Official OpenFlash Tablet FA20 Tuning Tech Discussion Thread

[Rombinhood@OpenFlash]

THE OFFICIAL OPENFLASH TABLET FA20 TUNING TECH THREAD

Greetings! First and foremost, we would like to say thank you for all your support of the OpenFlash Tablet. We're excited to bring it to you. And hopefully you'll be excited to use it for many years to come. So it's only fitting that we kick things off with a FA20 tuning tutorial. As some may know, we've contracted Vishnu Tuning to do map development for the OpenFlash Tablet. They have been tuning Subaru engines since 1999 and were the fist to even offer a stand-alone engine management system for the WRX precursor, the 2.5RS. In 2000, they released the only successful EJ25 turbo kit which can still be found on the road today. Then when the WRX came to the US they quickly quickly became the highest volume Unichip and Ecutek dealers in North America. And for the last 6 years they have been developing their own engine control systems for the direct injection turbo I-6 engines found in the BMW 1, 3 and 5 series. They have set every one of the the HP, 60-130, 1/4 mile and 1/2 mile airstrip world records records. So they are no strangers to ECU tuning and, more importantly, not against disclosing what they do and how they do it. So I'll pass the mic to Shiv@vishnu and he can get things rolling.....

--------

Hi guys,
I've spend the last couple of weeks developing tunes for the OpenFlash FR-S. This involved over 20 hours of dyno time (on both Mustang and Dynojet Dynos) and several hundred miles of road testing. During this time I've collected a lot of data. So today I had the unique opportunity to run the fully tuned car on the dynojet and, piece by piece, strip away mapping changes until we were left with the stock tune. I found this interesting because it clearly shows the effects of every change which is what we all want to see. As a tuner, I found this refreshing because I'm so used to tuning high hp turbo engines that have a typical run-to-run variance of 5-10whp which makes seeing the effects of subtle changes virtually impossible. So tuning an NA car that is perfectly repeatable does allow us to analyze the exact effects of each mapping change. I hope you find graphs/tables below tell an interesting story!

All dyno runs were done in 4th gear on 93oct on the same dyno (DNR Performance in Hayward CA). Car is 6MT. 900 miles on the clock.

First here is a dyno chart comparing a stock run of OpenFlash's 2013 Scion FR-S to a run with a full bolt-ons (4-2-1 catless header, cat delete, 2.5 complete exhaust and Air Raid cold air box):


All the bolt-ons made the car very loud! But it picked up some decent power. Top end power gains were decent but what became obvious was that the Air Raid aftermarket intake, like most aftermarket intakes, caused some changes to how the MAF read incoming air above 6000rpm. This caused the car to run very rich at high RPM which contributed to a rather lumpy top end.


From testing done earlier on another dyno, there is no doubt that the stock intake box is restrictive so upgrading it is necessary to extract max power. But there is also no doubt that different aftermarket intakes will induce different MAF errors with each one require different fuel mapping changes to compensate. In our case, we had to enlean the AFR mapping at high RPM. Once we did that, things looked a bit better up top:


It's important to remember that you want to run as rich as possible without getting to the point were you are starting to loose power. In our case, while we enleaned the AFR mapping above 6000rpm, we actually enriched it below that. We ran approx 0.3-0.4 point richer up to 6000rpm and saw no loss in power. You can run leaner and still make the same (if not more) power. However, running lean does make the engine more knock prone since there is less in-cylinder cooling. In the case of our FA20, the external wideband AFR meter we use showed the engine running in the 12.5-12.6:1 range up to 6000rpm. By 7500rpm, it would enlean slightly to 12.8:1. This yielded us the best compromise between power and knock resistance which is important in the next step....

