1988 Mazda Rx7 Frankenso - Halfway there!

[evan5599]

Hi all,


I've got a 1988 Mazda RX7 N/A Convertible. I saw multiple people have (supposedly) attempted this on this car before, but the previous projects all died out (https://rusefi.com/forum/viewtopic.php?t=1201). Figured I'd just go ahead and put my money where my mouth is and get started.

I've already deleted most of the emissions along the lines of this https://www.rotaryresurrection.com/2ndgen/tech_s4_na_emissions_removal.html so that should make things easier.

I have the board and the harness is mostly set up, all I have left is the ignition. To make things easier and avoid any sort of secondary spark mess, I've decided to run nosplit and have purchased a secondary trailing coil with the intention of doing something like this http://www.banzai-racing.com/FC_coil_install_how-to.htm.

Now that I've got the hardware mostly set up, I think it's time to start playing with software. I'll start digging into the simulator this weekend, but would really appreciate any ideas for setting up a simulated rotary. I'm no programmer (actually a sysadmin, i.e. know just enough to be dangerous) and this is my first foray into tuning, so all input, directions, ideas, and sanity checks (I already soldered up the harness once backwards, but caught myself halfway through) are greatly appreciated. Thanks!

Attached to this post is some proof that I've actually done something and my I/O diagram for the Rx7 harness. Sorry for the poor quality pictures, I can scan and upload if needed. If this actually ends up running I'll make a nice pretty harness diagram.

Note, I made a mistake on the first picture. On the Frankenso connector I have two pins (3A and 3V) written down as going to the same pin on the RX7 connector, 3C. 3C (ground) on the Frankenso connector goes to 2C on the RX7 connector. 3V (injector) on the Frankenso connector goes to 3C on the RX7 harness. Will sharpie in the correction soon.

WIN_20180914_01_12_57_Pro.jpg (1.7 MiB) Viewed 10122 times

WIN_20180914_01_12_21_Pro.jpg (2.05 MiB) Viewed 10122 times

WIN_20180914_01_11_49_Pro.jpg (2.47 MiB) Viewed 10122 times

[AndreyB]

You seem to be on track to be first rusEfi rotary engine! Welcome to the forum!

[kb1gtt]

Do you know what kind of signal goes on that "CPU" line of the coil? Does your coil have an integrated igniter? It was common in the late 80's that the igniter was in the CPU. There is a strong possibility you need an igniter between the Frankenso and the coil. Do you have an electrical schematic or similar for the coil?

About your harness adapter, I would suggest considering an adapter board. The flying leads are very weak, and are going to be problematic. Something like this is suggested.
https://rusefi.com/forum/viewtopic.php?f=4&t=507

We have several of these already designed. You might find a KICAD file already exists. if not then they aren't hard to make. I would just need to know the details of your connector. Hole diameters, spacing, etc.

[evan5599]

You seem to be on track to be first rusEfi rotary engine! Welcome to the forum!

Thanks, I'm sure we'll get it running eventually!

Do you know what kind of signal goes on that "CPU" line of the coil? Does your coil have an integrated igniter? It was common in the late 80's that the igniter was in the CPU. There is a strong possibility you need an igniter between the Frankenso and the coil. Do you have an electrical schematic or similar for the coil?

The igniter is part of the coil and not the ECU, wiring diagram attached. As for the signal, the car has a wasted spark system.
As for the CPU line,

The G and Ne pick-up coils are located inside of the crank angle sensor. Their output is sent directly to the PCM. IgT is the ignition timing signal (trigger) sent to the igniters from the PCM. The igniters will not toggle the primary without this signal. The leading coil (mounted in the LF corner of the engine compartment) is a true distributorless coil. Both leading plugs fire together, even though the rotors are 60 rotor degrees apart or 180 crank degrees apart. The two trailing coils are mounted in the LR corner of the engine compartment.

