Page 1 of 2 12 LastLast
Results 1 to 10 of 20

Thread: Custom Engine Swap

  1. #1

    Custom Engine Swap

    Its been talked about a lot before and when my 2007 SX4 threw a rod bearing I decided to go for it.

    I purchased a J23 from an ebay seller. Got a J18 crankshaft from an Australian market Suzuki Baleno (Esteem in US), also ebay.
    Had everything checked out by a local machine shop and engine bored slightly to accept 3.572in pistons.

    Pistons are from Racetec, flat top, 3.572in, 1.0in compression height, 0.866in pin. Originally designed for Ford modular v8 engines. Pistons coated with low friction material from Swain Tech coatings.
    Custom connecting rods to mate the J18 crank to the pistons from Pauter. Length is 5.981in. Designed to keep the deck clearance to 0.028in.

    I'm lazy at taking pictures, so I have none. But the engine has been assembled and piston valve clearance is good, in fact I should have gone with domed pistons to get a little more compression ratio. I'm in uncharted territory for me, doing all of the part spec myself and was worried about the interference design of the engine and the fact that the oem pistons are dished. At any rate, new compression ratio works out to 10.15:1. Engine displacement is 2146cc. Rod to stroke ratio is 1.83.

    My goal for this engine is high rpm, probably set it to 7500rpm, again just to play things conservative. At 7500rpm maximum piston speed is 6650fpm, at 8000rpm max speed is 7100fpm. For comparisons sake the original J20 design at 6250rpm has a maximum piston speed of 6070fpm (I'm rounding). So at 7500rpm I'm looking at a 10% increase in maximum piston speed. That should be OK right? (crosses fingers).

    Last weekend I took the J20A out of the SX4 and started swapping motor mounts and accessories. The J23 block exterior is identical to the J20A. I am using the J23 head and that is where the challenge lies. The top motor mount attaches to the exhaust side of the head with an aluminum bracket and to the intake side of the engine on both the block and head with a cast iron bracket. The mounting holes for these brackets line up on the block but do not line up on the head. Modifications to the aluminum bracket (sloting) allow bolting up on the exhaust side. The intake side is a little more tricky. One of the head bolt holes lines up if the cast bracket is bored out a bit. The other bolt hole is unusable, further, the cast bracket has to be massaged to allow the Aerio intake manifold to be bolted on. The cast bracket is in the way of the aluminum intake manifold bolt locations. The head and intake manifold for the J20A SX4 are different. You can't use the SX4 intake on the J23 head (and vise versa).

    So with all this talk of being risk averse I've decided that 4 out of 5 bolts on the exhaust side is good enough... I'll learn whether or not that is a mistake if/when I crack the cast bracket. If that does happen I'll probably get something welded up to replace the oem bracket. For now I'm going with the fingers crossed method.

    Where I'm at now is swapping sensors from the J20A to the J23 block/head which is where I discovered something that should have been obvious to me but was not. Everything is 100% bolt on and matching with the exception of the camshaft position sensor. The J23 head has a rotary sensor driven from the end of the exhaust cam. This sensor has 4 wires, +12v, ground, cam position, crank position. I was hopeful that I could simply ignore the crank position signal and make a wiring harness adapter but no go. The signal from the J23 cam sensor has a different waveform than the one from the J20A cam sensor. Technically it would be possible to cut a new sensor disk and install it into the J23 sensor housing. I have seen this done by the DSM guys for converting their cam sensors to support aftermarket ECUs. However I'm going to go the low tech route and swap the intake cam from the J20A into my J23 head. The intake cam on the J20A has teeth built into it for a hall effect sensor mounted to the valve cover. So I'll use the J20A cam and valve cover on the J23 head, keeping the J20A sensor identical to stock. This is important for me because for now I am using the oem ECU.

    I may or may not be putting this engine in over the holiday weekend. We'll see if I run into any other show stoppers. I guess I should mention the oil pan is different between the J20A in SX4 and J23 in Aerio. The oil pans in these cars are strangely load bearing and include mounting locations for the timing chain cover and the transmission. So use the one from the J20A.

    I haven't mounted the transmission up yet, but I don't see any showstoppers. I also didn't look too carefully...

    I mentioned that with the J23 head you have to use the aluminum intake from the Aerio. Thats sort of a problem because the intake manifold for the SX4 has the AFM sensor housing integrated into it. The Aerio intake also points towards the passenger side whereas the SX4 intake points to the driver's (US). I purchased an intake from Cobb for the WRX or something, I did this because I think the AFM sensor is identical (haven't verified for sure yet), but most importantly the Cobb intake diameter is 2.5in. The diameter is the critical bit because the sensor is calibrated to that pipe size. Cobb part number is 712100. Hopefully this will solve my airbox and AFM sensor mount in one go.

