NefMoto

HI,, Today i was using a program at my work which simulates a logical controler that i program and simulate my program on.. But I have heard that engine devolpers are using this methods to..  Could it be possible to simulate a WinOLS proram in a vritual Bosch Me7.x.x for example?
Crazy idea but would be awsome..

So you want to simulate all the sensors in the car on the bench?

So, speed, rpm, coolant temp, aircon on or off, comms with airbag, cluster, knock, ait, boost.

Why?

type virtual ecu into google , their are quit a lot of programs doing this , but looks like a lot of money .

by the way , I was on here at 3 in the morning , reading this forum and looking up virtual ecu on google  :o fookin spooky or what !! lol :D

Yes thats spooky jimmcgee5 :) by the way hi im also a Jim googling this right now.

I want to simluate an ECU connected to an engine with sensors and other stuff.. Why because it would be cool... I know there is some really expensive platforms tha car companies use.  I googling as hell right now.

You'd need to know the details and ranges of every sensor, it's default working value and a way of controlling all it's ranges (assuming you're going to do it all yourself).

The cost of a ready-built kit would be a fortune.  You'd be much cheaper building a frame/engine-support and having a standalone engine run in your garage.

Yes probably...  But i wonder what I could do with labview and a Labjack connected to an ECU. Generate a trigger signal and just use some static values on generating temps and so on to start with.. wouldnt that be cool :) I think i can make curves in LabView to simluate som temp resistors.

You probably could, but I don't really understand what you want to see that you can't log in realtime on the road? 

 :P what if its 3 in the morning and you cant be arsed going out in hurricane bawbag and want to test a tune on your ecu  :P , oh and someone just posted a pic of snow on the m8 , even less reason to go outside now ! ;D

Hehe well i have no good answer to why exactly.. Only i can come up with is that you can sit at home and tune ECU and drink beer at the same time :) But probably im a geek that can´t explain some stuff I do. Maybe I could test some functions that i dont want to test irl.

Tempsensors can be made in Labview and you can also create a resistance curve. . That can be used for every sensor. You can also form your own primary and scondary trigger pulse then you do some sort of environment called "Tune landia" with an ambient temp and so on.. Ofcourse this wont be water proof but..

It's hurricane Abigale now haha.  A big gale.  Shite patter.

But aye, it's snowing in Lenzie too, feck.  I'm gonna take the car into the garage tonight to flash, bugger flashing in the snow.

Snow? Heck, that's the main reason for having a quattro in the first place :)

Well snow is a good reason to buy a quattro... but i bought the hairdresser model (TT) it only for summer.. haha and my "Real quattro" suffered from a brun out.. I kinda burned the ECU.. and the whole car..

Yeah same here in Sweden... Its getting cold and dark.. = Stay at home drink beer and keep varm and do something geeky. :)

Well one can build an ecu with LabView

https://www.youtube.com/watch?v=UqRoWkUPvik

Nationl instruments Engine sim program
https://decibel.ni.com/content/docs/DOC-5684

Still wont have feedback from an actual engine, knock sensor feedback, power numbers, etc.. There are people that do this for Megasquirt I think it's called the "stimulator" (insert joke here).

read.. it has knocking simulated !!  :O)   Megasquirt stimulator I know about it... it sounds lika a dildo " (Joke inserted = TRUE)

I understand that, but not having an actual engine to produce the knocking isn't going to allow you to learn anything. You already know an engine that knocks pulls timing from requested.  All engines are different, and react different to conditions. In addition the knock thresholds, and frequencies they look for differ on different engines. How are you going to simulate octane rating? Mixture differences? Effects of timing advance on an actual engine?

audiguy yes all engines are diffrent so on and so forth...  Im not saying this will be a total simulated engine... I just want to start with something simple to se how the ECU react if im manipulating some values.

Imagine a PC connected to a LabJack and a harness into the ECU. In LabView I will simulate an rpm signal, AFR, MAF, MAP, IAT and what have you.
So lets say i crank up my Simulator with a start button.

1. The ECU see the RPM signal and initiates cranking and cold startup.
2. The simulator recives a pulse from injector outputs and the ingitionpulses and reacts after that.
3. Engine is now at idle thanks to the ECU giving the simulator ign pulses and fuel.
4. Throttle feedback into ECU is feed via either external potentiometer or an outbut from the labjack controlled by a slider in the program or something.
5. Simulate a road and build a simple model for car running on a road with ambient temp. If car runns faster cooler air and so on..

