So, I bought a turboed Miata... umm, learn me?

So I bought this guy. 1996, 81000 miles, hardtop, FlyinMiata springs and swaybars, tokiko iluma (sp?) shocks, kosei wheel with potenza tires, ok body with some minor rust. 

Has a turbo. GT2560 with a begi intercooler and coolant reroute. The computer is a megasquirt ms2v3. Has an innovate wideband o2 thingy. 

Now, he even gave me a laptop to interface with the megasquirt. But I’m kinda scared to link them up. I don’t want to mess any settings up, as it runs pretty well, but has a few misfires. Mild acceleration sometimes 2500-3000, and sometimes at steady cruise around 3800. The wideband gauge shows it’s fairly rich on accel but the plugs look like it’s running lean.  It’s set up for 8psi. He says it dynoed 172. It really rips. 

Im not sure where to begin, I do t want to do something wrong with the computer and end up with a car that doesn’t run!

any suggestions?

First, congratulations on your purchase. You're doing the right thing by trying to figure it out before you break it. 

I would suggest looking around for a shop that can tune it for you on a dyno. See if you can get some recommendations so you know who to call.

I would think from reading your post that you're more interested in a safe tune and drivability than dyno numbers for bragging and dragging. That's going to take the right person. 

You're gonna have to talk to the MS to diagnose your issues. It's possible that you have 'normal' problems that have nothing to do with the MS or the turbo system, but being able to watch different datastream items while the car is misbehaving is pretty critical to ruling things out so you can narrow down on the culprits. 

Stumbling during acceleration is probably a lean issue. Does your Innovate wideband require free-air calibration? 

IIRC if it's an LC-1 it will require free air calibration. I'm not a big fan of the LC-1, the one turbo'd Miata I owned had one and it regularly required resetting.

I'd start with hooking up the laptop to the MS and try to datalog the issue to see if there is anything obvious in the logs. The trick with tweaking ECU settings is to do it in small increments, as that tends to reduce the risk of blowing up the motor or getting to the point where it won't run.

The best way to diagnose the issues would be on a dyno, obviously, but you might be able to figure out some of the issues by datalogging on the street as log as it's done in a safe manner.

As far as screwing up the settings, when you connect TunerStudio to the ECU, select "save tune as".  Write the tune out to a file, and then go copy it somewhere safe (flash drive?  Email it to yourself?  stuff like that).  That guarantees you the ability to go back to where you started from.

Free air calibration is a good place to start.  Also check that the LC-1 is wired properly -- it's got two grounds, a signal ground and a power ground for the heater.  Many people will tie the two together, but the heater on the sensor is a square wave (on/off/on/off) and will introduce a ton of noise into the wideband signal.  The MS should have both a signal ground and a power ground, make sure to use the signal *only* for ground on sensors, not for anything that's sucking power.

If it's an LC-2 instead, those only have one ground, you're forced to gang them together.  This was bad enough on my car that I got an MSLabs serial-to-CAN converter box so that I could feed the digital signal from the LC-2 controller into my MS3 without contaminating the signal ground.  I don't think the MS2 supports CAN though.

One thing I'd suggest is to go through the wiring and document it.  I like using Excel to document wires, put each connector on there, with the # of pins, wire color, signal that it carries, and any other notes about where it's connected to.  That'll let you make sure it's all right, and also provide a good means by which to learn about what's in there.  It's also VERY useful when it comes time to tweak the wiring to add something else, and make sure you update the document when you do that.

Thanks guys!  I appreciate all the information, I’ll start with calibrating the o2 tonight. And sorting out the wiring for it, as I’ve found stuff like this:

that can’t be helping anything at all. 

Im wondering, (hey Keith!) since the turbo is similar in size, it’s intercooled, runs a similar level of boost, and dynos at about the same as an fm stage 1 setup, would I be ahead of the game (not for tuneability but reliability) if I went back to stock ecu and used an fm voodoo box?  I don’t care about getting the maximum ten more hp out of an engine, I want reliability and drive ability. And not blowupability... what do you guys think?  I could probably sell off the megasquirt anyhow. Or... sell the whole car off and do an fm1 on my 1.6 car???

Joel, if you get to the point where you're selling the Megasquirt let me know....cuz Challenge car.

