1 2 3 ... 7
codrus GRM+ Memberand SuperDork
11/15/16 12:42 a.m.

So earlier this year I picked up an FD. I wanted something low-miles and very close to stock, which is fairly difficult to find at this point, so I spent about six months looking for it. When I found one it was on the east coast, I'm in California, so I posted a thread about whether to ship it or go pick it up (here). I wound up having it shipped, it arrived in June. I've posted stuff about it a few other threads on a couple forums, but decided it really wants a build thread of its own. Since I don't have one on GRM yet, I'll put it here.

It's a one-owner car (well, two now), had 58K miles on it and was completely stock when I got it. It still had the original shocks, original brakes, original radiator, etc. I intend to keep it that mostly stock -- replace broken things, improve reliability, but fundamentally it's going to be a fun street car. I have a nutso track car already in the form of the Miata.

The day it arrived:

My first peek!

Out in the California sunshine!

With the rest of the fleet (minus wife's Odyssey). There are six turbos in this shot!

Tucked away in the garage:

codrus GRM+ Memberand SuperDork
11/15/16 12:51 a.m.

The first order of business was to get it registered, for which it needed a smog test. I was a bit worried about this -- CA smog tests aren't simple, and FDs aren't known for being particularly easy to pass. Turned out it was a big deal about nothing, it passed with flying colors.

The next order of business was to do something about the tires on it. They were full tread depth, but the date code was from 2007, and they were all seasons. Clearly it wasn't getting driven very much, but those tires were just not going to do. The stock size is 16x8 and decent tires are somewhat hard to find in that size, I wanted to upgrade them as well. I've been a big fan of 949 Racing's 6UL wheels on my Miata, and they offer a 17x9 +48 in the right bolt pattern. It's intended for S2000s, but it fits the FD nicely, and with RE-71Rs it's got a ton more grip than those ancient all seasons offered.

They match!

codrus GRM+ Memberand SuperDork
11/15/16 1:03 a.m.

Next up was to check out the motor and see how it was doing. I'd had a pre-purchase inspection done at a Mazda dealer in NC, they did a compression test and reported that it was good (don't have the actual numbers sitting around any more). From what I can tell from reading about them, the real question is whether the boost numbers line up properly. You're supposed to hook up a boost gauge, run it to redline, and watch the gauge and the tach at the same time while not hitting anything. Ideally, it should go up to 10 psi, drop to 8 psi as the second turbo spools up, then climb back to 10 and hold that. Well, I'm a software guy and have been playing with Arduinos on my Miata a bunch lately, so we decided to datalog it instead.

So, tee off a vacuum line, hook it to a MAP sensor, plug that into the arduino. Tap another line into the IGN output in the diagnostic connector and I've got an RPM signal. A couple pages of C later I took it out for a drive, logging the values to a csv file. Then feed it to MegaLogViewer (intended for Megasquirts, but it turns out it'll basically graph any csv file you throw at it), and I've got:

In the top chart white is map, red is RPM, both are smoothed. The raw signals are in the lower chart, as you can see there's a lot of noise in the data. The curve looks pretty good, that's basically exactly what it's supposed to be. I guess the rats' nest of vacuum lines seem to working as intended!

Next up was to do something about the muffler. The car had clearly not been driven a lot, and the muffler had rusted through on the bottom, presumably from condensation building up in it. There's very little rust on the bottom of the car -- more than I'm used to living in California, but "rust-free" by east coast standards. So I called up Goodwin Racing (one of my favorite Miata vendors, but it turns out they sell RX-7 parts too!) and ordered a Racing Beat muffler for it. Installation was pretty simple.


Obviously upgraded, but not the giant ugly tip a lot of aftermarket FD exhaust have.

codrus GRM+ Memberand SuperDork
11/15/16 1:22 a.m.

After this the car got a few more projects with no interesting photos. I replaced all the brakes with OEM-equivalents and Hawk HPS pads, flushed out the swamp-muck-colored brake fluid, and did a Bose-ectomy on it.

As I said earlier, I don't plan any track days in this car. That said, the stock cooling system is known to be marginal, and my FD had a 25-year-old radiator with plastic end tanks in it. That's about 18 years into borrowed time already, so I ordered some cooling system parts. A Koyo radiator, new upper and lower hoses, new thermostat & gasket, and an aluminum 'air separation tank'. The FD radiator is mounted low and about a 45 degree angle, so the radiator cap isn't on the radiator itself. Instead there's a fill cap on the thermostat housing and a small plastic tank with the pressure-relief cap leading to the overflow tank. These ASTs are also infamous for failing, mine was also 25 years old. The parts arrive and then sat in the garage for a month or so while I got the Miata ready for the annual Miatas@MRLS event at Laguna.

