'88 5-speed dyno log
#91
For my S4 chips, since I'm not worried about the coding plug, I am going to try to use two, maybe all three, ignition/fuel maps for extra special resolution.
(92 cat, 92 non, 89 cat / MAF ideal, cat, non).
There will have to be some overlap where they switch, but it should allow for virtual 32 x 32, or better, maps.
Using the 35-pin bridgit I made, I've been able to find most of the basic I/O ports in the code.
Mainly, I was able to figure out the in-gear input (29.2) for autos.
(It threw me for a bit, because it is wired to ground in the manual wire harness.)
I am going to connect this input to the clutch switch for No-Lift-Shift, maybe also a downshift rev...Launch Control?
Not sure what to do with it yet, except for a test light, but the unwired pin 26 (P1.7) can be switched (12V) in the code.
(Pin 26 would have been a CASIS shift light, but there's no code for it.)
Also, I didn't realize it before, but the LH is not coded for auto or manual.
(Coded for cat, code plug 6. There are two others, 2 and 4, but all model code plugs wire these to ground.)
Because of the torque converter, autos have an effectively different gear ratio.
Ideally manuals and autos would have their own fuel mapping.
(92 cat, 92 non, 89 cat / MAF ideal, cat, non).
There will have to be some overlap where they switch, but it should allow for virtual 32 x 32, or better, maps.
Using the 35-pin bridgit I made, I've been able to find most of the basic I/O ports in the code.
Mainly, I was able to figure out the in-gear input (29.2) for autos.
(It threw me for a bit, because it is wired to ground in the manual wire harness.)
I am going to connect this input to the clutch switch for No-Lift-Shift, maybe also a downshift rev...Launch Control?
Not sure what to do with it yet, except for a test light, but the unwired pin 26 (P1.7) can be switched (12V) in the code.
(Pin 26 would have been a CASIS shift light, but there's no code for it.)
Also, I didn't realize it before, but the LH is not coded for auto or manual.
(Coded for cat, code plug 6. There are two others, 2 and 4, but all model code plugs wire these to ground.)
Because of the torque converter, autos have an effectively different gear ratio.
Ideally manuals and autos would have their own fuel mapping.
#92
Hmmm! When I think about Ken's great work with the 86.5 and how it seemingly blitzes the S4, why did Porsche introduce the S4 instead of working the S3 with S4 sensors- what am I missing here?
If Ken can get the S4 to where his S3 is [like for like x-pipes etc] that will be quite some achievement in itself.
Regards
Fred
If Ken can get the S4 to where his S3 is [like for like x-pipes etc] that will be quite some achievement in itself.
Regards
Fred
#93
Getting there.
After weeks of decoding and days of cutting and pasting rows/cols...
Now a consistently flat-ish AFR line up to 6500 with no WOT map addition after 2400.
Tuning blog/FYI...
Those two rich afr (purple) drops w/o a corresponding injector duty (red) spike are knock tell-tales.
(You can see them in the logs, above, too.)
IIUIC, after 4900 there is factory warmup/cranking code which messes with the WOT fueling.
AFAICT, it's using straight MAF/RPM/coolant calculation at some point past that RPM. Not using the base map(s) much/if at all.
This code might have something to do with the fact that factory WOT is not properly tuneable (see scaling below).
After bypassing that code, the base map appears to be read and used properly all the way to the map limit.
(It is factory coded to bypass it by setting the RPM threshold constant to max/FF.)
Still working on that mountain range around the HP peak.
Sometimes it's kind of like whack-a-mole, moving the RPM cols and/or MAF rows to fix a AFR peak or valley.
Fix one, another opens where you moved it from. At some point, I just have to say FI.
I will probably add a little to the WOT map, 2400-up, to bring it to just under 13:1.
(A little WOT map addition will sometimes smooth the injector spikes, too.)
Scaling
I could not believe my lying eyes when I grokked the MAF scaling.
Stock S4-GTS LH chips only go to 324 MAF units! That's 3100 rpm at WOT!
Lazy engineers! It's the same range used on baby SAAB and Volvo engines.
It's as if the factory only cared about tuning for part throttle operation/emissions.
In contrast, S3 scaling goes to 648! Way too high for what a S3 can flow. Half the map is unused.
The ST has more headroom, but there is no resolution to be able to tune anywhere in the peak power band?
I chopped out some of the useless part throttle rows/cols and made new ones for heavy/WOT between 2800-6600, instead.
Let the LH interpolate (guesstimate between rows/cols) in the lower power regions (which are controlled by the O2 loop anyway).
MAF scaling starts with the highest (16 bit) number, then subtracting in increments of 8,16,24,32,48,96.
RPM scaling starts with the lowest (8 bit) number, then adding by any number.
RPM to number relationship is not linear, more like a NTC curve.
MAF units don't correspond to g/sec in a line, either.
