Issues configuring for boost control

zmanco · Post by **zmanco** » Mon Mar 10, 2014 7:51 am

I recently put a new actuator with a stronger spring on and am revisiting boost control again. I want to be sure that I have the correct configuration as things appear to have changed with 3.3. I think I'm a candidate for the new approach to managing boost control here: http://www.msextra.com/forums/viewtopic ... 34&t=53051 as I've tried multiple times without success to get to a stable system, but want to be sure that it's not a configuration issue. Firmware is 3.3.0.

Setup: I am using the DIYAutotune solenoid configured as recommended.

wastegate control.png

I started with open loop and have verified that a larger duty values increases boost.

Closed loop is configured as such:

close loop example settings.jpg

Here is a datalog. Note that now boost rises with falling boost duty, the opposite of what I believe it should be. I've tried changing from normal to inverted, and also swapping the open and closed duty values but that just results in overboosting.

Is this now normal for BC e.g. higher duty means more boost for open loop, but means lower boost for closed loop? Or do I have something improperly configured?

closed loop example log.png

slyrye · Post by **slyrye** » Mon Mar 10, 2014 8:44 am

You might wanna check-out how you put together the pressure hoses from and thru the solenoid valve, it should be connected in a "normally open" fashion, you can check it by blowing the holes with air and observe where the air exits from.

zmanco · Post by **zmanco** » Mon Mar 10, 2014 9:05 am

I did exactly that when I put it together, and verified the valve blocked flow to the actuator when I changed the open loop table value to 100 at idle. I also verified on the road by starting with the open loop table at all zeros and confirming that boost was limited to the mechanical actuator's spring limit. As I increased the table values above zero, boost increased.

Trumpetrhapsody · Post by **Trumpetrhapsody** » Mon Mar 10, 2014 10:13 am

You have closed at 0% and open at 100%. Since on an internal arm style waste-gate closing the valve increases boost I would think you should have that the other way around.

I have the same MS settings as you for my external wastegate, where opening the valve increases boost. Normal polarity, closed 0% open 100%. One of us must have it backward. My logs show boost increasing as duty increases.

MS3 Documentation, but with swapped normal/inverted description. Assuming it's out of date since the 3.3.x change:
Normal - 0% output pin duty means the wastegate is fully open.
Inverted - 100% output pin duty means the wastegate is fully open.
Closed Duty - This setting controls the lowest allowed reported duty. Normally this should be 0%.
Open Duty - This setting controls the highest allowed reported duty. Normally this should be 100%.

Sounds like if you set open to 0% and closed to 100% you'd still want Normal polarity.

WestfieldMX5 · Post by **WestfieldMX5** » Mon Mar 10, 2014 11:03 am

I was about to make a topic when I noticed yours because I have exactly the same behaviour.
Tried the same things as you, changing output polarity doesn't solve it.
Here's a log I did yesterday.
I wasn't sure if I needed to start a seperate thread, my apologies if I did.
Logfile was too big to upload so I zippend it up here: https://dl.dropboxusercontent.com/u/396 ... t4.msl.zip

zmanco · Post by **zmanco** » Mon Mar 10, 2014 11:15 am

Just went through the various runs I've done in the last few days and found one where I had swapped open and closed settings, albeit I used 40% for closed to try to help the CL algo (I know from the open loop table that 40% will always overboost). Is this binary nature of the boost duty normal? I thought it wasn't so assumed it was a configuration error. I tried several passes with a variety of values for closed duty and they all had the same pattern where boost duty was either closed or open, never anywhere in between.

MSQ for this pass attached in case that might offer some clues. Target MAP was 200kPa.

Screen Shot 2014-03-10 at 12.06.53 MDT.png

zmanco · Post by **zmanco** » Mon Mar 10, 2014 12:54 pm

I had a chance for a quick run at lunch and tried swapping the open and closed values, but with boost control set for open loop. Interestingly, it made no difference vs. not swapping them.

screen_shot_2014-03-10_13-40-28.jpg

screen_shot_2014-03-10_13-40-42.jpg

screen_shot_2014-03-10_13-40-05.jpg

Normally with just the mechanical wastegate actuator operating, I see boost falling off at higher revs due to what I believe is back pressure in the too small turbine. The rising values in the open loop table compensates as you can see making me believe I have the basic mechanical and plumbing setup done correctly.

