My eyes are darting shiftily back and forth between the 2002 and the latest family of BMW engines, particularly the 4-cyl version of the three in my 2015 Mini. More broadly, "new engine in old car" is a lot of what makes this world great.
Of course, in engine swaps as in so much of life, we live in interesting times. Direct injection; CANbus; cam timing in various forms... How far off is this summary of the options?:
Option 1: Try to convince the engine donor's ECU it's still in its old home, and not just a disembodied brain in a jar (or in an MGB). Lots of things it expects to be there that aren't any more, and any engine-related info that needs to get to old-car gauges either needs its own sensors, or some sort of seriously advanced adapter to get the CANbus info converted... Seriously involved.
Option 2: Megasquirt I think is still chewing on DI. Motec, Bosch, and perhaps other fairly serious-money ECUs have the capability to control modern engines, but what about know-how? Can they help you reproduce a factory-tune-like starting point? Outputs that get more directly to gauges of choice? At least you don't have to convince them they still have a factory-spec rear window defrost... Seriously involved.
Option 2 sounds more appealing, but it's not cheap, and apart from spending a bunch of dyno time watching what the stock ECU does with inputs and outputs in a working donor, I'm not sure where to start with operating parameters. But it seems like the more useful long-term solution to learn about given that the advancements in engines aren't going to come with port injection, etc...
What's the story with aftermarket ECUs and engine swaps using new engines? Links? Personal insights?
Did I miss additional options 3 through N? Leave out important details about the current state of things?
The aftermarket is way behind the OEs when it comes to engine management at the moment. Not just control capability (ie, DI control) but the level of calibration. With multiple variable cams, electronic throttles, multiple injection events per power stroke, etc, calibration is far, far more complex than it used to be.
Option 1 with a CANbus adapter is the way to go. There are a lot of people who don't know what they're doing in this space. We work with MRS Electronics and they can pull it off.
Ah, I was missing a bunch about option 1! Thanks, Keith!
Looks like I've got some study of the MRS site to do.
The MRS site may not tell you what you need, as it's fairly industrial and German. I can hook you up with an engineer to answer questions. Access to the CAN protocol for the engine and car helps a BUNCH, otherwise you're reverse-engineering.
Being able to program helps a lot. I forget the exact language - C? Something like that.
In reply to Keith Tanner:
Er, yeah, the Automotive section doesn't exactly have a subheading for "Engine Swap Adapters"...
That would be fantastic! I'm a software developer by trade, and am serious about pursuing this swap; I'm also acutely aware that I'm not going to become a major account for them anytime soon and want to be duly respectful of their time, (semi-series production of LS-powered Miatas seems to be covered), but someone who can tell me generally what the concerns are and how to pursue them would be phenomenal.
Thanks very much!
The mini engine is FWD only. That's going to be a serious hassle. I'd skip it.
Best 2002 swaps:
M20, good with a turbo. Been done by many people many times (including by yours truly). Individual Throttle Body kits have finally hit the market, making ~190whp on a stock unopened engine, so that's worth thinking about for the future.
S14 if you want to keep the heritage. It's somewhat overpriced to swap it for any other reason. Yes, the potential is high, but so is the price tag, $10k+ for a stock rebuild...
Honda F20/22C if you want S14 power without the S14 troubles. Tuned S14 power for probably 1/2 the cost of a tired S14 swap. The f20c swap kit can already be bought from CaTuned. K20/22 would be the next logical step up from here..
M42 has also been done, but with that amount of work, you might as well choose one of the others that nets more power.
Nothing wrong with a properly built M10 Turbo either. A proper build with modern parts will be way more driver friendly and probably way more powerful that an actual 2002 Turbo.
Either way, don't forget the swap the diff guts to make an LSD for the new beast.
In reply to mr2peak:
The 2002 is the next major project, but as much as anything, starting this thread is about readying myself for a future of engine swaps where everything coming down the pike is going to require this sort of work. I expect it to be some tough sledding especially on the first go, but I hope to learn how this stuff works so I'm not constrained to an ever-aging selection of port-injected engines for the rest of my life 
Also, the current Mini engine is a current BMW engine, so it's not FWD-only. See BMW B38, B48. I should've probably referred to it as a BMW engine in my initial post, but I have a Mini and that's how I was introduced to the series; oops.
