BDC

I'm about getting sick of seeing this stupid problem. It seems that any kind of interruption of the normal, sequential flow of power output from one of these older blocks (older than the 13BREW's) runs the risk of cracking both the front and rear upper dowel lands. Happened to mine on my front plate in Dec '06. This one occurred this past weekend on Luke Weber' FC3S up in OH.



In this particular case, we had a perfectly-good running car blazing along on the dyno, digging its rear wheels into the dyno drum along with its front-end rising sharply into the air. Did about 15 runs like this without the engine missing a beat, running cleanly without a hiccup. A dramatic boost spike to 28.7psi (from about the 20.5psi neighborhood), an odd jolt on the motor and driveline, then shazam -- oil spraying all over the windshield, hood, and engine bay. Upon disassembly, the front rotor also had a few flattened corner seal springs, nearly exactly like the failure on my engine.



The setup:



Series 5 Mazda Rx7 Turbo II

Half-bridgeport S5 engine (porting and assy by myself)

- recently replaced the rear plate due to this same failure a year ago

HKS TO4R w/ 1.32 A/R P-Trim hotside, divided manifold, dual Tial WG's

Alkycontrol system running twin M10 nozzles, 100% methanol use

720/1680 injectors, adjustable FPR, and Supra in-tank Denso pump

Haltech E6X



I first tuned this setup to 100% pump fuel at about 14.5psi of boost. It didn't take a long while as he already had a pretty decent street map from my prior trip out there. I mostly shaved off some of the richness up at the higher RPM's to keep it a bit more uniform in terms of AFR's. From there, I whacked 30% off the bar 22 (~15psi part of the map), advanced the leading spark from 13* at 15+psi to 15*, and skinnied the split from 10* to 7*. I also re-linearized the fuel curve from the 15+psi spot backwards to about bar 16 to help tune the transition for alcohol. My target was a 70/30 to 65/35 ratio of pump gas to alcohol.



EGT's (with the probe located approx 3" past the turbine housing in the downpipe) on the 100% 14.5psi pump run maxxed at 1450*F. With the 70/30 - 65/35 alcohol going, even at close to 21psi of boost, they were in the 1050*-1150*F area, although I am not sure if there's a practical difference in EGT's loading on a dyno vs. on the street where my car saw 1300*F consistently w/ the probe in the same location.



My theory on why it cracked the plate:



The car had been experiencing a steady amount of boost instability since its start. I would've never guessed that (for whatever reason) it would spike to the point of launching itself 10psi over where the target was. I had seen some comparably smaller spikes in the prior runs but didn't think a whole lot about it since they weren't major and nowhere near to this degree. In any event, I'd spent a good hour or two tuning and trimming fuel as we raised target boost from the 14.5psi spot up to the then 20-21psi spot. My goal was to get it to 25psi and stop there. We'd had one jolt to the motor several runs prior at lighter load but nothing occurred that was fatal to the motor. At first, I thought perhaps I heard a slight loss in compression in one of the chambers but never verified it. The motor sounded a little 'off' at low RPM's but as we continued with high power runs it seemed fine. I am curious as to whether or not my gut was correct on this and that the jolt that had interrupted that prior run did indeed compromise the spring pressure of one or two corner seal springs. On one run, I also noticed a 16000rpm spike on the tail-end of the datalog -- trigger noise. I'd never seen this while I was hammering on the car prior or even during my trip up there last year. While this is a problem and can certainly be detrimental to a motor, I don't believe this was the culprit.



I think the boost spike was like trying to skip ahead forward in time -- while it's theoretical that we can run that much boost with the dousing of alcohol in the charge that this system can yield, that portion of the map wasn't tuned even though the fuel curve was already set to hopefully run the motor at the target AFR's of high 10's to low 11's I was shooting for. I've run 26psi on my own setup but albeit with a smaller turbocharger. Plus, on mine, I'd worked my way up. In Luke's case, the boost was shooting up nearly 9psi in a split second. It could be that the stock Turbo II ignition (along with the MSD 6A on the leadings) can't light that charge, especially with that amount of alcohol. I honestly don't know; haven't done it yet. My theory is that the spark was blown out due to the mixture being mega-overrichened ( by both the EFI system as well as the Alkycontrol system since it uses the same MAP sensor 5V output for load-sensing). The blown-out spark caused a "nothing" or two event to occur in the engine, shaking the hell out of it, stressing the housings and causing the crack.



