Mazda Miata ITB Project

By way of introduction:

I write software all day and over the years I've found that I need to do something completely different to unwind at the end of the day. Working with my hands is that something.

For a while, my medium was clay...

But we all know that Miata is the answer...

I recently moved to the Bay Area and discovered the joys of TechShop, which is this chain of fun places with expensive tools they encourage you to use. I've been farting around doing this and that, teaching myself to TIG weld, powder coat, etc. I finally decided that I needed a project that would tie together the different skills I was learning.

Which brings us to the beginning of this story.

I don't quite remember the logic that got me here, but I started to think that the world of Miata ITBs was pretty sloppy. The solutions that I saw there all felt pretty half-assed and expensive to boot. Most off them start with these janky old throttle bodies off of a JDM-yo Toyota.

So I set about looking for newer vehicles that use individual throttle bodies, which sent me into the world of German over-engineering. Perfect!

Step 1: Does this have any chance of working? I ordered a Miata intake manifold gasket and a singleton German throttle body from eBay.

The initial "does this have any chance of working" answer was yes...

I took the 3D scanning class at TechShop, thinking I could scan the throttle body and start putting stuff together virtually.

The first rule of 3D scanning with the tool they have is "nothing shiny". I can fix that...

After some equipment struggles, I was able to scan the part in, with some help from the Machinist's Handbook...

With the throttle body scanned in, it was simply a matter of scanning the gasket (on a regular copier) and extruding that into 3D. The 3D throttle body scan did not turn out to be worth the trouble. I ended up doing a E36 M3-ton of measurements and drawing the throttle body flange by hand.

The good news is that this is pretty much the worst interference between the two parts. It's also a great indication of just how well the German throttle outlet matches the Miata intake port.

And with that, it was time to order the full set of throttle bodies, with all the extra bits of German OCD that I could hope for.

Here they are. And what's really nice here is that big silver (aluminum? stainless? dunno) tube that's not connected to the fuel injectors. It connects all the cylinders together, providing a single MAP port as well as a source for idle air (via that large connector in the center of the picture.) Also, a completely solid electronic TPS.

Here's a better view of the MAP port, PCV port (or something...), idle air source, and throttle arm.

While waiting for these to arrive, I took a first stab at designing the flange. I took my design, bought some 0.125" MDF, and used the CNC laser cutter to cut a template.

This was the first crack at it. Good in some ways, bad in others.

I swore I had done the math right, but somehow everything was just ... off. The template should have lined up perfectly with the Miata gasket. I screwed up somehow.

After some head-scratching and rework...

Great success!

To the CNC router!

This is the second try. The first time, I knocked the control box off the table halfway through the program and it (of course) landed on the E-stop button, which causes the machine to lose its place in the program, which means it's time to start over again. On this iteration, I learned that it's really valuable to preview your cutting path carefully before running it. The path I generated was great except that it went way too deep way too fast for the path for the outside edge of the part. But this was good enough to play with.

For instance...

One of the other neat bits about these particular throttle bodies is that they're attached (in a tidy German fashion) to these...

Despite being threaded on the circular end, these do not unscrew from the plastic plenum they sit in. It was a godawful mess extracting them without damaging them.

Here things are starting to come together... Take note of that throttle linkage arm. It's usually connected to a throttle-by-wire actuator motor, and while I did give some thought to converting the Miata to DBW (it's just electrons, after all), that's a bit more elephant than I want to chew on in this project.

While searching the local junkyard for some parts (the idle air control valve and TPS pigtail), I looked in a bunch of different engine compartments for inspiration. I found a nice USA-made Craftsman flathead screwdriver, and then came across this bit just sitting innocently in a Volvo.

Exact same arm, albeit slightly shorter. And actuated by a throttle cable. Score!

German/Swedish mashup for a Japanese car...

Time to tear apart the Miata to see what's going to fit...

Initial fit-up...

There's unexpected interference at the back, where this oil line for the VVT actuator runs way too close to the rear intake port.

