Megasquirt vs. Ancient Ford ECU... Help Me Pick

I'm at a bit of a crossroads with the rally XR4Ti- the tired stock engine is coming out, and another 2.3 which came with 80lb injectors, forged everything, roller cam, and a Holset HY35 turbo is getting ready to go in.  That new engine also came with a Megasquirt MS3 and Ford EDIS module running it, which has not been thoroughly tuned but did run well enough to drive around when I bought it.  Obviously, the MS3 has the correct features to let me harness the mighty power of MOAR BOOST, and therefore be faster than the old Thunderbird ECU and Cosworth T3 that motivated the rally car for many events, and this new engine is built to handle it.

BUT

That Thunderbird ECU, archaic though it may be, and in spite of its' many brittle old 80s connectors, is a magical berkeleying wizard box.  It starts the car at -15F.  It starts it when the engine is 250F.  It runs with nearly every sensor unplugged, a fouled O2 sensor covered in oil, and marginal signal from the few that remain.  So I'm tempted to keep it, because I know that I can get it to keep the engine running and the car moving forward long after most EFI systems would have given up.

Am I going to be able to get this level of reliability out of the Megasquirt, or would I be a fool to cast aside this invincible junkyard ECU in the name of frivolous features like diagnostics and tuneability?

Sounds like a question of usage? Are you going to be autocrossing/rallying/otherwise in a position to wish you were making all the p0wars you can? Or is it mostly DD?

I had zero issues I would call "reliability" while DDing an MSed car, but there's no denying that the cold/hot-start tuning is tricky and I never got it to OEM quality. Tuning software has progressed since then... And it was just an MS1, fuel-only.

I'm facing a similar question on the will-be-DD Ranchero, running a mostly stock SBF. The only differentiator is that instead of having both computers, I have neither starting out.

In reply to Ransom :

It's a rally car, so nearly 100% rough offroad race usage.  I don't need a crazy amount of power (what it has now at 18psi is adequate) but more wouldn't be a bad thing.  The biggest thing is that, unless the engine has physically come apart, we need it to keep running as long as possible.

You can't win races from the paddock. If you have a well sorted and reliable option available, I'd probably ride that as long as possible. If you're losing races due to a lack of power, and need the tunability of the MS3, then you can upgrade at that time.

It will take a good bit of time and a lot of work to get that level of reliability out of the MS3, but it is possible. I'm almost there on my 4AGE - although I'll never know if it would start in winter conditions

If the old ECU works and you don't need more power right now, maybe stick with it until you're ready to move ahead with engine upgrades.

GameboyRMH said:

If the old ECU works and you don't need more power right now, maybe stick with it until you're ready to move ahead with engine upgrades.

That's the thing though, I'm swapping the engine for this freshly built, already upgraded one- the one in the car makes more boost than oil pressure at this point 

So the question is whether to move the Megasquirt over at the same time.

I'd keep the MS3, personally.

They are orders of magnitude better than the MS1 in terms of ease of tuning and reliability.

My vote?  Go with the MegaSquirt and enjoy the ability to datalog, monitor data in realtime (if necessary) and fine tune the power output for more power under the curve, which is where the real speed is.

How much can the stock ecu compensate for improvements you make to the engine setup?

I feel like an intercooler and the ability to run a little bit more boost (say 23 psi?) would cover your power goals pretty well. I think you should do the IC either way for engine health/system heat reasons.

Some form of boost controller to be able to switch between stock and whatever the highest boost the stock ecu can handle might be the answer. 

Edit: how's driveability w/ the boost and stock ecu? Any improved response from the ms3 over stock ecu could sway my opinion heavily. Staying out of obstacles is as important as the motor itself not asploding

¯\_(ツ)_/¯ said:

GameboyRMH said:

If the old ECU works and you don't need more power right now, maybe stick with it until you're ready to move ahead with engine upgrades.

That's the thing though, I'm swapping the engine for this freshly built, already upgraded one- the one in the car makes more boost than oil pressure at this point 

So the question is whether to move the Megasquirt over at the same time.

Well in that case, yes you should go to the MS3 at the same time.

