Doing MS3x on a stock 91
 

So, I bought a MS3x with a DIYBOB from a member here. Should have it in my hands next week. Trying to be ready to plug it in and start playing ASAP.

The unit in question was set up for and installed in a '94. So, I'm looking to do whatever mods I need to do to convert it from '94 spec to 90-93 spec.

Studying the ECU pinouts between the two cars, since the 94 wasn't really all that different electrically, it appears that the only thing I'll have to change is the cooling fan output from pin 1L to pin 1R.

My first goal is just to get the car running on the MS with the stock AFM in place (I know I'll only really be using it for the AIT and to turn on the fuel pump). I want to see how much difference can be made strictly with tuning. Then I'll remove that AFM and play with sequential injection... and whatever else I feel like doing.

QUESTIONS:

1. Is the above assumption about 94 to 90-93 wiring differences correct?
2. There's some "fuel pump mod" that I can obviously avoid doing by just jumpering the fuel pump... is there any reason to consider DOING this mod? What's the benefit? (bearing in mind that I hypermile in this car, and frequently engine-off coast with the ignition on)
3. I know on the MS2 there is a recommended ignition output mod to keep from frying coils... and I know that the MS3x uses it's own on-board outputs. What's the status of that default high ignition output on the MS3? Is that mod required, or is it a non-issue?
4. The diode has been added to the IAC circuit on the DIYBOB. Any other similar Miata-specific mods recommended?
5. Please point me toward the best MS3x write-ups out there. I've found some of them, but which one is best? Idunno.
6. Any other tips that you think might help.

This will be the third MS that I've tuned, so I'm not a total newb. Just second-guessing myself a bit.

Working on answering some of my own questions, might as well document here.

#2 - Fuel Pump
When I remove the AFM, I'll wire this output to Pin 2o on the stock harness. This will send the FP switching ground from the MS through the unused AFM signal wire. Then, I'll jumper pins A & F on the AFM connector, which will send the ground to the Fuel Pump Relay. Simple. No added wires. Easy to revert to stock. I like it.

#3 - Ignition Output default state
In the thread about the output inversion mod, I found it was stated (by both Matt & Joe) that the mod is not required for the Microsquirt (and Micro based PNP) because those units use a low-current logic-level output, which should be low, anyway. Since the MS3x uses the same type of logic-level output, I'm going to assume that the same applies. No danger of burning out igniters or whatever due to a default high-current output being applied at power-up.

might help. used it to build my own ms3x (although I haven't got my car started with it yet...)

90-97 MS3 with MS3X | My Westfield MX5 90-97 MS3 with MS3X | & Megasquirt site

yes, use the writeup Frank (and I*) have out there.



*I say I because I contributed to it.

Okay, cool. Just read over that one. Pretty straight forward. Funny that I came up with exactly the same method for the FP switching. Must be the right thing to do!

Wish I could find some existing wires to use for sequential injection, but there don't appear to be any good options. Oh, well. First priority is just to get the car up and running on the MS...

there's the purge valve, or there's probably two more in the AFM harness you could use.

I wanted to leave the purge valve because I read that some people were using it as an additional "idle-up" valve to help with AC idle-droop. (and I have a Fidanza, so that could be an issue)

I did consider the purge valve and the PS pump wire (I have manual steering) as an option. Didn't consider other AFM wires... hmmm...

lol. you have a ms3x, that has idle up code bro.

a/c activation is better than stock.


dont read any posts that pertain to a MSI.

The MS3 actually implements A/C control directly. You wire the A/C switch into an input pin on the ECU, and configure an output pin to drive the A/C relay. When you turn on the A/C, the ECU idles the engine up, waits, and then engages the relay.

It also supports idle-up for the fan output as well.

Very slick.

(Edit: curse you, Braineack, for beating me to the point. Now I look like a Johnny-come-lately. You will rue this day.)

For the sequential conversion, I just ran six new wires (four INJ and two GND) through the little hole in the firewall which is behind the fusebox. Bored it oversize a bit with a unibit, used a wood bit to mangle a hole through the rubber pad behind it, and then used a coat-hanger to pull the new wiring bundle through. Put some split-loom tubing around it, and ran it in parallel with the stock harness that goes up to the back of the head.

Much easier than un-wrapping a bunch of 22 year old harness to sneak one or two wires out of it.

