WG step delay

Sy#26

New member
Is there any delay built into the code relative to stepping the WG solenoid to higher or lower values? Reason I ask is, even w/ the delays set to zero (in the data accessable to Promgrammer), I see only one WG boost correction step per ALDL frame, which occurs roughly every 100ms. Seems to me this is pretty slow response if the actual MAP is far off the target value.

Thanks -
 
B

Blake

Guest
Barry,
This isn't my info, but considering the source, I'm certain it's correct. I included the whole thing in the event there is another part of it you need for reference.
I hope I understood the question properly.
George Blake

The following is a quote from BG:

First, needs to be determined if closed loop boost is enabled

0x03B3 KCLBSTEH Close Loop Boost Mode Threshold 39.84 kPa

0x03B4 KCLBSTEL Close Loop Boost Mode Threshold 5.47 kPa

These are compared to the current boost level, enabled at 5.47kpa (as it increases), disabled at 39.84kpa (as it decreases)
0x03A5 KGDCTMPL Wastegate Low Coolant Threshold 20.00 Deg. C

0x03A6 KGDCTMPH Wastegate High Coolant Threshold 115.25 deg. c

these are checked for closed loop also. Enable if higher than 20 (warmup), disable if higher than 115.
0x03B2 KWGCLTIM Time Threshold for Wastegate Closed Loop 0.4 Sec.


these conditions must be satisfied, then delay of 0.4 sec til closed loop.
ok..start closed loop
0x03BE KWGDCLOW Starting Boost Level 68.75 %

this is loaded as wastegate duty cycle initially (we are dealing with 2 different values now, wastegate duty cycle and actual boost pressures).
i believe this is used when closed loop is not enabled as a base, then used as a basis that is modified to the desired boost (limit boost, overboost, stepup, stepdown boost etc).










The code is complicated, the major functions are below (the ones i understand enough to describe).

0x03C2 F70_TABLE Desired Boost Pressure Base vs NTRPMX and NTPSLD

kPa
| 37.5% 50.0% 62.5% 75.0% 87.5% 100%
---------------------------------------------

2400 | 44.5 44.5 44.5 44.5 44.5 44.5

3200 | 44.5 54.7 54.7 54.7 4.7 54.7

4000 | 44.5 64.8 64.8 64.8 64.8 64.8

4800 | 44.5 75.0 75.0 75.0 75.0 75.0

5600 | 44.5 84.4 84. 84.4 84.4 84.4

6400 | 54.7 98.4 98.4 98.4 98.4 98.4


Desired boost pressure base is looked up based on current TPS and RPM.

0x03AA KGEKCENA Boost Reduction Retard Enable Threshold 7.91Deg.

0x03AB KGEKCDIS Boost Reduction Retard Disable Threshold 6.15Deg.


the base boost looked up (at above table) is reduced when knock retard is above 7.91deg, then the reduction is stopped when knock goes below 6.15deg.
0x03A9 KGDCTIM3 Delay Time to Start Boost Reduction 0.2 Sec.


but it waits this long (above) to begin reduction, and reduces by these amounts:
0x03AC KBSTDLT1 Desired Boost Pressure Reduction Delta 7.03 kPa
0x03AD KBSTDLT2 Desired Boost Pressure Reduction Delta 3.91 kPa
(why 2? i forget)

closed loop functions:
compares actual boost level to the desired boost level if error greater than:
0x03B5 KBSTERDB Boost Error Deadband 3.13 kPa

it will make adjustments to the wg duty cycle. if the error is less than deadband, no change in current duty cycle.
0x03B6 KDCTIM8N Closed Loop Update, High Boost 0.2 Sec.
0x03B7 KDCTIM8P Closed Loop Update, Low Boost 0.2 Sec.
if a change is required in duty cycle (error larger than deadband) the code waits this long to do it (why delay is my question)

ok, so we need to change (error is greater than deadband, delay has been met) the set values we use for up/down of the duty cycle is
0x03B8 KDCSTEN Negative Step Adjustment to Wastegate DC 1.95 %
0x03B9 KDCSTEP Positive Step Adjustment to Wastegate DC 3.13 %
 

