Dartht33bagger

I've gotten my EBL P4 fairly well dialed in except for one issue plaguing me with warm starts about 50% of the time. Cold starts are perfect and I never have issues with them. 50% of the time, warm starts are fine as well. The other 50% of the time, the car will drop in rpms almost instantly and die unless I give it gas. Even while giving it gas, the car will run very rough until I can shift it into drive (still holding the gas), and then start moving. After I do this, everything is fine.

I've been able to datalog this issue twice now. Both time it occurs, the computer seems to think the car has died. MAP climbs and then stays at 103 KPa, IAT toggles from the correct value it has been reading to -40 for a few frames and then toggles back to the correct value, AE toggles on for a frame or two, and the IAC and injector pulse widths climb. The entire time I'm giving it gas to keep the car alive, the computer is outputting sensor values that make it seem like the engine should be shut off.

After a little while the computer gets out of steady state values and resumes like nothing happened. Could this behavior be limp mode kicking in since I have memcal plugged into the EBL P4 currently?

Attaching my two datalogs of this occurring for experts to look at. The first mark in each is the place of interest.

RBob

iTrader: (1)

Definitely strange, off-hand I'd say that it appears that the ECM is failing. Or possibly the EBL board is loose in the MEMCAL socket. Check that first.

RBob.

Street Lethal

iTrader: (16)

I only looked at the first file, it seems like the air/fuel is momentarily off when you start the engine when warm during OL. At 200 degrees, your target air/fuel should be 14.7, but yours lingers around 13.0 for awhile gradually making her way back up. May possibly try increasing the target AFR values in Open Loop during that RPM and kpa to average closer to 14.7...?

Dartht33bagger

Quick update: I checked the EBL connection to the ECM and made sure it was connected correctly. Everything looked okay connection wise. I've driven the car a few times since then and haven't done many warm starts to really test out if it's fixed yet. Today I plan to do a large number of warm starts to see if it still occurs.

Dartht33bagger

New update: slow progress on this with vacations and work being crazy. Finally started debugging this. Problem is still occurring after re-seating the EBL board connection to the mainboard twice. Everything looks good there. As a last ditch effort I'm going to try to see if leaning out the commanded AFR after a warm start will help things this week. If that does not help, then I'll be asking rBob for some more assistance

Dartht33bagger

I spent quite a bit of time trying to track this issue down this week. This is what I came up with:

1. I changed the Open Loop - AFR Multiplier vs CTS table for the 90C+ values to 2 and then 4 since the EBL documentation says that positive values = leaner. This did not seem to help at all. After a warm start my cAFR was still pegged at 13.5 - the VAC vs RPM base value. Is there another value I should be changing? From reading the documentation this is the table that sounds like it would the right place but apparently not.
2. Driving the car until it warms up, stopping somewhere and attempting to restart the car will not cause the issue. The car will start fine multiple times in a row.
3. Driving the car until it warms up, parking it for more than 30 minutes, and then attempting to start the car will cause the issue to occur where I have to use the gas to keep the car alive.
4. The EBL board is connected to the original ECM correctly.
5. I have continued to observe the odd sensor readings of -40C on the IAT when the issue occurs.

RBob, this looks like it may be an ECM issue unless you can think of something else...

RBob

iTrader: (1)

Most likely, ship the ECM to us and we'll check it & take care of it.

RBob.

Dartht33bagger

Sending out the ECM this week RBob (most likely tomorrow). School is starting up again so it won't hurt so much to have to park the Camaro for a bit

Dartht33bagger

I'm bumping this back up as I actually have time to track down this issue now. For the past year I've just been dealing with the warm start issue. I know that if the car sits for more than a few minutes I'm going to have to give it some gas to keep it alive until I get the car moving.

Updates from the last time I posted in this thread:

1. RBob was not able to replicate the issue when I sent the EBL back to him.
2. Both the engine and the chassis ground wires are connected. A 0.1V drop is observed between the battery and the chassis. The same is true for the chassis to the engine.
3. Earlier this year I replaced the battery connectors with new 1/0 gauge cables. The new cables also included a 2 gauge chassis ground wire and a 2 gauge alternator wire.
4. Increased the maximum AFR on crank from 13.5 to 14.5 so that it no longer runs so rich initially when starting at higher engine temperatures. This did not help.
5. With the car running, I was not able to find any A/C voltage in the system with a multi-meter. I likely would need to figure out a way to start the car when it is warm to observe any voltage oscillation in the system.