    On the topic of intake. The Aerio throttle body is drive by cable so will not work on the SX4. Further the mounting holes are not identical between the SX4 and Aerio throttle body. Since I was switching things up and things dont mount out of the box anyway I purchased a used throttle body from a Subaru BRZ / Scion FR-S / Toyota GT-86, these are 65mm throttle bodies with the same drive by wire system as the SX4. I have access to a Form Labs 3D printer at work and designed an adapter to be able to mount the BRZ throttle body to the Aerio intake. I hand ported the aluminum intake to match my adapter and throttle body to 65mm, so basically my whole intake will be 2.5 inches now with no restrictions. Happy about that. The Form Labs is capable of printing in a high temperature plastic that is good to something like 450F. I haven't done that yet, my prototype was done in the cheaper material and I will install that in the car to see how it holds up before I go using fancy plastics.

    Really the intake and head is the big challenge with this install. I haven't yet gotten all the sensors swapped so I'm sure there are more snafus associated with that intake that i haven't discovered yet. But the good news is that I haven't run into any major, "wow this was a bad idea" moments. Maybe its early still...

    So once the engine is in the car and running on the OEM ecu, cross your fingers, I'll start looking at stand alone engine management. It's the only way to increase the rpm limit. That and flashing the ECU but Dynotronics doesn't seem to be in the game anymore.

    What else have I done: I got the blue 440cc Subaru injectors second hand, the ACT 3000301 clutch recommended by AdR, and had the flywheel lightened locally (shaved 3 lbs). I'm also adding a oil to water cooler from a Cummins diesel, mostly because I like the idea of faster warm-ups, and I figure if it's good enough for a Cummins then its good enough for a Suzuki.

    When I do go stand alone ECU it will be a megasquirt. I like the DIY aspect of it though am a little afraid of finding someone locally that is willing to (and can) tune a megasquirt. The microsquirt module is capable of doing sequential injection and spark on a 4-cylinder engine and that is likely the route I will take. I'll be leaving the oem ECU in the car but driving the injectors and coils with the microsquirt. Getting this engine built and to the point where its "almost" in the car has taken me a year, so I wouldn't be waiting with bated breath for the results of the microsquirt conversion any time soon! I also won't be getting any dyno results until I start using an aftermarket ECU as well. At this point I'll just be happy for the car to be moving under its own power again.

    I blame Roadkill and MCM for this foolhardy endeavor. Always wanted to build an engine, and those guys just keep making things look easy!

  2. #2
    This is already a long post, but I wanted to add my plans for the car for the future. I use the SX4 (or used, really) for SCCA Rallycross. Its the reason why I want to get more RPMs and not necessarily more power (though more power would be nice). Somewhere on this forum I did some calculations that showed even with the stock J20A the ideal shift point is somewhere around 7550rpm (extrapolating the power curve). It turns out that for my wheel/tire setup 7500rpm on a typically rallycross course is equal to maximum speed in 1st gear. The courses are short and times are close, within a second between competitors, so eliminating 1-2, 2-1, shifts will help A LOT. Of course the J18/J20/J23 frankenengine is going to put me in the class with the big boys so I'll probably be destroyed and be back on the forums asking about turbochargers... They also changed the venue for my local rallycross since I've started this project and my understanding is that the top speeds have increased, probably meaning I'm back in 1-2, 2-1 shifting territory. We shall see!

    The SX4 is techincally AWD but it kind of sucks, a lot. It's really good in the snow and low traction environments but the rear clutch that sends power to the rear diff is really weak. Maybe its the 98k miles, maybe its just weak from the factory. But you aren't breaking the rear tires loose in an SX4 unless its really slick out. So what to do? Its been talked to death, there is no real way to convert an SX4 to true AWD short of swapping in the drivetrain from something else. Maybe I should have done that! People always say Subaru but the Evo has the engine mounted the same way as us so I think a Evo swap is best. Of course I'm not going to do that. What I AM toying with is the idea of blocking off the front driveshafts and converting the SX4 to RWD. The RWD guys have more fun at rallycross. The RWD guys have more fun at autocross. I'm thinking thats the way of the future for me and the SX4. Besides, who doesn't want a RWD, high rpm, hot hatch? I do!

    Thats a topic for much much later.

  3. #3
    Join Date
    Apr 2013
    Sherbrooke, Quebec, Canada
    Dude... this is awesome.... keep us updated... post videos and pics if you can. MCM and roadkill are my favorite internet shows..... especially MCM. Also evo brembos are compatible with the sx4.... both cars share similarities.
    Last edited by Samus; June 30th, 2017 at 02:56 PM.
    AWD 2013. CVT. Beast (RIP 2019)
    AWD 2013 6 speed Monster (stay tuned)

  4. #4
    Join Date
    Apr 2013
    Sherbrooke, Quebec, Canada
    AWD 2013. CVT. Beast (RIP 2019)
    AWD 2013 6 speed Monster (stay tuned)

  5. #5
    Nice rides! Guy at work just got a Evo X MR, pretty sweet.

