Jaynen wrote: Titus your car looks great. Good luck on the swap. This is sad but every time I drive our Odyssey right now I think of how great an engine it is.
Thanks!
Jaynen wrote: Titus your car looks great. Good luck on the swap. This is sad but every time I drive our Odyssey right now I think of how great an engine it is.
Thanks!
It is funny how the people who make the arguments against N/A BP builds only serve to solidify my desire for one. Going over my old rallycross videos, which I have to do because rallycross season sleeps during the winter, I come to the realization that I spend way too much time not on the throttle, which means I am thinking about the car and not the course. So clearly N/A 13B is too much torque and I need a smaller engine. BP or B6 would be just about right but I'd want to work the engine up so it still makes 230-240hp. So now I'm off to see what it takes to make a B6 live a reasonably long time (about 15,000mi) with an 11k redline. Answer: Nobody knows because people put big dumb turbos on instead.
I don't think that's true Knurled I know quite a few shops including the 949/supermiata guys who have done extensive NA builds. And flyin miata would not have a full 2.0 stroker kits etc if they had not gone into building NA miatas either.
If you want to build an 11k redline NA miata please do it, I'd love to see it and know the effort and cost involved. People should do what they want. The general consensus however is since miata's are "cheap" and the people who own them tend to be budget conscious most contention surrounds what is the cheapest way to get the desired 230-250hp power range
I'm struggling to find any on line info on the computer control options. AEM, hondata and MS don't list specific applications for the j series. am I just sucking at searching?
In reply to Jaynen:
2.0 would be too big - a good 2-liter makes about the same torque as a 13B (180-190ft-lb) and I want to cut that down a bit. The thing is, forging new ground is expensive. Other engines (4AG, BDA, B/D-series Honda) are known factors. It would suck to go through three or four bottom ends before narrowing down a failure to some obscure thing that never fails in low power or turbo spec engines.
icaneat50eggs wrote: I'm struggling to find any on line info on the computer control options. AEM, hondata and MS don't list specific applications for the j series. am I just sucking at searching?
AEM Series 2 30-6051. It is plug and play for AT harness CL/TL Type S (J32A2) and Odyssey (J35A4).
The heads just don't flow at super-high rpm. And yes, Miata owners are budget conscious for sure.
In reply to Knurled:
Crank won't take that. The data is out there, even in pretty recent threads. Cams available won't get you that far either ASSuming you want to make power up there.
Basically... Break out the wallet in a huge way.
Give it a shot, though. I'd love to read a thread dedicated to an n/a BP.
Knurled wrote: In reply to Jaynen: 2.0 would be too big - a good 2-liter makes about the same torque as a 13B (180-190ft-lb) and I want to cut that down a bit. The thing is, forging new ground is expensive. Other engines (4AG, BDA, B/D-series Honda) are known factors. It would suck to go through three or four bottom ends before narrowing down a failure to some obscure thing that never fails in low power or turbo spec engines.
Well at least 1.6 engines have to be pretty cheap? I find it interesting the miata is a good platform for rally cross. But I guess once you go away from AWD it makes sense as its small light and balanced
Kind of funny you actually DONT want torque lol
I am also doing the J swap in my 95M, along side the '90 w/ N/A '02BP swap…..so ill get to see which one does it for me. 
as said earlier, rogue, please please please do a build thread
icaneat50eggs wrote: as said earlier, rogue, please please please do a build thread
Yes, needs more build threads. Link to it in here.
Question for those doing the build or more familiar with ecus. The website says the engines can be run with stock computers, but performance will be limited. Do they just mean that performance gains over stock are limited?
I emailed them about this and they said they simulate a gear input and speed to the stock ecu. Is this a viable way of using the stock computer?
If you can get it running well without the cost of an aftermarket ecu the. I could really see this as a good option
icaneat50eggs wrote: Question for those doing the build or more familiar with ecus. The website says the engines can be run with stock computers, but performance will be limited. Do they just mean that performance gains over stock are limited? I emailed them about this and they said they simulate a gear input and speed to the stock ecu. Is this a viable way of using the stock computer? If you can get it running well without the cost of an aftermarket ecu the. I could really see this as a good option
My intention was to go with the stock ECU initially but I think I am slowly moving away from that position and may end up getting the AEM. At 1200 it sounds like a lot, but I am getting the feeling that getting the stock one to work will get you near that anyway without the benefit of the added power that comes with tuning a full ECU.
