kacperoooni
Full Member
Karma: +9/-2
 Offline
Posts: 124
|
 |
« on: November 22, 2019, 04:42:52 PM »
|
|
|
Hi,
I need to disable boost PID control (don't ask why, I'll publish when it's ready). For example - Ecu request 1600bar absolute of boost. I set in KFLDIMX 600hpa row for example 30% dc. My goal is to force ecu to output 30%dc pwm signal (n75 output) for 1600 absolute of boost. No change of DC regardless if my actual boost is meeting requested or not. Don't want to use KFLDRAPP.
LDRQ0 - LDR PID Q0 in dynamic operation (proportional term) - likely leave this alone
LDRQ0S - LDR PID Q0 in static operation (proportional term) - likely leave this alone
LDRQ1ST - LDR PID Q1 in static operation (integral term) - likely leave this alone
KFLDRQ2 - LDR PID Q2 (differential term) - adjust this to compensate for overshoot when your boost ramp meets requested. If you have overshoot, try decreasing KFLDIMX first. If you have undershoot, try increasing KFLDIMX first. Only change Q2 as a last resort.
Will it be sufficient to just FF or 00 this maps?
|
|
|
|
« Last Edit: November 22, 2019, 04:50:42 PM by kacperoooni »
|
Logged
|
|
|
|
|
nyet
|
|
« Reply #1 on: November 22, 2019, 08:11:01 PM »
|
|
|
zero ldrq0, ldrq0s, ldrq2, leave ldrq1 alone.
integrator windup will hit imx but you'll also have to disable imx adaptation. don't know the maps offhand.
There is probably an easier way but i'd have to dig around the fr to answer.
Note that DRL will have to be made linear probably.
Eventually you'll throw an over/underboost code unless you disable that as well.
Good luck.
|
|
|
|
|
Logged
|
|
|
|
kacperoooni
Full Member
Karma: +9/-2
Offline
Posts: 124
|
|
« Reply #2 on: November 23, 2019, 08:40:35 AM »
|
|
|
Thanks again nyet :p I'm really getting the feeling that you are the only person active here. Hope that setting STLDIA to 10k will be enough to disable I adaptation. But first I need to get my oscilloscope working :p
|
|
|
|
|
Logged
|
|
|
|
|
nyet
|
|
« Reply #3 on: November 23, 2019, 08:57:04 AM »
|
|
|
You also will have the issue of spooling up the integrator w/o p.
Not sure what you are trying to do, but I think its going to be more trouble that it is worth.
|
|
|
|
|
Logged
|
|
|
|
kacperoooni
Full Member
Karma: +9/-2
Offline
Posts: 124
|
|
« Reply #4 on: November 23, 2019, 10:04:03 AM »
|
|
|
You also will have the issue of spooling up the integrator w/o p.
Changing this KFLDRQ1: LDR-Reglerparameter Q1 (Integratorbeiwert) shouldn't zero intergrator term? |
|
|
|
|
Logged
|
|
|
|
|
nyet
|
|
« Reply #5 on: November 23, 2019, 10:07:20 AM »
|
|
|
Changing this KFLDRQ1: LDR-Reglerparameter Q1 (Integratorbeiwert) shouldn't zero intergrator term?
How exactly are you going to get any WG output? I think you should probably abandon this approach if it isn't clear why. |
|
|
|
|
Logged
|
|
|
|
kacperoooni
Full Member
Karma: +9/-2
Offline
Posts: 124
|
|
« Reply #6 on: November 23, 2019, 10:23:31 AM »
|
|
|
Well ok, I'm going to build compressor+turbo 1.8t build. My goal is to transfer by n75 output rpm and requested boost from me 7.5 read by external microcontroler (which would control n75 and compressor bypass valve). I've made my own boost controller some time ago so now programming it again shouldn't make any problems. That time it had crank sensor and tps input. Now I want to simplify it a bit so it could fit more Ecus like MED9 for example. Ofc I could use piggyback or standalone but i would be too easy haha.
|
|
|
|
|
Logged
|
|
|
|
|
nyet
|
|
« Reply #7 on: November 23, 2019, 01:04:16 PM »
|
|
|
the best you can send to n75 output is lde.
d=0 i=0 p=1, linear drl.
|
|
|
|
|
Logged
|
|
|
|
kacperoooni
Full Member
Karma: +9/-2
Offline
Posts: 124
|
|
« Reply #8 on: November 23, 2019, 01:40:17 PM »
|
|
|
the best you can send to n75 output is lde.
d=0 i=0 p=1, linear drl.
well for example 1600hpa axis look like this, 31% dc for 1k 32% dc for 2k 33% dc for 3k 34% dc for 4k etc for 1800hpa like this: 41% dc for 1k 42% dc for 2k 43% dc for 3k 44% dc for 4k etc then doing a simple modulo 10 i have rpm and (dc - modulo10) pressure (lda) |
|
|
|
« Last Edit: November 23, 2019, 01:44:47 PM by kacperoooni »
|
Logged
|
|
|
|
|
nyet
|
|
« Reply #9 on: November 23, 2019, 01:43:12 PM »
|
|
|
The input to the PID is not req boost, it is lde. You can't get the PID to follow imx unless can convince the PID to spool up, and even then, if you go under boost, DC will fall.
I don't really think you understand how the PID works if you designed around the idea that you can convince the PID to output req boost.
You're best bet (if you really are set on this foolish idea) is to asm hack the ecu to send req boost directly to the n75 output.
|
|
|
|
|
Logged
|
|
|
|
kacperoooni
Full Member
Karma: +9/-2
Offline
Posts: 124
|
|
« Reply #10 on: November 23, 2019, 01:57:55 PM »
|
|
|
The input to the PID is not req boost, it is lde. You can't get the PID to follow imx unless can convince the PID to spool up, and even then, if you go under boost, DC will fall.
from s4 wiki: lda ("actual absolute pressure" - "requested pressure") So I thought by simply adding actual absolute pressure to n75 output I would get absolute boost. Doesn't KFLDRL limit DC even under spool? Well I still can use RAPP which I wanted to avoid. I was inspired by pseudo feed-forward boost control posted here |
|
|
|
« Last Edit: November 23, 2019, 02:04:50 PM by kacperoooni »
|
Logged
|
|
|
|
|