This is not a typical distributorless system. You'll notice that there is only one IgT signal sent to the trailing igniter from the PCM. Well, how does the trailing igniter know when to fire each coil with only one signal? That is the purpose of the Select signal. The Select signal tells the trailing igniter which coil to fire.
Courtesy http://rx7.pw.cx/guides_manuals/collected/2nd%20gen%20ignition/demystifying.html


Rotary Engine 101: Ignition
Since the working chambers of each rotor fire in the same geographic location in the engine, only one set of spark plugs are needed per rotor housing. Due to the complexities of combusting a long chamber, two spark plugs are used in each housing. The lower one is called the "leading" spark plug, while the top one is called the "trailing" spark plug. Additionally, the front rotor housing is denoted as "1" and the rear rotor housing is denoted as "2". So, in short hand, the Trailing (top) plug on the front rotor housing is referred to as "T1", the Leading (lower) plug in the front rotor housing is referred to as "L1", and so forth.

As the working chamber approaches Top Dead Center (TDC), the leading plug fires first, starting the ignition of the air-fuel mixture and contributing most to the generation of power. The trailing plug typically fires 10 to 15 degrees later and effectively completes the combustion of the remaining air-fuel mixture above the minor axis of the trochoid housing. Additionally, the leading plug fires a second time ( Note: the Leading (lower) plugs always fire simultaneously, while the Trailing (top) plugs fire independantly of each other.) late in the power stroke, which is called a "wastespark". The wastespark is done to simplify the ignition system by allowing both leading plugs to be fired "on the same channel" (one coil, one signal) . Basically, both leading plugs always fire at the same time, so there is an extra (wasted) spark during the power stroke. The trailing plugs cannot be fired in wastespark mode due to the location of the plug above the minor axis -- when one trailing plug is firing, the other trailing plug is already in the next working chamber, where it would preignite the incoming air-fuel mixture.
As you may have figured out, there are three discrete ignition "channels" -- one for the leading plugs and two separate ones for the trailing plugs.
Courtesy http://www.vemssupport.com/forum/index.php?topic=665.0

As the factory 15° offset between the leading and trailing spark plugs firing is only for emissions, both plugs in each chamber can be fired at once. I'm planning on wiring up both the leading and tailing plugs to the same coil for each chamber. I might have to draw out the timing first, I can't figure out if the ECU should have two signals going out to fire each pair of coils, or something else.

About your harness adapter, I would suggest considering an adapter board. The flying leads are very weak, and are going to be problematic.

Understood. I'll probably redo it the way you suggested after it's running, or doing something fun like mounting the RX7 harness adapter straight to the board and jump as needed.

[kb1gtt]

Hmmm, an igniter with a condenser. Hopefully that's not as bad as it sounds. Generally the condenser catches high voltage spikes, so I fear there might be some high voltage spikes on your PCM signal's. But perhaps not.

Do you know if the coil is looking for 0-5V signals or 0-12V signals? Also do you know how many amps the coil signals require? Does the engine run now and do you have a scope which could be used to take some measurements?

As for your funky coils with the selector lines, would you consider a more traditional coil? The entire selector signals will be there own complications, a more traditional coil would likely prevent several issues with getting spark working correctly.

As well do I read this correctly that you have 4 spark plugs to control?

[evan5599]

I'm sure we'll make something work. 0-5V signals, not sure about the amperage but will keep digging. The engine does run (not so much in the way of idling cold due to the emissions removal and low compression from a recent rebuild). I have access to a scope and can take measurements.

I believe the condenser is to prevent a voltage spike bouncing back out of the coil from trying the ECU, not the other way around.

I'm not opposed to running different coils if the stock ones won't work, but the stock ones are reputed to be pretty solid and new ones are expensive

I have four plugs to control, two for each combustion chamber. It can run on just the leading plugs (one for each chamber, which I may do initially to simplify things).

