Hi rx8 peoples,

Firstly, my question is: Do you all see anything that screams "don't even start this thing, it's gonna blow"? Also, do you what stands out?

I recently picked up a 2008 rx8, manual trans, 6 port engine, 103k miles. I was told by the seller that it burns coolant. It started right up first try and blew a little white smoke, but to me it looked and smelled more like intake/engine cleaner (think sea foam). I only started the car one more time to pull it onto a trailer and there was almost no smoke. I have been reluctant to run it till I got a look with a borescope. I personally don't see much to make me not run it at least a little while to see if the white smoke clears up. Engine is easy to turn over with a wrench on the main pulley. Not even a torque wrench, just a 3/8 ratchet. I'll run a compression test before I close it up and follow up with the results here. Maybe I'll also see if they still sell the kits to test coolant for exhaust gasses. Planned on pulling the engine and doing a very basic rebuild with my son, but if we can drive it a bit before that's a huge win.


Car sat for a year or more (or so I'm told). line looks like where the engine rested or maybe nasty chatter?

Is this carbon build up or coolant?

Rear rotor, lower plug

 

Doesn't look terrible. I would run it, do a comp test as you say, etc. Don't do a basic rebuild, unlike a piston engine, there's no "basic" version that doesn't result in needing to do it again in a year. Do it right, check specs on everything, replace big parts, etc, or do it twice.

 

If a coolant leak is suspected, you can buy (or make) a cap with a Schrader valve for the degas tank, and pressurize the system (bicycle pump, not compressor.) and see if it holds.
Easy to do, and you will see if the PO diagnosis was wrong, and if its good to run as-is for a while.

 

Good idea. We generally did chemical tests to be 100% certain, but for myself this may make more sense.

This car is a salvage title so I'm not investing heavily till I can put it through its paces. besides, Mazda has made it clear they don't intend to support these engines. Replacing things like housings, irons, and rotors with new parts is not going to be reasonable soon if not already. Makes me sad to see my favorite engine configuration tossed aside, but making due with basic resurfacing and restoration for major components is the only way I can see keeping these cars going for years to come. I would love to do everything right, but that's not something my bank will allow.

 

Mazda still produces parts for it... And new engines. You can buy one from the dealer.

i generally discourage rebuilding on your own, but as a fun project with no expectation of a long running car, why not.

 

Mazda would disagree.

The Rotary Engine Lives On

And with the MX30 R-EV, the Wankel tech will keep on spinning.

 

So... I didn't like the idea of using a dial gauge to guess at compression. For under $15 and some left over parts I'll make what I want. Yes, I have a problem. Yes, this is unnecessary.

sometimes you want what you want and you can only afford to use the spare parts laying around.



They are required to make every part for the car for 10 years by US federal law. 2012 was the last model year so 2022 would end their obligation. The fact that they already downsized the engine manufacturing line to 10 people in an old factory 2 years ago (2021) indicates to me that the end is near. Add to that steep price increases recently and the outlook gets worse. This is just my opinion, to each their own.

 

They were still producing 2004-2008 engines though, which are different from 2009-2012. But yes, nothing is forever.

I have a similar DIY compression tester with a DataQ acquisition device rather than an Arduino (who knows, might be an arduino inside?). There's a write up in the DIY section for building one.

 

Found an Arduino project with a nested while loop inside a for loop that I like the way it handles the compression peaks and rpm calcs. I will definitely look at that write up here and add its technological distinctiveness to my own project. Though I don't like that the project I found is only using a single measurement for each face, probably going to use a 2d array to store multiple readings for each face and take the average. I don't find many Arduino boards inside commercial stuff, but I do find a ton of ESP32 SoCs in IoT stuff. I will probably make a separate post on the tester if people are interested.

 

Mazda just doubled the price for rotary engine kegs last week; new REW keg that was $5000 - $6000 is now $10,000+, it will be cheaper to buy all new irons, rotors, and housings and have it rebuilt now.
.

 

Quick update on the engine and compression tester

Engine: Borked
Started it up to get it warm for a compression test. Only a trickle of white "smoke", then it warmed up. Billows of white smoke filled the sky. I got in touch with the previous owner, and they let me know it had run low on oil to the point where it overheated. I'm guessing the oil control rings and bearings maybe damaged as well as the water seals. gonna have to come out.

Compression tester: going well
Tested the electrical set up and the transducer produced the right voltage per pressure. Tested basic pressure reading through the Arduino. It produced accurate numbers according to the gauge on my compressor. Wrote the heart of the code to: take compression readings 12 times on each face of the rotor, calculate the rpm for each of the 12 readings, average the 12 compression and rpm numbers, and finally spit out the average value for each through the serial output. Haven't tested the code yet, but Arduino IDE will at least compile it.

I need to build a sturdy version for testing on the car. Robust cable, mini board to hold the cable and capacitor, and a printed case. I may add a screen, some kind of input, some menus, and maybe an SD card.

 

Sad about tge engine, but nice work on the tester. What kind of sample rate do you get with rhe Arduino? I noticed my dataq was just on the edge of acceptable if i averaged a few seconds of spin time per rotor. The pressure max peak is very short, it would often get lost between samples, so I would see compression praks with a chopped top.

 

\

Haven't tested the code yet, but I'm not sure I can check the reading frequency without slowing it down. I may add an array to grab the milli() time each cycle of reading, but that will inherently add operations. I'm using for and while loops to minimize operations that run during reading. I'm only recording the peak values, dumping them into an array, and processing the results after readings are complete. I usually build custom sensors for data acquisition in bizarre situations and environments that never need a frequency greater than 1 reading every 2-5 seconds, so I don't have any experience building for high frequency. I'm just implementing design philosophies that should minimize cycle times.

 

With the extra code to swap an output pin on and off it's sampling at 6 KHz. lower than I thought it would be. I may switch over to an esp32 based board. The lower voltage will reduce the accuracy of the sensor, but 10x the processor speed should get that resolution up a ton. What kind of frequency do you get with the dataq?

Edit: Some quick googling leads me to believe the older DataQ based models clock in at around 240 Hz. I will accept 6,000 Hz... I guess

 

Yeah I think you're ok with 6 kHz That'll give you 1200 samples peak to peak at 300rpm. 240Hz for the DataQ sounds about right, and it's workable but just barely.