Approx. S4 manual MAF unit range: idle 50 <-> 415 WOT
stock
020 036 052 068 084 100 116 132 164 180 196 212 228 244 260 ___ 324
ST
___ 036 052 068 084 100 116 132 164 180 196 212 228 ___ 260 ___ 324 ___ ___ ___ ___ ___ 420 ___ 516
PK
___ ___ 052 068 084 100 116 132 164 ___ ___ 212 ___ ___ 260 292 ___ 340 372 396 404 412 ___ 428
3rd gear, manual S4, on road
Now a consistently flat-ish AFR line up to 6500 with no WOT map addition after 2400.
Tuning blog/FYI...
Those two rich afr (purple) drops w/o a corresponding injector duty (red) spike are knock tell-tales.
(You can see them in the logs, above, too.)
IIUIC, after 4900 there is factory warmup/cranking code which messes with the WOT fueling.
AFAICT, it's using straight MAF/RPM/coolant calculation at some point past that RPM. Not using the base map(s) much/if at all.
This code might have something to do with the fact that factory WOT is not properly tuneable (see scaling below).
After bypassing that code, the base map appears to be read and used properly all the way to the map limit.
(It is factory coded to bypass it by setting the RPM threshold constant to max/FF.)
Still working on that mountain range around the HP peak.
Sometimes it's kind of like whack-a-mole, moving the RPM cols and/or MAF rows to fix a AFR peak or valley.
Fix one, another opens where you moved it from. At some point, I just have to say FI.
I will probably add a little to the WOT map, 2400-up, to bring it to just under 13:1.
(A little WOT map addition will sometimes smooth the injector spikes, too.)
Scaling
I could not believe my lying eyes when I grokked the MAF scaling.
Stock S4-GTS LH chips only go to 324 MAF units! That's 3100 rpm at WOT!
Lazy engineers! It's the same range used on baby SAAB and Volvo engines.
It's as if the factory only cared about tuning for part throttle operation/emissions.
In contrast, S3 scaling goes to 648! Way too high for what a S3 can flow. Half the map is unused.
The ST has more headroom, but there is no resolution to be able to tune anywhere in the peak power band?
I chopped out some of the useless part throttle rows/cols and made new ones for heavy/WOT between 2800-6600, instead.
Let the LH interpolate (guesstimate between rows/cols) in the lower power regions (which are controlled by the O2 loop anyway).
MAF scaling starts with the highest (16 bit) number, then subtracting in increments of 8,16,24,32,48,96.
RPM scaling starts with the lowest (8 bit) number, then adding by any number.
RPM to number relationship is not linear, more like a NTC curve.
MAF units don't correspond to g/sec in a line, either.
Approx. S4 manual MAF unit range: idle 50 <-> 415 WOT
stock
020 036 052 068 084 100 116 132 164 180 196 212 228 244 260 ___ 324
ST
___ 036 052 068 084 100 116 132 164 180 196 212 228 ___ 260 ___ 324 ___ ___ ___ ___ ___ 420 ___ 516
PK
___ ___ 052 068 084 100 116 132 164 ___ ___ 212 ___ ___ 260 292 ___ 340 372 396 404 412 ___ 428
3rd gear, manual S4, on road
Last edited by PorKen; 04-12-2013 at 03:43 PM.
#95
Very nice info Ken - that helps explain the massively inconsistent dyno results on stock manual S4's (same car, same dyno, same day, up to 20hp difference in pulls depending on how warm the car is.. interestingly it took a couple of pulls to get the car warm enough to read nice high hp even at 25C ambient).
I know what you mean, but its worth pointing out that here in Oz, we have a different, more 928-appropriate, version
I know what you mean, but its worth pointing out that here in Oz, we have a different, more 928-appropriate, version
#96
This thought had crossed my mind - Porken, with the addition of a custom chip, will their be any safeguards on it when the MAF begins to fail, or age, significantly affecting the A/F ratios?
#97
What gear you choose also makes a difference, off-map. If you tune in 3rd, for example, 1st and 2nd will be progressively richer, 4th and 5th, leaner.
When on-map, if the rows/cols are close enough and adjusted properly, the AFR is the same, even though the MAF/RPM ratio changes with each gear.
Not any worse than the stock chip. Probably better, in fact, if it's done properly, as the stock chip is quite lean in the lower rpms at full throttle, and pretty much random at high rpms. Mostly stupid rich.
Interestingly, engines will knock when it's too rich, as well as too lean.
Worst case, the knock sensors will always be there to limit knocking.
When on-map, if the rows/cols are close enough and adjusted properly, the AFR is the same, even though the MAF/RPM ratio changes with each gear.
Interestingly, engines will knock when it's too rich, as well as too lean.
Worst case, the knock sensors will always be there to limit knocking.
#98
When it comes to smog here in CA, we are pretty strict. Do you think your chips will still keep us in the zone of the sniffer? I barely pass right now, but I think it's because I have high flow cats and a lean misfire I've been chasing down forever.
#99
After fighting with the area between 5600-6000 (peak HP) for far too long, I busted out the code to make a second MAF ideal map and use the non-cat addition map together for WOT fueling only. I haven't finalized it yet, but it will be scaled somewhere between 3000 and 6600 rpm, and 260-420 MAF units.
This is good as now I can keep the part throttle maps essentially stock, for emissions and response, and have a whole 'nuther 16 x 16 map just for full throttle (>75% pedal) precision!