I agree wholeheartedly about the LSD; even with a stock M10, it transformed this nose-heavy little car's ability to exit corners at autocrosses. But I digress... hoping to keep this about modern engine swaps, not my 2002 in particular.
Option 3 would be to remove some of the complexity - for example, putting a port injected manifold and cable operated throttle body on an engine that was originally GDI and drive by wire. That would make running a standalone ECU considerably easier.
MadScientistMatt wrote:
Option 3 would be to remove some of the complexity - for example, putting a port injected manifold and cable operated throttle body on an engine that was originally GDI and drive by wire. That would make running a standalone ECU considerably easier.
I'm not sure that would work. We need AlfaDriver to chime in on this one. I think that the combustion chamber is very different for Direct Injection with the super high fuel pressure for mega atomization and I'm not sure it would even work with 'normal' fuel injection.
Paging Eric to the Bat phone.
The shape of the pistons is definitely different.
Ransom, email me at keith@flyinmiata.com and I'll hook you up.
IMHO, what's the goal in mind?
What does a modern GTDI project get you? Is it worth 1) the work and complexity of hacking a system so that it runs OR 2) the cost of some kind of crate motor package that you can get?
Converting that to PFI/cable throttle- how much boost can you run on the elevated compression ratio designed specifically for the DI? And what do you replace the DI injectors with (there are holes there).
Next question- is there a slightly older engine that's simpler and easier to deal with? For example- making a 3.5l PFI motor from a Mustang work is a WHOLE lot easier than making a GTDI motor from a SHO work. If it's the power you want- are there packages that you can work with? Say the mustang motor- has anyone figured out a turbo package for them (I think so).
Of those two questions- given the point that tuning a super advanced engine is really difficult- when you try to adapt, are you going to give up anything? Especially compared to the less tech option with toys attached?
I'm going to stick with the Ford example, but say I'm putting a modern engine is a Cortina- I've seen a V8 put in one, but I4's are pretty straight forward. Now, I'm going to ignore the fact that I can get a GTDI crate motor with controller- since that's not all that common... But for the basic swap of a 2.0l or 2.3l turbo- you get 260-300hp depending on where you get it- it will run great, and be done. But no changes from that unless bolt on parts that might help. So how much cost/time will it take to adapt the system to work in the car w/o all of those extra car modules?
The alternative is a 2.0-2.3 (even 2.5l) duratc motor. I know there are cams available to that, I'm pretty sure there are turbo packages for them- and one can tune the crap out of that- as long as you have a VCT capable controller.
A vs B.
Personally, I'd choose B.
(and for a 2002, seems like there are some good 4 cyl motors that have huge aftermarket that would go into that hole and look right but are not DI.)
I'd like to underscore this point once more: This is not about choosing the most cost-effective package for the 2002. I think I need to learn how to deal with modern engines and modern engine management, not because it's the most cost effective or simplest package right now, but because that's where engines are going. There are very few port-injected engines still in production. DI isn't the future, it's the present (which makes it the future of junkyard swaps).
The M42 I have tucked away was probably badly abused, but they're already a bit thin on the ground. If you want to be able to pull a high-output package from the wrecking yard and plug it into an older car five or ten years from now, I don't think scavenging ebay for a Duratec is going to be the way to go.
The aftermarket has a lot to learn about modern engines, as Keith pointed out, but we're seeing a few things; uprated DI pumps and so forth. The ability to tune these engines will improve, and we'll go from having tidy packages that already have great power and driveability in factory trim, to real fire-breathers (not that the latter don't exist in factory trim...)
Duty now for the future 
In reply to Ransom:
Which is a different goal than anything I thought.
But it's going to take a long while, I think, before there's a good DI aftermarket tuner. Especially if you are starting from scratch.
Then again, the in-cylinder pressure transducer kits are now more on a scale for the aftermarket.
Still- the tuning of DI stuff is really tricky, between injection timing and pump timing (especially on a VCT camp) it's mostly done on a dyno. Modern ETC is it's own complexity all by itself.
In reply to alfadriver:
All true! Which is what seems to push me (along with Keith's advice) toward Option 1 from my initial summary; figuring out how to get factory ECUs to play nicely after being evicted from their originial homes.
I have a bad habit of phrasing my initial questions inside a wall of text. I could have been much clearer about my reason for offering those two options.
