PunKidd

OK, to clear a few things up from earlier in the post. None of the three issues listed on the first page are really valid.

The stock ECU DOES monitor the knock sensor up to ~5K RPM, and there isn't a microphone knock sensor system out there that can reliably monitor knock above 6K due to engine noise at those speeds. The only way you're going to get reliable knock control at those engine speeds is to use direct pressure measurement or an ion sensor (neither are cheap and the MIT auto lab is the only place I know of that use direct pressure measurement. A couple OEMs use ion sensing, but it's rare).

Yes, over-boosting does happen, and the most common cause is the WG line coming off or getting melted due to poor installation (if it was installed by someone who knows what they are doing, it should really never be a problem, but accidents can happen anywhere). It doesn't matter if the EMS you're running can cut boost in this situation because once that line is disconnected there is NO control over the waste gate and the turbo is going to push all the air it is capable of pushing.
Here's what I do on EVERY FI vehicle we tune here at UpRev. I always make sure that there is at least one extra column of head room in the fuel target and ignition tables and in that extra column I add a TON of fuel and pull ALL of the ignition. This doesn't provide for an actual limit like an ignition or throttle cut, but I doubt that an engine is going to blow running 10:1 AFR and -10 degrees of ignition advance. The engine usually will not even rev any higher under those conditions, but then again, I have not actually tested this at really high boost levels like you would see in an over-boost situation. I can say that at 18psi it will not rev any higher. If you do run into these columns, it's going to be REALLY obvious to the driver and they will know something has gone wrong.

You CAN run low impedance injectors off the stock ECU, but you will have to add an injector driver. We used the AEM one to run 1000 HKS injectors on a +600WHP G35 (they were WAY overkill, but that's what the guy had already when we pulled out his Fcon), and they idled fine, car started great hot or cold and ran just like it came with those injectors from the factory with nice smooth power all the time.

All in all, if the tuner is familiar with Osiris they will be able to tune the car just as well as with a stand alone and in many cases even better when it comes to drivability. When putting this kind of money into a build I would only recommend tuning with what the tuner is comfortable with, regardless of what they have access to. If they are comfortable with Osiris, then go to town. I'm sure you'll be happy with the results.

Jared@UpRev.com

ttg35fort

iTrader: (2)

Thank you for Chef-J for your support!!!!!

rcdash

iTrader: (18)

That's an interesting work around for an overboost failsafe Jared. Thanks for the detailed reply. I wonder if other licensed tuners do the same thing? The overboost issue is probably the only one that would give me pause in running Osiris alone for FI.

I needed Osiris anyway to get what I wanted out of the stock ECU even with the Haltech (speed limiter, idle, emissions/CEL control) so it wasn't a matter of choosing one over the other. I wanted the best solution at a reasonable price point. Haltech + Osiris reflash is a really nice combination if budget allows.

Just to clarify, the Haltech provides a complete fuel cut for overboost. It cuts exhaust flow and engine speed immediately. It doesn't try to just set target boost to 0 (like many EBC add-ons do).

Aced16

iTrader: (5)

Terry contributes on a totally different level! He goes beyond just the basics, and contributes with detailed information on a math/scientific level. I'm glad you got the job terry! i follow lot of your treads and recently just purchased your book recommendation and plan on reading it cover to cover! dont let the haters bring you down

djamps

iTrader: (1)

Reflash (osiris, or others) is still a good idea, standalone or piggyback. My reflash deletes the speed limiter, removes the advanced timing map (makes transition from ecu to utec consistent always), richens up the top end, and helps a bit with injector scaling at idle and cruise. In fact, the my utec doesn't even take over timing and fueling until 4psi -- reflashed ecu handles it just fine, afr and timing is in check and transitions perfectly to the utec every time.

ttg35fort

iTrader: (2)

^^^^

Thank you Aced16!

As with most everyone else here, I am fascinated with cars, and always have been. I was an engineer for alot of years (though in a different field), so I tend to look at things from an engineer's perspective. It is unsatisfactory to me to plan my build based on various recommendations without having a technical foundation to support those recommendations. To really know what is best for any particular build, we really need to understand the underlying technical aspects. I try to share that perspective with the forum. I'm really glad that other's are interested this as well.

Resmarted

iTrader: (2)

This is where I think re-flashes shine. When you are comparing them to similar cost level options they are very good. Utec is awesome (especially considering the vast possible applications) but it doesn't really come close in the fuel control department (ie increasing rev limiter) but it also can't disable codes. To me, if all were equal except Osiris can remove codes, it would be worth going Osiris. None-the-less, I still claim Utec is one of the best Piggybacks you can buy.

John at J&S

I respectfully disagree. google "giles aem log" to see a comparison of the AEM knock detector versus brand X.

The tuner said he was listening to the knock sensor with headphones, and the AEM could not reliably detect knock, yet brand X was accurately detecting knock and controlling timing at 7500 RPM.

SOLO-350Z

iTrader: (26)

^^ Sorry but no way possible. Not with that much engine noise period it can not detect knock. Believe the selling points on what things will tell you, I can promise you it catches any knock above that, it will already be too late.

rcdash

iTrader: (18)

Do you have personal experience with the J&S Safeguard? If not, I don't think you are qualified to say what is possible or not. Just because it's not possible for the stock ECU, doesn't mean it's impossible. LOL.

I'm always open to new ideas and new technologies. The J&S hasn't taken off on our platforms, but it has on others...

Chef-J

iTrader: (58)

I finally understood...

westpak

iTrader: (19)

well I am thinking of trying this again, this time I will probably skip changing injectors and get the AEM injector driver for the low impedance injectors and see how it goes

djamps

iTrader: (1)

Agreed...the comment about RPM limits to detecting knock is total BS.

