Racing related features

[jsmcortina]

Some features I have in mind to add:
Time (from launch) related boost
Switchable boost tables
Time (from launch) related nitrous stage 2 activation

Proportional nitrous (would need to use high-current outputs)

Adjustable launch rpm (2-step limiter) via analogue input

James

[UnaClocker]

These features sound good. Also
Back during the MS2-E development, I remember someone mentioning a trick some old Ford ECM's do.. Rate of RPM climb based spark advance.. So you get more spark advance in lower gears, and not as much when you're pullin a big hill in a higher gear or what not. I guess it'd work a lot like the MS1 traction control (and almost be the opposite of traction control).. But seemed like a good idea anyways..

[muythaibxr]

These features sound good. Also
Back during the MS2-E development, I remember someone mentioning a trick some old Ford ECM's do.. Rate of RPM climb based spark advance.. So you get more spark advance in lower gears, and not as much when you're pullin a big hill in a higher gear or what not. I guess it'd work a lot like the MS1 traction control (and almost be the opposite of traction control).. But seemed like a good idea anyways..

I do not like the idea of this feature, and would rather find out what the manufacturers are doing *now* instead. We also discussed that it was *probably* to account for slight spark retard when accelerating faster due to inadequate prediction.

I'm not saying we won't do something to address the problem you're talking about, but I'd rather do something more like what manufacturers are currently doing instead of what the "old" ones do.

Ken

[Crabman]

Now i have a suggestion: Why not having a 3D map for injector opening time, i mean "real fuel pulse" on the X axis and battery voltage on Y axis, "real fuel pulse" is the !!EXACT!! amount of fuel that the engine needs (in milliseconds) and cells in the map are that value plus deadtime. In this way we can define with precision the amount of fuel injected not only for small pulses (under 1-2 ms) but also for saturating injectors that over a certain value of pulse width start to close, so if the injector is used in a state of start-opening-and-sudden-closing (under 1 ms) it is hard to take account of the actual fuel injected ,in the main map in terms of istinctive tuning and things go a lot worse in blend mode, MAYBE with a function like this we can get rid of the second fuel injector rail and go straight to big injectors . BUT we definitively need a lot of info from those injectors, and we must be sure that in short openings there is no variation between one pulse and another the same duration, BUT the first guy that extrapolates some (reliable) data will certanly help a lot the others (especially using same injectors), at the end if you know nothing about dead time you can upload a standard map that will certanly work exellent in poor to mid applications.
If you want to measure the precise amount of fuel sprayed you can arrange a fuel pump, one injector, fuel press. reg. and a (big) T fitting from the fpr out and pump inlet and third spot of T goes to a filled graduate syringe or something equivalent, megasquirt in test mode and so on for finding dead time in every condition.
Whatdoyathink about?

ALX

[muythaibxr]

Ehh, I'm not sure I fully understand what you're getting at, but it sounds like you're saying get rid of VE table and just use pulse-width in the tables instead. If that's the case, that's the opposite of the direction we're going.

We do plan on having the ability to set opening time based on battery voltage instead of just assuming it's linear, but the VE tables will remain just that, VE tables. We will not be changing to PW tables.

Ken

[Paul_VR6]

Externally setup 2step rpm would be great, especially if boundary settings can be in software: min launch 5800 max launch 6500, then 'dial' in between.

I agree the proportional nitrous, and general timed features would be wonderful. I'm sure I can use that staged nitrous timing retard as a down track limiter for bracket racing too.. er, not that I would ever do that!

[Matt Cramer]

Here's another race oriented feature that I've seen requested a few times: A pit speed limiter activated by a button. For cars with no VSS, it could be yet another 2-step. Come to think of it, this would also allow a valet key function for less race oriented machinery.

Juan Pablo Montoya could have used it if he had any electronics allowed.

[jsmcortina]

3-step? Also used as a burnout limiter. Can do that.

James

[Crabman]

What i meant was: an ADDON map for deadtime, some ECUs have it only based on v batt (2D) megasquirt have a fixed correction based on voltage and acts somehow 2D, the third axis i was talking about is pulse width, because injector deadtime is different approaching 1 ms of LOGIC pulse. This 3D map can be usefull for having big injectors and smooth idle for boosted small engines (i have read that it's difficult to tune idle without a second rail of small injectors), you can enter 100 200 500 700 900 1100 1300 1600 microseconds for locations into X axis, and calculating with MAIN map raw idle ve values can correct linearity and errors in two ve values interpolation, it's only an idea, maybe it's useless as hell or will never work

why not having an analog input that for every voltage gives a boost pressure and maybe you can enter an rpm value that below that sets a standard pressure whatever the potentiometer position (i'm dreaming of a bicycle gear selector for boost knob with Shimano logo on it )

PWM nitrous via ECU is sweat, you don't need to make a circuit for sync with tach, what's the amperage of a 200 bar solenoid? is spike damping enough? maybe we can use a nice relay, i've never researched about relay response time but maybe will only increase solenoid deadtime a bit, maybe we need to set a deadtime for fuel and another for nitrous if we use wet system

[muythaibxr]

Ahh, I see what you mean. I already had plans to do something about this once I had a chance to do some research on the actual phenomenon that's occurring.

I think that having a 2D curve based on voltage will go a long way towards fixing these problems. I'll have to see if I can come up with a compensation for the non-linear behavior at low pulse-widths.

Either way it's going to depend greatly on people being able to accurately characterize their injectors, so we'll have to come up with a way of doing that as well.

Ken

[Paul_VR6]

Launch enable/disable by vss would be nice as well. I do like the separate burnout limiter too.

