it's definitely worse. I just made a new tune with less AE, and hopefully that will take me in the right direction. If that helps, I'll also try leaning out open loop at those temps.
I was trying to look at a couple new datalogs I've taken though, and I'm really having trouble seeing any sort of pattern in O2 sensor volts when TPS % goes up. I've attached them; can either of you tell me anything from them?

I tried leaning out startup, and that didn't help. It wasn't any better when it was cold, and it was a lot worse hot. Is it possible this is moisture related? I realized I kept it in the garage in CA, but with it staying outside here I'm wondering if dew could be causing a problem.

I've noticed the coolant temperature needle is always in the exact same spot when it starts running rough. Does this just point to a bad spot in my closed loop AFR table?

Attachment 13674

What would be the table to adjust for this - "open loop AFR vs temp" (0x45e) or "startup enrichment vs temp" (0x431)?

Lately it's been happening pretty bad, and I've been paying close attention to what it feels like.

It seems that when I add throttle, I get a momentary surge of acceleration, but then it completely bogs. I've run a few maps with very little timing in certain areas, and that's exactly what it feels like. After a minute, even if I'm just cruising at part-throttle I can feel it surge forward (I assume when it enters closed-loop).

Yesterday while it was running rough I noticed a pop from the engine bay while accelerating under maybe 60% throttle, which I guess must have been a backfire? It definitely wasn't just exhaust popping like you get under decel. AFAIK, backfiring can be caused by either a lean mixture or heavily-retarded timing, right? It's tough without the wideband hooked up (and I don't have time to put it on right now), but I guess I should just try adding some fuel on the warmer side of the open loop table, unless you guys have some other insights.

Again, can anyone point me to the difference between "open loop AFR vs temp" (0x45e) and "startup enrichment vs temp" (0x431)?

Here's a new datalog. For some reason I didn't see it running super rough like usual, although right after startup it was wanting to buck a bit when cruising in first/second gear.

The only change I made was increasing the "Closed Loop Proportional Step Size Gain Vs. RPM" (0x4D8) values by 10% across the board. Interestingly, cells 1600-2400 just reverted to their previous value due to the number encoding. (I also changed the "Injector Flow Rate for Instrument Panel" to fix my MPG readout, but I'm sure that is irrelevant.)

I was able to take a few more logs, and on the first two I'm certain I captured the problem. I haven't been able to log with someone in the passenger seat watching my phone, but I'm almost certain it occurs just before it goes into closed loop. Looking back to a log I took yesterday I believe it was running rich at that point - the O2 sensor shows about 850mV right before it drops into closed loop. This problem did seem to show up when I added headers, so I was wondering if I perhaps needed a heated sensor, but if my sensor were too cold I believe it should be reading a low voltage, not a high one.

I'll drop the AFR a bit around 120 °F and take another log as soon as I can. Although the weird thing is it's already tuned to run a bit lean in open loop - the table reads 97.3% change to AFR from 20 °C all the way up to the top of the scale, which would mean it should be pulling fuel? I wonder if I have an exhaust leak that's throwing my O2 sensor off; I have been noticing my exhaust tips getting really sooty lately.

I once had an exhaust leak at the header collector coupling, because both my narrowband and wideband sensors were mounted near by they were being effected by the leak

Quote:

---

Originally Posted by Kitch

I once had an exhaust leak at the header collector coupling, because both my narrowband and wideband sensors were mounted near by they were being effected by the leak

---

That's what I'm wondering. When I'm driving with the windows down, if I pass something that reflects my car's sound back to me it sounds like I can hear an exhaust tick coming from the left side.

Exactly how much are you supposed to tighten the bolts on a header ball flange? I got them fairly snug, but I wasn't able to find a torque spec online, and I didn't want to risk crushing the ball flange.

Crawl under it with it running and feel around the flange with your hands to see if you feel a puff of air indicating exhaust leak. If it's that bad you can probably hear it though.

Sooty black indicates rich condition. What do the plugs look like? I've had to clean/replace a couple sets of plugs on the TBI I'm working on dialing in.

And definitely get that WB on there. Will make it easier on you. The first time I put a WB on something I mounted it really close to the end of the tailpipe. The readings on the WB were all over the place from the atmospheric air that was getting to it.