I had some project ideas for my 2004 Mazda RX-8 that involved tapping power from its auto-dimming rear view mirror. After determining its power connector pinout, I realized I wasn’t excited to keep going. This marks the end of a chain of events that started with a puddle of coolant demanding my attention to find the leak. I think I’m done with car projects for a bit. I’m writing down my current thoughts which may or may not hold up in the future with more hindsight wisdom.

A big part of keeping an old car running is replacing or repairing old worn out parts. As a car of early 2000s vintage, my RX-8 has lots of plastic and rubber pieces that have become brittle and fragile with age such as the radiator I just replaced. Paradoxically, much older cars may be easier to keep running because more of their parts were made of metal. It’s not very satisfying to replace plastic and rubber pieces that were only designed to survive duration of the warranty.

I did learn quite a few things as I worked my way to the radiator, understanding more of what’s under the hood. I now recognize hoses not just between the engine and radiator, I also recognize hoses leading to the coolant reservoir and leading back towards the interior heater core. This is similar to the thrill I get from teardowns, but that thrill is diminished by many annoyances. The first is that age throws barriers in my way — like stubborn hoses — that are not illuminating and not fun to solve. The second is that, unlike most of my teardown projects, I have to put this thing back together afterwards!

21st century cars also tend to be very cramped under the hood, making everything hard to access compared to the relative wide open spaces under the hood of something from the 1960s. So many sensors, wires, and hoses are packed in there and every single one is a potential point of failure. Parts availability dry up as cars get older, but this problem is worse for low volume cars like a RX-8. Parts are expensive when they are available at all. This is a problem shared with most cars that date back to the 1960s, the difference is that it’s much easier to design and fabricate a substitute. For a 1960s car, it’s likely something mechanical that can be built in a home workshop. For a 2004 car, good luck fabricating a replacement sensor that can keep the engine control unit (ECU) happy. And there’s a risk of such a project going sour killing the ECU, which would turn a small problem into a huge one.

Some people work on their cars to make it uniquely their own and feel a connection to it. I can certainly understand that and I’ve played around with it myself. But projects like retrofitting Android Auto and a backup camera aren’t breaking any ground. (New cars just come with that stuff now.) Neither is a dashcam, so I don’t think it’s going to happen.

I’ve managed to replace the radiator in my 2004 Mazda RX-8, and now I have to put everything back together starting with new engine radiator hoses. Just like with the radiator, I did a quick comparison of old and new before installation and I’m really glad I did.

Cars typically have two hoses between engine and radiator, a RX-8 has three. Its top hose is divided into two parts joined by a metal mid pipe visible in the top left of this picture. The new upper front hose looks pretty close to the old one.

The new upper rear hose looks pretty close as well.

And this is where doing the comparison paid off: the lower hose is not an exact match. It looks pretty good on the engine side (left in this picture) but I see problems beyond that. The hose curvature is different but probably close enough to be bent into shape. The next two problems might be related: First, the radiator-side outlet flares out to roughly 5mm larger diameter than it should be, more than what the hose clamps can compensate for. Second, the flared section added overall length of the hose making it too long. I did a loose trial fit and jamming a too-long hose in place would cause unacceptable kinks in the hose.

Maybe it was a packing error? I double checked part number on the label (Continental 61098) and verified it was the intended product to fit my car. My next hypothesis is a manufacturing error. Somehow this hose wasn’t bent to the exact curvature and the radiator end was not trimmed to the correct diameter and length. The good news is that it’s a lot easier to cut a hose shorter than it is to cut a hose longer. I can fix it now rather than return this item and wait for a replacement to be shipped. I trimmed off the flared end so both the length and diameter is correct for a secure and kink-free fit against the lower radiator fitting.

The final challenge is compensating for curvature. The flawed shape wants to rub up against a sharp corner on the mid pipe bracket. That would quickly cut a hose open and not acceptable.

Fortunately Mazda engineers designed a white plastic clamp that clips between hose and mid pipe bracket so the lower hose is properly routed to avoid damage. Success!

I’ve been working my way to the radiator in my 2004 Mazda RX-8, which is buried pretty deep in its nose and set at an angle. It’s been a production to get to it for replacement. Once I removed the engine radiator hoses (which I will also replace) I have removed everything explicitly called out in the workshop manual. That manual called for removing the radiator with the cooling fan module still attached, but it didn’t explicitly say whether the assembly should be lifted up out of the engine bay or dropped down below it. People on RX-8 Club forums have reported success in both directions.