The week after the Laguna event, I was driving the FD to work and started smelling coolant. Oops, guess it's time to put those parts in. So in we go:

Airbox comes off. Wow, that's a lot of air box hoses.

Battery & tray come out.

Here's the AST:

Here's the thermostat housing. The smog pump to the left has to come out to get this off:

After getting the smog pump off I looked underneath and, uh oh. What's this?

After pinging a few rotary friends I concluded that this job had just gotten bigger, need to do the water pump too. I live about a mile from a Mazda dealer, so I hopped in the Miata and went round, but alas the parts department had closed 10 minutes earlier. Calling around to the local FLAPS got me a water pump & gasket available Sunday morning. So I can't finish the project, but there's more to take apart.

The rest of the intake plumbing comes out, showing the fancy multi-plate throttle body:

Alternator comes out:

Water pump comes out, leaving this housing:

codrus GRM+ Memberand SuperDork
11/15/16 1:30 a.m.

The water pump housing sits on the front housing of the motor. Unfortunately, it's not bolted down, instead it's sandwiched between the water pump and the front housing using studs. This means that when you pull the water pump off, the gasket between the WP housing and the front housing usually tears, and mine was no exception. The FLAPS didn't have that gasket, it didn't even show up in their computer, and Mazda was closed. So, technology to the rescue -- take a photo of the flange with a ruler in it for scale:

(yes, one could copy those keys from the photo, but it turns out that they're both obsolete ones going to locks that don't exist any more -- I really should get around to taking them off my key ring!).

Load the photo into photoshop, rescale so that the ruler is correct, then edit out everything but the flange. Print it and cut it out:

Inkjet paper is kind of a lousy gasket though, so trace it onto some proper gasket material and cut it out again.

Wow, that worked really well. I've never made a gasket that fit that well before.

After that a friend came over and helped me put it all back together and I forgot to take more in-progress shots, so I just have a few of various parts on the garage floor. Here's the whole WP housing:

How cute! It's just like an intercooler, only smaller! :-)

The Koyo is noticeably bigger than the stock radiator. Early versions of this rad had a reputation for not fitting very well, but this is a newer version and it went in fine without needing to cut anything. The intercooler bolts were moderately a PITA though.

BTW, these hose clamp pliers are a godsend when doing something like this. I picked them up on a friends' recommendation a few months ago, hadn't used them until this job. Highly recommended, I think I got them at Amazon.

codrus GRM+ Memberand SuperDork
11/15/16 1:42 a.m.

So that's got us pretty close to today. There are a few things I still need to do with it:

  • I have a set of plug wires to put in, should probably swap the plugs too.

  • If I drive the car, park it for 5-10 minutes, then restart, it doesn't idle well for the first few minutes. It will often stall while backing out a parking space in this situation, for example. I'm not sure if this is hot restart tuning or what. I have read that, unlike piston engines, rotaries get looser as they heat up, and that a rotary that won't start when hot is the first sign of needing a rebuild -- not sure if that applies here. The Mazda dealer supposedly did a compression test, but maybe they did it cold? (Supposed to be done hot). Sounds like I need to do my own compression test on it.

  • The AC was dead when I got the car. It had previously been converted to R-134a, so I filled it up with refrigerant (it was very low, but still above zero). After that it worked fine, but a month later it was dead again. Poking around showed it leaking UV dye out of the evaporator drain under the car, sounds like the evaporator leaks. Alas, nobody makes evaporators for this car any more. Mazda said they had one (1) in the country, in Texas, but they thought it might have been a return, and they couldn't guarantee it fit my car anyway. Apparently half the FD AC systems were port-installed and got a totally different compressor/evaporator/fitting/etc from the ones that were installed in Japan. A friend of mine just bought a crashed FD as a parts car -- hopefully it's got a good evaporator in it that I can scrounge.

  • It could do with some suspension work. It's not going to the track, but it's got 25-year-old shocks in it, and it shows. I'm just not sure what to do with it -- I really like digressive valving for shocks, but I don't want to pay that kind of money for something that's just a street car. OTOH, it would be nice to drop the car a little bit, just to close up the wheel gaps. So I'm undecided on what to do there.

  • It's not going to the track, but I might take it autocrossing. Unfortunately, I didn't pay attention to that when I bought the wheels, and the stock ones are 16x8. 17x9s would push it out of street, and AFAICT there's no STx class for FDs. That means I'd be running it in ASP or SM2. While I didn't expect it to be competitive, that's kind of a steep curve. :) Ah well, I'm not sure I'd fit with my helmet on anyway.

