PICS: of CHEAP EFI controller.

[eeaston]

I found a site on the net that sells a kit to build a programmable FI controller.  The kit includes all the components, PCB, enclosure, and software for programming.  It uses MAP and temp guages instead of MAF.  The site is"http://www.bgsoflex.com/megasquirt.html"  Of course it may not be in the same league as the Haltech or Electromotive, ;but it's not in the same price league either>  Price of the kit? $120!   Has anyone ever tried this controller?  If you know a little about electronics and a little more about engines, I'd look into it.  I'm buying one as soon as I can get my hands on it.  It'll run any number and about any type of injector and is completely programmable with a laptop.  One downside is that it isn't sequential, it's batch fire.  not a really big drawback for $850 less!  Oh, and it's triggered by the ignition, so there's no crank trigger to install.

[Blue1989TSi]

An FI system for that price? I would be doubious to say the least.

But if it work, then, great. Let us know.  :-[

[eeaston]

I'm sure that the folks who sell their systems for $1000 don't pay even 100 for components(for FI only systems)  I'll post results as soon as I have them.  For now I'll use it to run the stock intake, but I'll be building a 4-injector manifold this summer.

[eeaston]

NOTE!!  This isn't a system, just a bag of parts and an aluminum box.  You stuff the board , solder and build it yourself.

[Morphing89TSI]

!!!!  EXACTLY what I'm looking for!

 

What is the difference however between the firing.  Yes, I am looking to building one.  I have an aquaintance that is actually decoding the signals going to the ECU.  I was going to write the program, but looks like is has been done at a much cheaper price w/ the parts already!!!

 

Now that's a deal ... if it works!!! ha!

 

Morphing89TSI

PS  I'll do more research and let you know how it goes.  Are there any others out there?

[eeaston]

Looks great, doesn't it!

 

The difference between batch-fire and sequential injection:  Sequential injection controls each injector individually, allowing the user to time each injection to correspond with the intake valve opening on that cylinder.  This results in slightly better atomization in the chamber.  Batch fires all 4 injectors simultaneously at a given crank position.  This means that one cylinder may be timed so that the injector fires just before/while the intake valve opens, the other three could fire up to 1.5 crank revs too early, giving the fuel a chance to condense on the walls of the intake and pool up.  As the site points out, this isn't really all that serious, since the camaro's tuned port intake uses batch injection.  For the stock manifold and injectors, it's not a problem at all, oif course.

 

A partial solution:  The MS controller allows the user to run two seperate injector banks(usually used on V engines) and gives the option of firing both banks at once or alternating between banks(ping-pong style).  There is also an option to produce more than one injector pulse for each crank revolution.  If you were to use the two drivers to control two injectors each, ping-pong style and programmed two pulses per ignition event, you could have nearly sequential injection.  Ideally one pulse would fire on the power stroke and one on the intake stroke.  At the very least, this would give consistent A/F mixture from one cylinder to the next.

 

There is a discussion group dedisated to the MS at http://groups.yahoo.com/group/megasquirt/ that is fairly active with tons of archived posts from users.  I'm doing everything I can to get a kit ASAP so I can start programming!

[Joel]

I would like to point some things out about this kind of setup.

 

This ECU is using a tach signal to figure out when injection events should occur. That doesn't work, its far from optimal. Your car will run, it may even slightly give you a tune, but you'll still be in that "less than optimal" stage you are now with the stock ECU.

 

The stock ECU uses 3 magnetic poles inside the distributor for 3 things... Ignition timing (which is almost not required), injection timing, and fuel pump relay.

 

Tach signals are in frequency. There is nothing in that signal to tell you which piston is at TDC, and which valve is opening.

 

Batch mode would be much more optimal then trying a semi-sequential setup, as you have a greater chance of getting the valve timing right (although still not exact).

 

Likewise, there is no ignition control, so if you can get your fuel tuned in, you still have no control on how ignition is done. So you have to base all your fuel maps on how much ignition advance you have already.