Sy#26

New member
Thanks for the reply, George. Yeah, I have the same info that you provided. I'm learning a lot about WG control; here's where I am so far:

"The best WG control is the least control needed". So, what does that mean? Means that the closer you run to (above) the built-in 'base' pressure of the WG actuator, the better results you're going to get. The problems w/ WG control are two (mainly); 'spiking' on sudden throttle application (or forced downshift), and 'sagging', slow decay of boost once set pressure is reached ("creep" not included, because that's just a factor of a crummy exhaust system &/or WG). Assuming the optimal plumbing - I'm talking integral WG here - good exhaust system, adequate WG diameter, you'll get the best control by hooking the actuator directly to the compressor nipple. Spikes occur when the turbine bypass is inadequate, either due to being too small, or by the actuator reacting too slowly. Actuator will react most quickly w/o anything else in the path to it, i.e., the solenoid; it does present a bit of restriction (but that can be improved). Typically, with TC locked in 3rd gear, WOT is applied @ a speed high enough to prevent a downshift - 54mph on my Sy. What you'll see is a rapid climb to somewhere from 1-2psi above your base pressure, a fall to a 'plateau' (a flat section which is your base pressure) for 1.5k rpm or so, then a gradual fall in boost due to increased engine load & increased exhaust backpressure related to that.

I'm using a "19#" actuator, but its *actual* opening pressure on the bench is around 22psi. Backpressure acts along with the internal spring helping to open the flapper valve, so 22# bench opening pressure = 19# actual base pressure. This is operating a 1.45" flapper in a 14cm port-matched Mitsu feeding a 3" all the way, cat-less exhaust.

Solenoid-wise, its pretty non-linear in its output pressure. Applying 20psi input pressure, I got these figures (average of 3 runs). The solenoid operates @ around 20cps, duty cycle varies - 12v square wave.

Duty Cycle Output Pressure, psi
----------------- --------------------------
0 20
10 20 - (no solenoid action heard)
20 19.4 - *very slight* solenoid noise heard
30 18 - getting louder
40 16
50 13.5 - very noisy
60 9.7
70 6.0
80 3.2
90 1.0
100 zero

So, things don't *really* start happening till you're above 50% D/C. Where I am with this right now is setting closed loop boost control to begin at around 16psi, starting D/C at zero. Reason being is that boost rises *fast* and you don't want the open loop starting too early. The initial D/C is zero, but will begin increasing immediately until the software set pressure is reached. That way, you get zero D/C right off and just a positive step or two before boost target is hit; this steps back to zero almost immediately. My set pressure is just a little bit below base boost, set mechanically. What I see is zero D/C through the initial spike, through the 'platuau' and shortly into the following droop in pressure. Soon as the MAP drops below software set pressure (w/ real small hysteresis), D/C starts stepping up. However, it has to get to around 70% D/C before the downward trend in boost is reversed. To get from zero D/C to 70% takes almost a full second (pos increment @ 5.08%) - pressure has fallen a full psi in that time. The parameters in the Promgrammer data are all set to zero delays. By increasing the positive step in D/C to, say, 10%, gives a faster rise, but choppier boost control. BTW, boring the solenoid armature out to 1/8" greatly increases flow through it; will get boost signal to the actuator a lot faster. There's a small lip seal in the stocker that's destroyed by enlarging it, but action is seemingly not affected - same results on the bench.

How come the D/C takes almost a full second to get from zero to 70%? Maybe by setting the software 'desired boost' to just an RCH below the actual mechanical boost set point, I can get the D/C climb to begin @ the beginning of the plateau area. Cutting it pretty fine, but I'll see if that's do-able. Too bad actual measured pressures and what the ECM sees aren't *exactly* the same - fudge factor needed here.

Parts are ordered for a butterfly style WG added onto the Mitsu that hopefully will work acceptably and negate the backpressure effect problem. If the WG wasn't acted upon by exhaust backpressure, things would be a lot simpler....
 
Top