Yeah, I'd read about the MS3's great idle control and AC idle-up, and I'll be using it. Just wasn't sure I wouldn't need to do any other tricks, so I was going to leave the purge valve on the table. But, if you guys think I absolutely don't need to worry about that... I'll trust you. :) And I look forward to being impressed!

I hear ya, Joe. Was thinking of maybe doing something like that... maybe run a dozen "spare wires" through the firewall to cover sequential fuel and anything else that might come up later.

My box is supposed to ship today. I'm gonna be jonesin' all week.

You don't need to worry about it. The MS3 does everything except make you a sandwich, and I understand that they have that partially functional in the current Alpha code.

Mmmm... sandwiches.

Hey, what about pins 1d, 1f and 1k that go directly to the diagnostic connector that I'm not using? I'll have to check their size, but they could work for injector circuits.

1D is a ground. 1F and 1K could work. 2P 2O and 2K should all work off the AFM.

I'm using 2P for the AIT. 2O for FP switching.

Looks like there are a few wires hanging out there that could be used, though. I like repurposing stuff.

Update.

For the seq. fuel, I'll be using pins 1F and 1K to the diagnostic connector. I'm trying to retain easy backwards-compatibility with the stock ECU and systems for now, so I'll just tap into those wires and run a 2-wire harness for those two injectors that can be easily unplugged and switched back to stock.

Pretty simple stuff so far.

I also decided to go ahead and order a WB02. Bought one of those $105 units from 14point7.com, we'll see how well it works. I was going to try to tune with the stock NB, but thought better of it.

FedEx says I won't see my MS3 until Friday.

I'm looking forward to your review of this unit. I've read a lot about it, but never actually heard a first-hand account. Hopefully you will retain the stock NBO2 sensor as a calibration reference.

Not planning to install a second bung in the exhaust (though... I guess I could).

But, I can easily do some reference tuning and logging with the stock sensor... then switch in the new sensor and log the same points. (all with EGO off, of course) Would be a good reference test in lieu of having both sensors in the system.

Hey, is anyone familiar with the stock injector harness? Oh, I bet someone is!

I figured I'd go ahead and pre-configure my injector harness so that I can switch a plug when I'm ready to go sequential.

I know the injector signal wire is shared between pairs of injectors coming out of my ECU... anybody know where they actually split into two wires? Is it right where the harness splits to go to the individual injectors, or is it somewhere before that?

it's easy to determine when you rip apart the harness. I cant quite remember off the top of my head.

here it's just around the first injector:

https://www.miataturbo.net/attachmen...ine=1353361820

Yeah, but I was wondering if I could rip the harness apart somewhere other than at the fuel rail. If the split is right there, then I have no choice. But, in the unlikely event that the wires actually split somewhere near the ECU... well, that presents different options.

I could always rip things apart and hack away at stuff and ignore the experience of those who have gone before me. :) I try to avoid that when I can. Saves time, effort, headaches and money.

thus why: you must run two wires and redo the injector harness, is not an option.


you gotta pay to play, brah.

Ooh, nice ninja-edit!

Okay, that answers that question. Gotta do my connector-splice right there at the fuel rail. I can deal with it.

Thank you, sir.

It is inside the injector harness itself. The two injectors in each pair are spliced to a single wire, and that's how it comes out to the six-pin connector which is mounted to the back of the head. Specifically, that connector contains two INJ lines, +12 to the injectors, the dashboard CLT sensor wire, and the two wires (signal and ground) to the ECU CLT sensor.

Thus, you must run new wiring from the ECU up to the head, and also modify the injector harness itself to bring out the additional injector wires on a new connector.

This is what mine looks like:

https://www.miataturbo.net/attachmen...1&d=1353362406

https://www.miataturbo.net/attachmen...1&d=1353362406

I cut the two injector wires off of the original 6 pin connector, and added an additional six pin connector to carry four discrete injector wires plus two ground wires from the ECU (the ones terminating at the ring connector). This mates with a new harness which I ran through the existing grommet hole in the firewall behind the fuse box (after drilling it oversize).



edit: Beat me to it while I was tracking down the images...

You guys rock!

I have a better image of what's going on beneath the tape at home- I'll try to remember to dig it up tonight.

Here we go:

https://www.miataturbo.net/attachmen...1&d=1353377333

In this photo, I have already severed the two yellow wires at the end of the wrapped portion coming from the stock 6 pin connector, as well as separating the paired yellow wires where they join in the middle between two two sets of injector pigtails. I've then soldered an individual white wire to each of the four (now discrete) yellow injector wires, and brought them out in the same direction as the stock connector. It is these four wires which will be terminated at the new connector, along with the two additional ground lines (these are optional- I just have grounding fetish.)

