RomRaider

For some reason my initial logs didn't take, but here is one of it idling lean. This is open loop. Idle does seem to climb to get leaner than I initially thought, though AFRs are better when driving than with old MAF. This is now including the LC-2 AFR data. I will log more when I drive home and post those later. I think since reinstating the stock MAF scaling the car is building boost faster and overboosting a little bit from the boost curve I originally setup for it, was hoping to capture that in the logs but I will later.

It looks like the AFRs are 17% higher than they should be. If my new O2 sensor still fluctuates at idle and cruise then I am inclined to think it is due to trying to make such a large correction to the AFRs. If this is the case I am going to apply a 17% correction to the MAF and then try running the closed loop rescaling operation and hope it fluctuates less so that the tool can apply a proper scaling.

And if all fails then I will look for more leaks.

Will update later.

Made some logs, tweaked the tune, and stopped forcing it into open loop. Still a little stuttery and fluctuating during closed loop cruise, but much better than with the old MAF. I'll swap in the new O2 and see if it goes away.

3rd gear pull, I think my dip is more extreme than it needs to be to control the surge, I am going to mellow it out. Also ignore the weird spikes in the boost I don't think they were real.



4th gear responds well from a punch at 4k. My turbo dynamics pulls some duty cycle out to catch it and the boost does not drop too much. I have TD adjusted so it becomes really fine right near 0 boost error.



And a partial 3-4-5 pull to see how close to target boost it gets. I got WOT pretty high in the RPMs already so the turbo doesn't have much to give until I shift into 4th. TD saves me here again. Right around 4750RPMs. I think I can mess with the Max WGDC to match what the TD is making the WGDC hit.



I think I am really close to the Max WGDC curve that I need for my setup. It is also 10* out today so the ECU is pulling Max WGDC. I will take more logs this week when it is warmer and compare them to todays logs.

My AFRs are decent, ECU isn't pulling a ton of timing, I am going past 100%Duty Cycle with the injectors, but it doesn't lean out. I am sure that is not great for longevity though.

I decided to do an experiment to see if I could make the TD Continuous handle the boost without any restriction from the WGDC Max, so I set it to 99.6 across the board. I wanted to get the TD nailed down as well as have an ideal WGDC curve drawn by the TD for each gear. The TD barely needs to do anything even with the WGDC set to 99.6 in regards to pulling duty cycle. Today was a lot warmer than the previous graphs I posted, probably around 37F vs 10F last time. I also reduced the % that the IAT compensation pulls because it was lowering the max duty cycle when the turbo was not even hitting max boost. I may 0 it out, then re-tune in summer and see what % WGDC difference is and input that into the IAT compensation. I also set the boost higher from 0-3500 rpm to help it hit the target boost faster.

I wonder if the new T in the vacuum lines going to the WG and BCS is dirty again, because it seems like it should be boosting higher with the WGDC at 99%. Maybe the turbo is worn and it is making it boost lower than what I see in other people's WRXs. I will install my 3 Port BCS and see what happens, and I may buy the new CHRA I posted earlier if it still has trouble boosting higher.

Here is a pull from December 21st, which was a cold day, but 3rd gear hits over 21PSI, and I can't get anywhere near that currently with the WGDC at 99%. Makes me think that the vacuum lines need to be cleaned. Using the 3 port should help with this.

Overboosting in third:


Here are pulls from today.

Full 3rd Gear Pull:


3rd Gear Partial:


4th Gear:


1-2-3-4:


3-4-5:


Any insights? Let me know if my graphs are hard to understand but I try and clean them up and clearly label them.

Edit: looking at the final two graphs it is confusing me because the first graph is definitely a 1 - 2 - 3 - 4, I had a long empty on-ramp to myself, but the 3 - 4 - 5 is almost definitely a 3 - 4 - 5 because I do not remember shifting down to 2nd on the highway... but the boost curves are identical but offset... I will take some more logs and pick an order to record them in.

I did a few different tests yesterday and today to see why I am not hitting full boost in 3rd when before (December) I would overboost.

