944S2 Supercharger Build Thread

[Eric_Oz_S2]

So after a year of planning, thinking and procuring it is now time to start construction of my S2 supercharger system. My reasons for this project is an aim to move the car's performance up to where I can compete with GT3s for club sprints. I'll end up this year as a class E winner (7.5 to 8.5 kg/kw P/w) so I'm looking for another challenge next year - moving into class C (5.5 - 6.5 kg/kw). For the small cost associated with upgrading the S2, I think I can be competitive against cars in class C and also those GT3's in the classes above. I know I won't be able to beat the better GT3 drivers, but I'm sure it will be fun nonetheless.

The kit will be a Rotrex C30-94 running about 6-7 psi and utilising a 951 intercooler. Interestingly for some, I will be keeping the AFM in the first instance and perhaps later upgrading to a MAF depending on how the AFM fares. Other parts of the system includes 951 injectors, and a retune using an Ostrich II emulator, fast response IAT, MAP sensor and LM-2 for logging.

Anyway, preparation work is continuing and I have sorted out a few of the things necessary.

The first thing to do was to wire in the LM-2 analogue inputs into the DME harness. This is not such a fun thing to do having to use a certain amount of acrobatics to work in the passenger footwell to get to all the wires.
DME wiring splice. Note I have also connected wires to both the variant and coding plugs to allow remote switching of maps.



2 bar Delphi MAP sensor installed to intake manifold vacuum line and connected to LM-2 inputs.



LM-2 sits in the glovebox, all wires are routed through that convenient hole that Porsche put in the back of the glovebox.


Wiring for knock sensor has been connected to the diagnostic plug in the passenger footwell.


Rather than go to the hassle of getting a custom faceplate done, why not make your own? I made this from a faceplate used in equipment racks. It is anodised aluminium (we can't buy aluminum here ). The hole was simply cut using a jigsaw and the panel is held in with some double sided foam tape). The left side will eventually house a SPA design dual boost/oil temp digitial gauge. The knock counter is a Redlion CUB400 counter with a nice backlight powered on when the engine is switched (powered by diagniostic plug).


Next up is the install of the WBO2 sensor. I will be using the LM-2 's NBO2 simulation for tuning, so I can use the stock bung.

The WBO2 is interchangeable with the stock NBO2 thread.

WBO2 installed.


There is a convenient hole covered with a grommet that is next to the main harness penetration through the firewall. It is perfect for use for feeding the WBO2 lead through to the passenger compartment and certainly a lot simpler than the sometimes used shifter hole solution.

Some special order parts from Porsche Germany - in this case the 3 pin round plug and pins for connecting the LM-2 NBO2 output back into the DME harness. Effectively I am plugging the LM-2 output into where the NBO2 sensor normally connects - very neat.



Here is the plug assembled. I haven't connected the heater outputs to anything as it doesn't seem to matter if these are left disconnected. I have sean some in the 964 forum where load resistors are connected to the power feed to prevent errors being generated by the DME.
This thing was a bugger to get together, the plug is rubber and the wire and pin must be pushed through from the back quite a way. A small screw driver and some force were in order.


All wired up. The green wire is the output from the LM-2.

[Eric_Oz_S2]

Next up is to replace the stock injectors with turbo injectors. The S2 injectors flow at 264 cc/min at 3.0 bar and the turbo injectors flow at 370 cc/min (968 injectors flow at 296 cc/min). From a calculation of the S2 injector capacity, these would need to flow at 130% for the expected boost. Some have used a RRFPR for this, but on these old injectors the spray patterns break down at higher pressure and they become inefficient. The turbo injectors have been calculated to run at 90% and should provide enough fuel with the S2 3.8 FBR and expected 300hp target output (flywheel). I actually got reconditioned Mustang SVO brown top injectors which are identical to the turbo ones.

This is one job I always hate. Removing the plug on the fuel rail (the one with the ball bearing seal) always results in a heap of fuel spill. You can't get a container under so you use paper towels and stink out the garage.

Injectors removed from fuel rail:


Old and new.


New injectors in:


The S2 fuel rail has an unusual washer arrangement for the fuel rail connection. You need to be careful not to loose these spacer washers that sit between the rail and manifold.


Next up is to reprogram the DME to put in a new injector factor. Basically I'm changing the stock (Position 0) of the FQS address in the DME and reducing this value to compensate for the larger flows. Hopefully my estimated value will be OK.

[Eric_Oz_S2]

Thought I'd post a picture of the SC unit that is going in soon. Currently fitted with a 100mm pulley for initial tuning. Probably go smaller depending on boost levels achieved.
Progress update.

Removing all of the bits that need to come off for the SC install.