There's no way that just because engine noise is louder than knock that it cannot be extracted using modern DSP (FFT) algorithms. Take GPS for instance, where the actual signal is orders of magnitude smaller than the noise surrounding it. It depends on the engineers that designed it and the number of taps and quality of filters that are implemented to extract it.

PunKidd

First off, the stock ECU has a slew of DSP functionality built right into the processor, so if it was a matter of digital signal processing, the factory would have done it.

Second, the real problem is the sensor. The stock knock sensor just doesn't have the range. By the time you get to 6000RPM, the poor little thing is just plain maxed out. Remember it's just a microphone.

Third, at high RPMs the engine is already creating the same frequencies that you would be trying to listen for to detect a knock event, but it's normal for the engine to generate those frequencies at those speeds and the engine actually generates those frequencies at a higher amplitude than a knock event generates (the sensor is already maxed out and clipping the signal at that point). Now how are you going to filter out a signal of the same frequency and higher amplitude on a signal that is already getting clipped by the range of the sensor and find the knock event? Not going to happen.

I actually do have some experience with J&S. I used to run one on my turboed Impreza RS back in 2001. I'm sure their technology has come a long way since then, but I can tell you that I could tune the engine to knock at 5500RPM and the J&S didn't pull timing. Not only that, but the J&S DIDN'T use the stock sensor. You had to install the one that came with it. Maybe their sensor has a better range for high RPM? IDK.

Bottom line is that the stock sensor does not have the range to get a clean signal above 6000RPM, I don't care how you filter it.

Jared@UpRev.com

rcdash

iTrader: (18)

^ thanks for sharing your experience on the J&S and on the limitations you have observed of the stock sensor. I did a little reading on knock detection strategies and while some frequencies may be the same and the signal to noise ratio poor, a few advocates of third party products indicate that by recognizing frequency patterns in association with filtering and phasing (same frequency but only detecting during a short window where knock is expected to occur), there is improved discrimination of signal from the noise. Clearly such a system must be trained and the R&D that must go into it is not trivial.

I'm not saying it's easy and I'm not saying it's currently feasible. However, making a blanket statement that "the stock ECU has a slew of DSP functionality built right into the processor, so if it was a matter of digital signal processing, the factory would have done it" is just plain short sighted that overlooks innovation and new developments not to mention the complexity of DSP programming that continues to evolve. I think it might just be possible that the stock ECU has just a few limitations in its DSP implementation, eh? Granted maybe a different sensor might be required but I'd like to think that this is not such an insurmountable engineering obstacle if one is willing to look beyond the stock set up.

John at J&S

We re-calibrated to work with the stock sensor in '02. Our earlier Subaru kits shipped with a tuned GM sensor from the Saturn.

Scott at forcedperformance recently installed one of our new Interceptors and sent a data log, showing the unit retarding at high RPM.

He said he forgot he was using a high octane map which had two degrees more timing in the lower gears, but was running on low octane. He ran it again on the low octane map and there was no knock retard.

Link to the new eight channel Interceptor is on my home page.

djamps

iTrader: (1)

I'm sorry, but just because the stock ECU has the supposed capability to do alot of DSP doesn't mean that it is. It was never intended to be used in FI applications thus detecting knock that high of rpm wasn't worth the extra CPU power or possibility of false detection.

The simple fact is that my UTEC can and has detected knock that high and probably saved my engine with the MAP came unplugged. I would also argue that the UTEC isn't by any means a high power device.

In addition, J&S has supporting proof right here that it can be done.

PunKidd

Well, I guess I stand corrected. If J&S is doing it on the stock sensor now, then I'll have to believe it.

Limiting the monitoring to the 5-10 degrees per cylinder where you would expect knock makes a lot of sense. We were playing around with a Bosch OEM knock monitoring chip a few years back that did the same thing, but we never really got the thing working (no time).

Yeah, these computers were never intended to run FI, but it is the same processor they use in the GTR. I'm thinking we need to do some digging in the GTR code to see what they are doing different with the closed loop ignition control. Maybe we can port some of that code over into the Z ROMs somewhere down the road.

For now, my work around of dumping fuel and pulling ignition when the MAF goes above what is expected seems to be doing the trick.
Even if you're not over-boosting, but you get some bad fuel or an injector fails, the stock knock control is going to hear the pings down low and pull timing across the board.
We've tuned more than two dozen FI Z cars on the stock computer over that last 4 years (a hell of a lot more than that if you take our pro tuners into account) and we have yet to have one fail from detonation or over-boost. We've seen a couple of other failures like oil starvation and spun bearings, but no pitted/cracked pistons on any or our tunes on the stock ECU.

Jared@UpRev.com

Old School

iTrader: (28)

I use my osiris in conjunction with my VPRO. To just use the Osiris, I'd say you be much better off with that than the utec. I couldn't stand the hickup between boost and vacuum. you can feel the crossover point. osiris is pretty seemless. If you're looking for over 600whp, I'd wait n see what others do with that software or consider another ems. Anyone running more than 600whp on Osiris? I remember Uprev tuned a TT VQ @ 550whp. Not sure if they've done anything more than that.

Chef-J

iTrader: (58)

highest whp i heard is " rookie "'s FX setup. Tuned by Technosquare Reflash and Sard 800cc/RFS..etc blah.
But seriously, both of company are BEST for ECU reflash. Just diff type of tuning skill and way. I don't see why not.

Edit: I personally want to SEE that how much stock ECU can handle...