1) I replaced the vacuum lines from the turbo to wastegate and BCS. I had previously cleaned them but not replaced. I cleaned the T but it looked fine.

2) I disconnected the WG from the vacuum lines and saw how much boost it would make and where.

3) I bypassed the BCS and hooked the WG and Turbo pressure line directly to the turbo inlet to see if that would cause worse performance than in scenario #2.

I learned that:

1) Cleaning the T and replacing the lines made no difference.

2) Disconnecting the WG from vacuum I would spool fast and hit full boost in 3rd, but it stayed under 20psi without any input from the WG.

3) Bypassing the BCS performed nearly identically to disconnecting the WG from vacuum entirely. This is good news as I know I am not being limited by the vacuum lines themselves.

Possible reasons that it is not overboosting anymore are that the turbo is wearing out, the WG diaphragm is not pulling the WG closed as much anymore, or I have a pre-turbo exhaust leak. I will seafoam the car to rule out the pre-turbo exhaust leak. I will probably get a new CHRA as the turbo is old, and I am going to use larger vacuum lines between the BCS and turbo inlet so there is no restriction with getting pressure away from the WG. I am also going to start tuning with the 3 port solenoid. Have to make a bracket for it. I will also add a few turns to the WG arm. Might swap in a WG from one of my parts turbos.

Anyways, here is the data, I think I am going to turn my boost down to 17.5PSI because I am getting some feedback knock correction and I seem to be running the turbo at as much boost as it will make. It also doesn't really feel faster at 19.5 PSI than it does at 18PSI. I have my TD pretty well sorted, especially recovering from an overboost event, and I think it will work out nicely.

Test #1: Replaced Vacuum Lines

Clean 3rd Gear Pull to 6K



Test #2: WG sees no Vacuum

2nd Gear Pull



3rd Gear Pull



Test #3: WG and Turbo outlet Direct to Turbo Inlet

2nd Gear Pull



3rd Gear Pull to 5700



It can be seen there is a clear improvement in Tests 2 and 3, and the graphs for 2 and 3 look nearly identical. Also, I should have omitted the WGDC max and WGDC data from test 2 and 3 as it is not actually doing anything to affect the turbo in those scenarios.

I have done a ton of logging, graphing, tweaking, repeat. After my tests with the WG vacuum line disconnected and connected to the turbo inlet, I did a little thinking and I came up with this simple modification.



I replaced the original bleed line from the EBCS (yellow arrow) to the inlet with a line that is 1/4" the entire way. Any flow through any tube is going to cause back pressure, I think its hagen poiseuille's equation to figure out how much of a restriction a tube will be in terms of pressure drop. Post turbo, but pre EBCS and you want the lines as small as possible, this makes them react to the pressure change quickly instead of having to fill a volume with pressure and then vent that entire volume of pressure with the EBCS. However, after the EBCS would ideally be vented to open air so that the EBCS can do its job of preventing the WG from seeing boost pressure. It seemed to me that the smaller diameter line was going to adversely affect the EBCS from doing its job.

Well I think my modification succeeded as I am able to hit 17PSI in 2nd now which is awesome. I have very decent boost curves for basically having TD draw the boost curve. It is cool how the different curves look depending on the gear the car is in. I had to mess around with the tune a bunch because it was massively overboosting in all the other gears after the change. I was also getting a lot of partial throttle overboosting that I managed to tame as well. The car was driving great before, but now the boost hits hard. I need to pull a little more WGDC% at the intake temps that I am seeing as I am hitting right around/ under 20psi in 3-5 for a short window. I am not really getting any knock events and my AFRs are solid, but I worry about the turbo spinning too fast. I try and hold the boost right around 17 until it tapers. I think if I pull a few more percent at 30* intake temps that I can tame the 20psi hit. I can also make the TD a little more aggressive at pulling boost when it is over boosting, right now it pulls boost slowly unless its overboosting by more than 1.5 PSI.

Here are my current relevant tables:



And here are the boost curves and other data:

2nd Gear from 2.5K



2nd Gear from 3K



3rd Gear



4th Gear (ignore the phantom spikes)



5th Gear



As I said, note the orange WGDC curves per gear, pretty neat how they progress with the additional load of higher gears.