Each chamber's set of plugs can fire simultaneously. The trick is to not fire the trailing plug when the apex seal is between the two plugs, otherwise the plug will be firing during the intake stroke and preignite the incoming fuel.

As far as the stock coils, I'm definitely dumping the switching (trailing) coil. To the right of that one in the diagram is the leading coil, which is distributorless and fires both plugs at once. If we can get the timing right in software, each coil will fire both plugs in one chamber simultaneously.

It's a timing and geometry question now I think. Does the ECU need to be set up as a one, two, or four cylinder engine? If I just fire the leading plugs as initially planned, I think one cylinder should do it as (in the stock setup) both leading plugs always fire simultaneously. The trailing one switches the wastespark (which we'll be doing away with in this setup).

http://rx7.pw.cx/guides_manuals/collected/2nd%20gen%20ignition/demystifying.html
This link has an awesome diagram of the firing sequence and an even better description of the stock ignition setup, if it helps.

[kb1gtt]

I see 2 plugs for one combustion chamber. Do I understand this correctly? Both plugs fire at the same time, such that you fire the plugs once, at say 5 TDC, such that the leading sparks at 5 degrees, as well the trailing fires at 20 degrees because of it's located 15 degree's later. So you only need one spark signal per chamber, and that spark signal is fired once per rotation.

Also I see the 0-5V signal and it does not have spikes. You'll have to configure the hi/lo driver to 5V. It can drive 0.125 amps, which should be fine for our ignition driver.

[evan5599]

Good to hear the driver should work, thanks. You are correct on the spark order and needing one signal per chamber. The second spark is fired because the combustion chamber is so long.

Since we're dealing with the signals from the CAS, we can think in crank (as opposed to rotor) degrees. I think there are two firing events per rotation, not one.

In one rotor revolution, there are three firing events or power pulses.
The eccentric shaft makes three revolutions for every rotor revolution. This means the eccentric shaft rotates three times faster than the rotor. This allows the rotors to follow the shape of an epitrochoidal curve.
In a two-rotor engine, the eccentric shaft rotates through three revolutions (1080 degrees) for six firing events or power pulses. Therefore, and this is important, in one eccentric shaft (crank) revolution, an ignition event occurs every 180 crank degrees.
The front rotor leads the rear rotor by 60 rotor degrees or 180 eccentric (crank) shaft degrees.


So maybe I've been overthinking this. At every 180 degrees of shaft travel, both plugs fire in one chamber off the same trigger signal. 360 degrees == two firing events. So the ECU needs to be set up as a two cylinder engine. Does this sound right?

[kb1gtt]

To me that looks fairly straight forward. You fire rotor 1 at 0 and 360 degrees of output shaft rotation, and rotor 2 is 180 and again at 540 degrees of output shaft. I'm assuming your equivalent to crank / cam are on the output shaft. So that seems simple enough.

[evan5599]

Sweet, thank you. I have been loading up tunerstudio, and will give a quick overview of my settings to see if anyone sees any problems or anything I forgot to touch.

Started with tunebymaf.

Initial error:
1 Errors:
Error: Error in section [VeAnalyzeMaps], unable to parse ini row., Problem at:
[mainController.ini]:[Line:950]: veAnalyzeMap = fuelTableTbl, afrTableTbl, AFRValue, egoCorrection, { 1 }

Is this alright to ignore?

Base Engine Settings:
Cylinders: 2
Displacement: 1.3L (this is the actual displacement. It might have to be set up as 2.6L, https://grassrootsmotorsports.com/forum/off-topic-discussion/dispelling-the-rotary-displacement-myth/43487/page1/ top post has the reasoning). Not sure what part this actually plays or how much it matters, but definitely unsure.
RPM Hard Limit: 8000 RPM

Injection:
Mode: Sequential
Injector Flow: 460.00

Ignition:
Mode: Individual Coils


Everything else is left as is for now. I'd like to get the simulator set up for a 13B rotary (if anyone can point me in the right direction) before firing, but I'd really like to try and see if she starts on Saturday.