Its a stuupid idea but it would be fun to just experiment with. Not sying this will be a fraud or sucsess. Its just geeky fun..
But it requiers a pretty decent Labboard with a good I/O.

You could probably get this to work, but it would work, not be functional.  I don't see the point really, there's way too many variables to try and simulate accurately, and even if you manage to get one set of variables to work, you can't vary anything because you don't know how the variation affects other sensors.

Say, for example, the "injectors" are on for 13ms, how much fuel is injected?  What happens to the lambda reading?  Do you force a lambda reading, so say it "sees" 0.83, what happens to the injectors?  I'm not trying to shoot you down, but you'd be much better buying a donor and mounting it on a stand and venting the exhaust outside so you can tune inside.  Hell, it's easier to set up an emulator for the ECU so you can monitor/edit variables in realtime.

IDK about where you're from, but in the US I can find an ME7 car that is beat to shit for 1-200 bucks.

Likewise, can pick up abused 1.8t's for about £150/200 quid here.

Have fun logging analog signals from hall sensors for crank and camshafts and then simulate these in labview.

Pros and cons are both correct. You wont learn much but as far as testing what your mods will do to actuators it could be useful.  :)   I like the idea of the stimulator from DIY auto tune. You make it yourself. It would be easy to put wires from the stimulator to the ECM and have a vag com wired as well to monitor the signals. This way you wouldn't have to figure out how to create a signal to feed the crank or air flow. it is all on the stimulator board.

Yes that greate but.. if you use Arduino its even better and cheaper i have already started the lab. :)   There is already code for free out there and you can simulat any trigger wheel and so on... I have started with my first test but I changed forum for this project.

First laboration will be controlling a Volvo 6 cyl with a Speeduino just to get familiar with the system.   http://blogs.libreems.org/arduino-wheel-simulator/.

As well as analogue or pwn inputs you'd also need a can bus network I think or it will shut off.   Tester present even over kwp might be enough.

I think this is a great idea... Oh wait a minute. Here is one of the units on my desk. I use the siglent to generate the crank/cam signal and I use SPI powered digital pots and DACs to stimulate the inputs. Most cases I have the pot board plugged in as it's easier to work with. Each ECU I do has a different board that gets plugged into the backplane. I capture the injection PW with a scope.

Since each ECU has its own challenges (some DBW some not, some need to talk to many CAN components on the bus, some do not) there are a number of boards I can choose to plug in and use to keep a given ECU happy.

For simulating the ECU itself I've looked into that but for example to realtime simulate the ST10 series of processors using a lauterbach machine is about $50k USD. Outside what I can cost justify.

So I think I got this figured out...I set up my laptop on the left side, and on the right, I have a cordless drill cable tied to the table. I press the buttons on the left, and rev up the drill, and pretend I'm making 400 hp with slight knocking at 4k RPM on E85 with antilag. I then open the windows in the room (to simulate cold start) and then press some more buttons in the cold start map, (and use much less finger pressure on the drill) to simulate a colder engine. I pretty much saved 49K using this setup, and all I needed was to close my eyes and use my imagination. I can't justify spending 50k to simulate an engine virtually. Probably only 45k.

your being sarcastic aren't you  :P

Seems funny but back in the day I did successfully run many an "engine" using a distributor on a drill. It's a viable alternative if you're just doing a few simple things.

 ;D

right back on track here  :D, what we need is a windows emulator that will run an me7 bin file as if its really in an me7 ecu unit connected to a virtual engine ,  :)

ahhh cool project you got there hackish... I simulate the trigger via 60-2 simulated signal from an arduino now.. works perfekt :)   I hocked up VAG-com and logged the simulation. Now im ready for nextlevel :)  I´ll post pics soon.. Or maybe a video

While interesting that's a project that would take someone a year or so full time. Anyone looking to do a PhD thesis? A better approach would be to patch it so the RAM/ROM is off-chip then use that to run a ram based emulator. Why bother to emulate hardware that's readily available and cheap?

Think rpm...

Could you mot use moats?   Erm can the vag ecu not execute external code like Volvo?   Bootloader.   Or is it all coded into the ecu?