I left the white/red wire (the 12v supply) untouched.

Outstanding! Thanks, Joe.

pretty much exactly what I did too. that extra connector you see in my pics was to be able to revert back to batch, if connected, so I'd still be able to test out any ECUs I built.

This is child's play.

Yeah, it's just a matter of getting around to digging up the connectors and some wire and getting it done. I'm solidly good at procrastinating.

For the record, should anyone end up here from a search later:

The factory splices are located between injectors 3 and 4 in the factory harness. They are brass crimp splices, and without heat-shrink.

Finally got my MS3x yesterday, but didn't have a lot of time to play with it short of popping it open to inspect the quality of assembly and soldering.

Going to mod the DIYBOB wiring to fit my '91 tomorrow. Doing final planning on that now. Looks like I only need to reposition 4 wires, and add the no-VTPS resistor, then I'll be PnP ready.

I modded the injector harness Thursday and ended up putting the plug on injectors 2 & 3 (so I can remove them from the stock harness, and move them to individual wires). When I started really analyzing this, I realized that batch fire pairs up 1&3 (injA) and 2&4 (injB)... and to keep the injector sequencing correct when going sequential, I need #3 to be on injB. My solution there is to, in addition to switching the connector that gives me 2 additional wires, also switch the connectors on #3 & #4 at the injectors. Net result, original wires feeding #1 (injA) and #3 (injB) and then new wires going to #4 (injC) and #2 (injD).

Does that sound right? Is there an easier way to do it that I'm just not seeing?

I plan to run batch until my WBO2 shows up, I just want to get the wiring on the ECU set up correctly in one shot so that I don't have to mess with it again.

Lastly, unless someone has a better option, I'm going to fire it up on Frank's base 1.6 MSQ. I'll disable EGO (or maybe switch it to NB) since I don't have the wideband yet, otherwise, it appears to be configured correctly for my initial installation.

When running fully sequential, the correct wiring is:

INJ-A = Injector 1
INJ-B = Injector 3
INJ-C = Injector 4
INJ-D = Injector 2

Same as the firing order of the engine.

Okay, sounds like I'm doin' it right then. Let's see what I can accomplish today...

We have achieved first start!

Trying to figure out idle settings right now... it wants to idle at almost 1200.

AC is inop. Will have to check wiring on that.

Main cooling fan has yet to come on, but that's still triggered by a thermoswitch, right? It's cool today. Engine temp has yet to exceed 195 idling.

Anyway... so far, so good. Just need to learn my way around the new MS3 settings and options.

The idle issue was the idle screw on the TB. Backed it down and the idle is quite nice now. Having fun trying to tune with a NB sensor until my WB shows up next week.

Haven't looked much into the AC problem yet, but found a good thread where Brain explains the ins and outs of it. It all makes sense, just need to figure out what's up.

Tomorrow, I'm going to set REQ FUEL to something reasonable (it was 12.0 by default and has proven to be lean above 4k even at 12.5... and the calculator says it should be 11.8, I think) and then use the NB to try to tune the VE table to stoich. Once I find stoich, I'll add 10% to the top end and interpolate values for the area between there and the cruise area of the table.

Fun stuff!

Cool. I haven't tested it yet, but it looks like my AC isn't wired "wrong", just "different". It's using the N20 switch (29) instead of the Table switch (28), and it's using the N20 output (25) instead of InjG (1). Should work, I just need to set the soft options correctly. And I'll likely never use Nitrous, so no problem there.

I'm diggin' this whole MS3x thing! The configurability is great!

Still fiddling with the AC and the idle and startup settings. (and probably will be forever) But, I've got the car running really well now, even without a Wideband. I'm not saying that I tuned it with a Narrowband... I made a feeble attempt, but it was a pain.

What I ended up doing was searching around for a good NA MSQ. And, for anyone who is searching in the future, let me throw out some search terms: NA 1.8 1.6 stock megasquirt msq base tune normally aspirated

What I found was a dyno-tuned MSQ for a stock 2000 1.8 making 132 hp at the wheels. (Yeah, dynos lie... whatever. Take 5% off of it and it's still pretty damned good!) It's a 1.8, not a 1.6, but I figured the timing map should be really close, and the fuel map should mostly just need to be scaled.