Another thought: Shouldn't the open and closed values also affect open loop? Or perhaps the open loop setting just assumes one of the two possibilities?

pit_celica · Post by **pit_celica** » Tue Mar 11, 2014 10:57 am

If you get this "sawtooth" boost duty, it's because your Open Duty and Closed Duty are inverted one vs the other. As stated in the manual, set the closed duty at 0% and the open duty at 100%. In my understanding, the Open/Closed duty are only used when closed-loop control. In open loop, it follows the duty table, no matter the other settings.

As by your first post, I think it's just a matter of PID tuning. Follow Ken's tuning strategy in his new boost control thread : Set all PID at 0, set the bias table as if you were in open-loop, enable the PID parameters one at a time, slowly.

Sam

tutuur · Post by **tutuur** » Tue Mar 11, 2014 1:32 pm

This is exactly what your problem is, i didn't dare to say it before as i'm not that experienced but indeed the manual states it must be inverted for closed loop.

I asked why in the past and it had something to do with how the code is written...
After knowing that i changed my open-loop strategy to inverted too as i want to use closed loop in the future

zmanco · Post by **zmanco** » Tue Mar 11, 2014 2:12 pm

Thanks guys, in a parallel conversation with Ken, he said the same thing that closed duty should always be lower than open duty. The problem I saw is that with that configuration and polarity inverted, the engine still overboosted even with boost duty = 0. This was with 3.3.0.

Screen Shot 2014-03-11 at 15.10.20 MDT.png

I'm going to jump to 3.3 alpha 4 to partake of the new BC features and before I enable them, see if this is still an issue. Weather here in Colorado isn't going to allow me out till tomorrow though ...

tutuur · Post by **tutuur** » Tue Mar 11, 2014 4:27 pm

That's exactly the problem, when puting in inverted you'll have to make the whole table 100 and when lowering it you increase boost.

Don't ask me why, i've asked it before cause i found it rather confusing but i just learned to work with it and it really isn't that hard. Just work your way down from 100 instead up from 0.

No new firmware will help with any setup errors

Post by **muythaibxr** » Tue Mar 11, 2014 5:51 pm

tutuur, we now require more duty to give you more boost. If it doesn't, rewire, replumb, or invert the output.

tutuur · Post by **tutuur** » Tue Mar 11, 2014 6:00 pm

You mean it has changed with the newest firmware? I haven't updated it since the beginning of 2013

zmanco · Post by **zmanco** » Tue Mar 11, 2014 6:13 pm

I upgraded to 3.3.2 beta 4 and just used I/O test mode and my mouth to blow on the disconnected hoses. At 0% the valve passes boost thru to actuator which would reduce boost. At 50% it blocks all boost to actuator which would increase boost. This is regardless of the Polarity setting. So I have set Closed Duty = 0%, Open Duty = 50%, Polarity = Normal.

Do I have the proper definitions of "open" and "closed"?

If "yes", can I conclude that I have things wired/plumbed correctly?

zmanco · Post by **zmanco** » Wed Mar 12, 2014 12:02 pm

The issues I was having with 3.3.0 are gone with 3.3.2 beta 4. With polarity = normal, closed duty = 0, open duty = 50, the boost duty acts predictably. Obviously I have work to do setting up the initial duty table, but things are working as expected, which was not the case with the old rev.

One thing I noticed was that last night when I was testing the valve by blowing against it, at 50% it felt closed. Looking at how slowly the turbo is spooling today, I can see that is really not the case and I'll need to go with a larger duty to fully close it.

Another observation: last weekend I tuned the open loop table and expected that it would be a good starting point for initial boost duty. However, it was in the upper 60's then, and today it was barely over 30. With the same values it was overboosting by a lot.

graph.png

Megasquirt Support Forum (MSEXTRA)