Boost Control with RomRaider logger

Hello all!

I'm currently through my journey and at the point where I am tuning for boost.

I've watched Andre's lectures on Boost tuning, but I have a few questions and frankly; this seems to be the most difficult part of tuning that I've come across so far.

So I have a few logs that show target boost, boost error, TD proportional and integral (my maps do NOT have Derivatives), manifold relative pressure(corrected) as well Primary WGDC, RPM, and TPS (percentage).

I have a conservative boost target that I would like to reach and hold for my build; as well as a target RPM; say 3400-3600RPM to begin building boost. I have a FP blue with which I have calculated a good 76% efficiency with ~18PSI and at ~360crank HP. Thats my goal.

Now where I live is 5600ft elevation (1706m) and the pressure here is ~12.00psi (0.827Bar). This is important, lol!

My pressure logs are all over the place. I can not get an objective reading from which to adjust my tune. Furthermore, I honestly don't know what to adjust first. Do I adjust the Wastegate duty cycle? or the Max Wastegate duty cycle? TD proportional? or the TD integral?

I have uploaded pictures of both the compressor map of my turbo and my target boost from ECU Flash as well as a recent pull in 3rd gear to about 6400RPM.

I've gone through the videos and completed the quizes but still have no clue about how to adjust (or even what to adjust) when it comes to boost.

Can someone please help me :(

Hello J P,

Starting from a stock tune, I usually only have to change Target Boost and both Wastegate Duty cycle maps to achieve good results. Most OEM will have put a lot of effort in calibrating PID and compensation maps in most condition already. Unfortunately, I can't confirm with that much of elevation as I live near sea level.

First thing I noticed from your log is that you didn't use the same scaling for Target Boost and Manifold relative pressure in the graph, which makes it harder to analyze. A quick glance would tell you that you are operating below target when in case you manifold pressure is above target. The reason why your boost target differ from the MAP Target Boost is probably that you have compensation affecting the final boost target (baro compensation).

With that being said, looking at the compressor map you provided, you wrote that a pressure ratio of 2 would be 14.7psi, while in fact it would be 12psi. Like the name implies, it's a ratio, between outlet pressure / inlet pressure. As you have only 12psi of baro pressure, it will take only 12psi of boost to have ratio of 2. 14.7psi of boost would give you a ratio of 2.225 which changes things quite a lot. 18psi of boost would give a pressure ratio of 2.5, not a bad thing as it moves toward peak efficiency island of the turbo map.

Most problem with boost control comes from boost target being wrong. The boost control is a closed loop, constantly trying to achieve boost target. If you target a boost level that the turbo can't achieve, the PID will adjust Wastegate DC in order to reach that boost level. In the event that the target can't be reach (let's say you are asking for too much boost at too low RPM, during initial spool up), the PID will be maxed out and will overshoot the target and oscillate. It's not a PID problem, it's a target problem. Like anything closed loop, the less work the PID as to do, the better it is. Looking at your graph, the Integral is working pretty hard right now.

My process is quite simple, I log Manifold Pressure (corrected) and target boost, and Wastegate DC. During spool up I check what my manifold pressure and adjust my boost target accordingly. Then I check what my wastegate DC is, and adjust my wastegate tables so my PID system is not working too hard. I set my initial wastegate DC table 10% lower than my max wastegate table.

Perhaps provide a log. I'm not clear on what "all over the place" means. If you are overshooting target and then boost control over reacts, it's likely because you're using stock PID values with a 3 port solenoid. On that particular application, halving the PID gains is a decent start.

Thank you Frank and Mike!

I went ahead and decreased my PID about 30% from a previous resource.

I also zeroed out my initial and max wastegate duty in order to focus simply on Target boost.

I did this to target my turbo wastegate spring pressure which is 18PSI (the wastegate starts to crack open at 17.8-18PSI)

Since Andre mentioned in the lesson that open loop boost control is easier to set than closed loop. I think it best that I focus on that first.

I have smoothed out and input some conservative numbers for target boost in open loop; simply to see if my turbo can reach those targets at that specified RPM. Based on the research online for my turbo and my car I've found that a good starting point would be at about ~75% throttle, 3600RPM, with a target of 18PSI.

