OMG, there must be something left to do? I can't just stop...

EZ and LH fully rebuilt with new code and maps applied to stock images.
(Cleaned out all the forgotten changes after using the same files over and over for testing.)
All cat and non-cat base maps restored. Broken WOT switch reverts to stock.


EZ
-WOT 64-col map
-Octane adaptation (warm recall)
-Knock light enable with error indication
-Soft rev limit (retard just before fuel cut)

LH
-24# inj
-Sport Mode
-MAF smoothing
-WOT 64-col map
-WOT separate MAF ideal map
-WOT separate injector advance time
-WOT IAC disable + IAC full open on decel >4000 rpm
-Rev limit, lower rpm restore point (dual rev limit/NLS too)

 

I haven't thought of a way to do that. All the injectors fire at once, twice per revolution.
The LH doesn't have a cam sensor input* and the EZ doesn't have a good way of communicating that.

There may be enough RAM (and/or SRAM) to make eight ignition WOT maps, if the maps are limited to the problem areas, only.

I have considered making a learned WOT correction map by RPM (64-col, ~100 rpm resolution), at least.



*FYI, the hall sensor input is very simple. It goes through a Schmitt trigger chip, which goes to a flag pin on the processor.

Almost all the analog inputs are switched through gates, so only a few digitized inputs are read. (One reason the EZK is so compact.)

The knock sensors are switched as one input. It would be possible to code out one knock sensor (and the hall sensor) so a EZK could be installed on a S2/S3.

 

Can you get them to fire once every second revolution, in sets of four? I am not worried about sequential firing. Once per two engine revolutions would be enough for me. I just want two full fuel maps, one for the end cylinders and one for the center cylinders.

If that can be done, then one can also make dual pattern cams for these engines that will maximize the potential from the stock intake manifold. If it can't be done, then the dual pattern cams have to compromise to bring the AFRs in line between the two groups of cylinders.

 

Yes? The shift light output (pin 26) could be used to drive an external gate for each group.

The stock LH already has a single pulse mode. (Half rate. I delete it on my chip as 24# inj. don't need it.)




Speaking of which, I think the EZK could be made to drive wasted spark coil packs using the knock light (5v) or CEL (12v) outputs to switch the coil drivers.

 

If you were willing to give up automatic idle control, the IAC frequency could be used for a second set of injectors.

Or, EZK load could be replaced by a frequency MAP sensor, and the load frequency could be used.



In the LH, one 3 x 16 bit timer chip controls injectors, IAC, and EZK load.

 

You don't have to. There is a market for you to tune the aftermarket ECU units that various 928 and other Porsche people have. You obviously have a knack for it. That clear enough.

 

Next please:

http://www.adaptronic.com.au/

 

Tuomo, are you worried about 2nd and 3 harmonics? The pulse strenght difference of those are something like 3 - 4% max.
The flow difference between intake tubes is almost 10%(270 - 300cfm). Acceptable cfm difference is between 5 - 7cfm, what I have been told and read.
Static flow difference + dynamic (pulse strenght) together may turn things even worse. .( I'm sure you know these things better).

I have just finished my S4 intake flow enhance + balance project and found that turn from throttle plate to oval tubes which are leading to plenums are very very restrictive.
The good news here is that even the static flow difference is huge between intake tubes, it's s till not that big issue as there is another big limitation in flow path.
Therefore you cannot push the intake tubes to their limits, which would cause too large static flow differencies. Just again my simple engineer theories, but that's what i saw.

When I first worked with those TB plate to oval tube turns and tested car after this, engine started to knock right after switching to WOT. Knocking cylinders were like 2 & 6.
They started to flow so much more than others after reducing great bunch of frictional losses.

One idea of balancing static flow across tubes was to use LH and measure AFR only from one side and still able to push timing to as far as possible.
Also the MAF voltage smoothing algorithm is on table as well.

In your TT case the situation may be still different, since I was using my reduced intake port heads in my tests.

 

If you've got photos and flow numbers, please post them. Maybe in this thread: https://rennlist.com/forums/928-foru...-ideas-11.html

 

Surely, will take some time..

 

Ken,

That is very interesting. 5v or even 12v with a level shifter is pretty simple to interface a coil with. You could even use a quad-channel ignitor in-place of the stock distributor. My question is how would you be able to do per-cylinder knoc retard, as I thought the EZK relied on the distributor "windowing" to apply knock retard to a specific cyl? I would think it would retard the whole connected bank?

 

It would still know which cylinder was being fired when doing wasted spark and could adjust accordingly as it knows which cylinder is on compression, and which on exhaust.

 

As I understand it, the hall sender rotor flags the EZ only left or right (coil).
Per cylinder is counted from TDC #1 and #6.

The EZ code is very simple = fast, which conversely makes it more difficult to understand.
I would like to delve deeper into the per cylinder programming for per cylinder mapping...maybe for V2.0.




Driving around on real chips is a weird feeling. No adjustments!?
I think I may be suffering from postpartum depression.



The 'soft' rev limit (ignition retard >6300) feels funny - as if the car is getting sleepy as it struggles to hit the 'hard' limit (6650).

 

So, when can we buy them?

 

I guess as soon as the PayPal link is up, if you really want to be an early adopter.


They are at least as good as these chips, right? Same price? Half?




I have had 27256 EPROMs, little cases, all set to go, since...well, let's just say I thought it was going to be easier.
These are certainly polished, unlike the S3 chips when those first rolled out.