Next, we can see the effects of a slightly tweaked ignition map:


In this case, I advanced ignition timing by 1.5-2.5 deg (depending on RPM). I carefully monitored knock learning to make sure that the ECU wasn't detecting knock and pull back timing. This took a while to dial in since the engine would have different knock sensitivities at different RPM points. It definitely picked up power up top with each 0.5 deg of advance I gave it. However, it's easy to get greedy and end up running into knock at high RPM. Rule of thumb when mapping ignition advance is to advance timing 0.5-1 deg at a time and carefully quantify what engine speeds it helps a little, a lot, or none at all. You never want to advance timing unless you are picking up gains. So for those RPM zones that you see no gains, remove the added ignition advance you put in. For the areas where you see gains, keep it in and carefully monitor knock as you progress. Stop when you start to see knock retard. On pump gas, a high compression motor like the FA20 will never get to the point were it is making maximum power (MBT) before the occurrence of knock. So you need need to take your time if you want to do this yourself (instead of just using an off-the-shelf map provided by us). Also, when tuning the ignition map, it's useful to go back and forth between ignition and fuel map adjustments to find the right combination. Running slightly richer may result in a power loss. But running richer may also allow the engine to take slightly more ignition timing which may more than make up for it and result in a net power gain.

Next step shows the effect of a revised Intake VVT map. By selectively advancing certain areas of the intake VVT maps we were able to bump up power through most of the RPM band with biggest gains right in the infamous torque dip and in the useful 6500-7000rpm band:


Next step shows the effect of a revised Exhuast VVT map. By selectively advancing certain areas of the exhaust VVT map, were able to see big improvements to the FA20 torque dip:


Needless to say, the effects of these VVT changes alone completely transform the way the car drives. Another very good reason to use a ECU reflash over a piggyback that doesn't give you the ability to make any of these changes.

The next tuning change is pretty subtle but useful and takes us back to what we learned when we played around with the ignition timing maps. We learned the key is to run as much ignition timing as possible while maximizing torque and not experiencing knock. Well, it turns out that the Subaru engine tuners have ignition compensation tables for each of the 4 cylinders. The compensation tables for cylinder 1 and 2 induce extra ignition retard. Most likely because the engineers determined that cylinder 1 and 2 run hotter than cylinders 3 and 4. This could be because of how the coolant passages are routed in the cylinder head and/or possibly the effects of having a relatively closed-coupled catalytic converter in the header causing high back-pressure levels. Since the car has an aftermarket header (no cat) and upgraded exhaust components, I decided to play around with the cylinder compensation tables for cyl 1 and 2. I first tried zero'ing them out completely only to find that the car was picking up knock more frequently than before above 6000rpm. So after much testing, I ended up zero'ing out the retard below 6000rpm. Above 6000rpm, I cut the retard amount in half. Upon testing, the car didn't knock and it made more power, more smoothly. As one would expect when all 4 cylinders are now working more evenly:


The final step below shows the effect of remapping the Direction Injection firing angle. It's subtle but there were consistent gains. More importantly, adjusting the firing time (done through load vs RPM table) does seem to improve knock resistance as well. Which, in a street car running pump gas, is just as important as making more power:


Finally, here is a comparison showing where we started with the stock tune (but will still all the bolt-on mods) to where we ended once we remapped the above mentioned tables:


And here is were the car currently stands compared to where it was when it was completely stock (no mods, stock tune). Same gas, same dyno:



In Summary:
Keep in mind that tuning is a engine properly isn't as straightforward as what is described above. You will rarely go from tuning ignition, then fuel, then so and so. I didn't do it that way and neither will you if you decide to start from scratch. You will likely bounce back and forth between maps because there are interdependencies between different maps. For example, you can't dial in an ideal ignition map unless you have a good fuel map. And visa versa. Also, changes in VVT maps and GDI injection timing will have an effect on AFR and knock sensitivities. So think of tuning as a back and forth game where you try to eek out the best set of compromises. To do this, you will need to become very familiar with the effects of each mapping change and how these changes effect other aspects of the tune. The nice thing about learning on an NA engine like the FA20 is that it all follows textbook engine tuning theory quite nicely. So nicely, in fact, that I really enjoyed getting back to basics! It's also nice that you be learning on a fairly bullet proof engine instead of a high stress turbo/sc engine that is looking for an opportunity to crack a piston or bend a rod

It's also worth mentioning that more tables were adjusted in the tune that was mentioned above. Other modified tables include overrun, fuel enrichments, throttle mapping, rev limits, idle speed, etc,. While these changes don't have an effect on dyno power (which is why they weren't mentioned above) they do have a profound effect on how the car drives on the road. Making power is great but being able to improve drivability and response is even more important since that is a bit contributor to what makes a car enjoyable!