 

When it comes down to it, this unit really isn't for "performance" but rather to change away from a stock setup (carb or maybe even tbi) and see some difference in fuel economy and a slight performance gain. If your aiming for high hp, this is not a wise investment. If your looking for a learning experiance, perhaps a little tunability, and < 250hp, this might be an ok unit.

 

Also, keep in mind that the build quality of an OEM ecu is several times better than your average SDS, Haltech, and TEC-3. If you all haven't read the DIY-EFI list, you might want to (DIY-EFI.org), as those "electronics geeks" have gone over that several times. If you believe that you can build a better unit than the OEM as far as longevity, you may find yourself very mistaken.

 

Joel

[eeaston]

Joel,

 

I do tend to agree that batch firing will be better than trying to rig a semi-sequential system, not because there's a better chance of getting timing right(with batch you'll never have more than once cylinder timed "right), but because there's not a chance of having all 4 timed wrong.  As I pointed out before, for the stock manifold, it's all batch anyhow.

 

As for the performance capabilities of the unit, I definately beg to differ.  THis controller is functionally equivalent to the SDS EM-3D FI controller, and I don't think there's an argument as to whether that offers performance gains.  The only real weakness is as you pointed out that it doesn't control ignition, however this isn't remedied by the AIC and extra injector setup and lots of people swear by this approach.

 

The point is that for less than the price of that AIC, you can get a fully programmable FI controller and its capabilities are limited only by ignition, fuel system, and the user's ability to tune it!

 

Thanks for pointing out potential shortcomings, I do need the brain food!

 

As for longevity; If you think longevity is a primary concern to anyone here, then you must be looking for the Honda Accord discussion group!  ;D

 

P.S. What options are available for ignition control besides an all-in-one unit or the MSD 6BTM?  Thanks again for the discussion!

[Joel]

Joel,

 

I do tend to agree that batch firing will be better than trying to rig a semi-sequential system, not because there's a better chance of getting timing right(with batch you'll never have more than once cylinder timed "right), but because there's not a chance of having all 4 timed wrong.  As I pointed out before, for the stock manifold, it's all batch anyhow.

 

As for the performance capabilities of the unit, I definately beg to differ.  THis controller is functionally equivalent to the SDS EM-3D FI controller, and I don't think there's an argument as to whether that offers performance gains.  The only real weakness is as you pointed out that it doesn't control ignition, however this isn't remedied by the AIC and extra injector setup and lots of people swear by this approach.

 

The point is that for less than the price of that AIC, you can get a fully programmable FI controller and its capabilities are limited only by ignition, fuel system, and the user's ability to tune it!

 

Thanks for pointing out potential shortcomings, I do need the brain food!

 

As for longevity; If you think longevity is a primary concern to anyone here, then you must be looking for the Honda Accord discussion group!  ;D

 

P.S. What options are available for ignition control besides an all-in-one unit or the MSD 6BTM?  Thanks again for the discussion!

 

I think you miss the point I was trying to get accross with performance in regards to this unit.

 

The SDS units ALL use a trigger, this one does not. "Functionality wise" doesn't nessisarily say something about how well something performs.

 

Look at the SDS units, and you will find a triggering device. You will have a trigger device that (on a four cylinder there is 2 event triggers) so the injectors know when to fire.

 

This device on the other hand, does not do that in any way. It "calculates" when the motor will need fuel, and when you have your foot stuck in it at 20psi, I really doubt this 8mhz processor will calculate it fast enough to keep up.

 

I still don't believe this system is optimal in anyway about 250hp, if even that. Stock intake or not. This device has no idea when to start firing injectors, and its no configurable in that way. 3000RPM can be read in to start at #1 cyl TDC, and the very next time it could be exactly the opposite. I'm honestly suprised it runs.

 

If you've read up on how Multipoint injection works (Multipoint is batched), there is always a trigger. only CIS type injection systems (which use fuel valves for injectors) don't have a "trigger".