Any ideas of where to go from here? My only thought is that since I can keep the car alive by giving it some gas, maybe I need to adjust the IAC so it gives more or less air? I'm grasping at straws at this point since the car runs perfectly otherwise.

RBob

iTrader: (1)

Get a data log of the hard start. Either post here (zip it first) or email to me.

Are you still getting incorrect sensor readings, or just a hard hot start?

As for depressing the pedal a little, can increase the IAC steps on a hot start.

RBob.

Dartht33bagger

I'll drive the car to work on Thursday and get a log of it so I have the most recent data.

Description how to make it happen and what occurs:

1. Drive the car and get it to operating temperature.
2. Let it sit for 15-30 minutes. If I shut the car off and immediately try to restart it, it usually fires right up with no issues.
3. Start the car.4. If the car dies, I can fire it right back up but need to input a little bit of throttle to keep it alive. Giving it enough throttle to run at about 1000rpm keeps it alive.
5. Shifting into reverse/drive it will almost die but pick itself back up.
6. Once I get it into drive and get it moving it runs fine. It idles and drives in open loop for 30 seconds or so before switching into closed loop.

I may try adding a few more IAC steps as well as a test on Thursday. I'll try to get a datalog of both what my current BIN is and with the updated IAC values.

MrMac98

A long shot but:
when your sensors read off ie -40, check that sensor plug for supply voltage +5 and a ground,
I think those sensors have the same supply voltage origins

could be some type off heat soak issue with electronics or even a connector plug.

Dartht33bagger

I got some really good examples of it tonight. The first two times I tried to start it up when warm it died as expected. The third time I gave it some gas and kept it alive but it was still running like total crap until I got it into closed loop.

Looking over the datalog, it looks like the IAT may be the issue. I see in both cases it is reading ~450 degrees C on startup. Even later in the datalog when I'm almost home and in closed loop I see it spiking up to 800 degrees C.

Could you confirm that this is likely the issue Rbob?

RBob

iTrader: (1)

I'm not sure where you are seeing those high IAT/CTS readings. Also, they can only go to 151 *C maximum. I also don't see any strange sensor readings.

I do see the engine start then stall (too lean?). The INT is being used to create cross-counts, so need to bump up the proportional gains a little is those areas of air flow. Once the INT is working correctly disable the EGR & CCP and do some VE Learns.

RBob.

Dartht33bagger

Sorry my mistake. I thought the O2 sensor mv column was lining up with IAT. Thats what I get for thinking I made a big breakthrough late at night.

I thought at crank the car would be in open loop. How would the INT come into effect here? How will I know when the INT is working correctly?

Should I be modifying PRP - Gain Multiplier vs Airflow? All of the PRP values are the stock values that come with the 3005 (vortec) EBL tune.

RBob

iTrader: (1)

Had a thought, some Windows installs don't have the proper fonts installed. I've seen some WUD screen grabs where the text is off a little. It may be in this case off a lot.

Here is what it should look like...

RBob.

Dartht33bagger

My laptop has WINE on a Linux install so it is very possible that the font is funky. I will take a look when I get home.

Do you have some resources I could read about the proportional gain and how I will know if it is correct or not? I read over the EBL documents this morning and I still don't really have a good grasp of how it will help my issue.

Edit: I looked at my laptop and the fonts are screwed up. The columns don't like up with the text. On my desktop everything lines up right. Now I'm searching through threads and documentation to try to figure out what I should set the PRP values to.

Dartht33bagger

I came home from work and spent about 2 hours reading about prop gain. From what I understand, proportional gain works with the INT to cause the INT to hit cross points. At idle and all driving ranges, the O2 sensor should be oscillating between the lean and the rich O2 R/L cross points to allow the INT to adjust. In my case it appeared that my cruising proportional gain was okay but my idle was way off to the point where the O2 sensor would peg rich or lean.

With that in mind, I spent an hour tonight driving around and trying different values. Initially I tried to modify the proportional gain vs airflow table for the 16 and 32 airflw entries. After maxing those values out to 100% without any changes, I switched to modifying PRP - Gain Multipler for Idle since my prop issues were at idle.

