JB4 logging question: Target boost data unit convention

[i3igpete]

Installed a used JB4 and the first thing I did was update firmware to latest stable. For some reason, boost target is always clipping at 3 psi during the entire drive cycle in map 1. This was strange because looking at other people's JB4 logs (mostly absolute maps, map 4 or 6) target boost is the real, targeted value.

The actual DME_BT is reading correctly (which I believe is raw boost request from the DME), and AFRs are in the right ballpark, though the ignition timing is pulled way back (maybe some bad gas?). And the throttle valve is all over the place. But at least the other data appears to be accurate, even if non-ideal.

Can any JB4 veterans confirm that the JB4 will report additive +X psi as the "target boost" dataset when on map 1/2/3? Has it always been this way?

https://datazap.me/u/i3igpete/weathertech-pickup?log=0&data=1-3-5-9-11-19-23-34&zoom=369-616

Sponsored

 

[i3igpete]

Update, mystery solved I think. Might as well keep the thread up for someone else that is curious.

I found a log from someone using additive map 2, they also show a flat value of 4 psi for target boost:

https://datazap.me/u/kaj750/oct-1-e30-map2?log=0&data=1-3-9-11-23-34

Meanwhile, someone else using absolute map 4 shows a series of real values for target boost:

https://datazap.me/u/coloneladama/2023-supra-mt-map-4?log=0&data=1-4-11-14-18-25-26-27-28

It's probably way too late now, but I would recommend that Burger record these two datasets as separate things and null the unused signal when logging. One is a delta-pressure offset, the other is a pressure. They are not the same unit category, from an engineering perspective. +10 degrees Fahrenheit is not the same as 10 degrees Fahrenheit.

 

[zrk]

Relative maps show the delta as target, while absolute maps show absolute pressure as target.

It's dumb, but it's always been like this.

 

[ColonelAdama]

The actual DME_BT is reading correctly (which I believe is raw boost request from the DME), and AFRs are in the right ballpark, though the ignition timing is pulled way back (maybe some bad gas?). And the throttle valve is all over the place. But at least the other data appears to be accurate, even if non-ideal.

https://datazap.me/u/coloneladama/2023-supra-mt-map-4?log=0&data=1-4-11-14-18-25-26-27-28

hey thats me! Yeah, your timing is a bit all over the place and its probably just due to the throttle closures. Some throttle closure is normal and okay, but generally throttle dipping under 75ish means you are overboosting a bit more than the car expected which will pull timing.

I would recommend using Map 4 instead (93 oct+ only) and adjust your FF and duty bias to mitigate boost overshoot/reduce throttle closures. Read this if you plan to DIY: https://www.jb4tech.com/forum/model...g-review-ab/4229-jb4-duty-bias-ff-tuning-info

You can see in my log that the timing is pretty much perfect - climbing to 10+ deg as I approach redline. I don't let my boost exceed target by much. Definitely some throttle closure as boost finishes spooling and reaches steady state is normal.

Feel free to message me about questions. Send me a log on Map 4 if you want.

 

[i3igpete]

Hey Adam,

I liked your DOE approach in the 6MT thread and was going to attempt a step-by-step approach to this. two questions before I start:

can the JB4 completely override the dme wastegate signal when using absolute modes, or does it only modify the dme signal?

can the PID response get disabled by using a zero gain and zero FF offset?

do all settings in the user adjustment menu apply to map 6 only, or do some also apply to 4 and 5?

My idea was to disable the FF/PID functions and do a parameter sweep of duty cycle to build a lookup table of boost as a function of duty and rpm. this would be done for 2 different gears to see if the response differs.

then, do the same for FF and PID gain to evaluate adjustment range and transient response.

 

[ColonelAdama]

Answered your questions below!

Hey Adam,

I liked your DOE approach in the 6MT thread and was going to attempt a step-by-step approach to this. two questions before I start:

can the JB4 completely override the dme wastegate signal when using absolute modes, or does it only modify the dme signal?

