I'm going to use Windows because forum topic read Windows Fiero ECU

First off, BAH! At saying the car will blow up, the car is NOT going to blow up because of a malfunction of software or PC hardware. Because when the computer comes to a halt, the car will to. Try this, pull your PSU power cable out right now. Did your computer blow up? Didn't think so. The ONLY way any problems will exist is through the coding of the ECU software and real time readout software.

now for the hardware to drive the injectors, First I'd need to know the Voltage and current requirements. (I've seen some vehicle injectors that take 70V!)

Get me the specs and I'd be happy to help with a driver interface

Saturated or HiZ injectors just get zapped with 12V

Peak/hold or LoZ injectors get a 3A or so current spike to open, then are held open by about 1A current.

I think... =)

A harddrive mounted vertically should last longer than once mounted horizontally.
Also, how about using the new Commodore 64. It is powered by a 16 bit 65816 cpu that runs at 16-20Mhz. It's serial, parrallel and CIA(joystick) ports are easy to read in real-time and is a system easy to program for. Less cooling issues also.

The stock Fiero ECU is a dog. It only makes fuel adjustments every 100rpm and timing adjustmenst every 500 rpm. The big kicker there is the timing adjustments. Even if you keep the fueling the same, improving the timing adjustments would make for a smoother motor and would net you a bit more fuel economy and hp. Maybe splicing the injector harness you can also program in a sequential injection firing order... Also, finally add a knock sensor to the v6 block...

Years ago I came across a website that had programatically disassembled the Fiero ECU down to the raw computer code logic. Don't know what the link is anymore but that should be the starting point.

ps,
The old C64 had a cartidge port. There were cartidges that offered EPROM adapters. An eprom could contain all the code and lookup tables. This allows for future mods just by burning a new EPROM just like is done now but would also change/upgrade the program code too. As I recall, the stock Fiero ECU is still a dog compared to the original C=64.

[This message has been edited by lou_dias (edited 04-30-2005).]

Embedded computers don't have to be ridiculously fast. They just have to be fast enough to get the job done.

Not to mention the way factory computers interpolate data between points and make adjustments. Saying that it only adjusts fuel every 100 RPM isn't really true.

------------------

My Web page | The Turbo Super Duty Build.
You know that little voice that says it can't be done? I duct-taped mine's mouth shut and pushed it down a flight of stairs.

I'm not complaining about the fuel table, it's the timing table that needs more resolution... I doubt the interpolation, I don't think it can really respond quicker...hence the need for a faster ECU.

http://www.opussolutions.com/modules.php?name=News&file=article&sid=27
those nice folks make a computer powersupply specifically for automotive use, let me quote

How much timing variation to you WANT across 100 RPM?
Again, it's not a matter of how fast the computer is.

That would be nice. It would make for a smoother engine acceleration and more efficiency. Then you wouldn't have to play with the timing to max out your peak hp but lose off-idle power. 500rpm is 1/10 of the useable RPM range (1000-6000), that's a big chunk of error built-in.

I've been thinking this for a while... You don't know what you're talking about.
How much do you want timing to change across 100 RPM? HINT: not much.
How much do you know about the code? does it round RPM off to the nearest 100 to interpolate the table entries? It understands and reports RPM to the nearest 25 at low RPM and I believe the nearest 50 at high RPM. Not sure what the changeover point is between the two scales.

[This message has been edited by Will (edited 05-04-2005).]

No, I just didn’t want it to get to that point

I don't know if Lou_Dias numbers are correct, but what he is saying is correct. Quicker timing adjustments would allow a smoother engine and improve efficiency, although probably not to a major degree.

Will, you have posted against Lou_Dias a few times, but have yet to post any reason or evidence as to why it would lead to a negative outcome. Remember the point of the new ECU is to improve on the stock ECU on both algorithms and visual output. Although it probably wouldn't be a major increase to adjust timing or fuel at a lower RPM then what the stock ECU is, it would be an improvement none the less.

Thanks for your posts!

Both you and he are discussing a flaw that the Fiero computer DOES NOT HAVE.
Define "quicker timing adjustments".
Idle speed control is handled by an algorithm called a "proportional integral derivative" (PID) controller. It uses the derivative of the idle speed readings to curb changes in idle speed, the integral of the idle speed to return to set point, and the proportional term (I believe) to dampen oscillations. In this controller, data can only be recorded and corrections can only be applied with a certain frequency, called the sample rate. If the sample rate is too fast, the controller won't be able to control the engine properly as it will sample again before the engine can react to the last adjustment. Rapid and radical oscillations in idle speed will be the result. That's not what you want since your stated goal is to have a smoother engine.