I have been trying to get Turbo Dynamics to control everything in every situation, but I have found that Turbo Dynamics has a hard time controlling major boost spikes in higher gears around 3000 rpms, and if I make it aggressive enough to control that, then it makes it very hard to hold boost for a while.

I also found it is knocking the most in these instances, so I relaxed the Turbo Dynamics and smoothed it out and made a Max WGDC curve that looks like this down low.



by 2850RPMs MAX WGDC is back up to 97, so in second and third where the RPMs go by much faster, it still seems to spool quickly and be pretty punchy. In 4th and 5th where I may be passing people without downshifting and overboosting down low/knocking, it helps smooth the spool out and it will hit full boost right around 3000RPMs. If you punch it from 2800 and up it will slightly overboost and TD will control it and it will hold boost for a while. I am still playing with it and will adjust it as necessary. The way I drive, I don't notice it limiting me down low as I am generally over 3000RPMs if pushing the car.

Here is my TD curve, you can see it is very aggressive when boost error is over 3PSI, and is mellow when it is under 3PSI , but pretty much is always trying to make it go slightly over, there is a window where -TD will not do anything, and then it ramps up, and by -2PSI error slows down as by the time you have hit -2PSI of error, a lot of WGDC has already been pulled and it doesn't need to keep getting more aggressive.



Also still a work in progress, I may be able to pull more negative correction out since it doesn't overboost nearly as aggressively as it was without limiting MAX WGDC below 2850 RPMs.

But I have not been able to log as I think my turbo is screaming, it sounds like a metal shriek, like a bearing that has gone bad. Still performs fine. I ordered a new CHRA. There is a chance it is a leak, but I have had many pre-turbo as well as intake leaks and none have ever sounded like this or been nearly as loud. I will be replacing gaskets with the center section, so it should hopefully fix the issue. I will smoke test the car again too.

Edit to say it was a bolt that got loose between headers and up-pipe and it was a combination of the whistle of air escaping and the rattling of the bolt. Gasket blew out and performance dropped. Oh well.

Hoping that my ebay headers are trying to copy Tomei, ordered the Grimmspeed MLS gasket for 3 bolt up-pipe and I hope it fits/can be modified to fit. Of course the ebay seller had no clue where to source the original gasket from that came with the headers.

If all else fails I will order a sheet of 14 gauge copper and cut a gasket out of it, but I don't think it will get to that point.

Probably will put the CHRA in a spare turbo I have and install the new up-pipe gasket at the same time as the "new" turbo. Got new JIS bolts for the up-pipe with split lock washers and JIS flanged nuts in the stock thread to clamp everything back together for now, obviously still leaking so I am babying it for the time being, might 0 out the WGDC Max just to remove temptation.

I am considering getting the PLM ELHs and will mate them to my grimmspeed 2 bolt up-pipe that is not currently installed, but I have to swap oil pans to make it happen. I don't trust these headers long term, and they fit like crap. Sound great though and perform decently, but the up-pipe diameter is way too large and no way they are going to last. I want to start collecting parts so I have them when these prematurely break in half.

Finally figured out my lean idle and cruise, I pulled the intercooler off to release the turbo so when up-pipe was unbolted from the headers I could slip the new gasket in, and when I had the intercooler off I was able to get to and tighten all of the bolts that I used to plug the TGV shaft holes. I also made sure all the other hose clamps under the intercooler were tight and they were.

For the people of the internet, the Tomei 3 bolt up-pipe uses the same sized and shaped gasket as the Ebay ones.