Thanks for all the help so far!

[AndreyB]

Thank you for reporting the issue if tunebymaf is selected - https://github.com/rusefi/rusefi/issues/608
works fine if you do not select it and you are far from tuning so you should be fine

Speaking about number of cylinders. Please remember that rusEfi does not natively support rotary engines. But there are good news - I believe your rotate could be represented as a 4 cylinder piston engine. There are some manuals online it's just that they are related to another popular not open source ECU so I am not 100% comfortable posting a direct link if you know what I mean.

[kb1gtt]

I think he's an equivalent of a 2 cyl setup as a 2 stroke. He has one spark signal and one injector signal per rotor, and I understand he has 2 rotors. The thing that's a bit odd, is that each spark signal actually sparks 2 different plugs at the same time, which are both in the combustion stroke. As well what is a bit odd is that his ECU will claim 360, when it really only did 180. So claimed RPM are likely to be off by a factor of 2.

I wonder how hard it might be to include options of 2 stroke, 4 stroke and wankle. AKA 2 stroke sparks once per rotation, while 4 stroke sparks once every 2 rotations, then wankle sparks twice per rotation.

Hmmm, will wankle always spark twice per crank rotation? I kind of think that's based on the gearing in the rotor. I could see how different gear ratios could allow for a different number of sparks per rotation.

[evan5599]

Hmm... Doing research. What's the advantage of setting it up as a 4cyl with secondary spark instead of a 2cyl when I'm firing both plugs from one output?

[AndreyB]

I am not aware of any, i know very little about rotary. If its the same picture on engine sniffer it makes zero difference.

[evan5599]

Understood. Think the two cylinder way is easiest for now, I'll need more time to think about the 4 cylinder thing.


I think a four stroke, two cylinder setup is the way to go, basically a straight twin. The CAS will send a fire signal every 180 degrees of CAS rotation, and the ECU will switch it between two "cylinders". Just to be cautious, I'll only start with the leading spark plugs. Shouldn't blow up the engine that way.

I am travelling for the rest of this week and possibly this weekend, but I'll see if I can get it running and report back.

[evan5599]

It lives! Sorta


Setting it up as a 2cyl wasted spark works. I'm running the stock leading coils only, which fire simultaneously off one output from the ECU. The engine starts a little rough, but dies when it gets past the cranking RPM threshold.

Current issues:
-In rusEFI console, my TPS shows a max of 112% when the pedal is to the floor. Tunerstudio has no issues with this. It's possible the TPS is actually out of adjustment, but I don't think so. Is there a way to calibrate this within rusEFI console?

-My dashboard tach no longer works. Looks like I have to jump a wire, no issue.

-My IAT reports -273 degrees. There are two IAT inputs on the stock ECU. The one I'm using is labeled "Intake air temperature sensor (dynamic chamber) 1-2V at 176 degrees F", and the other is "Air flow meter (intake air temperature sensor) 2-3V at 68 degrees F". Not sure what the difference is between the two, but I assume I should be using the other one.

-My old ECU appears to be hot all the time, not just hot on ignition. Any issues running ECU power off the ignition? Also, I had a few issues with (what I assume were from) brownouts when starting. Considering putting a few capacitors inline for ECU power.

-Are there any advantages to switching to independent coils?

[AndreyB]

-In rusEFI console, my TPS shows a max of 112% when the pedal is to the floor. Tunerstudio has no issues with this. It's possible the TPS is actually out of adjustment, but I don't think so. Is there a way to calibrate this within rusEFI console?

console and TunerStudio show the same data on the gauge. TunerStudio calibration is the way to go. tpsinfo is the command to show details

IAT is probably a thermistor and you should have three points of calibration. Since that's a mazda probably matches other mazda of these years? Can you measure resistance with ECU disconnected to confirm? volts really depend of ECU pull-up resistors, it's all about resistance of the sensor.