So, I took those fuel and spark tables, dumped them into my tune. Set the Req Fuel to 12.3 rather than the 12.8 from the 1.8 MSQ. I did have to tweak the timing at idle a little bit. But, otherwise, the car ran fine. Better than fine. And surely better than I'd get it trying to tune with a narrowband sensor.

Before I forget, here's the thread with the MSQ that I stole:
https://www.miataturbo.net/megasquir...o-sheet-67115/

Pretty happy for now... waiting for my WB...

why didn't you use the 90-93 specific MS3x basemap on Frank's website that I gave him that matches the wiring diagram I made as well?

If you load the map I made, it's 85% of the way there and all you have to do is put in your req_fuel and then tune fuel.

running random basemaps for different model MSes and different model year miatas is silly at best.

Silly though it might be, it worked way better than Frank's map. Don't know why. Frank's MSQ started and idled well, but flat out would not run no matter what I did with the Req Fuel. Regardless, it's running now and running well. (well enough until I get the WB installed)

(digression: The biggest difference I noticed between the two tunes is the timing map. The dyno tuned non-turbo 1.8 map is about 10 degrees more advanced at/near/below 100kPa above 4-5k. My theory is that because you guys are tuning on turbo cars... you're never at 100kPa at those kinds of RPM. You'd be in boost by then and having fun in some other part of the table.)

BUT, I have a question.

I just spent some time out in the driveway fiddling with idle settings and the AC and damn if that's not some weird shit!

I verified that the main fan control is working. It's output is on Nitrous1 and it works as-advertised.

The AC output is wired to Nitrous2. For kicks, I put that output on the main fan to test it. That OUTPUT works as advertised. Fans and AC both come on when commanded in that configuration.

So, I went back to AC settings. The AC switch is wired to Nitrous In. I set the AC output to Nitrous2. This should work. (of course, it doesn't)

What's weird is this:
First, I figured out that it was working BACKWARDS. The AC was off with the button pushed in, and it was ON with the button out. Was able to replicate this several times, and thought I had it figured out. It was weird, but okay... at least it made sense and it worked.

But, then it decided to go back to what it was doing on me this morning: It simply stays on! The AC button has no effect. On or off. Whether you start the car with the button on or off, doesn't matter... it's staying ON. (actually, when you start-up, it doesn't come on until WUE is done)

WTF is up with this? Anybody got a clue?

I'm happy to switch to another input (I didn't choose this one, it was wired this way when I bought it) if that will solve the problem.

sound like you wired it incorrectly. I like using INJ outputs for stuff like that, since they are high current drivers.

Like I said, the output is behaving. The input is being weird. Is there something different about the Nitrous Input?

BTW, I have discovered intermittency that might explain the problem with Frank's msq. Car was running GREAT this afternoon. This evening, it's back to intermittent starvation and stuttering. I found a leaky injector seal... Maybe affecting fuel pressure?

yeah, the n2o input is a 12v input; so it's reversed. you want an input that goes to ground, like tableswitch or datalog.

Bingo, that answers the whole thing. Gracias!

I'll switch that tomorrow... and install some new injector seals... and be a couple steps closer to nirvana.

Another FYI tidbit:
There's a jumper JP8 on the MS3x that can be put in place to change the Nitrous Input from a 12v to a Gnd switching input.

Megasquirt-3 MS3 Hardware Manual

If I hadn't already rewired my connector to use the Launch Input last night, I might have done that instead. Good to know, anyway.

it's easier to move a wire on the db37 and change a parameter in the software than open up the case and add a jumper.

Probably. It's still a good piece of info for the person who may be using or planning to use all of those other inputs.

Hey, has anybody ever used the 12v supply from the purge valve as power for their WB02? I was looking for an easy way to wire that... and that looks like a no-brainer. It's on the same circuit as the ECU and injectors. Put a 5A fuse inline to keep O2 shorts from blowing the main ECU fuse... should work, right?

why would you want your WBO2 controller in the engine bay? or even on that note, as far away from your ECU and exhaust as possible?

Can't imagine why not. The wideband does not require a lot of current, and it doesn't matter is there's a small amount of offset in the supply between the ECU and the wideband controller.

Ideally, thou shalt ground the wideband controller to the same point as the ECU, and as near to the ECU as possible. For a MS, this is typically achieving by grounding the sensor directly to one of the unused GND pins at the DB37 connector on the main board.

In the case of the Spartan, this would be the black wire, which is the "electronics ground". (The white wire, which is heater ground, can go anywhere.)