  • The clutch pickup is quite high, suggesting there's a lot of wear on the clutch. If I'm doing that, I'll want a lightweight flywheel in there too, the stock flywheel is much too heavy for this motor. Speaking of the transmission, the the synchos aren't in great shape either. I'm not sure if that means it's been abused, or if I'm just too to a Miata feel...

  • When I bought it, the car was presented as having only had the nose repainted "for rock chips". Well, it turns out it's had a LOT more paint work than that. The whole car has been painted, some parts of it twice. The driver's door has a chip where you can see three different layers of paint, and the top coat of red is definitely not the stock Mazda color -- it's got a bit more orange in it. It's not a bad color, but it's not the original. Unfortunately, it was done cheaply, there are lots of fisheyes and other defects in the paint, bad masking lines near the trim, etc. It looks good from 5 feet away, but up close it's a mess. There's also overspray on a bunch of stuff inside the nose, like the old radiator.

codrus GRM+ Memberand SuperDork
11/15/16 1:49 a.m.

Pushrod New Reader
11/15/16 3:25 a.m.

Nice work replacing the water pump gasket. Thanks, I'll remember that tip!

wvumtnbkr GRM+ Memberand SuperDork
11/15/16 5:34 a.m.

These cars are soooo sexy! Good job on this one!

The clutch pedal can be adjusted on these.

SyntheticBlinkerFluid UltimaDork
11/15/16 8:08 a.m.

Awesome. Quite jealous. I hope to add a non-molested FD to the fleet one day.

JoeTR6 HalfDork
11/15/16 8:16 a.m.

This is one of my favorite cars. Looks like you found an amazingly clean one. Enjoy.

Fitzauto Dork
11/15/16 11:04 a.m.

So jealous. One of my favorite cars and in a great color.

jfryjfry Reader
11/15/16 11:23 a.m.

My fantasy car. Where in California are you? I'm by lax if you're close - happy to lend a hand

docwyte Dork
11/15/16 12:02 p.m.

That sucks about the paint. That's one of the reasons that I refuse to buy out of state now. I've been burned in the past with out of state purchases, despite paying for a good PPI and being very specific about what I wanted checked.

codrus GRM+ Memberand SuperDork
11/15/16 1:35 p.m.

Yeah, I've learned a lot about buying cars out-of-state from this experience. Unfortunately, the number of decent FDs out there is very small, in the 6 months I was looking I only found three other cars that I seriously considered. One of them was sold before I could even get in contact with the seller, a second got sold to a friend just as I started seriously negotiating (I think it was one of those "you keep saying you want to buy this car, now's your last chance" kind of things), and the third one was close but not quite right. It was the color I really wanted (a '93 R1 in competition yellow mica), but it had a few too many mods on it for what I was looking for, and the price was higher than I wanted to pay. The closest of those was Texas, I didn't see a single car in California that was even close to the condition I wanted.

I live in in the SF Bay Area.

codrus GRM+ Memberand SuperDork
11/19/16 8:58 p.m.

So as I mentioned earlier, I want to get my own compression test numbers for the car but, as you probably know, compression tests with rotaries aren't as straightforward as with other cars because of the 3 combustion chambers sharing a single spark plug. There are techniques for using a standard tester that involve watching the pulses on the needle and trying to judge the differences between them, but that seems kind of unreliable for me. There's a really expensive electronic factory Mazda tool, and there's a $400 electronic aftermarket tool, but blech.

It seems to me that an Arduino and a pressure sensor should be all you need, maybe $30 in parts! So I ordered a pressure sensor, specifically a Honeywell NBPDANN150PAUNV. This is where my inexperience with electronics shows up, because it turns out that unlike the MAP sensors I've used for data logging stuff in the Miata before, this is an uncalibrated and unamplified pressure sensor, so the output is measured in millivolts, too low to accurately read with the arduino. It's also a differential output, apparently centered on 2.5 volts. Ah well, in my defense there don't seem to be many pressure sensors at mouser that are rated to 150 psi!

So it needs amplification in order to be readable, and from what I can tell an lm234/nte987 op-amp sounds like the right answer. So I picked one up at Fry's (along with a few others just in case) and decided to give it a try.