I'll be posting screen captures of my tuned tables (compared to stock) in the next post. This map (along with other maps) will be fully viewable, editable and download-able from our website later this week. You'll be able to see every change I made and feel free to ask why I made it. I'll check in this forum from time to time and will be happy to provide tuning related info. Hopefully we can get some ongoing tuning tech discussion on this forum and make sure everyone as a well-running BRZ/FR-S!

Best Regards,
Shiv
www.vishnutuning.com

[Rombinhood@OpenFlash]

Reserved for stock vs modified map screenshots

[mad_sb]

W00T!

Finally I'm not the only one posting a tuning thread

Big thanks to the openflash guys and Shiv for the brain dump I really enjoyed chatting this past week about the findings and comparing notes on the base maps.

I have a few questions and comments after seeing the digest version...

Cam Timing
On stock exhaust, I found a good bit of power on the low rpm and higher rpm sides of the dip by advancing the exhaust cam. I assume you found the same thing? I did not find any difference in the 3,750 range though... I assume the back pressure is just so high there for whatever reason that you can't close the exhaust valve soon enough to avoid reversion??

Which way did you go with the exhaust cam on the header car to get those gains? My long tube should be hear late this week and I will be hitting the dyno to retune on pump before I move on to E85 again. So I would love to hear what you found. My assumption is less Ex advance with a better flowing header?? I may just roll your cam timing into my starting point tune to save some dyno time.

AFR
After chatting earlier this week, it really hit home that with 12.5:1 compression (even with DI) you need to treat this a bit like a turbo car and use some extra fuel for cooling so you can get more timing in to get closer to MBT.

I have lots of logs of my car, when stock, running 13.2:1 WOT through the midrange with zero long and short term trims. I struggled for a while with trying to tune it at 12.8 – 13.0. It would make great power for one pull, but the minute you tried to do back to back runs it would start pulling timing. My last pump gas tune I did before going to E85 (The tune I switched back to for base map testing this week) I tuned at 12.5 – 12.6 with a dip towards the rich side of 12 up top... this was by far the most repeatable and knock free tune I have done.

GDI Injection Timing
I found a little bit of power around 6,000 on E85 by moving the SOI (Start of Injection) earlier a little lower in the rpm range. I haven't had time to play with it on pump gas. Love to hear more about what you found as I think there can certainly be a sweet-spot where you get the best in cylinder cooling combined with the best charge mixture etc.

Really looking forward to seeing the E85 results soon too. I found that we are not able to get even close to MBT on pump gas above 5,000.... down low we are only off a few degrees once tuned. This is using the same method for timing tuning posted above. Keeping the advance where it makes power, removing it where it does not.

[mad_sb]

I almost forgot....

On E85 I dialed the per cylinder timing comp back a good bit and also found some power and smoothness, once i go to the header I will play around some more with reducing them on pump gas and probably fully zero them out on the corn

[RehabJeff86]

This is very exciting to see!!

[Acree]

.... Hell of a post.

[Seattle944t]

Sweet - good to see more tuning options coming to fruition on this platform!

[Dustin]

If this turns out to be as good as it promises to be, I will be purchasing one soon!!

[hmong337]

My prayers for a TRUE end user flasher has been answered! I think I know what I'll be getting.

I want to know, will this be tested on any sort of force inducted fa20 anytime soon? It would be cool to see how it does in that sort of situation. Finding the opinion of what the tuner thought of OpenFlash while tuning a turbo system would be excellent too.

Thank you for the genius product!

[Rockstar Lifestyle]

I'm right there with you. I've been following the progress on this. OpenFlash certainly has my attention as a potential tuning solution. This is impressive work!

[SPCorBUST]

looks like I have some excellent reading material for lunch today. super stoked.

EDIT: damn, couldn't wait. took 20 minutes, but it was a good read!

[Captain Insano]

Absolutely awesome post, thread and info for the community.

[Anorexic 86]

I first met Shiv 10 years ago... he had just got his hands on the new to U.S. market Mitsubishi Lancer Evolution and was already tuning it with some high lift HKS cams and a few other goodies.

I learned alot that day... he's a class act and a gentleman.

[regal]

With the final tune and bolt-ons is there less tip-in knock than stock ?