I believe the aftermarket and us GRMers will learn how to work with this stuff. It just takes time. Less and less is left on the table with regard to factory tunes, which makes me feel better still about just working on learning how to transplant without the ability to change a lot on the engine or its tune.
Anyhow, thank you! I know you know more about this stuff than most of us here, so any insights to what's involved are useful, even if the result is only scaring me further away from attempting clean-sheet tuning.
I think the secret will be keeping the stock ECUs happy. This, unfortunately, probably won't be a one-size-fits-all solution. But with more common swaps, the hardware exists. Someone just has to invest the time in programming. Ford and GM are doing that for us already, but swapping something like a Miata Skyactiv 2.0 into something else while retaining full function will require a lot of work on the part of the first guy.
Putting an old engine in a new car has the same problems. Tachometers don't just read pulses off the coil anymore, they interpret a CAN message. Heck, the fuel gauge on the ND Miata is insanely complex, not just a simple read of the resistance on the float. There's a good reason, it's just shifted the required skills for swaps to a different set.
In reply to Keith Tanner:
One of the problems with "keeping the modules happy" is that you need to know what information comes from where. Some ECU stuff comes from other modules over the CAN network. It's been that way for some time, as the ABS module started doing vehicle speed a LONG time ago.
Oh, I'm very aware. That's why you won't be able to find a one-size-fits-all. You'll have to eavesdrop on the CAN traffic, reverse engineer it and modify it accordingly. That's what I've been working on with the ND V8. For example, we had the EPAS rack sitting on the bench wired up to a laptop, logging the signals so we can replicate them using a different sensor and keep the rest of the car happy. Even the stop-start system on that car talks to a bunch of modules. The reverse lights go through three of them, including the ECU.
The hardware side is solved. The MRS boxes can do what needs to be done - block some signals, pass some through, modify some and create some. It's the software side that takes the work.
So yeah, it's easy to say "keep the modules happy". But in reality that's a difficult thing to do. Those that do it will charge for their R&D.
In reply to Keith Tanner:
I forgot that you guys did a recent swap. Still- it does bear being pointed out to others.
Cracked modern ECUs are pretty interesting, not to mention more powerful than I thought.
I'm in the process of having the MS3 (2.3L Turbo with DI) tuned to make the most of the modifications by Stratified Automotive Controls out of British Columbia.
With the Accessport the tuner has been able to account for all my modifications and find something like 60 hp outside of the safety tune. That's with the emissions equipment intact and nothing disabled. It's pretty amazing all the parameters which can be adjusted within what is basically a fancy spreadsheet.
I understand that it's possible to suppress specific codes and tune around missing controls like the EGR and Catalytic converter. I think the Accessport may qualify as a "high tech defeat device" though. I actually would be pretty interested to see what the result of transplanting it would be.
The electronics were the most interesting part of that ND V8 to me. Lots of clever packaging solutions and fun reuse of OE parts from various sources - but the first time we got the start button to light up was a big moment. We're also coopting parts of the display for our own purposes, such as changing the exterior temp display as an oil temp gauge if you double-tap a cruise control button. You can do very cool stuff if you view the entire car as an integrated platform, which is what the whole Skyactiv concept really is.
The0retical wrote:
I understand that it's possible to suppress specific codes and tune around missing controls like the EGR and Catalytic converter. I think the Accessport may qualify as a "high tech defeat device" though. I actually would be pretty interested to see what the result of transplanting it would be.
That's why it's labeled an off road device only. Sometimes, the people who sell the product need to be reminded of that.
(It may sound odd, but there are some GRM'ers at the EPA, couple of still born challenge cars, included. So they go out and look for this stuff, too. For their fun)
I got pointed in this direction by Keith and I'll try to give a broad overview of CAN...things.
The CAN bus on a vehicle is used for a lot of vehicle functions. Communication between different controllers, the way to reprogram controllers, and the way to get data from these controllers. My work with the ND car from FM focuses on the communication between different controllers. Rewriting the actual code stored on a ECU or BCM is something for the Cobbs, Mountunes, etc. I listen to what the modules say in every day operation. (And they talk a lot!)