 

Also, after learning some interesting things last night, there is very little need to use barometric correction factors on a speed density type system. If your using a MAP sensor, your fuel correction will almost never need to be adjusted for altitude. If you live at sea level, you'll get the "same" tune as at 5000ft, but you'll never get to the last 1 or 2 load bars at WOT. The fuel values will always be the same. So including that feature (unless I missed something and you can use the TPS as your load sensing device) is almost pointless.

 

Not being able to match ignition to fuel is one problem, not being able to fire the injectors during the correct time is a major drawback.

 

For the price, however, you can't beat it. But you DO get what you pay for. There is no magic in tuning, its all about having the right equipment, the right support for the equipment, and how much you care to tinker with getting it there. Even the Haltech units, SDS units, and TEC units have complaints that 1000-1200hp is pretty far fetched on these units. People have done it, yes. I would imagine one could possibly get this "up there", but I believe that will that limit the drivability and the safety of the motor.

 

Joel

[eeaston]

Joel,

 

The SDS units ALL use a trigger, this one does not

From SDSEFI.com "Installation of the D system is simple and quick, just 2 grounds, 2 powers and the tach wire hookup."  SDS also uses the tach signal for fuel "timing" and only uses the crank trigger on ignition integrated models, and only for ignition timing!

 

Apparently we have different undertsandings of the two injection types.  If I am off-base, by all means, correct me:

Batch FI by definition does NOT time injection pulses, except to deliver one or more pulses per crank revolution.  Since it fires all (4) injectors at the same time, there's no way to "time" the pulses to correspond with intake valve opening(since they all open at different times).  That's the difference between batch and sequential injection: sequential times injectors individually to correspond with the injector's respective intake valve.  Batch fires all four injectors at once; timing isn't an issue, since at the most one intake valve will be open at the time of firing.

 

It seems to me that with a throttle-body intake, it's all moot because the injectors are so far from the intake valves that I can't imagine distinct fuel pulses remaining intact and undistributed all the way past the throttle plate and through the intake runners.

 

Anyhow, you hit the nail on the head: for the price, you can't beat it.  I'm anxious to get my unit so I can start posting some real data that should prove useful one way or the other!

[tux]

i'm gonna be getting one of these too

 

 

as far as i see i can use stock ecu for fuel pump

 

the megasquirt for fuel injectors

 

and my msd for ignition...

 

 

why is it so bad to have 3?

 

stock ecu used 2

 

and i use 3 right now for my "aic" setup

 

wish i could use my greddy rebec III & simulator for mpfi  what a waste of money it

was.  i'd be happy if it used a map sensor or something

 

i wonder if i could use it for mpfi... only problem is it works on boost/rpm ratio...

[Joel]

Joel,

 

From SDSEFI.com "Installation of the D system is simple and quick, just 2 grounds, 2 powers and the tach wire hookup."  SDS also uses the tach signal for fuel "timing" and only uses the crank trigger on ignition integrated models, and only for ignition timing!

 

Apparently we have different undertsandings of the two injection types.  If I am off-base, by all means, correct me:

Batch FI by definition does NOT time injection pulses, except to deliver one or more pulses per crank revolution.  Since it fires all (4) injectors at the same time, there's no way to "time" the pulses to correspond with intake valve opening(since they all open at different times).  That's the difference between batch and sequential injection: sequential times injectors individually to correspond with the injector's respective intake valve.  Batch fires all four injectors at once; timing isn't an issue, since at the most one intake valve will be open at the time of firing.

 

It seems to me that with a throttle-body intake, it's all moot because the injectors are so far from the intake valves that I can't imagine distinct fuel pulses remaining intact and undistributed all the way past the throttle plate and through the intake runners.

 

Anyhow, you hit the nail on the head: for the price, you can't beat it.  I'm anxious to get my unit so I can start posting some real data that should prove useful one way or the other!

 

Batch injection happens 2 times per crank revolution. Not once. For emissions, thats bad, especially since you essentially doing it "whenever". Batched injection on an ECU that pulls the trigger from a hall effect or some optical/magnetic interface will fire when its "supposed" to, although it still fires them all..