In reply to bluej :

How much can the stock ecu compensate for improvements you make to the engine setup?

Seems to be just enough to work with what I have now plus a teeny bit, turning the boost up doesn't make a noticeable difference beyond about 20 psi which is why I set it at 18psi for a little buffer.

I feel like an intercooler and the ability to run a little bit more boost (say 23 psi?) would cover your power goals pretty well. I think you should do the IC either way for engine health/system heat reasons.

I run a Cosworth intercooler currently, always have.  Forgot to mention that.  The parts car came with an even thicker and bigger intercooler (custom?) which fits in the same location but takes much larger tubing.

Some form of boost controller to be able to switch between stock and whatever the highest boost the stock ecu can handle might be the answer. 

It currently has a manual boost controller (set at 18psi as mentioned) which I can bypass for 10psi (like we did for the last stage at Black River after my super sketchy oil line fix)

Edit: how's driveability w/ the boost and stock ecu? Any improved response from the ms3 over stock ecu could sway my opinion heavily. Staying out of obstacles is as important as the motor itself not asploding

It's pretty good, as good as any other 80s turbo car I've driven.

My somewhat limited knowledge of EFI systems is factory ECU's are stone reliable - provided the engine is running within a set of parameters it can deal with. As soon as you start moving outside of those parameters they can sometimes get flakey.

How "modular" are the two systems? How hard would it be to swap them in the paddock? 

Ian F said:

How "modular" are the two systems? How hard would it be to swap them in the paddock? 

Probably a 20 minute job to swap it over... if the engine is cold and everyone knows what they're doing.

I'm less worried about the Ford ECU itself getting flaky due to boost/fuel/whatever and more worried about the 30+yr old connectors and wires in the harness.

I've found MS to be pretty reliable, but getting the tune right for both really cold starts AND really hot starts is time consuming.  Not that it won't run, but it might take a little time to warm up, for example, before it really starts running great.  You might consider getting the engine running on the Ford ECU first and once you know that everything's working, then switch over to a MegaSquirt box.  Doing both at once might put you in a position where you're chasing through MS when the problem is something else -- I chased MS settings for a long long time before finally finding that I had an intermittent fuel pump.

¯\_(ツ)_/¯ said:

I'm at a bit of a crossroads with the rally XR4Ti- the tired stock engine is coming out, and another 2.3 which came with 80lb injectors, forged everything, roller cam, and a Holset HY35 turbo is getting ready to go in.  That new engine also came with a Megasquirt MS3 and Ford EDIS module running it, which has not been thoroughly tuned but did run well enough to drive around when I bought it.  Obviously, the MS3 has the correct features to let me harness the mighty power of MOAR BOOST, and therefore be faster than the old Thunderbird ECU and Cosworth T3 that motivated the rally car for many events, and this new engine is built to handle it.

BUT

That Thunderbird ECU, archaic though it may be, and in spite of its' many brittle old 80s connectors, is a magical berkeleying wizard box.  It starts the car at -15F.  It starts it when the engine is 250F.  It runs with nearly every sensor unplugged, a fouled O2 sensor covered in oil, and marginal signal from the few that remain.  So I'm tempted to keep it, because I know that I can get it to keep the engine running and the car moving forward long after most EFI systems would have given up.

Am I going to be able to get this level of reliability out of the Megasquirt, or would I be a fool to cast aside this invincible junkyard ECU in the name of frivolous features like diagnostics and tuneability?

 

The way I see your specific question is a little different.  And I'm not including any question about the stock calibration vs. the modified car- as that may answer itself.

I'd separate your reliabilty points quite a bit differently.

1) basic hardware reliability.  The OEM ECU has been through a gauntlet of testing, including some EMF testing near some very powerful radio stations near Dayton, Ohio.  The board has been beaten up, and after +3o years, it's probably starting to possibily see the PCB coming apart.  For sure the wiring harness is old and will end up being a problem.  So can the MS system be that reliable?  Some of it certainly should be, as the same electronics have been been put on smaller, individual boards that take less power and are more reliable.  And for sure, if the wiring harness is built correctly, it will physically be better.