I believe the boost spike (and general boost instability) was caused by the boost controller's solenoid being overloaded from the volume of air the turbo was able to output. It concerns me and compels me to question the efficacy of your average boost controller's solenoid that's sold for cars that conceivably run smaller turbochargers than ours. The solenoid might be able to handle x psi of pressure, but volume of air is a whole different game. I think the solenoid couldn't vent the excess volume of air out properly, which created a chain-reaction with the solenoid, the controller electronics, and the turbo itself.



I'll post up more pictures as well as the final run's datalog here shortly.



B

BDC

Datalog pictures



B

BDC

Let's try that again!

j9fd3s

well since the failures seem to be the same between this engine and yours, i think its the same cause. so i'm thinking the focus can be more on how to adress the problem more than why it broke.



so assuming again, that its a full engine misfire...



1. how do you catch it earlier? maxt is right, if you hook the car up to a 5 gas analyser (smog machine), a misfire will show up as a BIG spike in hc's (full chamber of air fuel mix goes out the tailpipe), i think NOx goes down. not sure how you'd catch it at 6000rpms though



2. why is it misfiring? weak spark? is it getting blown out? too rich? too lean?





solutions:



1. raise your price for tuning, and sublet steve kan, that way if it fails, you can blame the tuner



2. more spark! i dunno what you need here, zeus?



3. possible theres a mechanical solution, maybe studs, or a really thick oil pan baffle plate is in order at this power level. imo this is worth a look, but i wouldnt count on this to save my bacon

Cornfed

Hey Brian, why would the injector duty cycle drop to 49% when the engine reached 28.7psi? do you not think that would cause a lean condition for that split second? It just seemed weird to me that it would drop and not raise.

j9fd3s

thats weird, cause the actual MS goes up

BDC

I suspect lag time on the sampling rate of the datalogger, myself. Notice it's only 11 times per second. I've seen that same sort of thing before on other datalogs. The 49% DC is probably closer to the drop in boost when it went back down to the 17.5psi range.



B

BDC

Even if it were a split-second lean condition (however that would happen), it wouldn't pop a motor. Lean conditions take time to generate heat in the chamber to then present the possibility of a failure due to a type of pre-ignition. That's not what happened here based on the failure mode.



B

Maxt

That 16,000 rpm blip is ignition back feed from a grounding/shielding problem in the haltech harness.. I experienced that a long time ago, and finally solved it by grounding the shielding of the CAS at the Cas end of the harness.. All sorts of people told me that it would cause more issues having double grounding but it was the solution to the 16,000 rpm blips, and to this day I run the haltechs that way with never a problem.. The blip causes a misfire because it triggers the ignition cut..

The misfire is brutal to a motor, when its at full load, misfire in a rotary is far different than misfire in a piston motor..

2 things happen, first the torque of the motor drops off then go's higher considerably than it was under regular combustion, this rapid down/up is what cracks the plates, second and the cause of the torque spike, is the pressure of the combustion events following the misfire on the same rotor is in the neighbourhood of 30-40 % higher than the normal combustion event, due to a couple of factors, the misfire creates little energy so the misfired face sees reduced back pressure, and the face ahead of sees less charge dilution on overlap, the misfire face also has reduced high pressure re expansion so its VE is considerable higher, those spiked pressures is the spring killer..

Its funny though at lower rpms we use those misfire pressure gains to our advantage, to spool the turbo a little faster, the big bridgeport spool mystery is actually a misfire pressure spike.. Watch any turbo on a bridge motor, and watch the compressor wheel kick after every brap which is a misfire the combustion events following have much more exhaust pressure than static normal combustion, the overlap inducing the misfire has more effect than the actual added overlap at idle, low rpms.. The way to make it work for you without grenading the motor such as 2 step brake boosting is to retard the timing so the combustion pressure doesnt spike like it does with power timing..

I have some poor quality scanned mazda pressure charts showing misfire combustion pressures but they pretty much illegible, next time I am in Japan I will take actual digitals of the pages, but the pressure spike is impressive to say the least..