I ended up changing that value from 0.78% to 100% over several iterations. At 100% I was finally starting to get some O2 sensor oscillations. I couldn't remember exactly what the graph should look at that point since I was in a parking lot, but it looked better than when I started.

By this point my idle BLMs would creep up to 138. Whereas before my idle BLM was around 128. My idle also became a little rough at this point - oscillating up and down in RPMs. I decided this would be a good starting point at least to try some VE learns to see if I could richen up the fueling at idle. After 3 VE learns and a good amount of fuel being added, I didn't see much change at all. 138 was still the BLM I was seeing. I also wasn't noticing any easier warm start. There was still an oscillation before the RPMs settled down.

Looking at the examples of correct O2 oscillation in this thread, it looks like my oscillation is still too slow: https://www.thirdgen.org/forums/diy-...int-delay.html. Is that correct?

RBob

iTrader: (1)

When at idle two of the prop tables are substituted by two individual parameters. Also, be sure to get the latest XDF file (V1.2a):

http://dynamicefi.com/EBL_UpdateP4.php

The PRP thread you found and linked is decent.

The reason you had to max out the gain vs airflow is that the gain vs O2 error table values are low. The gain works by grabbing the gain vs O2 value and multiplying it by the gain vs airflow value. This can only act to reduce the gain.

So a gain of 4 times 100% is a fueling change of .061 milli-seconds (of injector PW). This isn't enough to move the AFR to create a cross count.

Increase the gain vs O2 a lot, say by x10. Can then work with the gain vs airflow to dial it in at various engine speeds.

Can also reduce the duration. As that can help prevent overshooting the gain required for a cross count.

When in idle the gain vs airflow table and the duration vs O2 Error tables are not used. These two values are used in their place:

PRP - Prop Gain Multiplier for Idle (Airflow)
PRP - Prop Gain Duration for Idle (O2 Error)

Note that the other two PRP tables are still used in conjunction with the two idle values.

RBob.

Dartht33bagger

I went out tonight for another 1.5 hours and did what you suggested. I multiplied all of the gain vs O2 Error values by 10. Then I drove around and decreased the airflow vs gain multipler values for the airflows that felt jerky. Idle was a similar story - I messed with that a lot until I got it a good feeling idle. After 8 or 9 iterations I felt like I got the car running pretty smoothly. There is a little room for improvement but I think it feels good and looks good from a cross counter perspective. The only part I'm not sure about is the INT perspective - it doesn't seem to change too much. Is that expected?

I've also noticed that with the proportional gain setup correctly (at least more correctly than it was) I'm now running lean pretty much across the board. In fact, I tried 2 VE learns with a smoothing factor of 10 on the way home and the idle BLM still wants to sit at 132 for some reason.

Why does the car run lean now? I would have expected the BLMs to stay the same as the proportional gain caused the O2 sensor to oscillate around it more.

Attached is the last two datalogs I took with VELearn running, a picture of BIN values, and a picture of my O2 value at idle and at throttle.

Street Lethal

iTrader: (16)

Do you have a copy of the bin handy that you can upload?

Also, be sure you are conducting your VE Learns in the temperature in which the engine will be used most. Many EBL users will do their VE Learns in a very wide range, which is the wrong way to do it, they will set it for say 150 degrees through 200 degrees, and the changes that are being made, say around 165 degrees are then Flashed in. But when the engine reaches 200 degrees after the Flash, the air/fuel will be way off from the desired 128 BLM, regardless of the O2 correction. Set your VE Learns for a 20 degree spectrum at most, then make sure your stoich setting is accurate in the bin, as well as your injector in correlation with engine size...

- Rob

Dartht33bagger

I am doing VELearns in a wide temperature range due to many stops and starts. I can narrow the temperature window. But I'm still seeing BLMs in the 138-142 range while driving which I not even close to before adjusting the proportional gain parameters. Before I was usually in the 120-130 range for BLMs when driving the same stretches of road.

I haven't touched the stoich settings from the stock value in the EBL P4 3005 bin. My injector offset values and flow rate have been set for the 21 lb Bosch III injectors (which requires the flow rate to be set at 19 lb since they are actually 19 lb injectors).

Looking at other examples of proportional gain threads, it looks like my O2 sensor values swing between .1V and .9V roughly, where other examples swing between 1.5V and 8V. Am I possibly overshooting my O2 sensor cross counts by too much?

