True programable ecu ?

[oldcorollas]

Why is what so inflexible?
The cost to write and test the software such that its 110% perfect and will never be the cause of engine failure, vs the demand of people wanting it.

thats exactly it!!!!!!!!!!!!!!!

for the MS, it's a DIY thing. if it fails and your engine blows up.. too bad.
if you buy a motec and your engine blows up because of a code failure, you get MoTec to buy you a new engine! simple as that.
MS aint 110% perfect, but it's not far off, and is reliable enough to use day to day (i was scared the first times i went thru M5 and ED tunnels ).. i clocked up at least 10,000km and never ever had a fault. Gazadaman did at least 20 or 30,000..

and as Chuck said, sll they sell is a cheap box of silicon with some fancy code
Wolf even grind the part numbers off their chips they are so worried about ppl getting the basic hardware design!

without IP, they have nothing. with MS it's open source, so everyone wins (esp dufus like me who couldn't program to save his life )

[Enchanter]

thats exactly it!!!!!!!!!!!!!!!

for the MS, it's a DIY thing. if it fails and your engine blows up.. too bad.
if you buy a motec and your engine blows up because of a code failure, you get MoTec to buy you a new engine! simple as that.
MS aint 110% perfect, but it's not far off, and is reliable enough to use day to day (i was scared the first times i went thru M5 and ED tunnels ).. i clocked up at least 10,000km and never ever had a fault. Gazadaman did at least 20 or 30,000..

Im megasquirting my boatanchor 5me first, once its up and running in a basic state of tune the 2jz will be dropped in.
Obviously Im not to concerned about the 5me

[thechuckster]

Being a tassie person you should hook up with an RA40 driver who's done the MSquirt to his 18RG. It would be a very similar install for a 5ME (unlocked dizzi, TPS required, lowZ injectors, fuel-pump switch in old AFM).

RA40Celica was chris' account on the old forums - last time i read, he was upgrading to megasquirt-n-spark.

[Starfire]

If you want a really open source ecu you could alway pick up a microsoft pocketpc developer kit and interface board.

700mhz of processor and a touch screen that is smaller than most manufacturers hand controllers should make for an interesting ECU...

[ViPeR_NiPPleX]

hmm, would be cool to see a Z28 powered ecu... pitty they are a bit old

Enchanter - for ms'ing the 5m-e, you will need to swap the TPS sensor for a potentiometer type one instead of the existing switch one. Other than that... should be ok, just make the fuel pump relay output goto the fuel cut off switch pins on the existing afm plug Otherwise you will find the car will start and stall after a couple of seconds.

Has anyone on here got ms running on all factory toyota sensors? triggering from a vr sensor and outputing to an ignitor? thats what i want to try

[chris_rg]

Being a tassie person you should hook up with an RA40 driver who's done the MSquirt to his 18RG. It would be a very similar install for a 5ME (unlocked dizzi, TPS required, lowZ injectors, fuel-pump switch in old AFM).

RA40Celica was chris' account on the old forums - last time i read, he was upgrading to megasquirt-n-spark.

Hi,

this is me formerly RA40Celica on the old forums.

Enchanter, I've had my 18rg running on megasquirt since early 2004 i think. It's been sweet, the whole way has been a learning curve. If your the sort of person who likes DIY projects, and know a bit about electrical stuff (or willing to learn) then you will love it. These days i just drive it pretty much, can't decide what to do next, whether to go down the N/A road with ITBs, CAMs etc or turbo it for fun!, or maybe even a 1ggte running megasquirt. It's very reliable, i've only had one failure on the megasquirt itself, which was my fault anyway.

Your welcome to take a look at my setup and see whats involved. I think your from the north-west, my old stamping ground. So next time im up that way we could catch up.

check out my members ride thread below

Cheers
Chris

[thechuckster]

If you want a really open source ecu you could alway pick up a microsoft pocketpc ..

isn't that something of a contradiction? Micro$oft and Open Source ?

[ViPeR_NiPPleX]

gotta admit, one thing i love about toymods, no bastard pipes up and tries to convince us all that carbies are better...

[sideshow]

the reason the ecus are expensive can also be due to copyright and patents

years ago i bought a new chip for an autronic it was 250 or so dollars
50 bucks was for chip 200 bucks was for rights to the software

what u shgould weigh up is how much yr engine is worth


good ecus can calculate a shitload of information every millisecond

and adjust at the same amount

so this is one difference between cheap and expensive ecus

another thing is interference

whether is in the wiring or radio interference in a race car with telemetry

if an ecu using telemetry in a race car gets any interference during a race it can easily alter the settings and stuff the engine

[timbosaurus]

Real-time software is an absolute necessity of an engine management system. If you want to use your 700MHz Pocket PC with a flash touch screen, the last thing you want/need is windows... even Microsoft’s current “closest-to-real-time” solutions like windows embedded are not real-time enough. And they certainly are not open source.

