A few queries for those who do their own Shark-tuning.

Whilst recently working on re-tuning my 928 I experienced a number of anomalies that I have not come across before. Our fuel here is officially rated at 95 octane but I have some suspicions as to whether the fuel is all it should be.

My 928 GTS chassis is powered by the motor from my late 90 S4. We have been inside the motor and although we did not disturb the pistons, the main dealers felt confident there were no signs of distress with respect to the pistons/rings etc. We replaced a couple of valves and all the valves were lapped.

I have an x-pipe, and a single bullet muffler in the mid section plus the RMB. The cams are GTS inlet/S4 outlet. Everything else stock. The engine demonstrates very even compression at around 185 psig.

When I went back to the stock set up I found the top end was running very rich like in the low 11's. I decided to run some new injectors and went for a set that was supposed to be 30lb but in fact is 28lb according to the research I did on the units sent. The top end/full load is now running nominally 13 to 1 and I have the fuel pretty much where I want it but I have had to pull the timing back a bit from that on the stock 98 maps. Not exactly what I was hoping for.

Thoughts on the following appreciated:

1. What AFR's do you find best suit the n/a setup? I understand top end delivers max power at 13.5 to 1. Thus when using sharkplotter I try to make sure top end cells show a value such that when considering the lean end of the std deviation range I do not go leaner than 13.5 [i.e. 13.3 with a SD of 0.2].
2. For mid range /full load I have tuned for the same AFR but maybe this is not optimal- any thoughts on this?
3. I experienced a big knock sequence at 6k rpm so pulled the timing back 3 degrees on a couple of cells- this seems to have calmed things down. When I initially ST'd this motor I was able to add a couple of degrees at top end, not 3 back. Short of fuel issues is there anything else that could cause the knock control system to want to retard the timing? I have green lights for the sensors and Hall trigger. The only differences before and now are the air filtration system and the plenum cover extensions. I also run on a 20w50 Shell Helix mineral oil as I cannot get the Redline stuff I used to run. Typically knock initiates on cylinder 6 loosely followed by cylinder 2. There then seems to be a noticeable gap [5 or 6 degrees] before other cylinders start to show signs of knock. Is this a typical characteristic or does it vary example to example?
4. Once you have the fuel sorted, what levels of advance are you able to dial in at different parts of the rpm scale?
5. When knock events occur it seems the knock retard is sometimes instantaneous returning to normal immediately [or so it seems] and on other occasions it seems to return in steps that are not consistent with the values selected. Is this normal?

Regards

 

I personally set my vehicles up so that the higher rpms and higher loads are between 12.5 and 13.0. I think that 13.5 is a bit too lean, for the small amount of gain to be had.

I also set the "mid range" full throttle about the same.

Until you put some sort of oil separator device onto the engine, you will be ingesting oil into the intake system, at higher rpms. This oil reduces the effective octane of the fuel and will require you to "remove" ignition timing.

I've done a considerable amount of research into this and have a "system" that does this, but also improves the crankcase breathing and thus the oil return from the cylinder heads, to help the engine "live" at high rpms. There's been lots of discussion, on this Forum, regarding these problems and solutions.

Others offer simpler (and cheaper) oil separation systems, BTW. These systems satisfy the needs of some people and may be adequate for your needs. It all comes down to how "hard" you want to run your engine....and for how long....

Keep in mind that a few knocks are desired. The best running cars will have knocks and allow the igniton system to do it's job and "remove" some timing.

 

Good reply!

Ive got nothing to add..other than working with the ST I find Im always learning something.

 

I also shoot for ~12.5:1 at high RPM. My ideal AFR spread would start at ~13:1 around 2,000 RPM and then smoothly go down to 12.5:1 at around 6,000 RPM. This is with boost, but I would do the same for N/A.

Dan
'91 928GT S/C 475hp/460lb.ft

 

Peak power is between 13.1:1 and 12.8:1

If hot weather shoot for 12.8:1 if cool climate shoot for 13.1:1
If you track car 12.8:1 nothing leaner!!!

Adjusting the mid range will do NOTHING if you have it in cat mode as it will always pull back to 14.7:1.

Cruise range set adv at 36 deg from 1400rpm up
WOT run as much adv as you can with a couple knocks.

 

Don't you want to keep the turbines hot? ;-)

 

Excellent answer from Greg...

Colin - even from moderate loads at mid range the LH goes open loop, so tuning AFR is an option...

 

...which is what the target AFR map is showing as well.

 

John,

What is the cutoff parameters for open loop?

Last we spoke regarding this topic I asked if we could only have close loop with idle switch. You had stated it wasn't possible.

 

I think what I meant was it is possible but we didn't have the time to do it.... also it might be dangerous for cats to be wholely dependant on accurate tuning at mid range cruise instead of the security of the O2loop.

Parameters for the change from closed to open loop are complicated. The rpm and load that triggers the change also depend on the rate of change of the MAF (load) signal.

 

With LH2.2, the O2 loop bypass is controlled by a 16 column lookup table, with the same rpm divisions as the base map.
If the MAF row# calculated for the base table in a particular rpm column is high enough, then flag for bypass.
The threshold changes in the different LH EPROM versions from '85 to '86.5.