 

Thing about it. Lets start the motor at #1 at TDC. The intake valve is open, and your sucking in your "mix". Without a trigger, the "batched" event could happen... after than valve is closed waiting for the next cycle, or could happen earlier, or later, or, well... whenever.

 

Since sure, you'll have fuel sitting there waiting, its not real optimal for a daily driver, you won't get good gas milage, and its important to not randomly run rich or lean to help with transient response time. The less "random" theses events take place, the more likely you are to be able to tune it out really well.

 

Now, you maybe able to upgrade to a 16G and run 20psi on leaded fuel, maybe, but I think you'll find the net result to be less than optimal, which as you should expect, is relative to how much money you spent.

 

Building the component yourself means your already prone to failure sooner than later. You will be hassled with troubleshooting a PCB that you built, and even when it does run the constant shaking, moving, shocks, and moisture will all effect its lifespan, etc.

 

By all means, don't let me hold you back, I'm explaining what I know, and would hate others that don't understand what they are getting into that this is not a project for your average electronically challenged enthusiast. Its also going to be a tough road to power with tuning due to the inefficiencies in the system.

 

As far as running 3 computers, its not quiet the same thing.

 

An MSD has nothing to do with ignition control... you can buy modules that will statically change the timing, but you cannot reshape the timing curve, or drastically change it. You would be better off keeping the stock setup unless you need the ignition amplifier for high boost.

 

Your stock ECU will do fine with the fuel pump, and ignition events otherwise, but you still don't have the flexibility you could otherwise have.

 

You could probably do decent with the stock timing maps, but pulling the times at a 1/4 mile track will be a bit more limited, as its much easier (and allows you do give more timing) when you on leaded fuels. You'll be able to run more boost, more timing, and all that.

 

If your goal is not to be a dyno queen, or not to run fast 1/4 mile times, then it could be worth a shot. You may even run decent 1/4 miles, but just be warned that if your looking to be faster than the guys with the full setups, save your $150 towards a real setup.

 

Joel

[jinx]

interesting thread.... I think?

anyhoot

its all good to go back and forth with all this "theory", but, bottom line... are people ACTUALLY using this system sucessfully. That will answer ALL questions

 

joel

Are there really ignition tables in the stock starquest ecu ?

I thought we use vacuum advance, centrifugal weights and boost retard

microphone "hears" knock, and control box pulls timing. Is that not how the system works ?  I was under the impression that we only got a staged throttle body injection - a solid state carb, so to speak

 

You say "An MSD has nothing to do with ignition control... "

well, to a degree, what does a btm do ?  what about the DIS programmable systems ?

I think tux was saying to trigger the msd directly from the pickup, as many are currently doing, and fire the coil. Someone posted even using a FORD ignitor apparently. The distributor would still function normally (advance & retard)... but, you won't have the oe knock system. Tonns of turbo cars run superb without a knock system. Fact. It is an added safety OPTION tho'.

 

Someone posted this low cost system some time ago, on ix.net

http://www.perfectpower.com   looks good

I've found someone actually using the PIGGYBACK unit on a toyota 1.6 factory mpfi system, that added a turbo - worked excellent!! altered the fuel and timing curves

[Joel]

interesting thread.... I think?

anyhoot

its all good to go back and forth with all this "theory", but, bottom line... are people ACTUALLY using this system sucessfully. That will answer ALL questions

 

joel

Are there really ignition tables in the stock starquest ecu ?

I thought we use vacuum advance, centrifugal weights and boost retard

microphone "hears" knock, and control box pulls timing. Is that not how the system works ?  I was under the impression that we only got a staged throttle body injection - a solid state carb, so to speak

 

There is rumor both ways on ignition tables. For instance. If there is no "ignition" maps, I wonder where the claim that the 88-89 ECU has superior ingition and fuel maps came from.

 

I do believe these cars will run fine without have a computer controlled unit in place, as the mechanical ignition can take care of most of that. Bottom line is that while you can run perfectly fine without help from the ignition controls of an advanced computer, when asking for huge power from a smaller motor, its certainly not a disadvantage to have the ignition adjustable and multiple levels (RPM and load based).