So hey, why not run open source software on your pocket PC? Surely you can find a Linux port for your platform! Sorry, but unfortunately linux is not real time either. Well, lucky some smart cookie has invented real-time linux! And all that would require is to write your own port, all in days work! So, ok, maybe you do have the time and skills to write a real-time-linux port for your pocketPC. Then i guess, for you, it’s just all too easy to stumble across a circuit diagram and write countless device drivers for all of the application specific IC's and the graphical touch screen that your PPC uses! And all that software only took one weekend to write!

After all that, you now have a super expensive and very powerful gadget that replaces just ONE chip on your engine management circuit board. You’ll still need all the input and output circuitry and protection to condition the sensor signals and MOSFETs to drive the high current devices like injectors and ignition coils.

MegaSquirt is a fantastic project. I've never built one, only because I personally find pleasure in designing my own things from easy to buy parts. A cheap Atmel microcontroller with some sample automotive semiconductor parts from large companies with some homemade code can certainly run an engine... even interface with a PC for programmability. Trust me, i've done it. Is it a pretty green circuit board with extruded aluminium case and machined endplates? no. Does it have countless hours of quality and reliability testing? no. Will i be given a new engine on warranty if it kills an engine? no. did it cost me under $50. yes! Did i smile like a MoFo when it first ran an engine? You betcha!

I guess i just wanted to say that anything is possible. Maybe if what you want hasn't been shared on the web, it is bloody hard and time consuming. Perhaps you could take it upon urself to learn some special skills and put in some extra hours to do it youself and share in return. I can gaurantee it will be satisfying!

[Merudo]

i can see a windows box with the right amount of tampering POSSIBLY becoming a suitable ecu, but the fact is there are far simpler options out there that work just as well

I can see it now! Your ECU BSOD's and stops feeding fuel, engine grenades.

[Starfire]

I can see it now! Your ECU BSOD's and stops feeding fuel, engine grenades.

If it stops feeding fuel it will just stall.

[MWP]

I have actually seen photos & code for a ECU that was made from a normal home PC with some home made PCI cards added for the actual data aquisition and outputs.
It controlled ignition & fuel injection.

I have also made my own timing, fuel & ignition (CDI) systems (PCBs and software) from scratch, which i ran for about 10min before i pulled them due to unreliability and then lost intrest in fixing them.

So these things can be done.
Just they are not done often due to the amount of time needed to make them reliable and practical.

MegaSquirt is one of the only examples of where a home built system has been developed enough for other peoples use.

[crostek]

This is so true,
[QUOTE=timbosaurus]
MegaSquirt is a fantastic project. I've never built one, only because I personally find pleasure in designing my own things from easy to buy parts. A cheap Atmel microcontroller with some sample automotive semiconductor parts from large companies with some homemade code can certainly run an engine... even interface with a PC for programmability. Trust me, i've done it. Is it a pretty green circuit board with extruded aluminium case and machined endplates? no. Does it have countless hours of quality and reliability testing? no. Will i be given a new engine on warranty if it kills an engine? no. did it cost me under $50. yes! Did i smile like a MoFo when it first ran an engine? You betcha! [QUOTE]
I too have been also busy playing with PIC micro controllers and trying sometimes without luck to get a motor to run, or run well, I also am trying motors on PICAXE controllers.
The PICAXE system is quite easy to program, and have real time response and can be over clocked to run faster. the clock speed is also selectable when programming.
Which ever way you go don’t forget the problems with the EPA and getting the car road legal. this is one of the main restrictions that have stop many programmable ECU's

[Squid]

1) PC's can be real time if you use the right OS, QNX for example is a trusted real time OS in the PC world (it will also compile and run on almost any known hardware and I think is often used in motorola MCU's)

2) Knowinf a little bit of C and attempting to program an ECU is a very very bad idea. You must have an in depth knwoledge and understanding of both how C works and the particular hardware (PLL loops A/D conversion etc). The majority of VERY GOOD PROGRAMMERS are unable to write mission critical real time apps that are reliable. My best effort was code that ran for 12 hours with only 2 errors (both were trapped and handled gracefully). This seems like a good error rate, but could kill an engine or if it is on medical grade equiptment a person. Hence why all the best programmers work in these fields...