Who's talking about idle speed control? I'm talking about WOT programming. Also, with a quicker sample rate, smaller adjustments need to be made for a given interval. Also, since this PID is computing an algorithm, why you couldn't you re-engineer one to handle a faster sampling rate? In the end, you can only make adjustments down to a rate of a piston firing. It's pointless to go faster. We are no where near that, nor do we need to be... However, today's ECU's use 32-bit chips. The Fiero is a through-hole design (vs. surface-mounted). It's old and it's slow compared to today's vehicles. Do you think it's by chance that today's motors are more efficient? It's the systems controlling timing advance and fuel mixture that have become better, more precise.

The engine long block hasn't changed. The systems controlling the engine has. That is why despite tougher emissions requirements, hp has still gone up over time. Otherwise, we'd be fine with the carbuerator.

Do you consider the engine rough at WOT? Does it miss? What do you hope to gain with a faster computer?

Modern computers aren't faster to control more advanced engines. They're faster to deal with the enormous load of bullshit dumped on manufacturers by the EPA (IE, the OBD-II interface standard) as well as the increasing dependence on car area networks. The extra software required for all of this extra responsibility is what demands higher levels of computational power, not necessarily more advanced engines.

Now before you go on about sequential injection and coil per cylinder ignition, realize that I'm talking about what you're talking about... how often, relative to engine operation, the ECM/PCM recalculates timing. It's more often than you think. After all, every time it receives a distributor reference pulse, it has to calculate how long it will wait before triggering the next spark.

In trying to make the 2.8 a world class engine, you're trying to make a silk purse out of a sow's ear. In trying to make a 2.8 a world class engine by going to a faster ECM, you're trying to make a silk purse out of a sow's ear, but starting with the hoof.

I love to see people do more with a PC than play video games!

Are you trying to discourage this whole project?
If someone can build a better mousetrap, why not let him?
If more/faster timing and fueling adjustments can be made than currently are, why not do it? Who's it going to hurt? What's the extra cost? (a few more kilobytes, big deal). Any cpu IEatRice uses is going to run circles around the Fiero ECU. As long as he does his research:

which it seems he is going to, so be it.

I saw the website that decoded the ECU logic back in '96. Back then I was using a Commodore Amiga CD32 game machine with an SX-1 expansion module (8 meg card, ide connectors, serial port, parrallel port addon) running Amiga OS 3.1 and IBrowse web browser to get on the world wide web. So needless to say, I don't have the link. And if I did, chances are that the site is no longer valid.

But some people like arguing for the sake of arguing...

PS,
I loved my Commodore 128D!

[This message has been edited by lou_dias (edited 05-07-2005).]

The timing tables have values every 200-400 RPM... so what? Look at the variation from one cell to the next. Look at the change in slope from one transition to the next. Overall, the map is pretty smooth. At wide open throttle in the meat of the engine's power band (say from 2000 RPM up), there is very little difference from one cell to the next. There is NOTHING that a higher resolution map will give you that the interpolation routines already coded into the ECM will not give you.

As for the timing rising and then falling again... feel free to run 45 degrees of ignition advance at wide open throttle. Just let us know how many pistons you hole...

I saw Cooter's way back when then Shaun got booted and if he's got some dyno charts on the Pennock outcast site, please pass along the link. Funny thing about those manifolds that Cooter used is that at my current job (dealing with home heating systems) I see those manifolds used on boilers. They vary in length but all seem to be about 2" in diameter.

Oh and a TPI v8 was developed at the same time as the Fiero. Let's look at the LT1 or better yet, the LS1 maps for high rpm hp. And any pinging that might occur - that's why we'd like a knock sensor provision.

Also if you look at the table I posted, it does go up to 50 degrees at 4800 rpm in the 30-40 KPA range. Since the table stops at 4800 rpm...what happens after that? How can you interpolate between 2 points with out the end point (6000 or more rpm)?

[This message has been edited by lou_dias (edited 05-08-2005).]

30-40 kPa is light throttle. The mixture is sparse and has low energy density. Large ignition advance angles are best in that situation. The ignition advance decreases as MAP approaches 100 kPa. This is the same phenomenon that requires retarding of ignition timing for positive manifold pressures. The denser charge is more difficult to light and burns more slowly, but it is closer to the volumetric energy density necessary for detonation.