Removed header panel, alternator and bracket, left side wheel well liner, batwing, AFM, airbox, alternator cooling duct, receiver dryer, ac pipes etc.

I also worked out where to install the oil cooler - it will go in the left hand side behind the fog light area. This will be a mirror image of the S2 oil cooler location on the right side. The two tone horns need to be swivelled out of the way so they sit further forward so they can clear the oil cooler and pipes.

I also installed the intercooler and air duct. Luckily the S2 has four of the little rubber intercooler mounting feet already provided, 3 are used to support the airbox and one supports the AFM intake pipe. Simple matter of reusing the AFM one for the intercooler support. Very handy that Porsche used the same part so you don't need to buy these. It is interesting that I needed to wind up the rear support feet and put spacers between them and the bracket/support. Without these the intercooler pipe will hit the headlight linkage bar. Not sure how the turbo gets this to work?

I also installed the Rotrex and bracket so I can now measure up all the pipework. Some dremelling was required so the bracket cleared the plastic belt covers. Very minor clash here. Here are some pics of where I am at:

Intercooler in:


Rotrex in:

[Eric_Oz_S2]

Spent some time trying to figure out where to mount the AFM. My first thought was below, but there are two problems with this. Firstly getting the outlet to run in front of the steering tie rod and between the alternator and sway bar is a really tight fit. I recall Henk had to install a smaller alternator to get this to work. Secondly, mounting the AFM below and then getting the outlet to go up through the hole in the bottom of the "well" behind the left headlight is also tight.
Thought I might need to go MAF, but one last shot at installing the AFM - this time from above.

And thanks to Pauly for the Autobarn tip for the AFM plate. Found a blank one I can drill to suit for $25.

The AFM j-u-s-t clears the hood by about 10mm. Will need some foam padding there. Intake pipe to AFM will utilise the hole that is there for the 951 with an airfilter placed between the fender/guard, chassis and wheel lining.

[guru944]

Eric_oz_s2:

Great project, keep me posted. Currently working on AFM thing myself for my K27 Turbo setup 16V..

Latere

[Arizona_928]

I like....

you might go with maf for better response and tunning ability. Can't wait to A/F ratios, boost, afm compatability.

[Eric_Oz_S2]

MAF is certainly a consideration, my reservations are with respect to how the Motronic will deal with the much faster MAF signal. The AFM is damped so the sampling rate is different in the code than what would be used for a MAF.

I'll wait to Joshua (Rogue Tuning) sorts this out, and see how the AFM works first.

Got some silicone bends etc today and test fitted. I think the AFM will fit where I have it shown above with about 20mm clearance to hood. So I'm very relieved about that!

[North Coast Cab]

Great project. I would get the SC up and running with as little other mods as possible. Then look into the MAF set-up to help the motor breathe.

[Eric_Oz_S2]

Here is the stock Air Fuel Ratio from the LM-2. Have also logged MAP, RPM, TPS, AIT.



I'll tune in "No-Cat" maps (using code plug) and then copy this map to the normal cat maps. This way I can get around any issues of the DME (potentially) doing odd things when no o2 signal is detected (haven't seen any evidence of this, but just as easy to be sure).

The plan is to limit closed loop to lower RPM (up to 4000rpm).

New injectors worked perfectly (tested before dismantling and SC install began), ended up using a 0.8 scaling factor in the DME FQS factors. Also needed to reduce the idle maps, as these don't seem to respond to the FQS setting (interesting).

[billy mild]

So do you have the blower hooked up? where did you get the bracket at for the supercharger? I was unaware of a bolt on solution for the S2? Did you have to delete A/C?

Looks like a very cool build.

[Eric_Oz_S2]

I am still waiting for some pipes to complete the installation. I will then need to get some aluminium welding done for fittings, etc.

Also still need to mount and hook up rotrex oil cooler.

It is not exactly a bolt on solution, I have procured all of the parts individually and designed the system myself.

I deleted the AC to get this to work. I have considered (maybe) running the AC in the power steering location and switching to an electric ps pump.

I got the bracket from Holland - there are 2 companies that fabricate these for 968 SC kits.

[Dougs951]

Watching....