The JB4 applies a bias to the wastegate signal. The variable WGDC will show 50 when no bias is being applied to the stock position. >50 means the wastegate is staying more closed than the DME requested. <50 is the opposite.

can the PID response get disabled by using a zero gain and zero FF offset?

Yes, and I've done it in my Map 6 thread. You can get it dialed in like that, but when conditions change, your car likely won't run as smoothly. That's what happened to me.

do all settings in the user adjustment menu apply to map 6 only, or do some also apply to 4 and 5?

All settings under User adjustment are global variables that apply to every map, except FF is saved per map iirc.

My idea was to disable the FF/PID functions and do a parameter sweep of duty cycle to build a lookup table of boost as a function of duty and rpm. this would be done for 2 different gears to see if the response differs.

This approach could work, but I would recommend leaving PID gain on. It will actually help you build your lookup table:

WGDC = actual wastegate position bias
FF = target wastegate position bias

WGDC = FF + PID offset

PID offset = (Boost target-boost actual) * some function * PID Gain

All you need to do to build a perfect FF curve for your chosen boost curve is try to match FF to WDGC across all RPMs to reduce the amount of work the PID has to do to achieve boost target.

In my MT car, I can't let boost exceed target or I get huge throttle closures. Therefore, my FF will always remain below WGDC because (boost target - boost actual) must be > 0.

then, do the same for FF and PID gain to evaluate adjustment range and transient response.

In short - Once you are nailing your boost targets perfectly, just look at the WGDC curve - that is exactly what you want the FF curve to look like (for AT cars). Adjust FF variable to change the height of the curve, and Duty bias to change the shape of it (by RPM). Read this: https://www.jb4tech.com/forum/model...g-review-ab/4229-jb4-duty-bias-ff-tuning-info

 

[i3igpete]

I think we're on the same wavelength on this part. My idea was once the boost lookup table was mapped out, choose a WGDC vs RPM curve that was slightly below target boost (say, 2-4 psi low). WGDC curve would be picked so that it's below target by a constant amount for all RPMs. Then use FF and PID output to bring that up to optimal. If I understand correctly, FF is a slowly self-tuning global adjustment, so the WGDC should be set to be consistently below target boost at all RPMs.

And of course, large swings mean the PID can have integrator windup.

Since the DME can use the following levers to reduce boost:
Open wastegate
Close throttle valve
Retard timing (with practically no floor)

But the DME realistically only has one lever to increase boost:
Close the wastegate
Advance timing (but cannot advance past the baseline map)

The idea of using doing a baseline map was to eliminate as many variables as possible. And I think Jesse's technique of never letting boost overshoot target is essentially forcing FF and PID response to stay positive (if i'm not mistaken).

I'm guessing you're a MechE as well?

 

[ColonelAdama]

I think we're on the same wavelength on this part. My idea was once the boost lookup table was mapped out, choose a WGDC vs RPM curve that was slightly below target boost (say, 2-4 psi low). WGDC curve would be picked so that it's below target by a constant amount for all RPMs. Then use FF and PID output to bring that up to optimal. If I understand correctly, FF is a slowly self-tuning global adjustment, so the WGDC should be set to be consistently below target boost at all RPMs.

Be wary of underboosting by this much as your trims will be sky-high. Say your Target at 4000 is 19psi, and DME_BT is 14 psi. Underboosting by 3-4 psi will send an ECU_PSI value of ~10 psi or less, while your actual boost is ~16. The car won't be willing to give much fuel if it thinks its below 10 psi. FF is self tuning unless you disable it by turning Bit 1 on. You won't achieve underboost without turning off wastegate adaption. It will just keep bumping up FF till you hit target.

The idea of using doing a baseline map was to eliminate as many variables as possible. And I think Jesse's technique of never letting boost overshoot target is essentially forcing FF and PID response to stay positive (if i'm not mistaken).

Jesse's technique that you're talking about only applies to MT cars. AT approach is pretty different. But yes you are correct. In my car, the PID response is always positive when underboosting. With an AT, there's no reason really to start from scratch tuning the car. My approach would be to start on Map 4 for some baseline runs and just see where to improve from there. Underboosting shouldn't be needed on your car.