I'm running my old tune in the WOT section for now, but I'll get to moving all the rows and cols around here in a bit.
I can't wait. I feel drunk with power. Muhahaha.
So nice to have so much memory to use. The 16 bit MAF ideal and it's scaling maps are ~600 bytes, which doesn't sound like much, but S3 LH2.2 has only 4000 byte EPROMs and they are completely filled with code/data.
S4 LH2.3 have 16000 bytes to work with, but code is a bit scattered around. Fortunately, there was a 700 plus chunk of empty bytes at the end.
(LH2.3 actually uses 256 bit = 32K byte EPROMs, but only one 16K page can be accessed. AFAIK, only '87 ECUs use the first 16K, all others, the second.)
This is good as now I can keep the part throttle maps essentially stock, for emissions and response, and have a whole 'nuther 16 x 16 map just for full throttle (>75% pedal) precision!
I'm running my old tune in the WOT section for now, but I'll get to moving all the rows and cols around here in a bit.
I can't wait. I feel drunk with power. Muhahaha.
So nice to have so much memory to use. The 16 bit MAF ideal and it's scaling maps are ~600 bytes, which doesn't sound like much, but S3 LH2.2 has only 4000 byte EPROMs and they are completely filled with code/data.
S4 LH2.3 have 16000 bytes to work with, but code is a bit scattered around. Fortunately, there was a 700 plus chunk of empty bytes at the end.
(LH2.3 actually uses 256 bit = 32K byte EPROMs, but only one 16K page can be accessed. AFAIK, only '87 ECUs use the first 16K, all others, the second.)
#100
interesting.....very interesting......I have been wondering why Porken's S3 32V cars make SO much power.... Given the smaller valves, older ECU, non resonance intake, etc vs the S4's..... I have faith that Porken can figure it out
#101
Not perfect, but good enough to start work on the EZ.
Stable fueling allows more advance which means more power.
Changing the ignition timing will change the AFR, so there's no need to make it perfect.
As timing is advanced at a particular rpm, the area around that rpm will usually start going lean, if the advance is useful.
Injector duty is no longer locked to RPM so AFR is going lean near max rpms.
Leaning it out above 6000 is not really possible to do with the stock maps.
3rd gear, manual S4, on road
Stable fueling allows more advance which means more power.
Changing the ignition timing will change the AFR, so there's no need to make it perfect.
As timing is advanced at a particular rpm, the area around that rpm will usually start going lean, if the advance is useful.
Injector duty is no longer locked to RPM so AFR is going lean near max rpms.
Leaning it out above 6000 is not really possible to do with the stock maps.
3rd gear, manual S4, on road
#102
Ken,
Interesting stuff- you really are quite a clever bunny!
If I understand what you are saying correctly- there is in effect a "hidden" part of the brain that is throwing fuel in no matter what you do to the mapping at top end?
Regards
Fred
Interesting stuff- you really are quite a clever bunny!
If I understand what you are saying correctly- there is in effect a "hidden" part of the brain that is throwing fuel in no matter what you do to the mapping at top end?
Regards
Fred
#103
Hmmm, that's very intresting find. If I recall corretly it's the LH which is hitting it's limit first rather than MAF. Someone told recently that MAF/LH combo will hit the limit at 440ehp. Depends on what rpm is the engine top HP numbers, we could get some benefit of larger scaling, by more accurate fueling.
With the larger scaling, there isn't still any benefit of going further that 440hp, with stock LH/MAF combo, right?
It's the physical voltage limit of either of the component, rather than LH scaling, I think..
With the larger scaling, there isn't still any benefit of going further that 440hp, with stock LH/MAF combo, right?
It's the physical voltage limit of either of the component, rather than LH scaling, I think..
#104
One other trick that I used for my S3 chip, I have had to use for the S4, too.
In order to get control of the top end of the RPM range, I change the injector constant to a smaller number while under WOT.
This lowers the 'floor', so I can add fuel back using the map. (It's also an easy way to program for automatics, as they require more fuel.)
You could use a Super MAF, which compresses the upper voltage range.
If you are (turbo)supercharging, I could imagine using one of the unused variable inputs into the LH for a MAP sensor to compliment the MAF.
#105
I'm not quite sure what the code does. At some point, I will enable it again. Maybe it will be more clear compared to the nice, flat graph I have now.
One other trick that I used for my S3 chip, I have had to use for the S4, too.
In order to get control of the top end of the RPM range, I change the injector constant to a smaller number while under WOT.
This lowers the 'floor', so I can add fuel back using the map. (It's also an easy way to program for automatics, as they require more fuel.)
One other trick that I used for my S3 chip, I have had to use for the S4, too.
In order to get control of the top end of the RPM range, I change the injector constant to a smaller number while under WOT.
This lowers the 'floor', so I can add fuel back using the map. (It's also an easy way to program for automatics, as they require more fuel.)
I remember Louie once opining that max power on the 928's well designed heads was to be made at 13.5 AFR and that to get more than stock at top end you had to get from low 12's to this number-there or thereabouts somehow.
I look forward to reading more!
Regards
Fred