Frank; you bring up a good point about barometric pressure having some sort of compensatory effect on the logged manifold pressure. Moving forward, I will make sure that I log the proper units for pressure (be it absolute or relative), see if I hit my targets in open loop, and if not adjust accordingly. I like the idea of having the PID system do as little work as possible while in closed loop. That, to me, seems like an efficiently tuned map and I would like to have a goal of that.

I would like to keep this dialog going in order to get this boost thing figured out. I have also attached my current boost target table and will post up a log tomorrow.

Thanks again, gents ^_^

Frank, how do I "scale" correctly the Manifold Pressure and Target Boost so that they align?

Yup with any system, boost or otherwise, you always want your feed forward values to get you as close to target as possible, leaving less for the PID system to clean up. That will result in best and smoothest operation.

Frank or Mike:

If Andre has the example of a 3D table using KPA as Manifold pressure by my car has Throttle position instead; how do I reconcile adjusting the duty cycle in open loop for throttle position in the increments that Andre explains in the boost open loop video when I only have throttle position and not manifold pressure.(Im using ECU Flash)?

JP,

Your WG duty values and your boost targets are both based on throttle position, so you adjust the duty cycle values until you hit your boost target at a given throttle angle.

Most folks won't actually tune the entire table, and interpolate downward, but that's up to you. For example you could tune 59% throttle, then the higher points, and then interpolate your duty cycle values below 59% throttle down to zero at the throttle angle where you only want to achieve wastegate boost pressure.

This was on of my WOT pulls (no duty cycle), simply the internal wastegate.

The manifold relative pressure(corrected) is 14.33psi @ 99.75% throttle.

Actually, all four of my pulls maxes out the manifold relative pressure(corrected) at ~14.3 psi @ ~95-99%

But I also have an AEM Boost pressure gauge on my dash; and I vividly recall it flashing 18PSI every time I did a WOT pull...

(Thats the same pressure that I have my internal wastegate spring set to)

So why does my ECU show ~14PSI when my boost gauge correctly is showing me my boost pressure at 18 psi on wide open throttle pulls?

(screen shot attached)

To me, the descriptions of the channels doesn't make me think the sensors are measuring the same thing. The AEM is probably located post compressor, but before the throttle body -- this is boost pressure. The ECU is measuring the absolute pressure in the manifold, and possibly correcting it for the barometric pressure at your altitude (12psi according to your first post) or the correction might be simply subtracting 1 bar - thus is might read 2.5 psi or so low.. I can believe there is a pressure drop across your throttle body even at wide open throttle. And your two gauges are measuring that.

Want to know for sure, install two sensors in those two locations and connect them to the same system for data logging.

Hey David!

Very good observation; and one that I may have totally missed if you didn’t mention it.

I currently have my boost gauge/line installed directly on the intake manifold (after the throttle body). I’ve attached a picture of the location.

Where can I relocate the actual boost gauge to reflect the most accurate boost level; both for logging and viewing?

Also, how come my actual manifold pressure is measuring ~14PSI when I have a boost target set to 18PSI @ 6500rpm and 99% throttle angle?

I have installed the boost pressure sensor on the hose going to the boost controller.

Both your sensors are reading from the same source, the intake manifold plenum.

Have you verified your sensor multiplier and offset are correct? It seems that may be part of the problem.

I would compare pressure at ignition on engine off to simplify things and verify any issues.

I don't recall which Subaru you have off the top of my head, but there are various compensation tables related to barometric pressure that function in different ways depending on the year/model.

07+ STI or 08-14 WRX adds 14.7 psi to boost target, but then the baro comp value gets applied to that boost target + 14.7 psi value.

For example if baro is 12 and there's a 15 psi boost target +14.7 psi = 29.7 psia target

-20% comp at 12 psi baro is applied as 29.7 * .8 = 23.76

23.76 - baro 12 psi = 11.76 psi boost relative to baro of 12 psi, and less than that relative to sea level.

Depending on how your gauge reads that may also confuse things. Gauges often aren't baro referenced, so they assume standard sea level pressure is 0 and ready +/- above and below it.

1st gear, coolant temp, intake temp boost target comps all impact the end result as well.