Street Lethal

iTrader: (16)

I am halfway out the door and can't look at the bin until later tonight, but thank you for uploading it. Remember that GM purposely triggered Closed Loop to enable closer to 200 degrees to work with the stock programming. Example, my VE Learn was initially set to filter between 170 degrees through 200 degrees, this is to work in conjunction with my thermostat and fan settings of course, as that is where the engine stays and never deviates from. However, during warm up, my BLM will read in the low 130's below 170 degrees, then gradually reduce to 128 as it gets closer to its' target. It's important to look at your data at the targeted temperature range, otherwise your VE Learn will constantly correct due to the temperature change. Ten degrees makes a huge difference in terms of the denseness of the incoming air. Yes, try to narrow your Learn range down whenever possible...

- Rob

Street Lethal

iTrader: (16)

Bin looks essentially untouched. Never played with an L31-R, so be sure that the base timing, whatever the engine comes with, matches your Initial timing in the bin. As far as the datalog goes, I just watched run 11 and it looked pretty darn close. You never, or hardly ever, went over 2000-RPM, but despite that I did not see the BLM go over 140 once. The INT did, but that is irrelevant because it was just meant to drive the BLM down a tad. Seemed to have averaged low 130's with the target air/fuel in the low to mid 14's, which isn't that bad. Try increasing your acceleration enrichment some to even it out. As far as where the BLM is hovering at idle, I didn't see an idle clip so it is hard to comment on that, but the injectors you are running have been debated for years in terms of what they actually flow...

Dartht33bagger

Its true - I haven't really changed the bin much. I took the EBL P4 3005 bin, copied over the SA tables from the vortec bin, updated the injector flow rate and pulse width values, bumped up the idle speed, and did a few VE learns. For a year I've been driving on that with essentially no issues other than that I haven't been able to keep the car from stalling on a warm start without me giving it some throttle.

The datalog I had there at the end wasn't the greatest. I was about a mile from my house and really didn't get that long of driving stretch in. Earlier in the night when I was cruising at 50 in OD I was seeing the BLMs hover around 138-140. Lower speeds may be dialed in better. I'll go out for a long drive after work tonight to get a better datalog.

I struggled with the injectors for a while. Initially, I was using the Bosch III injector pulse width numbers I found in another thread and a 22#/hr flow rate. With VE learns I was getting close to 80/90 in the idle and low RPMs ranges to reach 128 BLMs. Switching to an injector flow rate of 19#/hr and resetting the VE tables fixed that. A VE learn or two later and I just drove it.

Street Lethal

iTrader: (16)

Double check with a timing light that your base timing by hand with the EST disconnected matches the Initial 6* BTDC in the bin. Switch back to the L98 (3005) SA Main Table temporarily as well, as after comparing it with the SA tables from the Vortec bin you might like the L98 SA Main Table better, along with a quick VE Learn. Give it a try. If you don't like it you can always switch back...

Dartht33bagger

I had a chance to drive the car to work on Friday. Other than a lean idle that I manually fixed in Tunerpro, the BLMs were actually pretty good. I guess it was just that one night I was seeing very lean numbers.

Unfortunately I haven't had a chance to actually re-test the warm start condition since I drove straight to work and then let it sit for 8 hours before restarting it. I had planned to go to the gym so that I could test it after it sat there for 45 minutes but meeting got in the way of that.

On a positive note I ordered a LC-2 wideband O2 sensor off of Amazon last week for 20% off. With luck I'll have that installed in two weekends so that I can do some open loop/PE tuning and confirm that my AFR is correct on my current narrowband setup.

Dartht33bagger

I've been driving the car a few times a week and I thought I finally had the problem solved. All thoughts of that were crushed on Sunday afternoon when I attempted to start the car when warm and it stalled out. Looks like I'm back to square 1.

According to this thread I may have a false lean condition happening causing the stall: https://www.thirdgen.org/forums/diy-...ome-those.html. The only thing that doesn't match with that thread is my cam setup. I am running the stock L31 cam in my car and it only has about 28 degrees of overlap.

RBob

iTrader: (1)

It may be the MAT heat soaking. Try setting the CTS/IAT Blend table to all 100% CTS. See what kind of difference that makes.

RBob.

Dartht33bagger

I will try that this weekend. Wouldn't this mean that air temperature would no longer affect the AFR in closed loop though?

dfarr67

iTrader: (2)

I really wish these threads would tie off with a fix.