Random Boost spikes and Ignition Cut

[Ricco]

Hi, I've been playing with and trying to learn/understand Boost control but I'm getting some odd results, or at least ones that I cannot logically figure out.  I tried to follow the "Closed Loop Boost Setup Guide" in the help file but I just wasn't able to derive the expected results. 

Here is my list of questions I've come up with.

1. Why can't I ever seem to hit 180kpa? Whenever I adjust the WGDC values to provide more boost, I just seem to hit ignition cut but never boosts more than 170-175ish.  
2. Why is the WGDC sometimes different to the 'Boost Base DC' even with the PID's zero'd?  I thought with PID's zero'd, WGDC would always just use the WGDC Table, once it hits stage 2 boost control (assuming trim tables are not coming in to play)?
3. I'm getting random Boost spikes. I've seen it jump over 100kpa in 1/1000th of a second which seems implausible.. Could this be a symptom of an old/failing/bad stock Map sensor?
4. I understand that Ignition cut is kicking in because of the boost spikes but I'm also sometimes seeing Ignition cut without a boost spike.. I even see Ignition cut, cutting in at ~170-175kpa with a boost target of 180.  My MAP limit is 205 and Control range is 15 so Ignition cut shouldn't even come in to the equation until 190kpa? Or am I misunderstanding this somehow?
5. When I use a Boost target that is zero'd out in the WGDC table, I see only spring pressure and everything happens like I'd expect.. The car is just slow  

I have a bunch of log files from my evening of playing around with the settings but I've tried to limit the uploads.. I think you should see examples of everything I'm questioning in the attached and linked log files.. I also attached my Map file, although I was making some adjustments to it as I went along - Mostly just to Target boost and WGDC Table with PID's zero'd - per the setup guide but I think I have PID's set per the Help file in this map as I just set it there towards the end of my runs.

Car is SW20 MR2 with stock Gen2 3SGTE with AFM removed and IAT sensor in place of the CSI. MAC 3 port boost solenoid. Aftermarket Catback.

Thanks.

https://drive.google.com/file/d/1HXx1EncOBJaSu5iTZduGHXzO0shUDQVN/view?usp=sharing

mr2 Boost Control 1-27-23.pclx PC Datalog IgnCut n Bspike ex- 2023-02-15 8;23;24 pm.llgx

[Adamw]

You are definitely hitting the MAP limit a lot in both of those logs.  Most likely you arent seeing all the detail due to the slower rate of PC Logging.  I would suggest you set up ECU logging to log MAP, Ign cut, MAP limit status etc at 500Hz and I think you will find it is hitting the 190Kpa limit range threshold. 

 

5 hours ago, Ricco said:

Why is the WGDC sometimes different to the 'Boost Base DC' even with the PID's zero'd? 

 That would have to be one of the WG trims.  Being an MR2, most likely IAT trim?  I cant really see an example in the logs without knowing what was in the DC table at the time.   

 

5 hours ago, Ricco said:

I'm getting random Boost spikes. I've seen it jump over 100kpa in 1/1000th of a second which seems implausible.. Could this be a symptom of an old/failing/bad stock Map sensor?

Those are intake backfires.  Hard to tell if they are due to the MAP limit ignition cut or whether that came just as a consequence of the backfire and the backfire was caused by something else.  Faster logging will help confirm that.  A couple of observations: In the short log there is an ign cut at 0:17 and 0:20, there was a MAP spike on the one at 0:20.  Just before these two cuts the battery voltage had dropped quite low, ~12.4V on the first and 12.8V on the second, so potentially a misfire as distributor ignition at 4000RPM and 12.4V and probably some of the hardware 30 years old isn't going to be making great ignition energy.  Intake backfires are often an indication of a lean fuel mixture, there is no lambda set up in these logs so I cant rule that out but it is something to keep in mind.

 

A couple of unrelated comments, just stuff I noticed when looking at the logs:

The warm up enrichment was still active for most of that longer log.  If that was at normal operating temps you might want to lower the last enrichment cell so it is not working when warmed up. 

The rough/oscillating idle in many places is caused by the idle control kicking on and off due to the MAP lockout being too low, bump this up to 65kpa or so.  

 

   

[Ricco]

Thanks for the reply Adam.  I've already taken a look at the MAP lockout and Warm up enrichment - Your recommendations make sense, thank you.  

I'll configure some ECU logging as you suggest and hopefully I'll be able to give it a try tonight. 

WRT to the WG trims, what I'm seeing would need to have a negative trim value since I'm seeing higher WG DC than I think I am commanding.   Take a look in the big log file at around 2:24-25 to see what I mean (just one example).. I'm sure I'm just misunderstanding something but I see Base WGDC commanded at 20% but I'm seeing an actual WGDC at 40% - I assume its related to trying to hit the 180kpa boost target but idk what is causing it to operate that way. I don't see any ECT or IAT trim in effect and the PID's are all zero. 

Incidentally, if you look at the 29-30 second mark in the same big log file, you will see the WGDC follow precisely the Boost Base DC which is what I'd expect to see every time. I don't think I changed any parameters in the configuration between these pulls as I was driving. 