I'll give a rather well known example of how the communication is pretty important: In the Subaru BRZ/FR-S/FT-whatever chassis vehicles, if the power steering control unit does not see a CAN signal from the engine computer, it does not turn on. This most likely is to prevent the power steering unit from over heating. If you are in the car with the key in the On position but the engine isn't running, you most likely wouldn't need power steering (and wouldn't with a hydraulic setup!). This becomes a problem for engine swapped individuals as they will never have power steering without the proper CAN signal they can never have power steering! Along with this, the power steering controller ALSO never sends a message saying its running which the instrument cluster looks for to turn the warning light off. Yuck!
Now, if you know what that message is and can broadcast it, the power steering unit comes alive, and the warning light goes out! Yay! This can be done with any device that has a CAN transceiver. Most low cost solutions involve some sort of Arduino based solution. These will work for a great bit of people's application, but aren't a production quality (or ready) solution.
I'll watch this thread if anyone has any generic questions I can help with. I like to think I know a thing or two about this CAN stuff. :)
Thanks for stopping by, Colin! For those who haven't figured it out, Colin is the savant who is responsible for getting our large engined bastard ND integrated.
The0retical wrote:
With the Accessport the tuner has been able to account for all my modifications and find something like 60 hp outside of the safety tune. That's with the emissions equipment intact and nothing disabled.
The equipment may still be there, but I'll bet it would fail a measured emissions test. Comparing the torque curves of stock 5.9, later stock 6.7 and chipped 6.7 Cummins engines, there's obviously an area of concern that gets ignored in the chipped trucks. There's a big hole in the (apparent) torque curve on the stock 6.7 compared to the other two, and it'll be there for a reason.
alfadriver wrote:
(snip)
That's why it's labeled an off road device only. Sometimes, the people who sell the product need to be reminded of that.
(It may sound odd, but there are some GRM'ers at the EPA, couple of still born challenge cars, included. So they go out and look for this stuff, too. For their fun)
The big players in the industry are very aware of the new attitude about emissions. We're taking it seriously - expect to see a whole bunch of new CARB-approved parts from us, and emissions compliance is a major factor now when we evaluate a new product. But this is another separate conversation. An interesting one, but not really applicable here.
In reply to Keith Tanner:
Thanks for answering those questions Alfadriver and Colin. I was curious how you eavesdropped on the CANBUS. It interests me that the communication isn't encrypted, although I suppose that would require decryption at the component end, since manufacturers seem to be interested in locking down the cars further and further. I thoroughly enjoyed reading about the integration into the CANBUS you guys at FM did when Keith first mentioned it.
Since it was the topic of the tread. How far off into the future would you guys say a consumer type solution is? Like Ransom I'd love to take advantage of a modern engine in an older car sometime in the future.
Was the process more of a rote learning process where you were picking out packets and routing info then reproduced them to see what they did? Or was there some documentation that gave guidance?
As for my car. I'd no joke send my car off to be compared to a stock one just to see the what the diffence is if someone wanted to actually publish the results.
I'll let Colin tell you what he wants, but here's the close observer view.
It's a bit like trying to break Enigma It's not encrypted, but it's non-documented and some of the values are twisted about in weird ways. First you identify what a particular signal is - say, the outside temp - then you have to figure out how it's being passed along. 255 minus the temperature in Kelvin converted to hex, for example.
So we'd start by logging the car and doing things to it. Spotting stuff like the cruise control buttons was fairly straightforward as they're a very discrete and binary signal that are easy to pick out. Things like instantaneous fuel economy were a whole lot harder to figure out. It helped that Colin went home from FM and immediately bought an ND 
Mazda USA couldn't give us any info, they say that Mazda Japan won't even tell them.
I believe I was checking out some of the signals from the ABS system here. Tapping into the module, powering it up and seeing what it broadcasts.
There already are consumer solutions for the LS engines as well as four of the Ecoboosts thanks to GM and Ford. For mainstream engines that are irresistible for swaps, it's basically just up to someone doing lots and lots of work and then selling a module that somehow hopefully keeps a stock engine computer happy - such as the PS spoof for a FRiSBrZ that Colin mentioned. It will be very specific to a given engine and not generic, though.
Theoretical, the tuning done to your ECU is a bit of a different thing. A log of the CAN system would probably be indistinguishable from a stock car, from what I understand, as the CAN chatter is communication between different modules and not the ECU's internal processes. The tuners, if they've cracked the ECU enough, will be able to turn off various inputs the car might need to run. That's an internal solution. An external solution would be something like another box in the system that's providing the needed signals.