[jsmcortina]

Launch enable/disable by vss would be nice as well.

Explain?

James

[Paul_VR6]

VSS < 3mph enable launch limiter (no switch) VSS <= 3mph, disable launch limiter (3mph chosen for no particular reason). I've never had luck using a clutch switch to launch off of, and using the ebrake mounted switch is a pain. I know there are systems that work like that mostly for Hondas though like Hondata K-Pro and S300.

[jsmcortina]

Clutch switch works great for me (on MS1.) I've not yet hooked it up to MS3 to confirm behaviour, but I will do soon.

James

[Paul_VR6]

Doesn't work too hot for me, I use the clutch switch for flat shift and to get that right it's too low to work for launch also. Have to preload the clutch at launch to make sure the drivetrain leaves with me!

[Black99rt]

I could use the VSS input to the launch code. With an automatic I would normally have the launch input pin wired to my brake light switch and in series with a momentary switch and a normal switch. Without the momentary switch, braking on the top end results in back fires if the system is not turned off mid run (dangerous)

I tested my standalone two-step that designed for my msd6al before driving to the track one day and left it on. (was just on a regular switch and my brake light switch) Since I was never braking above the launch rpm I didn't notice I left it on until I was on the highway. Next to a bus I rode the brakes for 3 seconds, when I got back on the throttle I had a serious backfire and all the school kids screamed I'd like to avoid that in the future.

[Maglin]

Isn't the main issue of large injectors and <1 ms of open time not about pw precision but actually the mechanical deficits that come with large injectors. I was running .7 ms of pw at idle on a set of 1000cc injectors on a 2.0 4 cylinder. AFR's swung between 16.2 and 13.9 at idle which I had to raise from 750 to 1000 rpm. This isn't using a MS btw. It was on a modified stock ECU using a 5v WBO2 input and sequential injection. Which was better than the stock computer at idle on the NB O2. It took some time but it was all about getting the injector dead time nailed down. The large pintle shrouds the fuel charge and since below 1ms of pw you are maybe seeing like .3-.6 ms of actual fuel injection it's really really hard to get a steady clean idle.

I would think the best way to get around this is to go to a returnless fuel system. Have the MS drive the fuel pump driver module to raise and lower fuel pressure and of course you have to use a fuel pressure sensor and input for it. Then at idle you could lower fuel pressure to like 25 PSI to get almost 35% less injector and more idle injector pw. The costs are pretty high for the system components and probably not worth it for most as many people don't idle down the track.

I'm going to run 1650cc injectors now with a MSII/extra sequential and soon MS3 once I get my hands on one. I definitely like and want a good clean idle, but can live with a slightly rough idle. I'll but putting my 2cents worth in.

[Quan-Time]

@JSMcortina / muythaibxr
Concerning the previous question about "factory" traction control. Unfortunately because of the rapid demise of GM (here in aust where i live) i cant hit up some contacts to get exact specs with detailed info, so ill go from memory / experience of what i have. (i will en devour to try and get specifics if i can, but many ppl have now left. Ive worked on fords and holdens from a (mechanical) engineering point of view for several years).

These examples are for the Holden "Commodore" range. Americans probably know them as the 2 door "pontiac coupe".

Traction control is manually selectable via a control switch in the dash / console, except for utes (pickups) which is a "full time" basis. The system is implimented along side the ABS braking system. There are sensors in the gearbox, diff, and front wheels.
Each rear wheel can be controlled via this system.
If any one wheel looses traction, the ABS will apply the brake to that slipping wheel only. It pulses it around 20times a sec.
If you remove the fuse for ABS, the system still works. But expect to "fell" the shunting thru your drive line.

The ABS and TCS (traction control system) uses I2C interface to be controlled.

[Quan-Time]

----------------
Off topic part
----------------
Ive been playing with a 10 GPIO data logger which is about the size of a creditcard. Its I2C enabled, USB, and microSD in fat16 (ask me about that fat16 formatting program and its uses ;)
http://www.sparkfun.com/commerce/produc ... ts_id=8627
This easily fits inside a MS case. Bring 2 wires in, bingo.. You have I2C running in. Use the logger to output data on another channel (almost act as a data converter of sorts) and you could possibly pass it to a spare MS pin, and have the log the data as if it were its own.
The big problem lays with multi input devices. You would need a channel thats able to accept data (8n1 ?) and the logger could then be used for some fairly cool purposes. Long as you write the firmware (bootloader) to log the part you want, and log it the way you want, and even pass it out to something else the way you want. Your set. Most I2C devices have 12k on it. You need to send a request to a memory addy asking its current status. You can also say "change your status to this", such as turning on stepper motors, resetting position (for accel meters), etc.. Many go higher, but the least you will see is usually this amount.

99% of car buttons are I2C nower days. The buttons on your steering wheel, cruise control stick / Stereo control stick, etc..
If you were able to hook into that loom, which is only 2 wires for the whole lot of buttons, imagine the possibility and endless fun you could have.
Im currently playing with accel meter (3axis) and about to put a pressure sensor (front brake line) on. All of which are I2C and provide some awesome tuning options, provided i can have MS "react" to the inputs. Being able to retard fuel and / or spark based on whats happening is the holy grail of bike tuning.

[jsmcortina]

The large pintle shrouds the fuel charge and since below 1ms of pw you are maybe seeing like .3-.6 ms of actual fuel injection it's really really hard to get a steady clean idle.

The individual injector opening curves and non-linear small pulsewidth curves should help here. (Provided you can figure out how to tune them!)

James