I will let you know certainly!

another option for all this that quasi just pointed out... I have a very solid 1992 that I swapped a 1.8 in. Still 1.6 computer. I wonder if I should move the turbo to that car, and use the voodoo box?

I need to learn more about the voodoo box. Been looking for a non-standalone way to boost my 99.....

Dusterbd13 said:

I need to learn more about the voodoo box. Been looking for a non-standalone way to boost my 99.....

Check it out man!

https://www.flyinmiata.com/voodoo-box.html

If those are the plugs you pulled out, they are completely wrong for the application.  Anything 5 heat range is way too hot for a turbo car.  In NGK language, small numbers = hotter plugs, big numbers = colder plug. 

See if you can look up the same plug, but with a 7 in place of the 5.  (i.e. BKR7E). 7 is the heat range, 11 is the gap, in millimeters (1.1, or about .043").  I'd also consider closing the gap down to .035" or so, see if that helps the stumbles.  

edit* it looks like NGK 4644 (BKR7E) is pre-gapped at .036" and available nearly everywhere.  Bout $3 each.

stylngle2003 said:

If those are the plugs you pulled out, they are completely wrong for the application.  Anything 5 heat range is way too hot for a turbo car.  In NGK language, small numbers = hotter plugs, big numbers = colder plug. 

See if you can look up the same plug, but with a 7 in place of the 5.  (i.e. BKR7E). 7 is the heat range, 11 is the gap, in millimeters (1.1, or about .043").  I'd also consider closing the gap down to .035" or so, see if that helps the stumbles.  

 

edit* it looks like NGK 4644 (BKR7E) is pre-gapped at .036" and available nearly everywhere.  Bout $3 each.

Oh sweet!  Thanks!!

Hi guys!

Some notes in no particular order.

Plugs: I'd go with a heat range 6. 7 is more prone to fouling. We use ZFR7F-11, but you'll probably have to close them up a bit to avoid misfires at torque peak.

The Voodoo Box would certainly work on this setup. You'd have to go back to stock injectors and the stock ECU along with a MAF - check to make sure there's a space in the intake for one. That will probably give you better driveability and no weird cold start/AC/idle behaviors (those are the hardest things to set up, and the MS2 isn't as capable in those areas as the newer units) but it will cost you a margin of safety. The MS can react to knock and should be able to run closed loop fuel, which is preferable to the stock ECU.

Misfires while under cruise is not likely to be related to plug choice. I'd be looking at fuel numbers in the MS logs and doing a comprehensive wiggle test of all wires with the car idling. Looking at that...splice...I'm going to guess you're quite likely to find a couple of dodgy connections.

I would not drive that car hard until you figure out Megasquirt, get some data logs and do some tuning.  Lean = detonation = motor carnage = sadness.

Yes to cooler plugs.  Gap them to .025" to start with.  The stumbling may be spark blowout, although I doubt it at 8 psi.

The install is probably full of janky wiring connections like that, I'd sniff them all out and re-do them with proper crimp connectors.  It sucks to have to un-do other people's sloppy work.

Are you on stock injectors?  You may be running out of injector headroom, which is bad.  EV14's are the way to go here.

Megasquirt is the right tool for this job.  You're taking a serious step back to go to stock ECU and Voodoo or any of the other dummy boxes.  Been there, done that. Yes there's a learning curve with MS.  Yes you need to put some time and effort into properly tuning.  Yes you can drive a turbo Miata 100,000+ miles with few/no issues if you do the job correctly.  Logging and tuning is 80% of the work with aftermarket turbo cars; the actual turbo stuff is just nuts and bolts for the most part.

There are forums out there that have lots of information on this.  Do some Googling.

In reply to joey48442 :

Just go to the Miata.net website

joey48442 said:

OMG!  That clearly should be green wire, who does that?!?!

fwiw- You can likely answer all your turbo Miata questions on miataturbo.net without having to ask any questions. 

Yeah the plugs don't lie. If the reported fuel mixture is showing it as good, that's a problem. Even with the other suggestions, I'd recommend getting a new brand name O2 sensor and go from there. That, or buy a freestanding O2 sensor/readout as a reality check. Innovate sells them.

You might just look at stealing a harness out of a scrapped car? That one looks like you might chase problems all day, everyday. Cool looking car otherwise.

