Hello, I did the first tests on circuit, I have 2 questions.
- First, I tried to set the PID Boost, if I put only (P) the boost curve follows the target, but does not reach the pressure or takes time, normal.
I then set I and D or I reach the pressure, but it exceeds or follows the target pressure above and the (I) remains at 5% positive, but not every time. I’ve tried lowering or raising the (I) but nothing changes, the only thing I haven’t tried is raising the D.
-The second is the Knock, I have a lot of noise on the 1 and 3 cylinders but only when the throttle is off or very little, not under full load, this seems astonishing to me…
Could you please help me?
Best Regards Yann
log pid 04.02 Dijon.emuprolog (6.4 MB)
0.4P 0.2I.emupro (32.5 KB)
Hi,
Regarding boost:
- Start with D gain at 0. Only use D if you have problems with overshooting the target. Otherwise, don’t use D.
- You are sometimes hitting the limits of the PID. The output min/max is 15%, and the Integral limit is 10%. That limits how much change the PID can make. You might want to increase those limits so that the PID can do more.
- You want the highest values of P and I that don’t generate oscillations or an unstable control. If you increased P or I and didn’t see the difference, you didn’t increase it enough. At some point, the system will not be stable and will start to oscillate. That means you have too much gain somewhere.
Regarding knock:
Why did you set the frequency to 15.07 kHz?
Audi 5-cylinder has a bore from 102.9 to 110.7 mm.
This gives a base frequency between 5.18 and 5.57 kHz.
You can also use harmonic frequencies, which are multiples of the base frequency. So, base frequency * 2, or base frequency * 3.
You want to select a frequency with the most stable signal and the least noise.
Hi
Thanks for your reply,
yes I tried the P max 0.5 after oscillations and I tried the D from 4 to 0.2 but I still have an overshoot of the Target, maybe I should have more I and more D, does (Integral Rat of change max) influence much on the speed and change?
According to the logs, I’m often above Target.
For the Knock.
I’ve got an 82.5mm bore, so normally 6.9Khz, I’ve tried 6.9 and 13.8 -15.2 but it doesn’t make much difference to the result.
Normally for the 2 Harmonics it’s about 2 to 2.5 times, on the EMU black I had these settings and I had about 3V but on average of course.
-What bothers me is the noise on these 2 cylinders, but not at full load, or is the noise stable?
-Is there a problem with the advance?
I’d like to find the exact noise of each cylinder to have a top setting to work at the limit of the advance.
From experience what do you think or how would you practice?
Thanks Best Regards Yann
What engine is that?
First, make the changes I described for the boost and get new data.
Now, I think there is something different going on.
There is a significant difference before and after the 5-minute mark.
Before, all signals looked very close, but something happened.
I would suggest checking the car mechanically.
Maybe something got loose and is creating additional noise.
Unless you have another explanation for this?
But I don’t think it’s mechanical, it’s on sequences and then it disappears, I’ve checked the other logs same thing, you don’t think the EMU is sporadically lost on Knock control per cylinder (Triggers Audi maybe the cause), it’s mostly after gas cut that it starts and then it stabilizes again.
I can’t put other Logs too big, I can if you give me your mail, thanks for your help.
Best regards Yann
Hello,
You may be right.
I’ve analysed the other logs and it always happens in the same place, braking and cornering, I think it’s the oil rising in the engine block and making noise… it’s going to be complicated to fix.
I also wanted to analyse the GPS values (I have the V1) in the EMU Pro the CAN always goes from OK to not OK, whereas in the ADU it works fine. Have I set the right parameters?
Thank you for your help.
Am I right that you have IMU fusion disabled in the GPS?
I think we set the timeout based on a 20 Hz refresh rate.
We didn’t consider that it could be 10 Hz without IMU fusion in GPS V1.
We have to fix that.
Yes, I’ve deactivated the merge, but I think I read that it should be deactivated.
Is it better to activate it?
No need to deactivate.
It’s a valid configuration for a high-vibration environment.
We have to change the timeout value for the next PRO update.