aidandj

iTrader: (5)

We won't know without more testing. But it explains why the log dives to minimum when trying to spool.

Only thing it doesn't explain is why it holds at max duty cycle before you reach the delta, and then drops to zero to try and spool it.

Savington

iTrader: (31)

This is from an early pull with polarity inverted (I copied Soviet's settings to start).

aidandj

iTrader: (5)

Ok. Based on that its not the issue.

Does basic mode get you to boost target?

aidandj

iTrader: (5)

Just looked back at some stuff Ken wrote.

Defaults in basic mode are P:100 I:100 and D:100.

Try those in advanced and see if you get the same results as in basic.

Savington

iTrader: (31)

Post #55 has charts with the slider at 250 and 330 in basic mode. Neither one created overshoot.

aidandj

iTrader: (5)

Something weird is going on. Those huge changes in boost duty should have a much larger effect on your MAP than they are having. Looking back at post #55 you go from 10% to like 50% and your change in MAP is like less than 5kpa. On my car if you did that you would go from like 250kpa to 120kpa.

Where did you get the solenoid frequency information?

Savington

iTrader: (31)

Shamelessly stolen from Soviet. The valve is a Pierburg valve, it's OEM for GM from Opels and Vauxhalls in the UK.

codrus

What solenoid is it? The usual MAC one?

Your chart in post #82 actually looks pretty reasonable as what it should do. It would be easier to read if you forced the boost target to the same scale as MAP, but it looks like it's turning on boost duty too early and too hard and just not ramping it down fast enough to avoid the spike.

--Ian

(Edit, I see the solenoid type now. If you want to try a MAC, I have a dual port one sitting on the bench that I took out of the car when I went to the three-port.)

EO2K

iTrader: (37)

Interesting. 18psi and I (mostly him) fiddled with the BW provided valve for some time before giving up on it and going to a MAC. We were under the impression the BW valve was dead or just straight not working correctly. I'm also on a Reverant MSLabs MS3-B, probably an early unit as well. I think we also stole the freq info we used for the BW valve from Soviet.

From the emails we've traded I believe Sav is still playing with the BW supplied Pierburg valve. Ian, you are on a MAC, correct?

aidandj

iTrader: (5)

You sure? He is dropping to 10% duty and boost keeps climbing. Thats backwards.

codrus

Yeah, but if it were inverted, the boost wouldn't stop climbing when the DC hits zero, and yet it does. At least, it looks like there's a couple data points between when the boost starts to drop and when TPS goes to zero, it's hard to say from that graph.

VSS starts to drop when boost drops as well -- were you getting wheelspin?

--Ian

Savington

iTrader: (31)

Made a couple of tweaks. I searched around the Intarwebz and found references to a couple people running the EFR Pierburg valve at ~35hz, so I changed from 26hz to 39hz. I also reduced the delta value so that the algorithm only starts working within 10kpa of target. The first pulls resulted in substantial overshoot and big oscillations in the boost curve, which is on the right track. I increased the delta to 15kpa and backed the slider down from 250 (big oscillations) to 125 (still oscillating) to 62.5, which gets me here:



Forgot to drop the control interval to 26ms to match the valve frequency, but that looks a lot closer than anything else I've tried. I can't decide whether the reduced delta value is a bandaid fix or just how it needs to be.

e: It's still not better than open loop, which is more than a little frustrating.

aidandj

iTrader: (5)

VSS drop would be the opposite of wheelspin. His VSS is off the tranny.

Its still crazy overactive. I had this issue on mine and Ken fixed the firmware because the slider was 10x (maybe 100) sensitive.

You really should switch to 1.4, it makes a big difference is is the current release firmware.

codrus

As the delta value gets smaller, the PID loop gets turned on closer to the the boost target, which makes it much less forgiving of errors in the initial value table. A value that's too big will result in the boot spiking, because there isn't much time for the PID algorithm to fix it before the spike happens. A value that's too low, OTOH, can actually lead to oscillations where it dumps too much boost too quickly and the MAP drops below the delta range puts it back into boost-assist mode.

My delta value is 20 kpa. I wouldn't say 10 is a "band-aid", there's a huge amount of variability here. I would think the lightweight turbine wheel is going to make the EFR a lot quicker to drop boost than a Garrett, so you have to go with what works for your car.

I think 10% as a min DC may be too high, because EBC is yanking it all the way down there and seems to want to take it lower. Did you set the min value based on where it starts making noise when cycling it while parked?

--Ian

codrus

Sure, I meant the VSS drop would be associated with the wheels regaining traction as the boost level came down.

--Ian

aidandj

iTrader: (5)

Something else is going on.

Why is it pulling it to his min duty cycle before he even reaches his target. PID doesn't work like that.

aidandj

iTrader: (5)

Savington

iTrader: (31)

I will switch over to 1.4 tomorrow.

aidandj

iTrader: (5)

You will be pleased.

EO2K

iTrader: (37)

I feel like I should have done this before having the car tuned