 

FE3tMX5 said:

fwiw- You can likely answer all your turbo Miata questions on miataturbo.net without having to ask any questions. 

I was going to suggest this, but wasn't sure how it would go over on this forum.

The answer to every single question you can think of can be found by googling:

<search terms> site:miataturbo.net

Yes and no. The signal/noise ratio can be problematic, and there are certain strongly held beliefs in that forum that may prevent you from learning about good alternatives.

Ok guys!  Thanks for all the advice, I’m going to try to get to all of it. I recalibrated the o2 the best I could. I couldn’t get it to break loose so I disconnected the down pipe and blew fresh air through to get it so it could calibrate. It does seem to run better, but still misses at steady cruise around 3800 rpm. Gauge says air fuel is right around 14.4 when it’s missing. 

I like my cars to just work, and I don’t think HP is a big enough draw for me to be worth the trade off of not understanding what’s going on lol. Maybe I’m a bad person but I don’t particularly care to learn megasquirt 

The miataturbo.net forum has a reputation for being cliquish and unfriendly to newbies.  I think that's a bit of an overgeneralization -- it definitely IS unfriendly to newbies who don't bother to do any research ahead of time, or to people who come in and directly challenge the conventional wisdom of the community.  That said, new people who show up having done some background reading and who don't take a confrontational attitude don't seem to have problems.

As with all forums, it has its strengths and weaknesses.  Despite the weaknesses, I find it to be pretty useful when carefully navigate -- it's the main nexus of people with budget/homebrow/DIY miata turbo kits like the one in this thread.  As mentioned, you can get access to all of that information without ever creating an mt.net account or making even one post.

In reply to joey48442 :

You don't need to "learn" Megasquirt to really sort out what's happening and to potentially make it "better" to just "drive"

The wiring issues are pretty basic and should be something you'll want to repair no matter what you use for an ECU.  I would start there, especially when buying someone else's project, you generally want to give it a good once over and make sure the wiring is in good shape (it isn't), the tune up parts are known to be good (unknown, how's the timing belt?  Is the timing correct?), the fuel system is in good shape (unknown, check the fuel filter and injectors for cleanliness and flow rate) and the ignition system (plugs are too hot for a turbo car, how are the plug wires, the coil?).  I would check the TPS with a volt meter to ensure there aren't any bad spots and that there aren't any vacuum leaks while I was in there since Speed-Density can be thrown off by this.

Anyway, think of the laptop as an OBD scan tool and use it to log data off of the MegaSquirt while you drive. 

Connect the laptop via serial or usb to the ECU, plug it into the cigarette lighter (hopefully it came with a power adapter you can plug into a lighter) turn on the laptop.

Strap the laptop down (or have it sit on a passenger's lap) so that it won't move around while you drive.

Start the car and let it settle into an idle, launch Tuner Studio and make sure it is connected to the ECU, save the tune file someplace safe and name it something that makes sense to you.  If you get any errors at startup, etc. take a picture to share here and don't make any changes.

Now, you'll want it to start data logging, follow the prompts from the menu bar and then go for a drive.

When you're done and you've created a log file, upload it someplace where we can see it and we can parse it and let you know what we find.

There are lots of great videos on YouTube that describe the process and what to do with the info afterward.

https://www.youtube.com/watch?v=LeqAzDNOeaI

Honestly though, if you want to just have a car to "drive" ?  I would sell this on, you'll be sorting through the PO's work for quite a while and I suspect you'll always have the thought in the back of your head that something is going to fail or act up.  A stock ECU and a VooDoo won't provide you with the same good feelings due to its own limitations.

Stefan said:  A stock ECU and a VooDoo won't provide you with the same good feelings due to its own limitations.

I'm not sure if that's a typo or not, but it's not correct. The Voodoo setup gives nice smooth power and has right about the same maximum potential as the stated 172 rwhp the car has already. The big advantages to the Voodoo setup are that you get the well-sorted factory calibration for all the hard parts of the map and that you have no emissions concerns. Sure, it's not going to make 250 hp, but it's well suited to the power levels and ease of use targeted here. The car will start better and idle better and just plain work. The trick is to use the best tool for the job, and for this particular job the Voodoo is a good match.

Once the current wiring demons have been exorcised, anyhow.