 

You say "An MSD has nothing to do with ignition control... "

well, to a degree, what does a btm do ?  what about the DIS programmable systems ?

 

An MSD unit does not tell the distributor when to fire, it doesn't do that even on a DIS setup. The BTM just has a vacuum connection that will let you retard based on a per pound of boost setting. I.e. you set it to retard 2 degrees per lb of boost, and you've just retarded your timing 20 degrees at 10psi.

 

I'm going to be using a DIS setup on my car (using a mix of RX7 parts) and a hall effect sensor on the crank.

 

 

I think tux was saying to trigger the msd directly from the pickup, as many are currently doing, and fire the coil. Someone posted even using a FORD ignitor apparently. The distributor would still function normally (advance & retard)... but, you won't have the oe knock system. Tonns of turbo cars run superb without a knock system. Fact. It is an added safety OPTION tho'.

 

Most standalones sold to date (except the AEM, and possibly the ProEFI since the AEM is based off the ProEFI) do not have ANY type of knock control what so ever. Most of the professional tuners claim that using knock control as a tuning tool as a very bad idea. Especially if you plan to drive the car around any where other than the track.

 

Tuning by a knock sensor is basically tuning to the "ragged edge", and if you were to get some interesting gasoline at the pump, or just some change in weather, you could be putting your motor at risk that way.

 

Someone posted this low cost system some time ago, on ix.net

http://www.perfectpower.com   looks good

I've found someone actually using the PIGGYBACK unit on a toyota 1.6 factory mpfi system, that added a turbo - worked excellent!! altered the fuel and timing curves

 

That was me, I posted that information.

 

The Mic-3 by perfect power is an alternative to the more expensive units, don't let that fool you. This unit is mostly a two-dimensional setup and poor resultion (i.e. not as many adjustable points). It also does not support as diverse of a setup (i.e. I don't believe it will support a motronic trigger, which doesn't matter for us anyway).

 

Taking the timing from the coil (i.e. motor position) is not an accurate way to do it. If you think about it, RPM and motor position (where the piston is in what cylinder) has no relationship at all. While you can make a motor run pretty decent firing injectors, emissions won't be as good, and you can still get a decent amount of performance out, especially when going "all out" on the tune for max power.

 

Also, if this unit was to ever control ignition, it would be nessisary for it to use another triggering device (obviously it cannot just guess the values itself for itself).

 

Joel

 

 

 

[eeaston]

Joel,

 

Do you know anything about the perfect power ignition computers-

http://www.perfectpower.com/products/ims.asp

 

i.e. how many RPM/load points it uses, what "coil amplifier" refers to...

 

Thanks in advance.

[Joel]

Joel,

 

Do you know anything about the perfect power ignition computers-

http://www.perfectpower.com/products/ims.asp

 

i.e. how many RPM/load points it uses, what "coil amplifier" refers to...

 

Thanks in advance.

 

The IMS uses Thottle position load sensing... not a great way to do it, but it can be done.

 

I'm gonna try to answer the question the best that I can, as I don't know if this is what your asking..

 

"Coil Amp".. Well, I believe this unit has a built in ignitor, so you don't need a separate one.

 

As far as the load points go, you have 8 throttle position "load points", and 16 RPM steps. The RPM steps they use are not linear, more like a gradient.

 

They have the map start at 200 rpm steps, and by the time you get to 4500 your going in 1000RPM steps.

 

This gives you a total of 128 points of adjustability.

 

The problem is you wont go over (without a good built high rpm motor) over 6000RPM if you like your engine.

 

That means you loose 3 of those RPM points. Now you only have 104 points of adjustment, and the software simply sucks.

 

With the haltech I have upto 10500RPM, adjustable in 500RPM increments (21 RPM points) in which I will only use up to 6500 (as opposed to only the ability to use the 6100 on the IMS as the next RPM point on it us at 7200), so I will have 13 RPM adjustment point there (all in equal proportions).

