Boost duty drops low before initial value (closed loop)

[varg]

I've been playing with this for a while, and this problem has been bugging me for ages and I can't figure it out. As you can see in the log, boost duty initially does what it should and stays at 100%, but when it reaches the threshold for CL activation, it dramatically drops off (to 14% in this particular log) before climbing back to the initial value where it goes back to behaving as I would expect. Until this issue is solved it's impossible to get a good CL tune, and boost always climbs slowly because of the drop in duty so I've resigned myself to using open loop. I've searched for other instances of this problem - no luck. Any ideas? I've never seen this behavior mentioned or in any car I've tuned though admittedly I have never had the occasion to tune closed loop boost control on anyone else's car - standalone or manual boost controllers are the norm. Hopefully I'm missing something obvious, I've been tuning megasquirts for a while but I'm no expert and nobody is beyond missing something stupid

Firmware version 3.3.3 20141112
DIYAutotune boost control solenoid and mod kit
MS2 V3 with Zeal daughterboard

Thanks

[prof315]

I see 2 issues.

1: Old Firmware update to current release pretty sure there were improvements to boost control

2: Boost values don't match between target and initial duty table

[varg]

I see 2 issues.

1: Old Firmware update to current release pretty sure there were improvements to boost control

2: Boost values don't match between target and initial duty table

Bummer, I was hoping to avoid a firmware update. Not that it's hard to do I just was hoping it wasn't that. As for 2, this should not cause any problem unless the code is faulty as MS should interpolate the values like it does on every other table. The reason the target and initial duty are different is because I changed the boost target and never changed the initial values to match it - but it has always done this, even when they matched, so it isn't the issue. I'll try a firmware update and report back.

[prof315]

Why are you trying to avoid a firmware update???? It's not hard and with TunerStudio 3.x the loader is built in. Don't know about you but I want the latest and best on my ECU(s)

[varg]

As I said, it's not like it's hard. I haven't updated because in my view there's no reason to update if there isn't a problem. With the first suggestion being that boost control was improved in later releases it's a potential solution so I'll do it, but if boost control wasn't changed at all in the latest releases I would keep what I have. It's not specific to megasquirt, I don't compulsively update firmware or software in general because too often I've updated to find something broken.

[prof315]

3.4.2 has been out there and stable since July 0f 2015! I get wanting to wait awhile to make sure new firmware doesn't have bugs and all but it's really a good idea to stay on top of things.

[varg]

You seem to be misunderstanding me, but it's not a productive discussion anyway; the merits of using old firmware vs new. I will continue to believe that there is no inherent value in updating if there aren't any problems with what's there and I'll update since the the problem may be caused by the firmware when I don't have anywhere to drive. I'll report back as to whether it fixes anything or not.

[varg]

Firmware update didn't change a thing. Still has the exact same issue.

[prof315]

Try a MAC solenoid. I have been having troubles lately with other brands. https://www.amazon.com/MAC-35A-AAA-DDBA ... noid&psc=1

[varg]

I appreciate your input and am not trying to be rude by shooting you down, but the solenoid can not change the behavior of the megasquirt output, it is a passive device and only reacts to what the ECU gives it. The output duty cycle is determined internally at the processor without feedback from anything other than the normal MS in puts - engine speed, MAP, etc.

In the log, the duty cycle drops off from 100 to some unusually low value before going to the initial value table, I do not believe this is normal behavior - AFAIK and from what I've seen the boost duty is supposed to drop from 100 to the initial value, as opposed to radically undershooting before climbing back to the initial duty cycle value. I have combed through my logs and played with my tune and PID values extensively and haven't been able to stop this from happening in closed loop boost control. I'm honestly baffled here.

[ashford]

in the log you posted the actual map and target boost never makes it into the initial duty table( max 150ish map and initial duty table has a minimum kpa of 176), of course initial duty isnt going to work if it is outside the table.

[varg]

Does it at full throttle too, will post a log.

[varg]

Just an update, I gave up on this. I've logged and tried to figure out what is going on here, but the boost duty will not stop dropping off before going to the initial value and it slows response noticeably vs open loop. I've tuned my duty cycle table well enough now that on a standard florida day I get the boost I want.

Hopefully someone else figures this one out some day, because I'm not wasting any more time on it.