JekyllandHyde

iTrader: (3)

I was wondering if somebody knows of a link or article where limp home mode is described in detail? such as what exactly happens and what is affected? I read Grumpy's article on Limp Home Final Answer, but it did not really have all that much technical info.

The reason why I ask this is because I've been trying to nail down a surging idle in my car for months now. It's an L98 short block with Trickflow Heads and a TPIS ZZ409 cam running 36lb injectors. I've tried pretty much every trick I could find on getting a stable idle and while I can reduce the surge to 100rpm, it will still do it when idling for an extended period of time and eventually, the idle would just degrade and the surging would go up and down up to 700rpms until it finally got itself sorted out. I have checked multiple times for vacuum leaks and am going to check again.

Which leads me to my second question. The other day, I had the car out, and was datalogging while emulating my latest bin, when my laptop shut down. All of a sudden, the idle stabilized at 800rpm and the car actually drove quite decently for a block or two before I noticed my check engine light was on and my laptop was shut off. (So much for thinking I had nailed the tune.) The idle was rock steady at 800. Now I'm thinking, if the car did have a vacuum leak that was affecting it in normal run mode, it should show up in limp home mode as well right? Surging idle and all?

RBob

iTrader: (1)

Odds are need to work on the proportional gains. They are responsible for making the AFR oscillate while in closed loop. Between the larger injectors and the O2 moved back (headers), the timing/gain/duration is now off.

First is to reduce the gains by the ratio of 22 #/hr divided by 36#/hr (injectors). That will bring the volume of fuel back in line.

Reducing the duration and possibly increasing the INT delay before update will take care of the O2 sensor being further from the engine.

The other piece of the puzzle is getting the proper injector compensation values in place.

RBob.

JekyllandHyde

iTrader: (3)

Thanks Rbob. I've actually tried messing with the proportional gains as per this article:

https://www.thirdgen.org/forums/diy-...ues-lucas.html

I multiplied the amount I got from the formula to the table located at L84DE. By injector compensation values, you mean the injecter offset vs. voltage right? I was able to enter in some values for that from a data sheet that I obtained for the injectors I am running. Had to do a little bit of creative math on MS Excel to interpolate and extrapolate values for the wierd GM voltage values. I have yet to try increasing the INT delay. It does seem to idle better in closed loop than in open loop though. I'm going to check vacuum and exhaust leaks again because the BLMs rise to 160 in closed loop idle and just stay there... though a fouled O2 sensor would do the same thing as well, right? (when I first did injectors, BLM was pegged at 108 and plugs fouled. VE values in that idle area were leaned 10%)

But why is it that the car idles very stable in limp home mode and not in closed loop? with all the mods I made, I would think it would be the opposite way? on limp home mode, the car drove with much better manners than it ever did, though power was way down.

RBob

iTrader: (1)

Ah, now if IIRC, you posted in the injector compensation thread about how to convert the Ford specs. Good deal on getting those set up, it makes a difference.

I have an idea that the real difference between limp and closed loop is the SA. In limp mode the SA isn't going to move at idle.

It will also run richer as it is calibrated for the stock injectors.

There are idle compensation/stabilizer SA tables. May need to work with those. They add/subtract SA to keep the idle RPM steady (or try too).

> exhaust leaks again because the BLMs rise to 160 in closed
> loop idle and just stay there...

That is a symptom of the O2 sensor cooling off. If heated make sure that the heater is powered all of the time the key is on or the engine is running.

It is also a symptom of cam overlap. The charge comes into the chamber and out the exhaust valve. NB O2 sensors are funny about this. Most read that as lean while others seem to read that as rich.

Back to prop gains, check the log and see if the INT is oscillating. There are two ways it will do this. First is that it is slowly rising and falling. In this case there isn't enough proportional gain/duration.

If the INT is rapidly oscillating that is usually a sign of too much proportional gain/duration.

RBob.

JekyllandHyde

iTrader: (3)

Thanks RBob. Yes, I was the one posting on the Ford specs. I just checked for exhaust leaks again (started car when cold, felt around the collector/y-pipe area), found nothing. I AM running an Autozone Bosch single wire O2 sensor. I have a heated sensor and the harness waiting to go in. Just to experiment, I tried idling in open loop, still had surging idle, (IAC counts were zero, which is why I suspect vacuum leak somewhere) but I noticed that the o2 sensor mvolts were under 200 perhaps 90% of the time with no cross counts occurring. I was able to get the car to idle OK in open loop but she still had the same problem of surging 100 rpm if idling too long and the surge would grow until it cleared itself out. MAP readings showed 45-50 Kpa at idle when it was stable, naturally when it started surging, Kpa decreased and increased correspondingly. I am going to install the heated one and see if it improves CL idle.

I basically zeroed out all my idle SA compensation tables. I will try 1-2 degrees on the tables and see if it helps.

As for the O2 constants, I have them set to 350/350/300, the cam calculates to 63* of overlap. The stock AUJP settings were 600/600/550. My friend gp90gta runs the same O2 constants and he has a cam with slightly higher overlap.

Thanks again for the suggestions. I will check out the INT on my datalogs and see what's going on.

JekyllandHyde

iTrader: (3)

I got a chance to fool around with the tune a little bit more. My friend gp90gta has been helping me out too. Heated O2 sensor is in and it cured the BLM pegged at 160 problem, however, this pig still likes to surge at idle. Prop Gains were pretty much cut in half at L804DE. It seems to help with the idle a little but the INT is still oscillating. Looking at the graph of my data log, it looks like the INT peaks when the O2 volts are swinging up and then bottoms out as the o2 mv approaches the bottom value. I read the fueling logic article for the 747 ECM and I think I follow. If the Integrator does not follow the proportional gains then it is mostly governed by the O2 signal, correct? I was interested in the INT delay that is talked about in the article, and RBob, you also mentioned that I may have to increase the INT delay as well. I have the O2 sensor in the collector of the Dyno Don headers. In the 8D mask, the only values I could find that vaguely resemble INT delay are:

"Loop closed Param, Min Error to make INT correction" at L8049E
"Loop closed Param, Add Correction INT Delay at idle" at L8049A
(This is a scalar zeroed out)
"Loop closed Param, INT Delay Multiplier vs Error" at L80504

Am I on the right track? I've also attached a copy of my data log. in this log, my O2 constants are set to:

Error Upper - 500 mVolts
Threshold fast - 500 mVolts
Error Lower - 450 mVolts

Essentially, 100 mVolts lower than the stock AUJP O2 constants. We raised the O2 constants, because it looked like the O2 sensor wasn't swinging at all and my cross counts were pretty non-existent.

Here's a copy of my data log if anybody would like to check it out.