[flash968]

i too am watching. i think it's great that anybody steps up to do this kind of thing. it's a huge undertaking.

as one of the companies who developed a rotrex kit for the 968, and the only one to retain the AC, and the only forced induction kit of any kind for the 968 to be CARB certified and 50 state legal (26 of them out there now, some running now for over 2 years, some with 10k miles on them or more, and none with any issues not installation related) i know what it takes to get this right.

the tuning of the motronic is complicated to say the least. it's very easy to mess up and end up with a mixture that causes either detonation or cylinder wash. WOT is simple, but not important.

i had to generate new maps to handle part throttle loads properly, and so far have only been able to get them to work in the later ECU. the early ECU does not accept the coding.

bigger injectors creates all sorts of issues too.

the other limitation to boost on the 968 was the dual resonant intake manifold, the S2 only has a single resonant manifold and may well turn out to be better for forced induction, even though the runners are smaller, as are the ports.

it will be interesting to see how the S2 differs from the 968 in how that all works out. the only reason i haven't done the S2 is that there isn't one near me that wanted it, and after the 200 hours tuning the 968, i really didn't want to spend the time tuning it, and i insisted on my kits being complete, including tuning. i may eventually get around to sending out hardware only kits, and let others buy the head units themselves and do their own tuning. who knows?

[Eric_Oz_S2]

Yes it certainly is a large undertaking. I expect a lot of tuning is required.

I think the biggest challenge is the PT fuel tuning and dealing with closed loop limits. There are some fallback positions/opportunities if this doesn't work sufficiently to maintain the right AFRs in the PT maps:

cat delete and run conservative PT AFRs
adjust acceleration enrichment maps
zero out lambda adjustment range maps above certain rpm
aftermarket ECU

It is interesting that the S2 (and I presume the 968) PT maps are used in conjunction with the WOT maps for WOT - ie the WOT maps adds fuel trim to the PT 'learned' values and effectively are delta maps over stoichiometric. You can actually leave the FQS at zero with bigger injectors and it will have sufficient range to adjust the AFRs right back down to 14.7 at PT.

[flash968]

yes. the 968 ECU corrects mixture at 1/4 to 1/2 second intervals, depending on where you are in which map, and returns it to stoich. this is precisely what makes PT tuning so challenging and time consuming. you have to do a LOT of driving and tuning to create the conditions so that you can adjust things everywhere.

where it will have problems with bigger injectors is down low. there is code there that prevents leaning it out enough to be right really low, and the spray pattern isn't right either when it's trimmed down that far. it runs too rich (down around 11), and will eventually fuel wash the cylinders. not a real problem for a race car, since you tear the motor down every year, but for a street car, this is a very real problem. it's similar to running with the coding plug installed. that leaves things in the upper 12s and 13s too long. these engines don't like that. you can smell it in the oil.

good luck.

[Eric_Oz_S2]

Already tuned the idle fuel, and I can get a smooth idle with AFRs wherever I want them. The injectors are only 36 lb, so not that much greater than stock. The trick with bigger injectors is adjusting the idle fuel map separately from the injector scaling factor. Above idle, (low throttle) the AFRs are adjustable, and have got these around 14 with no issues.

If you simply threw in larger injectors without map changes it doesn;t work so well....

[flash968]

tried that. oddly, on the 968 ECU, regardless of where i set the idle map, it made no change in the mixture. it was still way too fat, and especially at cold startup. it was as if the injectors were just dribbling, and not spraying, and wouldn't go below a certain point of flow

what software are you using?

[billy mild]

Why not switch the system over to a MAP based system so it can adjust fuel based on boost?

[Eric_Oz_S2]

Flash, i'm using tunerpro rt with an ostrich emulator. I presume you adjusted the injecor dead times for the larger injectors? The s2 has 2 injector dead time maps iirc. If these aren't right the afrs will be low and the car will struggle to start.

Billy - map speed density system requires extensive modification of the ecu to work correctly. Map would be used with a standalone ecu.

[Eric_Oz_S2]

A little bit of progress, but still awaiting delivery of some pipes and silicone hose.

Oil cooler has been installed in the left hand side using the existing support that mirrors the support for the engine cooler on the RHS. Simple brackets fabricated using 40x3mm galvanised steel flat bar and utilising the existing holes in the body. The cooler is a very tight fit but almost looks like it was made to go there. Cooling ducts in the front bumper cover line up nicely with the cooler. Some rubber grommets help isolate the cooler from vibration and give it some give between the radiator support frame and bracket. There is about 2mm clearance from the modified horn location to the cooler.


Also mocked up the hoses for the intake. It is amazing how much you can outlay on the various hoses, couplers and clamps. It would have been cheaper I think fabricating up most of this as hard lines, but at least this way I can easily adjust the setout.


The intake area is really tight. I am having to use a series of 45 degree bend and a 90 degree bend to get the pipework between the AFM and behind the front guard/fender. The cone filter will sit in the wheel well in front of the inner wheel liner and behind the front bumper cover. The hoses shown are not necessarily the right size, but show how it will fit together. I couldn't find anywhere to put the rotrex oil reservoir except in this busy little corner. It just squeezes in and is supported partly off the AFM. It may require a single hole drilled in the body for the lower bracket support.



From here on there won't be much progress for the next week or two as I am awaiting additional parts.