Obviously, the ECM can not interpolate beyond 4800 RPM. It is not to my knowledge programmed with extrapolation routines to figure out where the map was going from there. It just uses the 4800 RPM entry for RPM beyond 4800. This is perfectly acceptable because there's basically nothing happening up there. Beyond 4800 RPM, the 2.8 is making more noise than power. VE is falling, cylinder pressures are falling... there's just not much worth worrying about going on beyond 4800 RPM.
Also, look at the table. The bottom two entries are the same in most columns. The map is very smooth approaching the lower edge. Continuing it with those same values as engine RPM goes outside the map is nothing to worry about.

[This message has been edited by Will (edited 05-08-2005).]

? Posistive manifold pressure increases density. A denser charge is more likely to pre-ignite, that's why turbo'd cars have knock sensors.

And that's why we need a better ECM.

JazzMan

¿

Most of the components inside a vehicle ecm / pcm are not the same as you or I can obtain off the shelf. The reason for this is that the hardware had to be designed specifically to operate in the outdoor environment of a motor vehicle. The operating temperature range of the components found within a personal computers motherboard for example, will typically be 0° to 70° C at best (consumer grade). The components used in a PCM are all "at least" industrial grade. Industrial grade operating temperature range is -40° to 85° C or better.

In most cases the automotive industry has had components specifically produced to obtain an even better operating temperature range than that of an industrial grade component. This is the reason you can jump into your EFI controlled vehicle in the dead of the coldest winter and still have it start at all. Replace that level of design specification with a common PC grade motherboard, and that will no longer be the case.
..
Overall system operating temperature is another big factor here. Water-cooling the processor for over clocking is a great idea for a home pc, but this provides very little benefit to overall system temperature which is a much more important factor in the vehicle environment.

Memory, video processors, hard drives, etc, all develop significant heat, which must be quickly and efficiently dissipated. Ambient operating temperatures above those that provide sufficient heat dissipation will typically result in a system shut down. The best way to address this problem is by designing a system to operate with extremely low current demands. This design paradigm rules out using components such as full sized 3.5" hard drives, huge amounts of ram, high performance / high current processors, and even video subsystems that generate too much heat to operate without direct fan cooling.

A car-pc or vehicle network is fine for some "non-critical" aspects of powertrain operation. An example of a non-critical operation would be something such as switching on a nitrous solenoid. If the system fails you simply don’t get a nitrous charge, so no harm done. The task of controlling of any critical operation should always be dictated to equipment that has been specifically designed to perform these functions flawlessly under any environment that it is typically subjected to.

[This message has been edited by Mick (edited 05-10-2005).]

And that's why consumer grade CPU's and GPU's can be cooled using liquid nitrogen.... which is somewhere around -346*F? Ovbiously you are drasticly underestimating consumer grade hardware. The only problem would be hard drives at low temperatures but they have hard drive heaters for such a thing. The laptops the military uses has them, and they have tested them to work at -80*F and 160*F. Also you can buy waterblocks for everything that makes heat nowadays, memory, GPU's, hard drives, CPU's, and chipsets. I'm working on a watercooled power supply currently - drilling holes through the heatsink on a 'silent' psu.

For a fraction of the cost you are looking at to even develop the system you are proposing, one could have a purpose built vehicle network already proven to be reliable, together with it's on-board artificial intelligence interface that can monitor and tune the performance of your EFI system as you drive.
Kind of nice having the systems motorized in-dash touch screen deploy itself and then being greeted by its AI announcing the current weather forecast and your route specific traffic report.

[This message has been edited by Mick (edited 05-10-2005).]

I'm always open to new ideas, so please point me in the direction of this purposebuilt vehicle network.

[This message has been edited by Mick (edited 05-10-2005).]

Negative, ghost rider. At a given temperature, a denser charge is more likely to detonate. A less dense charge is more likely to pre-ignite. Yes, there is a (very big) difference.
It has to do with energy per unit volume and flame propagation rates...

The point I was trying to make in explaining why the Fiero ECM behaves the way it does is that it would be UTTERLY POINTLESS to map the tables much beyond 4800 RPM. The Fiero engine is DONE by the time it goes off the top of the table.
Before anything you do to the ECM to give it better high RPM control would be worthwhile, you need an engine that makes more than noise at high RPM...
And when you've built such an engine, you can rescale the stock tables in the stock computer anyway...

[This message has been edited by Will (edited 05-11-2005).]

OK, but who stays stock forever? I am stock until the motor dies then I go all out. Honestly, why get an aftermarket ECM if you are going to stay stock. Just getting the aftermarket ECM means you are no longer stock anyway.

Okay I've stated like three times now that the Ecotec is the motor that I will be using for this project.

Why get an aftermarket ECM at all?
For everything but RPM beyond 7K or odd control relationships, a remapped GM computer (not necessarily the Fiero computer... 7730 is probably a better choice) is better than anything on the aftermarket.