Either choice requires careful maneuvering. Lifting up means I have to be careful not to damage large stubborn wiring harnesses and refrigerant lines for air conditioning. Dropping down means being careful not to damage the air conditioning condenser dangling by refrigerant lines. The radiator’s protruding top mounting brackets are especially dangerous to condenser fins.

I first tried lifting the radiator up but got tangled up between refrigerant lines and wire harnesses. I then tried going the other direction and had better luck carefully lowering it below the car onto the floor.

I removed the fan module and found several pieces of debris that has fallen in there over the life of this car. The largest piece was a rubber bushing that I didn’t immediately recognize. I would later figure out it was part of the cosmetic engine cover. I lost both bushings over the years and my engine cover could only sit loosely in its position. Now that I’ve recovered one bushing I could reinstall the cover more firmly when I care to do so.

Some of the debris bounced around fan blades before settling down off in a corner. The old radiator on the left showed two arcs of damaged fins that lines up with the lower edge of fan blades. I set the replacement radiator down to the right for a quick side-by-side comparison to verify it looks largely correct. All the hose fittings seem to be the same diameter and located in the same places, as are all the attachment points for the fan assembly, the AC condenser, and chassis structure.

Transferring the fan assembly and mounting hardware, I had a new radiator assembly to install. I tried to reverse what I did earlier but had trouble lifting the assembly from the ground into the engine bay. This assembly is not heavy, but lifting it means bending over the front of the car and reaching through the engine bay out the bottom. It was an awkward pose with poor lifting leverage.

Next I tried to drop the assembly in from the top, but the same wire harness and refrigerant lines were in the way. This time I got smarter: I rotated the assembly 90 degrees so I could lead with the narrow side edge of the radiator, and that allowed me to clear refrigerant lines and wires. Once the leading edge is poking down under the car, I carefully rotated the radiator back into proper orientation.

So my opinion for the top route vs. bottom route discussion is: both! I think it’s easier to let gravity help instead of fighting gravity. Which means removing old radiator out the bottom, then install new radiator from the top. This is especially true for someone working solo as I did. If there are more hands available, it should be easier to maneuver the assembly and lift against gravity so a different path may work better when installed by a team.

I’ve learned that the radiator in my 2004 Mazda RX-8 was long overdue for replacement. Conventional wisdom says if a radiator is getting replaced due to age-related issues, it’s a good idea to replace hoses at the same time. Partly because they are flexible parts subject to wear from age, and partly because we have to remove one end anyway during radiator replacement. I added hoses to my radiator order.

First lesson: my locking pliers are my allies when battling hose clamps.

Second lesson: radiator hoses that have been clamped to the same fittings for over twenty years are reluctant to move. Once I freed the hose clamp, radiator-side hoses came loose with a hard tug. But the engine-side fittings were more challenging. It took far more effort to remove the top radiator hose and, once freed, I could see the problem: a layer of buildup acted as glue between hose and fitting. I think most of this is dried coolant, but some of it may be corrosion. Most of it crumbled when I scraped with my fingernail which I then wiped away.

The engine’s lower radiator hose is buried deeper. I chose to remove the secondary air injection pump to give myself more elbow room but I still couldn’t pull the hose free. Again I exercised my dial-an-expert option and Emily Velasco suggested working a screwdriver in there to free up the hose. This makes sense: I’m pulling parallel to the adhesion surface, the direction where it is strongest. I need to apply force perpendicular to the adhesion surface, like how we peel off a piece of adhesive tape.

However, tight quarters meant I couldn’t get a tool in there, but Emily had a simple solution: bend a screwdriver. I chose a small one that was bundled with my Creality Ender 2 and bent it into a tiny little pry bar. Now I could push the tip into the gap between old hose and engine fitting, and I could hear and feel the tactile sensation of those surfaces peeling apart.

Once the adhesive bond between hose and fitting was broken, I could pull these old hoses free clearing the way for radiator removal.

I decided to go ahead and install a replacement radiator in my 2004 Mazda RX-8 despite the fact it arrived with some minor damage presumably in shipping. I’ve temporarily patched my old radiator but I expected it to fail again soon given its age. This turned out to be a good call! As I started preparing to move the coolant reservoir tank out of the way, its associated radiator fitting broke off.

