do you have boost logs?

Here's something from today's morning (dump, log and map pack).

VCM's sampling is too slow to get useful data for quick changes - the first value is time in seconds. I was logging 4 values and got a sample for every 1,5 seconds. :problem: If I logged N75 duty cycle, it would be even slower.

That shifting-problem occured at time "1773.187000", row 46 in a .cvs file - then I released the throttle to keep the engine on its mounts.. :wtf:

Today I got into the dump again and searched for additional maps. I hope my predictions are correct: there are 3 maps that I assume are PID factors for N75 control - 6B718, 6B838 and 6B956 (in the attached map pack).

N75 is closed-loop controlled, N75 map is for open-loop control (more or less for quick changes only). Closed loop regulation is made using PID controller, of which D parameter is, as I assume, too high.

Will check, recalculate, try and report.

I extended PID maps and increased I and D parameters above 60 mg/str. It feels a little better, those surges are now felt only in 4rd and 5th gears when holding WOT for some time above 3500 rpm and then quickly shifting to higher gear.

Now I found out that my N75 is totally out. I will do a rough tune for now, in hope things will get better, and fine tune it after SOI maps are finished.

Now I experience massive overboost, but throttle response is quite a fair bit smoother. :thumbup:

Code: Select all

t [s]      MAF [g/s]      MAP [kPa]   IQ [mg/str]   rpm   VGT [%]               
1395.265      24.388535   107      19.937695      1169   65.000381               
1397.406      24.11833   112      47.102802      1166   66.062408               
1399.531      33.075138   129      51.900311      1262   80.003052               
1401.656      39.439976   160      67.41433      1426   75.376518               
1403.781      77.979263   229      78.878502      1722   50.751507               
1405.906      101.507149   250      89.096573      2096   41.873806               
1408.046      121.702492   250      95.700935      2445   30.811016               
1410.171      132.991074   250      95.700935      2745   22.998398               
1412.296      135.112686   250      91.09034      3003   22.000458               
1414.421      113.566307   226      80.124611      3225   38.001068


Is anyone else having trouble connecting OBD Car Doctor (Android) to ELM327 Bluetooth interface after last update? :( Too bad, that app is having really nice real-time graphs.. :problem:

do you have n75 vacum valve, or electrical actuator?

I have a N75 vacuum valve.

you could have problem with valve itself

The valve is fine.

I roughly recalculated PID parameters and did some changes to N75, today I was a bit underboosted, but things are getting better.

I could get it into surges only once this morning - had to rev it over 4200 rpm and shift it really quick, so I guess I'm on a good way to make this car driveable. :mrgreen:

I was so happy that I just ordered 18" Tomason rims and lowering springs.. :mrgreen: :mrgreen:

great infos!

checcoiamma wrote:great infos!

Yeah, the rims and springs make this car look like a beast.. :lol:

It's been a long time... ...since I've done last significant mods.

Had 2 breakdowns this year, the first is broken flexi-pipe (after catalyst) and the second is broken pipe between intercooler and EGR - no big deal, both cost me 35 eur to fix. A flexi pipe is 3 mm wider than original. I also made a model for de-cat pipe - I am about to make and mount it next week.

Anyway, I've tried many different setups, and always got smoke above 90 mg/str. It became a bit annoying. Turbo boost is somewhere around 2.35 to 2.45 bar, I don't have guts to go above that with a MAP of 2.5 bar measure range.

It came time to do something about it, I am looking for a MAP linearisation map. Not sure about it, but I found something very similar to what I am looking for:

Code: Select all

00003 04480 00010 00250

That could mean 10 kPa at 3 mV and 250 kPa at 4480 mV - it would seem a bit logical if there weren's 2 things:

first: the sensor data is 250 mV at 36 kPa and 261 kPa at 4750 mV. That would mean that there is 4530 mV at 250 kPa, not 4480. And it should be -270 mV at 10 kPa, not 3 mV.

There is also measuring range problem - a sensor measures from 36 kPa up, not 10 kPa.

Otherwise these values seem quite similar - I can't find anything closer to what should be a linearization map:

Code: Select all

Sensor:
mV   kPa
250 36
4750 261
function: f(x) = 0.05x + 23,5

Program:
mV   kPa
3    10
4480 250
function: f(x) = 0,0536x + 9,84


This part of code is at address 6311a, surrounded by this;

Code: Select all

03000 10000 03000 10000 05000 03000 02000 00000 00003 04480 00010 00250 00000 04097 00256 00256

Any experience to share? Please? :)