Old gasket on top of new MLS extra thiccc Grimmspeed gasket for Tomei application:



Put in new bolts with lock washers too, and I replaced the O2 sensor. The O2 sensor I used is from fastwrx.com https://www.fastwrx.com/products/denso-o2-sensor-front-2002-05-wrx-2-0?_pos=1&_sid=924db554d&_ss=r. It is not the stock sensor but it supposed to be superior. I have actually ordered the same sensor from Amazon many times for much less money, it is Denso 234-9011. The sensor from Amazon would always come in the correct box, so I never thought much of it, but the sensor from fastwrx.com was in a bag that also had the correct number printed on it, and the sensor had a couple minor physical differences from the one from Amazon. In the future I will order directly from Denso or from fastwrx.com as I believe that the cheaper sensors may be counterfeit. I had installed 2 of the Amazon ones since getting the motor back together and neither made the car run correctly.

The new O2 sensor causes the AFR values to be much more stable, and without an intake manifold leak, it idles and cruises right around 13.8:1 according to my wideband. It is much nicer to drive without all the little misfires and stumbles caused by the lean condition. I am unsure why it never really goes above 13.8:1 unless I am engine braking, I need to see what the ECU thinks is happening. Possibly I need to adjust the voltage that correlates to the 14.4AFR that it should be targeting. Also a possibility is the LC-2 is out of cal. I have about 5000 miles on the wideband, it says it needs recalibrating at 10,000 miles, I may recal early just for peace of mind.

And finally I need to start accessing the FLKC values to see if my timing needs further adjusting. I think it is logworks I have to install but I need to do some research.

I have done some open loop MAF scaling but have avoided closed loop due to the lean condition, so now I can finally tackle that and get it dialed in, and then install the new pink injectors and scale for those. I haven't had any issues with running lean, but I am over 100% duty cycle and I am not comfortable with that long term.

Car is nice and quick again, and finally running right at cruise too for the first time since performing the TGV deletes.

Turns out when I stopped forcing the car to run open loop, I accidentally adjusted the closed loop AFR target down to 13.36. There are some other compensations occurring that brought it up to the 13.8 I was seeing. I have since adjusted it to 14.1:1, which ends up being around 14.5-14.8:1 after some compensations are added by the ECU. I thinkkkkk these compensations are due to my rear O2 but I am not positive, however I think this because just recently the car has been throwing a code for the rear O2 saying that the heater circuit is bad. I bought a new Denso one to throw in there. I know I could have run an output from the wideband and stopped the codes from being displayed but all my wiring is inside the cabin and I did not want to deal with running a wire from in the cabin to that connector. The old rear O2 lasted a very long time. I may put it in a spark plug spacer to further help protect it like the wideband with its extra long bung. Once the rear O2 goes in I will be scaling the MAF correctly finally.

Been playing with Turbo Dynamics recently, I just hate limiting what second gear can hit for boost with WGDC Max, but I also don't want the TD to be so slow that boost comes on more slowly. I have it super aggressive until 7psi boost error and then it mellows out very dramatically to help the WGDC climb more slowly as it gets close to target boost. It gives that nice strong surge feeling while overshooting boost less. I need to maybe move it to 9psi of boost error because even at 7psi error I get a little too much overboosting in a 5th gear passing scenario (Pull from 3K). Now that I don't have any more leaks in intake or exhaust, I hope my tuning can be more consistent.

I have also been taking lots of logs on days of differing temps so I can nail my WGDC IAT compensation.

I also got an STI intercooler that I will install so I can push a little more boost.

Oh and I picked up a 2.5L crank from an STI to complete my parts gathering for a forged stroker build

I have been logging, graphing, and retuning like crazy, I think I have had over 10 iterations since the last data I posted. I would get good logs and then go out the next day and I would be over and underboosting suddenly due to different factors. I wanted to get TD nailed down so that I did not have to rely on MAX WGDC for controlling boost.

I still have MAX WGDC progressive below 3K RPMs because that has been working nicely for me to limit overboosting in 5th and 4th down low. Those numbers have not changed since I last posted my graph of that.

I have had a tuning goal to be able to hit full boost (18.5psi) in 2nd without peaking more than 1.5psi above that in the remaining gears. I think I have finally achieved this.

I noticed that while my car was really punchy, I was not holding boost as long as I could because if the turbo loses some spool to recover from overboosting, it never really gains it back.