2) software robustness.  The car you are asking about is before my time, a little, but I've seen examples of the code- as people who started before me kept the old code as an example of the code "blowing up"- we went from 30 pages of simple ladder logic just before I started, to hundreds of the same before OBDII got going in '96.  And now books are so big, you can't print them.  But the point is that the code in your Merkur is pretty simple.  And it's highly likely that much of the MS code is better.  The areas that I don't know is robustness- I'm sure the OEM code makes sure that it will run from -20 to +130 ambient, and that there is compensation for some of the various fuels available in the US market at the time.  Does the MS have that?  Does that even matter?  The most important thing is that the software should be robust to enough outside your operating conditions that you will be able to deal with field variability.

3) Calibration robustness.  This is where a lot of aftermarket systems run into problems.  They are *capable* of running a big range of conditions.  But the calibrator doesn't do the work, whereas the OEM calibrator is required to complete a full process of that so that the customer does not notice any issues.  Some of this YOU can deal with- as long as you are willing to constantly tweak the calibration as you experience different conditions.  You need to constantly be logging your driving, you need to make sure that your car is equipped with enough extra sensors to detect what needs to be done, and you need to analize the data so that you can constantly update the calibration.  OEMs put hundreds of thousands of miles on prototype development, go to Alaska in the middle of winter, go to Arizona in the middle of summer, go to Colorado and Death Valley for altitude work, constantly drive cars to tweak things for 2-3 years.  Of course, you don't need to repeat that- but some of that work will be required.  

I'm no MS expert, but IMHO, it seems to me that you *can* get the level of reliability you want.  But it's going to take work.

¯\_(ツ)_/¯ said:

Ian F said:

How "modular" are the two systems? How hard would it be to swap them in the paddock? 

Probably a 20 minute job to swap it over... if the engine is cold and everyone knows what they're doing.

I'm less worried about the Ford ECU itself getting flaky due to boost/fuel/whatever and more worried about the 30+yr old connectors and wires in the harness.

On that note- as long as you can find a source for the pins, the original harness can be rebuilt.  Takes some time and effort, but you can do it.

I've started MS'ed vehicles at 10 below F.  If you tune it, no problem.  I've put 80K on the MS I drove to work yesterday. 

If it is modded, stick with the MS and tune it.  If it's stock, then a stock ECU is probably the overall easier way to go, depending on wiring.

I have been field-testing that Ford ECU (the EEC-IV, I assume) to see how idiot-proof it is.  I ran it for months with my throttle position sensor wired in backwards.  So it was transmitting WOT when it was really at idle, and idle when it was WOT.

Ran surprisingly well.  I mean, it had a variety of undesirable symptoms, but I could still flog the car at the track without stalling it or blowing it up, for example.

I had no idea it was possible to use an EEC-IV to control boost.

Anyway, if you would like more programmability, functions, and sensor monitoring and still keep your EEC-IV, many people seem to like the Moates Quarterhorse daughterboard. 

In reply to alfadriver :

Thanks for the detailed reply.  Before I started pulling it (the parts car with the MS3 and fancy engine) apart, I spent a snowy morning tweaking the tune for startup and idle to get it to function well- but that's only 20F or so, and if we attend a rally colder than that I'll need to make sure I have the ability to alter the tune accordingly.  Same for hot, same for a bunch of scenarios... and that's what scares me on the calibration end.

JBasham said:

I had no idea it was possible to use an EEC-IV to control boost.

It doesn't- berkeleying thing has no idea what the boost is, it just depends on the fuel pressure regulator to take care of it.  I'm pretty sure all it actually needs as an input at WOT is a signal telling it how many RPM the engine is turning.

¯\_(ツ)_/¯ said:

JBasham said:

I had no idea it was possible to use an EEC-IV to control boost.

It doesn't- berkeleying thing has no idea what the boost is, it just depends on the fuel pressure regulator to take care of it.  I'm pretty sure all it actually needs as an input at WOT is a signal telling it how many RPM the engine is turning.