I learned from RX-8 Club forums this fitting becoming fragile with age is a known common problem, but I was surprised at how little effort was involved. I estimate it took less force than what it would take to break a piece off a chocolate bar. I might have been one bad pothole bump away from this thing breaking on me. In fact, looking at the pattern of dried residue around the break, it was already degraded enough to seep coolant for some time.

My next problem was how to extract its remnants from the hose. Consulting my friend Emily Velasco’s experience of working with old machinery, she expected this segment to be just as fragile as the rest. A firm squeeze with some big pliers should be enough to break this apart into smaller pieces for extraction.

Her expectation was correct, it didn’t even take much of a squeeze for the thing to fall apart. I had expected to feel some resistance or maybe a “snap” marking the point of breakage, but there was nothing. I squeezed and it collapsed. I pulled out the pieces and tried squeezing the hose again with my pliers. The effort level was basically the same.

I picked up one of the pieces and tried pushing it with my finger. I was able to fold it in half with about as much effort as a graham cracker. Yikes. This thing is long gone.

And that’s not the only thing showing severe degradation. Here’s a picture I took of the top engine hose fitting after I removed the radiator. It showed discoloration and material degradation similar to the snapped-off reservoir hose fitting. The worst discoloration corresponded to where the hose clamp applied pressure. The base of this fitting was where I found a crack (and applied J-B Weld) and it’s pretty obvious that crack was merely the first clear failure and many other potential failures were close behind.

Bottom line: Yes, replacing the radiator was a good idea! And with these broken fittings, I’m past the point of no return. The car is not going to drive under its own power again until the new radiator is installed. Next step: struggling with old radiator hoses.

I ordered a replacement radiator and associated hoses for my 2004 Mazda RX-8 before I found out my engine is terminally ill. I briefly contemplated not bothering with the radiator but then I decided to look on the bright side: If the engine is on its way out anyway, it doesn’t matter as much if I screw things up. I can dive into the learning experience of performing my own radiator replacement with less pressure. So far I’ve done oil changes, brake pad replacements, ignition system upgrade from Black Halo Racing, and replaced a water pump that had started weeping coolant. A radiator replacement seems like an appropriate next step up in difficulty.

After the radiator arrived I opened the shipping box and was mildly disappointed to see radiator fin damage both front and back. The radiator had fallen off of its packing foam blocks and the air pillows fell out of place. I wished they had taped the foam and pillow to the radiator instead of loosely draping. Inspecting the radiator, I saw no other signs of damage. All mounting points seem intact and accounted for, and all three hose fittings appear to be in good shape. The smallest and most fragile fitting for coolant reservoir connection even got a separate protective rubber boot!

Besides, the fin damage is pretty trivial compared to visible condition on my remaining heat exchange units: the air conditioning condenser in the middle and oil coolers on either side. Their fins are pretty bent up from twenty years of road debris, and they’re still sufficient. Besides, as a clumsy beginner, I think it’s pretty likely I’d bend a few radiator fins as I try to install it. If this radiator is not going to stay pristine for long anyway, why worry about a few bent fins now? It will have enough cooling capacity to keep the engine from overheating. And if I’m wrong… well, the engine was on its way out anyway. I’m going to jump in with an enthusiasm for learning.

My 2004 Mazda RX-8 has failed its compression test. This fact did not come as a huge surprise given its age, but it’s still disappointing news. Long term I will have to either get a rebuilt engine or move on to another car. In the short term I will live with this condition and I’ve been reading up on what to expect going forward from RX-8 Club owners’ forum and sites like RX-8 Help. This post is my summary notes to future self.

Poor compression degrades power output which in turn degrades fuel economy. Beyond that, it is expected to slowly degrade my car’s drivability until I reach limits of my tolerance for frustration. For example, what kicked off this investigation was the fact my car would occasionally stall instead of idle. This symptom will get worse. Increasing in frequency from “occasionally stall” to “frequently stall” and my patience will run out somewhere along that spectrum.

Low compression can also make “hot start” difficult: Unable to start an engine when it has already been warmed up and things are hot under the hood. Most commonly noticed at short stops such as when stopping off at a gas station. (… or when the engine stalls at idle…) I’ve noticed my starter motor cranking an extra second or two before the engine catches but so far it has always started. As compression degrades, it’ll take longer and longer until at some point it just won’t catch and I have to let the engine bay cool down before I can start the engine. Not a big deal if I’m sitting in a parking spot, but if I’m at a gas station I will have to push my car out of the way so other people can refuel. If this happens I will be highly motivated to act on one of my long term options.