If you don't have one of those model years, please tell us what you have so we know what boost control system you're working with.

Hi Mike,

Yes, I see what your saying. And yes, I have spoken to Calib over at AEM regarding the installed 3.5bar pressure sensor.

(offset -6.25, multiplier 12.501) this was how it was explained to me:

"The 3rd attachment you provided, labeled “Voltage Output” was mistakenly typed with “PSIa”. It is an absolute pressure sensor, hence the negative PSIg values. -14.7psig = 0.0psia. To alter the transfer function from the 2nd attachment for PSIa instead of PSIg, you would simply change the offset from -20.95 to -6.25, a 14.7psi difference to eliminate the assumed barometric pressure. If you want your reading to show the actual barometric absolute pressure without the engine running, set the offset at -6.25. If you want the reading to show ~0 without the engine running, set the offset at -20.95.

Absolute pressure simply does not assume a barometric pressure so that it is universally understood at all elevations as true and correct. “Gauge” pressure just assumes barometric and uses that as the “zero” point.

To answer your question, I have a USDM 2004 Subaru STi with a 3 port boost controller. My AEM boost gauge has never shown what the log parameters show; regardless of what manifold parameters I use...absolute or relative. I did go over the compensations on ECUflash for my specific model over the past two days, racking my brain as to why the logged parameters show something different than the actual gauge.

(You've helped me shed some light on that). The compensations I have for target boost are: ECT, 1st Gear, and ATM Pressure. I haven't touched any of them and to complicate matters worse; the 1st gear comp is set to disable after 25mph. I'm leaning towards the ATM compensations (which I've attached below) that may be the main culprit. I am also thinking of relocating the boost gauge from the manifold to the 3-port boost solenoid is as I think that may give me a more accurate representation of boost.

I've also checked for leaks with this Amazon smoke/leak tester I have; there are none.

So, as it stands now on my pulls yesterday. The log shows a max of 14.33PSI, but my gauge is showing 18PSI on the screen. I can tell you that the pull did NOT feel like 18PSI. (I've also attached the .csv file; just in case you'd/ or any one would like to take a look)

I've been following the videos, and I feel somewhat confident regarding my AFR, fuel tuning but regarding boost is intimidating and no matter what I try I don't get the results I'm looking for.

It seems like I'm spinning my wheels (pun intended).

~Julio

04 STI is different. Same 15 psi boost target table value, baro of 12 psi like I used above.

((baro * boost target comp baro multiplier) + boost target comp baro offset) * (boost target table + 14.7 psi)

((12 * 0.0521) + .25) * (15 + 14.7)

0.8752 * 29.7 = 25.99344

25.99344 - 14.7 = 11.29 psi

The final boost target is loggable in COBB, but perhaps not in open source.

TD boost error will show you how far you are from the target though, so you do have that and can always use it to see what the target is. Target = actual + TD boost error.

I think a good test would be to bolt/plug the stock MAP sensor back in, set the sensor scaling to stock, and see if you get the same readings or not. Then you'll know if your other MAP sensor/install/wiring/scaling is working correctly or not.

04 STI ECU is pretty barebones, doesn't have as many comps. Also the stock coolant temp table is already zeroed out, so you really just have the 1st gear comp and baro.

Oh bomb!! Having you work through the calculation like that really helps me see the comps! Thank You!

(and you have a great memory! The stock ECU does NOT have ECT comps; its a zero'd out table stock) I forgot to list that in my last post but you picked it up regardless! Neat trick!

I like the idea of putting in the stock MAP and compare the readings; that's easy to do. I really don't want to mess with the TD tables outside of adjusting them for a 3-port in place of a 2-port since Andre's boost lecture has us setting up open loop prior to setting up closed look and adjusting tables.

Ugh...I'm honestly, considering going with a Link G4+ plug and play since I keep running into limitations and problems with this stock ECU. Oh well, Its a labor of love I suppose.

Thanks again for the quick responses, and the help. I appreciate you and your time.

Yea sorry I didn't mean change the TD comps, I was just saying when you look at the logged TD boost error, you can derive the final boost target from it and the current logged MAP value, both of which I know you can datalog. It was a suggestion in case you can't log the final calculated boost target directly with open source...I don't know if you can or not.