ECU Options:
Manual stock ECU- If you go with the manual, you need a manual engine and harness (as the sensors are different). They are pretty rare, so they cost more than the AT ones. You need to deal with the immobilizer, which means getting and wiring and immobilizer sending unit and key from the same car the ECU came with, or sending the ECU off and paying $100+ to get it modded. If you want it to operate code free, you will likely have to do other trickery and/or add additional sensors like the multiple O2 sensors. Regardless, there is going to be some wiring changes to make it work, and that will cost more money if you can't do it yourself. You also don't have the easy upgrade path to the AEM, as the AEM is designed to work with the AT harness and sensors. If you wanted to upgrade, you would be buying another harness and sensors, along with the redoing wiring changes you made with the manual harness.
AT stock ECU- The AT ECU/harness/engine is easy to find and really cheap. It is also easily plug and play to upgrade to the AEM later. You still need to deal with the immobilizer, which means getting and wiring and immobilizer sending unit and key from the same car the ECU came with, or sending the ECU off and paying $100+ to get it modded. The bigger problem is the fact that it expects to see various sensor signals from the AT which isn't there. From what I understand, initially this will keep you in limp mode with a 2500 rpm rev limit. I hear this is medium difficult to remedy through wiring changes, unused transmission sensors plugged in so the ecu sees something, or black box stuff to send fake signals. This gets you past 2500, but then you hit the wall where the vtec would kick in (5000ish I think). Still researching but I think this is because of the lack of a speed sensor from the transmission. I guess it wants to see speed before going in to vtec, and that sensor isn't going to work with the Miata transmission. Your average back yard mechanic is not going to get past this on your own, so this takes you to the speed simulation you mentioned. There are a few shops building black boxes for it, but again it is more $$$. Like the manual ECU, If you want it to operate code free, you will likely have to do other trickery and/or add additional sensors like the multiple O2 sensors. Again, there is going to be some wiring changes to make it work, and that will cost more money if you can't do it yourself.
AEM- No immobilizer issue. No black boxes. No codes. Uses the cheap AT motor harness. Less sensors to buy and hook-up. Less complicated wiring work. More power potential. $1200, plus $150 if you want to add a wideband, plus tuning cost if you are not able to do it on your own.
That's kind of what I thought. Not really a viable option
Soooo this may be somewhat of a hammer looking for a nail, but does anyone have an idea of what sort of signal the "black box" is emulating? Basically taking one signal from the Miata box, and translating it to what the Honda ECU is looking for, right?
It can't be THAT hard, right? I'm learning how damn cheap electronics components can be if you're willing to do some soldering. I think it could be a fun problem to solve.
From what I have seen, those that have come up with black box solutions are hoping to get your money and are therefore not willing to share much information. This does not really surprise me, as I am sure it takes a lot of time and knowledge researching and with a stock car testing each sensor to figure out what the ECU normally receives in different circumstances, what changes or circumstances cause codes or limp modes, then figuring out how to emulate things. That is all before they build a thing, which then needs to be tested on different cars/engines to make sure it works as expected.
When I first heard about this swap, I thought who would want to do that...now I realize that person is me... 
Yes, very interested in a build thread.
Titus wrote: From what I have seen, those that have come up with black box solutions are hoping to get your money and are therefore not willing to share much information. This does not really surprise me, as I am sure it takes a lot of time and knowledge researching and with a stock car testing each sensor to figure out what the ECU normally receives in different circumstances, what changes or circumstances cause codes or limp modes, then figuring out how to emulate things. That is all before they build a thing, which then needs to be tested on different cars/engines to make sure it works as expected.
I bet we could come up with a cheap small package for this swap with some open source stuff. We just need someone with an oscilloscope and access to both cars to do some waveform sampling for us. 
Your assuming you know what the logic is that is reading the signals.
Not really, I don't care what's doing the reading, so long as we can emulate the sensor it's expecting.
Say the Mazda sensor is outputting a 5V square wave at 20 counts per revolution, and the Honda sensor outputs a 12V square wave at 30 counts per revolution. It's entirely possible to have a small, cheap microcontroller counting the 5V pulses, converting it to an RPM, then pulsing another output at the frequency and voltage that corresponds to what the imaginary Honda sensor WOULD be doing in that situation.
It's pretty trivial to do that with an Arduino.
I agree that mimicking one signal is doable, but how do you know what signal to mimic, or that is just one input? They could easily have logic with multiple if thens.
Vtec = on if speed > 10 mph and front wheel speed = back wheel speed
Only one way to find out.
is there a reason a megasquirt isn't in the list?
You'll need to log in to post.