My TD was way too aggressive before, I had it at 7 until 3psi error, then 5 psi error, then 7 psi error, then 10 psi error. I kept moving it back to stop from overboosting, but it spooled so fast with such an aggressive value that I wanted to be able to enjoy it when punching the gas from cruise. But I found less is more and went back to even more conservative than stock TD values for positive boost error.

For negative boost error my scheme of having a dead zone where no WGDC would be pulled actually has been working really well, so I continued that forwards.

Here is where I came from vs where I am at now, less is more in this case. Spool is less punchy but it actually holds boost now so overall it feels much faster.



I also have been iterating the WGDC IAT compensation, I started out conservative and have been pulling less and less WGDC as IATs drop. I had the car tuned to control boost with WGDC in order to come up with the current values.



And finally some pulls, I have some minor tweaking to do but the car seems pretty happy all the time and holds boost really well. I hit 18.5 psi in second which it holds from 4-5k and 19.99 psi in 3rd thru 5th where it quickly and smoothly settles down to 18.5 psi.

2nd


3rd


4th


5th


I have learned a lot from this process, most importantly to not try and implement too many changes at once, and not to change anything too drastically. It seems like it will take longer to make minor changes one at a time and log, but in the end it takes less revisions to where you are trying to get to. I am glad I was kind of all over the place in my tuning because I figured out how all different strategies work to get you similar results, but only a few strategies will work in all situations. I wish that TD had an IAT multiplier because I feel that would make the tune even better, but it is what it is.

Also a huge thank you to everyone who has contributed to making this opensource tuning possible, it is something I would not be able to do without all the hard work of the developers.

Tune is holding up in various conditions, no overboosting, still hitting over 18psi in second, car feels very happy. I will update with my current ROM, and disclaimer that I borrowed the timing table from Testes tune, but updated it to reduce some knock in certain high load areas. I would love to get on a dyno and really dial in the timing table because I am sure it is not optimized as well as it could be.

Picked up a much newer transmission today as mine gets louder by the day. Really glad I picked it up ahead of time.

Definitely have noticed a little overboosting in fourth starting from lower revs, I think my WGDC IAT compensation is probably so aggressive because the overall WGDC needs to be dropped, and now that temps are warmer I am experiencing no WGDC compensation and getting a little overboosting. It is nothing crazy, just overboosts to the point where boost drops more than I want and I get that dip in power before it comes back up to target boost.

I am going to do a little math when I adjust WGDC so that I can adjust the IAT compensation values to keep the WGDC what they have been in colder temps because I really had it nailed when it was colder out.

Bravo mate.. What a great read!!
Im about to go on this exact adventure, but i fear my road will be longer as my undedrstandings are not to the same level of yours.
Ive been slowly returning my Bug eye GG to stock.. Last thing is to remove the stupid manual boost T adjuster... But it is fun winding it up to around 18psi!

Ive finally got my logging setup sorted.. Here are my first two logs, i share them not for advice but for general use and comments..

Thanks for the comments, glad documenting my progress can help you get started. I was going to update this soon as I am going to be getting familiar with Airboy's spreadsheet to really dial my timing in, and I will record everything here. My car ran great for 40,000 miles since starting tuning it and then blew a headgasket when the radiator let go and dumped all the coolant out very rapidly. I have an EJ206 I just swapped in which has a higher compression ratio than stock, which is why I am going to take a deep dive into tuning timing.

If there is one thing I recommend to you, it is to get a 3 port boost solenoid, I bought the one from Mach V motorsports, it was well priced and is high quality. It can control boost way better than the stock solenoid, very easy to hit target boost quickly and it can react quickly to over-boosting without causing fluctuations. Tuning boost is definitely one of the most annoying parts, turbo dynamics is where it is at, I only use waste gate duty tables to limit boost below 3000 rpms, otherwise turbo dynamics fully controls my boost.