And the direct-to-velocity-stack isn't going to work, unless I want to be sucking in fuel line... (note that this is a best-case scenario - there's no adapter in use here)

Well... onward. Here's how things look using the Miata injectors and fuel rail, with the German throttle bodies and air tube.

And after a trip to the plumbing aisle of Home Depot... here's one option..

I tweaked the flange design and did another session on the laser cutter. Excellent.

And now you're all caught up. More posts soon, including some questions for the hive-mind...

Awesome.

Awesome! That would bbe nice if a shop like that exists here... Great mashup of international engineering :)

So you're in Cal? How will you smog this beast?

simon_C wrote: So you're in Cal? How will you smog this beast?

Return to stock. It costs me a few hours and a few dollars in gaskets.

In reply to peter:

So you'll be using the stock ECU?

Not without an air box.

What's the plan for filtration?

Keith Tanner wrote: Not without an air box. What's the plan for filtration?

The plan is fuzzy there. I intend to have as large of a plenum as I can fit without cutting things. I will feed that plenum from outside the engine compartment, likely from the space in front of the right front wheel, where the current intake draws air. There will be paper filtration somewhere along that path. I expect packaging will be an issue.

Right now, I'm trying to nail down a design for throttle actuation. Once I know how I'm doing that, I'll know where the plenum can and can't be. Once I have that nailed down, I can start to figure out where to sneak the intake tube through.

In my day-to-day as a software dev, we use Kanban boards to track progress and make sure things don't get forgotten. Since this is a complicated project, I'm attempting to keep track of it the same way. So I present to you my Kanban board for the ITB project.

Stuff in the "Backlog" column is a list of things that need to get done, but aren't being actively worked on.

Stuff in the "Blocked" column is a list of things that can't be worked on because something else needs to get done first.

Stuff in the "In Progress" column is actively being worked on. As in, today.

Stuff in the "Done" column keeps track of progress that's been made. It's motivational - showing you how much of the elephant you've eaten already. I haven't been using the board since day 1, so a lot of "Done" things have not been captured there. I'm OK with that.

Personally, I don't move things to "Blocked" until they've been "In Progress" at least once. I allow myself to put things back on the "Backlog" after they've been "In Progress". Some people don't do that. I'm OK with it.

Now, my question for the HiveMind™

What sort of engine bay temperatures should I expect to see in a normally-aspirated Miata?

I ask because TechShop has a nice vacuum-molding machine that can use. It would be really nice to be able to make the intake plenum from plastic sheet that I molded with that tool. I would like some confidence that the plastic I choose to use will hold up in the engine bay and not deform on a spirited drive, autocross, or track day.

ABS sounds like it would be a likely choice, but its heat deflection temperature is only 98℃ at 66Mpa load. Am I underestimating how intense that load is, overestimating the engine bay temperatures or what?

I can get polycarbonate sheet easily enough (140℃ HDT), but it's expensive. And somehow doesn't seem like the right choice.

The original airbox that came with these throttle bodies was made out of a glass-reinforced plastic of some kind, but which I'm not sure I can thermoform. TechShop's injection molding machine is far, far too small for an object of this size :)

Wow! Very interesting & great potential in this thread! McMaster-Carr has good information on the properties of various plastics that may be helpful for you --> About Plastic

The underhood temps aren't consistent. I've never measured them, but 98C is basically the boiling point of water. Most of the temperatures underhood aren't that hot, but you'll be up there near the header and behind the rad.

Fiberglass is a good option for this sort of thing as well. You can't use magic 3D tools to build it directly, but you could make a plug.

I'd probably do some initial testing to find out what length of runner you want. I've done this sort of thing with ITBs on a Miata engine and you can have some real effects. Chances are the answer will be "as long as possible" and packaging will be your enemy.

