ptuomov

I think the different cam timing for the two banks when the engine is cold has one and only one rationale. The aluminum block expands when the engine heats up, which impacts the two banks by an unequal amount. I don't believe the belt stretch has anything to do with it, so I wouldn't time the cams any differently with the racing belt.

Some ignorance-exposing questions about your system below:

If you smooth the MAF signal, at what stage do you smooth it? When it's still mass air flow per hour or when it's mass air flow per cycle? My first instinct is that smoothing it in the units of mass air flow per cycle would give a more stable true local mean and would reduce any adverse effects of smoothing. Are you concerned about additional smoothing making the system slower to respond and missing the goal at transitions?

You talk about the sport mode. I assume that this is when your system goes to the WOT fuel map. How do you trigger that, the same way as the stock LH triggers the WOT map? Is it just TPS to WOT, or is there also some load and/or rpm trigger? And what triggers the transition away from this mode?

I don't think one needs to be concerned about knock events at low load. Knock can only be damaging when it packs some punch. If the cylinder isn't filling and there's a knock, which causes the knock sensor activated system to pull timing, this should be completely harmless. It's only at high loads where knocks are very damaging.

How do you implement the octane sensing adaptation? How did you do this? It's as if you're blending two ignition maps based on the historical knock counts? Does all this code live in the EZK?

Hilton

LOL, we better hope the volvo guys don't all club together and buy you a free volvo with LH2.4

Interesting - so you're updating the ignition map dynamically based on knock events? Does that mean that it'll compensate for the high-compression piston variants too (89-91 S4/GT).

90% injector duty cycle with 24lb injectors? Crikey!

PorKen

I have kept the bank to bank difference to compensate for block expansion. I retarded both cams a touch because the reading for both banks is the same at #6 and #1 with a 'racing' belt, when measured with a PK32V'r. With a 'regular' belt, the reading is different because of valve spring pressure(?).

After the MAF voltage is converted to a frequency, then collected/counted by the processor. Slight variations in MAF voltage and/or problems in the A/D conversion create spikes rich or lean in what is used to calculate fueling (and EZ load). Past 5400 rpm is often completely random.

Lean spikes are knock magnets. These seriously limit the amount of ignition advance that can be run consistently at peak HP, 5800~6100 rpm.

For cruise, I use the factory averaging code, but changed it to average all the time (old + new / 2). For WOT, I made new code which follows the trend of the A/D value, not the actual (bad) data. The data has to trend plus or minus a number of times before the MAF variable is incremented or decremented once. Mostly, the variable can't keep up with the spikes, and has the 'true' value. The MAF variable is then averaged with the base map(s) + WOT map value as normal.
I was worried initially that it might be too slow, but it logs just fine. WOT fueling doesn't change very quickly, the code is cycled hundreds of times a second, so it has no trouble keeping up. Going WOT downhill is a bit rich for a time, because plus is coded to be favored over minus.

More info here - https://rennlist.com/forums/928-foru...maf-input.html


Stock, the LH goes open loop over 3500 rpm or a MAF value over 280(?) units (= somewhat heavy throttle). It then uses whatever is in the base maps (maf ideal + no/cat addition) averaged with the MAF value as normal, without any O2 correction, but still around >14.7 AFR.

For Sport Mode, I lowered the thresholds, still use the base map, but change the injector constant so the AFR is down in the mid 13's. It can only leave this mode when the MAF units are back to around a highway cruising level, RPM is ignored. Hitting the WOT switch will also trigger this mode, so if it is on the edge of the switch point, it won't drop directly into the O2 loop, which will often trigger a knock.

AFAIK, it will make more torque at lower rpms with the throttle closed somewhat. Maybe that's why, with proper fueling and more ignition advance, it feels like the engine has a bunch more cubes in Sport Mode.

I am retarding 0.4° at a time if there are knock events (within a time period counts as one event). The regular EZ retard procedure still handles the knock(s) as normal.
When the engine is started the maps are initially advanced to where cold weather or 93 octane would allow. After a few (unheard) knocks, the global level ends up being just below the knock threshold, which as a side benefit, is roughly where it will make the most power.
It works surprisingly well for its simplicity (of the concept). I figure it should handle 89-93 octane (AKI) and piston differences just fine. (From previous testing I have found that timing needs to be changed roughly 1° per octane point at WOT.)
Next fillup I will be putting in 89 to test for sure.