I also recommend forcing the car into open loop by changing the open loop targets from anything below 13.6 (stock) to 14.7, and then lowering open loop fueling targets that are 14.7 stock to 14.4. This will make the car ignore the O2 sensor, but you can still log it, and if you just unplug it, it can damage the sensor. Change your fueling map to be like one of 3 values, either 14.4, 12ish, or 10.5, then tuning your MAF scaling becomes wicked easy as the value is always going to be one of three, and if it is not you know which way to tweak the MAF scaling until it is always very close to what it should be. Once dialed, return your open loop transition to 13.6, and return cells in the fueling map modified to 14.4 to 14.7.

Timing wise I have only logged FLKC and FBKC as well as used learning view to see places it consistently pulls timing, and to go in and pull timing in those spots myself. I am starting to tune timing using Airboy's spreadsheet, det cans (headphones to hear knock) and I think I am going to get an EGT sensor to log EGTs to make sure my timing is not too low anywhere.

Good luck!

Also since my previous posts, I ported the TGVs, I had left the vane in there and just deleted the butterflies originally, installed some JDM pinks, STI top mount intercooler, and I am running a set of HKS equal length headers which greatly smoothed out spool and gave me more power pretty much everywhere with around the same power up top. The motor feels way smoother with the ELHs, I do miss the rumble of the long tube UELHs which sounded nasty, but the characteristics of the motor are so improved that it does not matter, and it still sounds good, doesn't sound like a Honda. Tuning the pinks went well, but I still need to tune the ECT compensations because it is way too rich during warm up and the AFRs fluctuate as a result. Did not matter much in the summer but as it is getting colder it is getting annoying. Also have an STI pan and a Killer B pickup now, but I wish I had got the windage tray too, was tight on cash. And most recently it is now an EJ206, piston slaps like crazy when cold and totally normal once hot, but my research says this is normal, they have different pistons.

Looking at your logs, your IAM is not happy sitting there at 12 (unless you just reset the ECU). It did not pull any timing, only added, but still it should be at 16 if there are not other areas in your map knocking badly. I would figure out where it is pulling timing and address those areas.

You are only running about 15lbs of boost here btw. I would change your logging parameter to the corrected manifold pressure and not absolute, I would also log fuel injector duty cycle over latency and bpw.

You are also logging a lot of extra values you don't need to that are lowering your resolution. A/F correction, A/F learning, battery voltage, coolant temp, MAF sensor voltage, CL/OL status, Closed loop fueling target, Fuel Injector latency, MAP voltage, all unnecessary. Your fueling is quite rich as well, it seems like your wideband does not read lower than 9.6, but I bet it is dropping possibly into the 8s. Not sure why your targets are dropping into the low 9s, I never saw lower than 9.9 with a stock map, but you are safe up to 10.7:1, and some people push as far as 11.5:1. I personally target about 11:1 at/near max load columns.

Just pick the necessary parameters to log for the situation you are trying to tune, my suggestions are related to logging pulls. If there were other issues you are chasing down, then you can log more parameters and remove some of the ones being logged currently. For example to monitor my car right now I am mostly concerned with knock, I am monitoring IAM, feedback knock, fine learning knock, rpm, and engine load. I also have a wideband gauge that tells me if my AFRs are correct, otherwise I would log those too. But what I am looking for is what cells in the Load vs RPM timing table would need to be adjusted, if there are certain cell locations where I see fine learning knock or often have feedback knock. I don't really need to care about anything else as it is not affecting the table I am looking to adjust. If I was doing pulls currently I would definitely be logging AFRs and boost as well to see if either of those were the cause of knock, too much boost at low RPMs always results in knock for me, and if my AFRs are a little leaner than they could be, I will adjust that first before adjusting the timing. I can also look at the timing map in 3D to see if there are any weird high spots where I am having my issues. Also targeting too much boost at midrange or higher rpms will put the turbo way out of its efficiency zone and it will be getting superheated air prone to detonation, there is no way to log air temps post intercooler (Stock) so sometimes just lowering the boost in those spots and logging again to see if it helped is the answer. You can find efficiency maps for the TD04 online, I try and hold 14psi at redline which is pushing it IMO and probably why I have had to pull a few degrees up top in my old map. Once again another reason to use Airboy's spreadsheet and see how much my HP is being affected by running a little more timing and a little less boost up top to try and find a happier medium.