Increased workload on the starter motor may burn it out. I could then replace it with a unit from a later “Series 2” RX-8 which runs faster. Since rotary engine compression improves with speed, a faster motor is more likely to start an engine suffering from bad compression. Some aftermarket starter motors can give me this effect as well. A faster starter is only a workaround that delays the inevitable, but it is an option if I want to spend a hundred bucks to postpone a far more expensive decision.

How long might I have before encountering such drivability issues? That is unfortunately murky. Poor compression means some of the air+fuel mixture, either before or after combustion, are escaping past one or more worn seals. The exact escape route will vary from engine to engine because every one wears out slightly differently. It is possible for certain escape routes to cause erosion that starts a destructive feedback cycle of engine wear. In which case my car’s gradual descent becomes a nosedive.

Speaking of escape routes: if some of that mixture leak past side seals, they would end up in engine oil. Such an increase in oil contamination will demand frequent oil changes in order to avoid triggering bad oil problems. I’ve done oil changes every 3k miles for as long as I’ve owned this car and this is one more reason not to slack off now.

And finally, the catalytic converter might get caught up in the aftermath. Even in a healthy engine, some small amount of fuel would escape unburned out the exhaust port. The catalytic converter exists to catch those molecules and convert the nastiest bits to less-bad compounds. But if unburned fuel is leaking past worn seals, the catalytic converter may receive more than it can safely handle and fail. Usually catalytic converter failure would illuminate my Check Engine light and the car would fail regular state-mandated emissions test. In the rare case, though, it is possible for an overloaded catalytic converter to catch on fire. At some level I admire a car that decides to go out in a blaze of glory, but that’s not the kind of excitement I want in my life.

For better or worse, none of these things demand immediate attention and I can still drive my car as-is for an unknown period of time. Which means I should get back to my project to-do list now that the replacement radiator has arrived.

I replaced battery terminals in my 2004 Mazda RX-8 because they were corroded and poor electrical continuity might have caused the occasional engine stall at idle. The new terminals look great, but my engine would still stall, so I investigated other possibilities listed in the workshop manual and found a likely explanation: my engine is very low on compression.

A compression test was one of the listed diagnostic items but not something I could do myself. While some auto part stores offer engine compression tester rental, they’re for piston engines. Rotary engines require a specialized testing tool. The service department of my nearest Mazda dealership doesn’t have theirs anymore so I searched for a local rotary engine specialist shop. I found Lucky 7 Racing offers rotary compression testing and took my car to their new City of Industry location.

A rotary engine tester measures pressure value for each chamber around the three-sided rotor and also engine RPM. The nature of a rotary engine means compression pressure rises with RPM. (Oversimplified explanation: centrifugal force flings apex seals outward against the housing, improving compression as it spins faster.) So the tester also normalize/compensate its readings to 250 RPM for comparison purposes. According to “Compression Inspection” section of Mazda’s workshop manual, standard compensated pressure is 830 kPa and minimum acceptable value is 680 kPa. For my engine, compensated pressure values for one rotor measured 581, 556, and 554 kPa which are well below minimum. The other rotor measured 393, 392, and 347 kPa which aren’t just below minimum, they’re less than half of standard pressure! My engine is in sorry shape.

Low compression is a common symptom of old engines as various parts wear down. For a piston engine, valves and piston rings are the usual suspects. For a rotary engine, their apex seals are usually the first things to go. Dropping below 680 kPa is cause for engine replacement under Mazda warranty but since the powertrain warranty has long since expired, this is now entirely my problem.

A new engine is not an option, Mazda production wound down years ago. A proper fix now means taking the engine apart to replace worn out bits. Lucky 7 Racing performs rotary engine rebuilds, but that is typically done for performance-focused customers who want to install more durable (and expensive) race parts to wring out maximum power. For a daily driver car like mine, its faster and easier to swap my engine core with another core that has been remanufactured at a dedicated facility. Either option will be significantly more expensive than the market value for a 2004 Mazda RX-8. In other words, it would cost more than I could possibly ever get back and makes no financial sense. Plenty of people are passionate enough about their rotary-engined cars to pour money into a hole, keeping shops like Lucky 7 busy. I will have to decide if I’m one of them. Until I make that decision, though, I will have to live with the problem.

Header image: “Rotary Man” sculpture sitting among parts and equipment at Lucky 7 Racing. His head (painted blue) is a rotor and his torso (painted white) is an eccentric shaft.