On the other hand, you already have to run some wire from the sensor to the ECU, and it might just be simpler to run all wire from the sensor to the ECU, rather than splitting the harness out and grabbing different signals at different places.

Brain, I'm using the new Spartan sensor that has the controller circuit built into the connector. There is no separate controller.

http://www.14point7.com/Support/Soft...ser_Manual.pdf

And I just realized it has two outputs. I thought it was WB only. Nice!

I'd still run it up the shift boot like most people do and splice it into the ECU harness for simplicity. Wiring that into the engine bay makes little sense to me.

My initial logic was that there's already an O2 sensor wire going to the ECU, and sensor grounds going to the ECU from under the hood. I could reuse those easily enough, just need a source for power from under the hood. No running of extra wires, no splicing into the ECU.

However, since my new WBO2 has an NB output, I'll probably go ahead and use that for the NB output to retain stock ECU backward compatibility. Which means I need to run a sensor wire to the ECU at the very least.

Joe's idea of just running a bundle of spare wires through the firewall for various and sundry purposes is seeming like a better and better idea.

My initial logic was that there's already an O2 sensor wire going to the ECU, and sensor grounds going to the ECU from under the hood. I could reuse those easily enough, just need a source for power from under the hood. No running of extra wires, no splicing into the ECU.

However, since my new WBO2 has an NB output, I'll probably go ahead and use that for the NB output to retain stock ECU backward compatibility. Which means I need to run a sensor wire to the ECU at the very least.

Joe's idea of just running a bundle of spare wires through the firewall for various and sundry purposes is seeming like a better and better idea.

Ditto that.

From the downpipe, it's a pretty straight shot around the transmission and up the shift boot, and from there it's just straight across the dashboard to the ECU.

Sounds promising. Still waiting on the WB to show up.

Thinking about heading to the salvage yard and raping a 12-pin connector from something to plug into one of the unused slots on my DIYBOB connector. Would make for much easier "non-stock" connections.

(why didn't I think of that last week when I was there?)

Even after switching the input, my AC is still acting wonky. Sometimes it works, sometimes it doesn't.

But, I think I've figured out why. The stock 1991 wiring switches both the AC relay and the fan relay on activation of pin 1R. My setup (please don't give me any more shit about this, *I* didn't wire it this way, and it was supposedly working in a 1994... though I never asked if his AC was functional) has 1R connected to a "medium current" output, which surely has enough current to trigger one relay... but might not be sufficient to trigger two in series.

A project for another day. I've got injector seals to swap out to stop a fuel leak and stabilize my fuel pressure. (I knew #4 was leaking... but didn't realize how much until I opened the hood with the engine running a while ago... it is SPEWING fuel!)

I told you to use an injector output.

I feel like putting all this time/money/effort back in March of 2010 into being one of (if not) the first to run ms3x on a miata and finding out best practices to spread wisdom to all my minions was just wasted effort.


also, 1R is the main cooling fan. 1J is the a/c compressor and a/c fan relays. could THAT be the issue?

I asked you not to give me any more shit about it.

If I had built the thing, I'd have followed all of your best practices. And as I'm fixing things, I'll be following them. But, this thing was built by someone else and I was led to believe that it was 100% functional, so I saw no reason to change anything other than what was required to wire it to my 91.

As I'm learning that is not the case, I'm fixing it where it needs fixing.

Lest you think that I don't know how to search, or who the "go to" guys are, I've already found this thread with great AC and Fan wiring info:
https://www.miataturbo.net/megasquir...-verify-67046/

I'll get there, man. Your efforts are appreciated.

you're lucky...this time.

:brain:

















have you seen this:

https://www.miataturbo.net/attachmen...ine=1354048142

this is how my unit is wired.

Wiring pinout saved. I've seen several similar to that one, but I'm not sure they were exactly the same. (there's one that came in the box with my MS3, even)

Oh, and I have 1R and 1J correctly wired. I just had 1R on the brain because that's the one I had to move (it was on 1L for the '94).

Will make some decisions on how to pinout the 12-pin connector for the WB02... and extra injector outputs... and... what other future-planning should I do?

At this point, I'm not planning on doing sequential spark. I haven't found a good reason to do it yet. I won't be adding boost. Is it worth doing for an NA application?

I guess it would be wise to go ahead and pin-out the remaining extra input and outputs just to have them at the ready. Hmmm... maybe I should use the 16-pin rather than the 12?