[Adamw]

7 hours ago, Ricco said:

I'm sure I'm just misunderstanding something but I see Base WGDC commanded at 20% but I'm seeing an actual WGDC at 40%

This is because you have the Base DC Mode set to "Stage 2".  This means for stage 3 the feed-forward (base duty) will be fixed at the last calculated DC value when it exited stage 2.  The WGDC table will have no effect.  If you want the base duty to be updated from the WGDC table through stage 3 you need to set the Base DC mode to Stage 2-3.  Typically stage 2-3 mode works the best for road car sized turbos, while stage 2 mode works best for large turbos that have reasonably stable spring pressure control.

[Ricco]

I tried ECU logging last night but when I try to download the log file 2-3MB, i only get a ~35KB file downloaded.. The file will open in PCLink but it only has about 0.5s of data in it.. Going to try and figure out what I'm doing wrong later. 

19 hours ago, Adamw said:

This is because you have the Base DC Mode set to "Stage 2".  This means for stage 3 the feed-forward (base duty) will be fixed at the last calculated DC value when it exited stage 2.  The WGDC table will have no effect.  If you want the base duty to be updated from the WGDC table through stage 3 you need to set the Base DC mode to Stage 2-3.  Typically stage 2-3 mode works the best for road car sized turbos, while stage 2 mode works best for large turbos that have reasonably stable spring pressure control.

OK understood. I foolishly thought that setting the Mode to Stage2 meant that there was no stage 3 any more. At least that explains it, thanks.

[koracing]

This is due to connectivity issues with the USB while downloading.  If the connection drops off it will try to save whatever you have downloaded to that point.  Make sure your ignition is on, engine off while downloading and have the laptop and cable outside of the car when doing this to get the best results.

This is an all stock gen2 engine?  Stock CT26 turbo with the factory internal wastegate?

Why aren't you using a wideband with the ecu?  Seems like feeling around in the dark blind folded with your eyes closed.

[Ricco]

2 minutes ago, koracing said:

This is due to connectivity issues with the USB while downloading.  If the connection drops off it will try to save whatever you have downloaded to that point.  Make sure your ignition is on, engine off while downloading and have the laptop and cable outside of the car when doing this to get the best results.

This is an all stock gen2 engine?  Stock CT26 turbo with the factory internal wastegate?

I don't remember if I had the engine running or not when I was trying to download. Surely it shouldn't make any difference?  Anyway, I'll try it again tonight with the engine off. 

It's a mostly stock Gen2. It has AFM delete with cone filter, IAT in the CSI location and an aftermarket carback. Stock turbo, WG and downpipe for now. 

[koracing]

While it *shoudln't* - I have found it certainly *does*.  Any hiccup in comms will abort the log download before you get the full data.

I have a 1995 MR2 Gen2 3sgte nearly all stock I've been using as a test bed for a "stock" calibration base map with modelled fuel strategy.  It does have the following mods: gutted cat and 2.5" exhaust, AFM still in place (so I can revert to the stock ecu easily), 14point7 wideband in the factory o2 location, and a plug in 4 bar map sensor with a G4X ecu.  Got 28mpg average driving 350miles to and from Portland, OR last October.  No aftermarket boost control on this, however, as I just set the output to an on/off ouptut on the factory turbo VSV based on coolant temp.  

Most mr2s I tune are aftermarket wastegates with Mac solenoids though.  I've tuned one with the G4+ plug in link on an upgraded factory style turbo with the internal gate and it was much less sensitive than the external wastegates - required much larger duty cycle changes to get to the desired boost level.

I think when you're seeing that MAP spike going to the 260kpa - that's your error value for exceeding the high voltage limit on your map sensor.  You can try raising the limit on that to 4.9v and see if it helps.  The stock map sensors are pretty terrible range (2-2.1 bar maybe) and you really should get a map sensor with more range.

[Ricco]

6 hours ago, koracing said:

While it *shoudln't* - I have found it certainly *does*.  Any hiccup in comms will abort the log download before you get the full data.

So I managed to get the logs downloaded from the ECU.. Even with the engine off, it took several attempts to get it to start working as it should..

6 hours ago, koracing said:

Most mr2s I tune are aftermarket wastegates with Mac solenoids though. 

Yeah, my bad.. I do have a MAC 3 port installed too, that's why I'm trying to play with and understand boost control.

Anyway, here is some high res data from a pull - I see that there are small map spikes that are hitting the ignition cut ceiling, just like Adam surmised - it seems to be happening so fast that I find it hard to believe it's a real boost spike, and more likely to be a small back fire, as Adam suggested.  I'm planning on pulling the plugs and gapping them to 0.30 (currently set at 0.32) to see if that helps but I'm probably clutching at straws and need to get a wideband setup so I can tune fuel and ignition tables (I'm guessing).

Here is the link to the log file.

https://drive.google.com/file/d/1H_1Aoiz024AxO0fKGy-KW-ern7SF0nWL/view?usp=share_link

 

[koracing]

Or perhaps raise your map error high voltage as I suggested.

[Ricco]

Just to close this out for now.. I re-gapped my spark plugs from 0.032 to 0.030 and the Ignition cut issue is 95% or more, gone. I suspect the spark was getting blown out by a number of contributing factors 

90's Spark distribution

Rich fuel mixture in the Link Basemap

Spark plug gap too big

I don't have any more time to play with it for a while as I'm in the process of buying a new house.