Well, apparently I'm doing it wrong, because it doesn't work. At ambient the sensor puts out around 16.5 mV between the two differential pins, and engaging the organic lung-powered mity-vac I can get it down to about 10mV (can't find the actual mity-vac, but this will do for now). I hooked up the op-amp to give around a 20x gain (so as to put it into the range the arduino can read with good resolution), and tested it with a 4.5mV input signal -- got a 26x gain. Larger than I expected, but OK. Unfortunately, when I feed it the inputs from the sensor, the 16.5mV ambient signal becomes a -17mV ambient signal, the lung power brings it down to about -23mV, but the op amp output is stuck at 578mV, which doesn't make any sense at all.

Still trying to figure out why it doesn't work, argh! :)

wvumtnbkr GRM+ Memberand SuperDork
11/19/16 11:26 p.m.

A video camera / smart phone and traditional compression tester work well.

JtspellS SuperDork
11/20/16 12:03 a.m.


Seem like good guys, and worst case you sell it off to someone else and get most your money back?

Look forward to the build on your FD and still have it in my mind to own one some day.

codrus GRM+ Memberand SuperDork
11/20/16 12:54 a.m.
JtspellS wrote: http://rotarycompressiontester.com Seem like good guys, and worst case you sell it off to someone else and get most your money back? Look forward to the build on your FD and still have it in my mind to own one some day.

Yeah, that's the $400 one I mentioned -- OK, I guess it's $340. Still, if I can build one for $30 in parts, I'd rather do that. :)

Jarod GRM+ Memberand Reader
11/20/16 8:39 p.m.

I love the car! Sucks you are uncovering some undisclosed stuff, but looks like you're really bringing it up to speed.

What lift is that in the garage? I'mean looking into a residential four post for my garage.

codrus GRM+ Memberand SuperDork
11/21/16 2:37 a.m.


The lift is a Bend-Pak HD-9ST. It's a 9K pound lift, with the short (length), short (height), and narrow dimensions. Being a 4-post it's not as convenient for service as a 2-post, but I got it with the RJ-45 roller jacks, so I can lift the car off the runways to change tires & work on the suspension. The garage isn't wide enough for a 2-post lift, they take a lot more space on either side of the car than a 4-post does.

I spent some more time playing with the pressure sensor today and decided that I really need to buy one that's amplified & calibrated, rather than trying to DIY that with this one. So it'll have to be a compression tester with $60 worth of parts, instead of one with $30 worth of parts. :)

fidelity101 SuperDork
11/21/16 4:12 p.m.

I never use paper gaskets for the waterpump, RTV works great! I think your hot start issue could be resolved with fresh plugs/wires.

also if you can build an arduino rotary engine compression tester I will give you money for one right now haha

I use the piston compression tester and cell phone video.

hold the schrader valve down and watch it bounce for about 8-12 times then let go of the valve and let it build max then stop engine compression test, the way the needle can bounce you can tell sometimes which seals are bad as they will affect how even they bounce and etc.

its very hard to find these stock and in good condition, thanks for saving this one. They are a blast to drive as is but a few fuel/cooling reliability mods don't hurt.

Before you get willy wild on the exhaust be sure your fuel system is up to date but if you want to daily it in cali you are limited in what you can actually improve I bet.

codrus GRM+ Memberand SuperDork
1/4/17 8:55 p.m.

I have fresh plug wires sitting in a box, ready to go in, need to get some plugs to go with them.

I bought the proper amplified pressure sensor, need to get back to that project. Now that the holidays are over, hopefully I'll have some more spare time. :)

I'm not daily driving it (I have an Audi for that), but it does get driven to work occasionally. Smog checks in California limit what I can do with the turbo system without jumping through a lot of hoops, so for right now I'm sticking with the stock computer. Downpipe is really the only other exhaust mod I'm pondering, and that only for heat reduction.

I did get something done on it recently though, took it in for a proper alignment. It's now running a degree and a half of negative camber up front, a degree in the rear, 6 degrees of caster, and a smidgin of toe-in at both ends. Basically treating it like a large Miata. It's much happier now than it was before.

84FSP Dork
1/5/17 6:43 a.m.

Beautiful ride - good to see it being saved and keep true to form.

4cylndrfury MegaDork
1/5/17 9:41 a.m.

I know that unmolested FDs are rare as unicorn farts. But honestly, I really want to find an FD in mediocre shape, and go crazy with a sawzall, welder, large doses of caffeine, and a boosted piston engine.

All that said, yours is gorgeous. And I love the wheels, great choice! And I am a huge fan of DIY tuning and maintenance approaches. I applaud your efforts - the photoshop gasket is really smart!

1 2 3 ... 7

You'll need to log in to post.

Our Preferred Partners