After installing a backup camera for my 2004 Mazda RX-8, I wanted to look into an engine stall problem I’ve been having. As I approach a stop sign or red light and release the throttle pedal, engine speed will drop as expected. Usually it’ll settle at idle speed, but every once in a while the engine would drop all the way to zero RPM and I would have to restart the engine… which sometimes would stop again immediately unless I use a bit of throttle to keep it at a higher RPM.

According to the workshop manual, there are many potential explanations for this symptom. The cheapest and easiest thing to check is my battery connection. I guess the alternator doesn’t provide enough power at idle so the battery needs to pitch in?

Well, my battery terminals are pretty old. There have been several batteries throughout the car’s 20+ year life and a few of them have leaked battery acid to corrode these terminals. The plating is long gone, and some of the core metal has been eaten away as well. They are due for replacement whether they are causing this problem or not.

I first tried a set of “universal” terminals from AutoZone. I think they probably suffice from an electric continuity standpoint, but they are a little bit longer than the previous units.

For the positive terminal, they’re too long for the protective cover to close. For the negative terminal, the extra length pushed wires against the battery strap and I worry about abrading wire insulation. I don’t think these are going to work well long term.

Now that I have learned length can be an issue, I measured length of the old terminal and found these units on Amazon with the correct length. (*) I confirmed the length with a quick test fit, but the test fit found another problem: they are too loose to clamp tightly on my battery posts.

To solve this problem, I cut away the center red plastic tongue to give it more room to clamp down. I used the cut-off piece of plastic as nut washer/spacer so my socket wrench could grasp the hex nut for a properly tight fit.

Once installed, these terminals appear to give me the same fit as the earlier units. The negative terminal no longer pushes wires against the battery strap.

And now the positive terminal protective cover closes properly.

I’m happy I fixed this particular eyesore under the hood, but soon after this swap, my engine stalled again. Old corroded battery terminals weren’t the cause, so I need to continue working down the workshop manual’s checklist of potential causes.

(*) Disclosure: As an Amazon Associate I earn from qualifying purchases.

I noticed a coolant leak on my 2004 Mazda RX-8 and tracked it down to a cracked radiator housing. The good news is I think I caught this problem before it caused any serious problems, the bad news is that it’s hard to get parts for a low volume car that’s over 20 years old. I had no luck finding a (reasonably priced) replacement radiator locally, so I ordering one to be shipped to me.

What do I do in the meantime? I suppose I could continue driving the car, frequently topping off the coolant reservoir with water. I don’t think that’s a good way to go, though, because hot pressurized liquids have a history of making small cracks bigger. Possibly turning this radiator failure into an overheated engine failure. So I decided to try plugging the leak with some J-B Weld. I learned of this miraculous product from watching 24 Hours of LeMons video summaries, helping many teams get their car back out on the track.

This will be my first opportunity to use the product myself, and a great practice opportunity! Since I plan to replace the radiator soon, it won’t really matter how well the product works. But I will gain the experience of using it, and a short period of seeing it in action. For some guidance I enlisted my friend Emily Velasco who has J-B Weld experience and here’s what we did:

For surface prep, the surface around the crack was cleaned up and sanded down to give the epoxy a rougher surface to grip on to. The coolant level was drained below the height of the crack so it would quit weeping out which would interfere with epoxy adhesion. Once the area was clean, roughed, and dry, the first batch of J-B weld was mixed and applied.

For patch jobs like this, Emily recommends putting in something to help add structure to the blob of epoxy. A thin sheet of metal is common, and I remember seeing LeMons racers cutting up old license plates for this role. Unfortunately, we are dealing with a compound curve here so bending a sheet of metal to shape is difficult. As a workaround, our Home Depot run involved picking up a sheet of fiberglass cloth (Bondo 20128) along with J-B Weld. Emily recommended using the original formula and not the quick-cure type I saw in LeMons videos, because unlike them I am not in any pressure to get back out on the track.

A small square of fiberglass cloth was cut out and pushed into the still-soft first layer of J-B Weld. After the first layer had set (4-6 hours) but before it cured (15-24 hours) a second batch was mixed and applied on top. The result is not pretty, but it held well enough to stop further coolant leaks. The awkward position of the crack meant we got some J-B weld on adjacent components though I don’t think they’ll get in the way of future radiator replacement. I hope this old radiator’s brittle plastic doesn’t crack somewhere else while replacement radiator is still in shipping.

While I wait on arrival of that radiator, I decided to tackle other RX-8 projects on the to-do list starting with installing a backup camera.