Use Datazap and set up some filters. I use 2 specifically: WDGC and boost. then Fuel and timing
Notice that the gap between FF and WGDC is pretty steady all the way through. As is the gap between boost and target. This means my FF curve is a decent fit since PID feedback is pretty uniform.

I'm guessing you're a MechE as well?

Good guess! ?

 

[i3igpete]

out running some errands, was able to get a fourth gear pull with map 4. Haven't taken a look in detail.

Side question, is there a more streamlined way to upload directly from the jb4 mobile app to Datazpp? At the moment I'm emailing the file to myself and then upload from my downloads folder

https://datazap.me/u/i3igpete/old-navy-return?log=0&data=1-4

 

[razorlab]

out running some errands, was able to get a fourth gear pull with map 4. Haven't taken a look in detail.

Side question, is there a more streamlined way to upload directly from the jb4 mobile app to Datazpp? At the moment I'm emailing the file to myself and then upload from my downloads folder

https://datazap.me/u/i3igpete/old-navy-return?log=0&data=1-4

You ignition timing looks horrible. Car isn't happy at all. Also, very bold running 14+ AFR with 18psi out the top, which is probably why the timing is bad.

 

[i3igpete]

Also, very bold running 14+ AFR with 18psi out the top, which is probably why the timing is bad.

if I'm not mistaken, the afr's are at 14 because that's the default factory fuel map (?). the jb4 has no fuel or ignition control on these cars, correct? so there's really nothing I can do other than turn down the boost.

but yeah, the map 4 run was at the request of @ColonelAdama. i was seeing terrible timing numbers even with map 1 (original post). I think we were using map 4 to see if the ECU was trying to use timing as a boost lever.

 

[razorlab]

if I'm not mistaken, the afr's are at 14 because that's the default factory fuel map (?). the jb4 has no fuel or ignition control on these cars, correct? so there's really nothing I can do other than turn down the boost.

Correct, one of the big issues with the JB4. I was merely stating that kind of boost at that AFR is bold. That is why you always hear people saying more octane is better with a JB4. It's basically a crutch.

The timing in your log looks more like preemptive knock correction, not torque correction. Torque correction is much smoother.

 

[ColonelAdama]

out running some errands, was able to get a fourth gear pull with map 4. Haven't taken a look in detail.

Side question, is there a more streamlined way to upload directly from the jb4 mobile app to Datazpp? At the moment I'm emailing the file to myself and then upload from my downloads folder

https://datazap.me/u/i3igpete/old-navy-return?log=0&data=1-4

That's how I do it. If you figure out anything better, let me know!

You ignition timing looks horrible. Car isn't happy at all. Also, very bold running 14+ AFR with 18psi out the top, which is probably why the timing is bad.

Mid 14s is 100% normal for this car, even at 20+ psi. Don't worry unless you see anything 15+. Anyone at BMS and also Jesse will tell you the same. Direct injection car with very precise throttle control to prevent lean conditions.

Timing is pretty bad yeah, what fuel are you running @i3igpete ?

Its hardly running extra boost at all in this run- its like 18 peak compared to 15 stock.

The FF curve as standard seems to fit your car well. Let FF adapt up some while driving and I bet the log cleans up nicely. I expect to see FF adapt to 58ish.

The timing in your log looks more like preemptive knock correction, not torque correction. Torque correction is much smoother.

Timing shouldn't be increasing until 4k+ rpm. It's crashing because ECU_PSI is too low on account of underboosting. Car sees low boost - it adds timing too early. boost kicks in - timing crashes.

 

[i3igpete]

93 octane with E10 is the usual mix here in Chicago, though I did go to a different station than normal - usually shell/mobil but this time I went to a speedway.

 

[ColonelAdama]

93 octane with E10 is the usual mix here in Chicago, though I did go to a different station than normal - usually shell/mobil but this time I went to a speedway.

93 is perfectly safe for using Map 4. Your timing crash is not an octane issue. I think its because of underboosting. Let FF adapt up or set it to 58 yourself. Set boost safety to 23.

Sponsored