On LH2.2 also, use of the WOT map addition can only be triggered by the WOT switch. Not by load/rpm.

 

Colin,

Depending only on the WOT switch for open-loop would be foolish, in my opinion.
Two reasons: First, WOT switches are not all that reliable: the idle switch is a proper microswitch, WOT is just a couple of copper contacts. But the other issue is that even 75% throttle (about where the WOT switch closes) is heaps of load especially at mid-RPM's, and you do not want the LH diddling AFR's up and down trying to average 14.7.

My take on it is that the LH does a pretty good job deciding when to go open loop, perhaps early but that's OK with me. The important thing (assuming cat's are fitted) is that it runs closed-loop in the normal "cruising" loads and RPM's, to keep the cats happy.

Fred, I am confused about your injectors. What were they before you changed? Larger than then new ones, or smaller? Did you change ST's injector-size setting?

Are you using the WOT fuel map, or is it zero'ed? If ST's injector-size and opening-time settings are set appropriately, then fueling should be fine across the map without the additional fuel.

 

Jim,

First of all thanks to everyone for sharing their thoughts to date- much appreciated.

My injectors were stock 19lb units. They were cleaned/flow checked etc and were OK flow wise with barely discernable variations. Having read some perceptions of performance improvement with 4 hole injectors and some preferring 24 or 30 lb units I took a look on e-bay to see what was on offer and "5-O" had a set of what they called brand new balanced set of 30lb injectors, 4 hole, design 2 [metal bodies] for $280. I figured what the heck and purchased the set complete with appropriate O rings.

For once I did a "before" benchmark and this was running way too rich on top end so I decided to do a completely new set up from scratch. When you have to remove the injectors you are the best part of the way to having the manifold off so I pulled the thing to give the system components a good check out and then put everything back including the new injectors and the stock FPR so that I had a meaningful benchmark for comparison with others.

I used the stock maps for starting and then set the injector rating to 30lb. Before this I looked up one of those injector chart lists on the internet and this suggested the injectors I received were in fact 28 lb units. when I fired up the motor things seemed OK at first but after an initial cruise to collect some data it seemed the motor was running quite lean. A quick calc suggested a 28lb rating was more likely correct so I reset the rating using the custom size feature and set it to 28 lbs. A few test runs later and the whole plot seemed much more credible albeit still running somewhat rich on top end so I set the AFR's to what I feel is appropriate.

Now the fun bit. If you remember the initial dialogue I had with Louie, John and yourself when I purchased ST2, I said at the time the throttle switch was not working and indeed I believed this to be the case. However, following suggestions from the list about the correct butterly opening, I cross checked this as I was replacing the manifold and found it was not fully opening. I also checked the throttle switch whilst the manifold was out and found it was a bit hit and miss. Checked my spare one and that did not work at all. No problem, put the lot back and reverted to my original strategy of zeroing the WOT map just in case it did kick in and ticked the box telling the EZ to ignore the full throttle EZ map. It also dawned on me that if you want to see the larger bubbles in sharkplotter it would be a good idea to data log throttle position!

Thus I am running open loop using the cruise maps only and I work on the principal that you can cover the full throttle range using two vertical load cells for each rpm band and at lighter loads you lean the mix to average 14.7 for the typical cruise band cells.

The motor feels to be running nicely, I pick up a few minor knocks in the 3700/4000 rpm cells and am ready to start playing around with the ignition timing. The down side is that at 6k rpm I had this significant knock issue that concerns me in that it went full scale on two cylinders. Never seen that before. I am assuming that the cylinders that knock do so because they are flowing more air thus run leaner but how much more will cause this? I also wondered if I had backed off a tad instantaneously resulting in transient leaning- the MAF load values at the point of knock initiation suggested this may have been the case [6k rpm in third is moving along quite nicely]. I have other runs that have gone through this point without knock so this makes me wonder.

I have been concerned for a long time about what Dr Brown says about oil lifting at high rpms. Indeed I started to project to build an external oil separator/improve breathing but never finished it off. Maybe the impact of this with the mineral oil I am running is more than it was with the synthetic Redline stuff that I prefer. Maybe this makes a difference? I did convert my S4 motor's breathing system to the GTS system utilising the GTS filler neck. The breather in the S4 filler neck is pathetically small. The GTS internal diameter is better but is still too small for effective breathing and that before you consider the oil slosh problem.

Based on what this thread has produced I will try to adjust AFR to give nominal 13 to 1 for low to mid revs and 12.8 for top end. Then I will start to increase the advance. Given I am at minus 3 degrees [relative to stock map advance] at 6k rpm I doubt there is anything more I can do up there so will have to live with that I guess. To be fair, on the auto box anything above 6k rpm seems a waste anyway given it wants to change gear at about 6.2k with the pedal mashed and it will not get up there in 4th [top] gear unless you drive over a cliff.

No sure I understand Colin's comment about setting the advance to 36 degrees for lower rpms- I presume the suggestion is to set all high load cell values at 36 degrees and back off until knock is at a usable level [I consider up to 3 degrees of knock retard acceptable.

whereas I do pay some attention to cells below 2k rpm I consider them relatively unimportant as long as they are roughly correct given the torque converter operation. Perhaps I need to rethink that a bit?

Many thanks

Fred