The stock system measures air flow, and at WOT, makes a decent guess on how much fuel to inject based on that measurement.  As long as you don't overly saturate the air meter, it will work- all the tables will just run off the top line of .9 load.

If the car is running on 80lb injectors, I would make a decent guess that the air meter has been upsized to match that flow increase.  (I don't think the car originally came with anything larger than about 40lb injectors)

In reply to alfadriver :

The old engine runs a stock ECU with an air meter and 35lb injectors.  The new engine runs MS3 with a MAP sensor and 80lb injectors.  Sorry if I didn't make that clear.

alfadriver said:

3) Calibration robustness.  This is where a lot of aftermarket systems run into problems.  They are *capable* of running a big range of conditions.  But the calibrator doesn't do the work, whereas the OEM calibrator is required to complete a full process of that so that the customer does not notice any issues.  Some of this YOU can deal with- as long as you are willing to constantly tweak the calibration as you experience different conditions.  You need to constantly be logging your driving, you need to make sure that your car is equipped with enough extra sensors to detect what needs to be done, and you need to analize the data so that you can constantly update the calibration.  OEMs put hundreds of thousands of miles on prototype development, go to Alaska in the middle of winter, go to Arizona in the middle of summer, go to Colorado and Death Valley for altitude work, constantly drive cars to tweak things for 2-3 years.  Of course, you don't need to repeat that- but some of that work will be required. 

This sums up the biggest complication with using the MegaSquirt or other aftermarket EFI. There are a number of cold start tables and the like that are set at an attempt to make a reasonably average default. But sometimes they need fine tuning. Once they're dialed in, they generally stay that way. The catch is you may only have one or two shots a day at cold start tuning.

MadScientistMatt said:

alfadriver said:

3) Calibration robustness.  This is where a lot of aftermarket systems run into problems.  They are *capable* of running a big range of conditions.  But the calibrator doesn't do the work, whereas the OEM calibrator is required to complete a full process of that so that the customer does not notice any issues.  Some of this YOU can deal with- as long as you are willing to constantly tweak the calibration as you experience different conditions.  You need to constantly be logging your driving, you need to make sure that your car is equipped with enough extra sensors to detect what needs to be done, and you need to analize the data so that you can constantly update the calibration.  OEMs put hundreds of thousands of miles on prototype development, go to Alaska in the middle of winter, go to Arizona in the middle of summer, go to Colorado and Death Valley for altitude work, constantly drive cars to tweak things for 2-3 years.  Of course, you don't need to repeat that- but some of that work will be required. 

This sums up the biggest complication with using the MegaSquirt or other aftermarket EFI. There are a number of cold start tables and the like that are set at an attempt to make a reasonably average default. But sometimes they need fine tuning. Once they're dialed in, they generally stay that way. The catch is you may only have one or two shots a day at cold start tuning.

Three, actually.  Which makes the cold weather development trip the worst one, by far.  Sleep isn't as important as getting that 3rd start.  Ugh.

I have been running different Megasquirts in different cars since 2003 or something like that.

The one concern I would have about putting it in a rally car, is that it is a pretty hostile environment. If you have  a standard MS3 with standard DB37 connectors that is very much a risk. You need to extremely careful to solder or preferably crimp all connector pins in the right way, use correct cases and tension releifs and even so you may well have issues with wires coming loose. Reparing a DB37 connector on course will not be fun.

As much as I like the MS, I really think there is merit to staying with an ECU that has automotive grade connectors.

Regards

Gustaf

This is the first time in my memory where alfadriver actually conditionally recommended a standalone over a stock ECU.  interesting.

as another data point, the micrsoquirted motor in my rallyx e30 has been reliably running a very mediocre tune for 1.5 seasons now. it's certainly not the same abuse as a stage rally, but it was running well enough to compete pretty much immediately. It took me longer to dial in the idle since I couldn't legally drive it on the road, but you could do it in much less time since the xr4ti is/has to be road legal. a week of commuting should get you close enough to be acceptable.

edit: I know you already have the ms3, but maybe a cased microsquirt is the answer for durability/simplicity.  might be worth comparing the MS lineup.