GT cams would require new fuel and ignition maps.


Yes. Lots of room on the EZK EPROM, BTW.

It was originally tightly written to fit on a 8K chip, but then it was ported to a 16K for additional diagnostic code.


Dual rate, ie two openings per cycle. Pretty much the same as my S3 which peaked at 321 rwhp @ ~12.8 AFR, uncorrected.

Single rate is only good for a 3% or so reduction in duty, I didn't want the transition, I wanted the codespace, so I deleted single rate.

BC

When available Ken? I can put 24s on this white 88 I am putting back together. Just have to put your tensioner on and it will be running.

Hilton

It does strike me that the GTS owners could really use an eprom with your "MAF smooothing", "octane adaptation" and "sport mode" functions coded. Should reduce the knock issues a lot with the richer AFR under throttle and retard with knock events (i.e. oil ingestion).

I think you should get Rob Edwards to destruction-test your chips in another S4 motor at the track

PorKen

002 stock
006 X-pipe, 24#, stock EZ

020-023 new chip (two runs, cool down 5 mins, two runs)

It could do these same numbers all day...consistency, yay!

STD 319/330

MjRocket

Impressive! ..Where do I sign.

PorKen

S3 oder S4? Die 5-liter 350 PS macht.

MjRocket

S4 Auto

DKWalser

Ken - That's about a 13% increase in HP and torque from the pure stock output. Very impressive.

Question: Were runs 020-023 with or without the X-pipe? (I assume with; I'm just trying to confirm.)

PorKen

020-023

S4 RMB
SLC X-pipe
big resonators
no cats/airpump
PKT w/racing belt
Ford 24# metal body
NGK BP7ES (1.0mm gap)

stock airbox/air filter/tubes
stock crankcase venting

M1 15W50
Redline Shockproof
Chevron 92 octane (E10)
18" Cayman S, 255/40 Conti DWS (40psi)

simos

Hmm, 24" injectors, 350PS and 90% injector duty cycle, is there something wrong with these numbers or do I misunderstand something?
I have got an impression of little over 400PS with 24" injectors on S4

13% increase on NA engine is indeed impressive number.
Ken, you can add "no carbon build up after hours of WOT tuning" to your list

PorKen

With a S4 FPR, 24# act as roughly 27#, which supports about 350 CHP at 80% duty, according to the internets.

It depends on the AFR, too. I probably could have squeaked out a few more HP by leaning out the top end, but I was so happy to make >310, and that it was consistent, I didn't make any changes. An AFR close to, or over 13 would have put the injector duty into the 80's at 6000.

Factory (minimum) specs are 316 CHP @ 6000, 317 CTQ at 3000. There are places to get E0 around here. It would be cool to make 316 RWHP.

It's great to know I can make tuning changes and have a real idea if they work! (Instead of trying to divine the meaning of random spikes.)


021 - AFR, MAF volts, inj duty, knock, RPM

BC

Those valleys in the knock volt are knocks?

Hilton

Ken - question for you on behalf of the non-cat-S4 owners around the world

First, a likely redundant question, but just in case.. Will your chips work without closed-loop operation? By "work" I mean will AFRs at cruise, along with the sport mode & octane compensation all be fine without any of the O2 Adjust parameters etc?

Assuming you've tuned for open loop, my other questions are:

Which exact injectors are you using? Are they the Ford Racing M-9593-A302 (aka F1TE-D5A, 0 280 150 947)? Injector opening time and cc/min varies a fair bit even across different disc-style 24lb-rated Bosch injectors, so having behaviour be exactly like your tuning setup would seem the most reliable way of making sure AFR's are as expected, in the absence of an O2 loop like you 'muricans have. My main concern here is I live in a hot climate, compared to the PacNW

If my current Ford Racing 24lb ones on the shelf aren't the right ones, I wanna start watching fleabay for a cheap set of 24lb injectors for my '87

What AFR should hot idle be set to using the mixture-pot, to provide a baseline for MAF-aging compensation/calibration for mixture adjustment? (this one is more just a prompt to stick this info in the "manual" - don't want/need an answer right now)