Ah! yes, I see what you mean now!

RomRaider DOES has a parameter for Target Boost; so that can potentially make it more straight forward.

Regardless, having both options gives me hope!

Thanks again.

You're welcome!

@From_Zero79

Might be a bit late to the party here, but I was struggling with tuning boost control on my Subie as well initially but some stuff that helped me get it working was:

- To start off I just completely ignored PID control when tuning boost. If you have a 3 port yeah probably best to just knock all the PID values down by like 30% to 50% to start off since the boost control system will be much more sensitive with a 3 port than a 2 port, but outside of that I made my first goal to try to figure out what WGDC %age values in the WGDC table I could expect to yield what boost target for whatever set up you have. From there you can construct a WGDC table that should closely approximate the target boost table. If you have these two tables set up well... you should see PID control in your logs being less active over-all as it wont be having to work as hard to dynamically adjust WGDC above or below whatever you set them at to reach your target boost.

- Once you have a good relationship between your WGDC tables and your target boost tables... its just a matter of looking at your logs and seeing if your PID (PI only in subaru's case) are over-reacting or under-reacting to whatever (hopefully minor) boost errors are occurring. For example if you are under boosting a bit and you see the P or I values spike up a lot and then it causes an overboost... probably the PI values are too high since you just want them to push you up to target but not over it. Inverse applies if you are initially over boosting with your specified WGDC values. If you see the I value consistently staying at significant + or - values in some cases... this is a good indicator that some cell in your WGDC table is probably too far off to begin with and you should go back and correct that first rather than relying on the PID control to get you to your target boost.

One big caveat to all of the above though....I think you mentioned the spring pressure on your turbo is 18PSI ... but also looks like your max target boost in your table is also around 18PSI. I believe spring pressure is generally specified as gauge pressure which means 18PSI above atmospheric pressure. In terms of tuning boost on your car... this basically means you dont actually have any control over boost regardless of what you do with your WGDC table or your PID control as all of the boost targets in your table are actually less than your spring pressure. If you want to be able to actually control boost between 0 and 18 PSI you would need a softer spring, otherwise the turbo is just doing whatever it does in that range.

@GSum,

This!! ^_^

This is the exact information that I was looking for! Thank you so much for taking the time to reply to my post...and no, never late to the party my friend!

I was wondering about the erratic nature and inconsistency of my boost gauge then you mentioned atmospheric pressure and it seemed to click.

I am currently at ~12psi ATM; give or take, and yes; my internal wastegate spring pressure is 18 PSI. So it would make sense that the wastegate would act like a manual boost controller. One thing I could do is mess with the wastegate lever arm. Screw it clockwise to DECREASE the spring pressure. The TiAL wastegate is MAX of 18 PSI so I have some room to play, thankfully.

As for the AEM digital boost gauge read out in the car, I switched the pressure sensor to a Honeywell (gauge pressure) one and now it reads 0.00PSI; which is what I was after all along.

I feel that I'm making progress, albeit slowly but your post gave me a good indication of how to go about tuning boost. I'll do a couple of runs with the PI shut off and just measure the wastegate spring pressure and see what sort of numbers I get, whilst making sure I keep my DET cans connected and listen for knock.

If I have any further questions I'll be sure to reach out again. This is a long road but I feel that I may in fact finally be getting the hang of it.

Thanks again GSum! :)

No worries!

Just to clarify though, I wouldnt recommend completely 0'ing out your PI(D) values. Those will protect the car if your WGDC values are way out of whack for what your boost targets out. I just meant that when tuning boost, IMO you can kind of ignore trying to tweak the PI(D) values and just use them as a guide to help you correctly set up your WGDC table as best as possible. Then later on you can start tweaking PI(D) values to dial it in.

JP the preload adjustment on the wastegate arm is just that. It doesn't change the spring pressure.

Loosening it will slow spool without significantly changing boost pressure up top, and over tightening will cause a boost spike, but boost will settle again to around the same pressure.

If you have a need to run less boost pressure than the spring currently allows for, changing the spring out to a lower pressure spring, and preloading the gate normally, is the best bet.