 

Then I get 32 load bars based on a speed density (as opposed to TPS). I have 416 adjustable points to use now.

 

The biggest thing is that the IMS is using a TPS base for load mapping, in which TPS isn't exactly a great device to use for that, although it can be done. Additionally, since its going to already be a challenge to tune using a TPS based map, now couple that with really confusing and under-developed software to tune with.

 

Joel

[Blue1989TSi]

Any progress with the install.

The raw cost sounds good. Maybe you could make some money selling a complete unit once you figure this ECU out. Just an Idea.

 

F R A N K

 

^{B  l  u  e  1  9  8  9  T  S  i  @  h  o  t  m  a  i  l  .  c  o  m}

[eeaston]

Well, trying to get  a kit is a chore.  It's the GP from HELL.  I'm not sure what the deal is, there are tons of people trying to get kits, but they schedule the GPs and trying to get one sooner is next to impossible.  I'm still trying a few avenues and I'll definately keep my progress posted.  If it works well, I'll try and share the wealth and help to get others set up...Just don't hold your breath waiting!

[eeaston]

Well, finally a minor update- I finally tracked down and got a partial kit.  I have to track down a bunch of capacitors and finish the board, then the task of installing the sucker in the car(a pretty big job in itsself).  If anyone is considering trying this out, the GP ends around the end of July, so get on it, as it'll probably be another 6-8 months before the next one closes.  There is a discussion board at http://groups.yahoo.com/group/megasquirt/  with thousands of posts.  You'll see both success stories and frustrated people.  There are a lot of very informed folks who are very willing to help "the rest" of us out.  If you're even remotely considering this system take a good look at the board.  If you buy one and decide later you don't want it, people never have trouble selling their kits on the board due to the ultra slow GP system.  I'm stoked to get the thing in my car, I'll keep you posted...  There's even a user who is finishing the development of a PDA interface program for Palms so you can tune the thing with a Palm Pilot instead of a laptop!

[tux]

i wonder if you can just buy all this s*** at radioshack?

[eeaston]

Many of the components are available at Radio Shack, but there are some very specialized parts(especially the processor and the printed circuit board) that you can't get there.  All the parts(except the PCB) are available through electronics parts distributors, but the processor has to be programmed, that's where all the work is.  They do publish the source code online if you know enough about electornics to use it, and they sell the PCB by itsself for $12, but the kit would save you a lot of time tracking down parts, and you wouldn't save much money buying it all seperately.

[tux]

ah ok, coz i was just wondering if i could do that instead of waiting

[eeaston]

Well, the computer is completed and installed in the console under the armrest(I removed the box, cut a little plastic, and gutted the ashtray(using the door only as a snap-on access panel to get to the connector in back)).  The GM water temp sensor fits where the vac temp switch was, and I brazed a boss in the OVC pipe for the air temp sensor.  The whole mess is nearly wired up, and if I hadn't started school yesterday, I'd be tuning it right now.  As it is, the car's at my dad's shop in Idaho, and I'm at school in Salt Lake City, UT.  It'll be two weeks before I get home to  finish up.  The install looks cool, rasie the armrest to access the comp and plug in the PC.  Hopefully I'll have more good news to post in a week or two!

 

PS, if you don't know what I'm talking about, see:

http://www.bgsoflex.com/megasquirt.html

and the yahoo group gets a lot of action:

http://groups.yahoo.com/group/megasquirt

[conquestpa]

Hey

 

I've been reading allot about this tonight... Sounds like it will run great....

 

Ive read about people running a max of 21 Psi with a 1 bar Map sensor. You can get allot more PSi then that with a 3 bar GM map sensor.

[eeaston]

All kits now include a 2.5bar map sensor and are loaded with the code for it.  You can use any pressure transducer you want if that's not enough.  It just requires a simple mod to the board and a change in one of the included files in the code.  I'll use the 2.5 bar, and if the day comes hat it's not adequate, i'll change it.