I am entirely unfamiliar with developing itb stuffs and in no way should I be allowed to do so without adult supervision. That said....could you machine the base/adapter for the tbs' at angle? I'm thinking if you could have the throttles angle up at say 30 degrees it may help the packaging?

IIRC the off-the-shelf TWM throttle bodies for 1.8 Miata do that.

PeteD wrote: Wow! Very interesting & great potential in this thread! McMaster-Carr has good information on the properties of various plastics that may be helpful for you --> About Plastic

Thanks, I spend a lot of time pouring over that website. Once I know the temperature range I can expect, it's easy-peasy to go there or walk into one of the local plastic shops and get what I need.

As an aside, one of the things that I didn't appreciate about Silicon Valley before moving here was how much "real world" stuff goes on here. It's not just nerds like me pushing electrons around - there's a serious manufacturing base here. Not heavy-industrial like cars, but doodads and prototypes. I bike past an aluminum supplier on my way to work and two different plastic wholesalers are just a couple blocks away. The old-school silicon companies that created this place still have a presence here - Fairchild, Tyco, others whose names escape me. It's just impressive, I think.

This looks like fun! I wish I had access to the big boy toys.

What is your plan for dealing with the TB bolt interference/overlap?

Keith Tanner wrote: The underhood temps aren't consistent. I've never measured them, but 98C is basically the boiling point of water. Most of the temperatures underhood aren't that hot, but you'll be up there near the header and behind the rad. Fiberglass is a good option for this sort of thing as well. You can't use magic 3D tools to build it directly, but you could make a plug. I'd probably do some initial testing to find out what length of runner you want. I've done this sort of thing with ITBs on a Miata engine and you can have some real effects. Chances are the answer will be "as long as possible" and packaging will be your enemy.

Huh, I had expected far higher temperatures than that, but this is why I ask. I figured that the coolant, at 100℃ (what I think of as boiling), was keeping the combustion heat in check, but not necessarily balancing it out. I figured the engine could get significantly hotter than 100℃, but the coolant kept it from reaching a critical temperature. As I think more about it, that makes sense though, at least on the non-header side of the engine.

Yes, fiberglass has popped into my mind. It would be a heck of a lot easier to just make a mold and vacuum form the piece if I can.

appliance_racer wrote: I am entirely unfamiliar with developing itb stuffs and in no way should I be allowed to do so without adult supervision. That said....could you machine the base/adapter for the tbs' at angle? I'm thinking if you could have the throttles angle up at say 30 degrees it may help the packaging?

Yes, it's possible to machine the adapter at an angle. But I don't think I can machine an adapter at an angle like that.

Crackers wrote: This looks like fun! I wish I had access to the big boy toys. What is your plan for dealing with the TB bolt interference/overlap?

I'll be replacing the stud in the head with a bolt/screw that has a narrower head. The surface for that head will be a step below the throttle-body-to-flange mating surface, so the TB can overlap the bolt by a bit. So far, it's actually looking pretty decent.

Keith Tanner wrote: I'd probably do some initial testing to find out what length of runner you want. I've done this sort of thing with ITBs on a Miata engine and you can have some real effects. Chances are the answer will be "as long as possible" and packaging will be your enemy.

One of my initial thoughts was to use the Lotus Engine Simulation (LES) software to get some ballpark estimates and work from there. Super helpful folks those Lotus engineers, they answered an email of mine in just a few hours when I had some trouble getting their software. They may have taken it down now, but for a while there the first hit that came up when searching for this software was an ancient, dead version of their download page. They set me straight very quickly & pleasantly!

I quickly learned that to get even a ballpark estimate, I'd need a heck of a lot more information than I have access to. I could build myself a flow-bench approximation or something, and that's still an option, but without a spare VVT head to muck around with, it's just a bit too much right now.

I remember at least one of your experiments with ITBs (on the Seven), though I know you've had others. I'd love to get some solid results, but realistically, for the first pass, I'm hoping not to lose any power. I figure after I get the basic mechanics sorted out, the runners and so forth are much faster to iterate on.