    OK, So I took some pictures. The plan today was to get that intake cam swapped from the J20A head into the J23. Long story short, that is a no-go. The cam lobes are significantly larger in the J20A than in the J23. The cam journals are identical though, so actually if you wanted to get regrinds for a J23, starting with a J20A cam would work really well. However it doesn't help me with my problem of the cam position sensor from the J23 being incompatible with the J20A ECU from my SX4.

    I have two options I think. The first is to take both cams to the machine shop and have them press off the cam position doohickey from the J20A intake cam and press it on to the J23 intake cam. It looks like it is pressed on to me, but I don't have a press so didn't attempt to remove it (plus I didn't want to mess with the orientation of the part). The J23 cam would have to be turned on a lathe what looks to be about 0.010in for the part to press on. Since its a holiday weekend I'm not expecting the machine shop to be open until after the 4th. So have to wait until then to get some expert feedback.

    Option two is what I mentioned earlier, cutting a new disc to install in the J23 cam position sensor. That will take some thinking. I have to reference waveform from the J23 an the J20A sensors so I ought to be able to make it work. I'll pursue this separately for fun, and more urgently if the machine shop says no go on pressing on the cam sensor lobes from the J20A.
    Image shows difference between J20A (left) and J23 (right) valve covers. The J20A has a hall effect sensor mounted to the cover that reads teeth on the intake cam. The J23 has an exhaust cam driven sensor that mounts outside of the engine.

    Image shows the J20A head and cams. Notice teeth on the intake cam (right), this is what the hall effect sensor on the cover picks up.

    Here is the J20A intake cam installed in the J23 head... before I realize that it was actually not going to work due to cam lobe size differences. Also shown is the last journal cap from the J20A exhaust cam installed on the J23 cam in the J23 head. You have to swap this over to get the J20A valve cover onto the J23 head.

    This is the inside of the J23 cam/crank position sensor. The inner ring is what provides the cam signal. In the J23 it is a series of increasing width pulses. The ECU in the J20A is expecting a pattern of equal size pulses, one one two two. So I figure it is as simple as getting a disc laser cut out that has the right pattern in the right orientation, easy right? ha.

  6. #6
    Some other pictures.

    Engine out!

    IMG_20170701_075159.jpg IMG_20170701_075208.jpg
    Exhaust side top engine mount bracket on J23 head. Had to notch the lower bolt hole to make it work.

    IMG_20170701_075310.jpg IMG_20170701_075256.jpg
    Intake side top engine mount bracket on J23 block/head. The block bolts all line up perfectly. None of the head bolts line up, however one of them can be made to easily work by boring it out a little bit. The top of the bracket has to be ground down to allow the Aerio intake manifold to be installed.

  7. #7
    IMG_20170701_075350 (1).jpg
    Aerio intake manifold with adapter. Adapter converts mounting holes to be compatible with the BRZ throttle body.

    Comparison of the SX4 J20A throttle body (left) with the BRZ (right). Actually, they don't look that different in the pictures, but trust me, the BRZ is way bigger.

    Lightened flywheel. Originally 17lbs. Now 14lbs.

    Cobb intake. Oops. Got my MAF sensors mixed up so it isn't a drop in replacement. Thinking of making an adapter. We'll see.

  8. #8
    Join Date
    Jan 2013
    East Hills, Wa
    All this reminds me of Honda mini-me swap I've done some years ago. It was good fun and great learning experience.
    How hard can it be?

  9. #9
    Join Date
    Nov 2010
    North of Montréal, Canada
    you should consider getting an aftermarket ECU and get rid of the cam sensor. second best thing to do would be a laser cut disc that matches the SX4's cam trigger wheel.

    Why would you want to use the OEM ECU anyways? it has a rev limit of about 6500rpm...
    former Suzuki Gold Master Tech

  10. #10
    I spent the day comparing the cam and ignition timing signals for the J20A and the J23. Came up with a disc that I think will work. Got the order in so now it's just waiting for it to show up. I am probably going to wait on taking my cams to a machine shop now, the disc is way cheaper than machining time.

    On the plus side I designed in a 72-2 pattern for the secondary output which I'll be able to use when I do eventually go stand alone.

    The reason I'm sticking with the oem ecu for now is cost. One thing at a time. I'll get the engine in and broken in, then start down the megasquirt path.

    It's been a long road already to get to this point even though I just now started posting about it. Spun the rod bearing around Nov 2016, attempted to hand polish the crank which worked for about 5 months until the bearing completely went again. The SX4 used to be my daily and rallycross car. I got a cheap ride in the meantime off Craigslist. Going to completely miss the 2017 season I'm sure. Hope to back up and running for 2018. Definitely won't be touching stand alone until 2018 if not 19. Slow and steady something something?

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts