Beautifully clean work. In inspiration to me for sure. I will be following closely.

I made some good progress the past few days.

*Supermiata crossflow radiator is installed

https://www.miataturbo.net/attachmen...ine=1421514377IMG_5158.jpg by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1421514377IMG_5154.jpg by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1421514377IMG_5155.jpg by Ryephile, on Flickr

*Radiator upper hose made from 3 pieces of silicone from Pegasus Racing and a section from the Escalade OEM hose. 1 3/8 to 1 1/4 45 degree from the re-route Kia neck, followed by two 1 1/4 45 degrees down beneath the intake manifold and meeting up with the S section from the Escalade hose going to the radiator. Two 2" aluminum joiners and a 4" joiner needed. The entire hose length doesn't hit or rub anything, even accounting for engine movement and harmonic vibrations, something most aftermarket shops can't even understand. One of you tuning shops should be making this hose a 1 piece silicone product and sell the crap out of them, duh.

https://www.miataturbo.net/attachmen...ine=1421514377RYE_9505.jpg by Ryephile, on Flickr

*Setrab 619 oil cooler mounted where A/C fan used to be. We'll see how the oil temps play out, but with Michigan's mild climate and my modest power goals it should be adequate.

https://www.miataturbo.net/attachmen...ine=1421514377IMG_5150.jpg by Ryephile, on Flickr

*Oil cooler lines built from Aeroquip socketless fittings and matching -10 hose. Oetiker clamps used as redundant insurance. The hose assemblies are surprisingly light.

https://www.miataturbo.net/attachmen...ine=1421514377IMG_5167.jpg by Ryephile, on Flickr

I used some sitting-around aluminum stock and leftover stock Miata hardware to make mounting brackets for the oil cooler.

https://www.miataturbo.net/attachmen...ine=1421514377IMG_5176.jpg by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1421514377IMG_5175.jpg by Ryephile, on Flickr

^^^Here you can see the shim I used to space the IM brace away from the block to account for the Hondata phenolic gasket. Everything just clears, you have to install the oil fittings before installing the brace. Below you can also see a portion of the upper radiator hose. It's about an inch above the oil lines at its closest point.

https://www.miataturbo.net/attachmen...ine=1421514377IMG_5184.jpg by Ryephile, on Flickr

...and a forward looking engine bay shot, showing the radiator, fan, oil cooler, and shiny red RB sway bar.

https://www.miataturbo.net/attachmen...ine=1421514377RYE_9510.jpg by Ryephile, on Flickr


I have to do a once over/double check to make sure all my lines and fittings are clearanced and snugged properly, and then I can add fluids to the car and begin the 1st start-up checklist. First crank is very close and I'm hoping Dimitris got more right with my base file than wrong.

Everything look clean and well done. But I have a question, wouldn't it be better to install the oil cooler rad in front of the water radiator to have it get colder air through it?

Go peek in Jeffbucc's build thread for an explanation as to why you're wrong. Just had this conversation within the last week or so.

I'll check it out thanks!

The Cliffs Notes is essentially, put the heat exchanger you want to be coldest up front, then warmer, then warmest.

Intercooler up front [closest to ambient]
then Radiator [180-200F]
finally oil cooler [230-ish F is fine]


I still have all my aero ducting to do. I'll be sure to post that when I finish it up. None of the theory matters if all your available airflow blows around the heat exchangers.

So, I'm getting real close to cranking it. The crappy old battery that's still in the car went flat so it's back on the battery tender.

*Brake fluid in via Motive pressure-bleeder with Castrol SRF
*clutch fluid in via Motive pressure-bleeder. BTW, the whole car only took about 3/4 liter, most of which is in the brake reservoir and my drain pan. Good old m.net/garage indicates 2 liters for the car, which is waaaay off.
*Coolant added via vaccum lift. About 1.3 gallons-ish, as I'm sure there was still some coolant in the block and heater core. 3 quarts distilled, 2 quarts HOAT, and half a bottle of AMSoil coolant boost. I don't quite need 50/50 for my climate, as record lows are about -15F and my garage is always warmer than that from ground warming.
*AMSoil 10w-30 Dominator oil added. 4 quarts for now. In theory I'll have to add about 0.1 quarts once it fills the oil cooler and lines and turbo feed.

So far no leaks. Filling the coolant with the vacuum lift allows you to hear for potential leaks, and I didn't hear any.



I compared my old MS1 cranking and warmup settings to the MS3-Basic, and wow they are totally different. Since I know the car starts and ASE on the MS1 tune, I changed those figures.

I was also able to test-fire my injectors individually, proving the sequential wiring is correct. When the battery is charged then I'll do the RPM sync test, then spark test, and finally try to start it.

I got the car started last night! :party:

I obviously have some cranking pulsewidth and ASE curve to tune, but otherwise it's in the realm of functional.

My HLA's are horrendously noisy after the engine's been sitting since late October. The couple minutes I had the engine running (only got the coolant to 122°F before calling it a night] they didn't shut up. They sound like direct injectors. I'm going to take the valve cover off and double check my timing belt and cam sprockets, and I'm considering taking the cams out and rebuilding the HLA's.

Thoughts on that?

Congrats on getting your car started (that's half the battle)!

No tips on rebuilding HLA's ('99 w/ solid lifters here).

Thank you very much!

I ran the car some more tonight. More cranking pulse width and ASE helped start it quicker and keep it running, but there's more start-up tuning to go for sure. I let auto-tune work on my warmup enrichment, and based on how much it wanted to push it upward even close to fully warm it seems that the VE numbers in the idle cells are too low. I'll start by fully warming the coolant, dialing in the VE in the idle cells, and then rescale my ASE and WUE and try it again so the WUE auto-tunes to end up wanting to be at 100% enrichment at fully warm.

I suppose there's a silver lining for doing this in winter. I get plenty of opportunity to nail down my winter cold starts and warm up enrichment.


Oh, and a big plus: the HLA's have massively quieted down. Hooray!

congrats man.

tuning cold start is very time consuming.
you only really get 1 try every morning or at least every dozen hours or so.

everything else isn't nearly as time consuming.

I got through my 2nd full heat cycle. Turns out I forgot to fully tighten the 1/8" NPT plug in the block where the OEM oil pressure sensor was, so it was just barely weeping. I had to pull the oil sandwich plate stackup to get access to it, but I found out that pulling them does not drain the oil cooler, which is nice. It's all sorted out now, better than ever.

VE cell at idle and neutral revving to 3k RPM are now about nailed. From there I reverse inferred the ASE and WUE by keeping the EGO correction off and just watching the gauge, and the car is starting quicker and better.

I also accidentally verified my rev limiter settings. I temporarily set it to 3k just to see how it behaved and check its hysteresis. I forgot I had the override TPS at 90%, so in an amusement to emulate anti-lag I ended up giving the engine a 6600 RPM WOT revving. it ended up making 8.9 PSIg just by revving, which I find very encouraging for low lag. It revved from 3700 to 6300 RPM in only 0.3 seconds before I realized what I did and lifted the throttle. I'm running totally untuned X-Tau AE right now and it's already a million times better than the MS1.

...then it died because I think my fuel-cut over-run and idle settings are wrong. More to tune obviously. Here's those few seconds on the MegaLog viewer:

https://www.miataturbo.net/attachmen...ine=1422332091Screen Shot 2015-01-26 at 11.11.11 PM.jpg by Ryephile, on Flickr

...and my finished engine bay circa this morning with terrible lighting.

https://www.miataturbo.net/attachmen...ine=1422332091IMG_5217.jpg by Ryephile, on Flickr

Stupid clean.

yep, love how pretty that engine bay looks. very nice

That's it. THAT'S IT! I'm getting rid of power steering and A/C. It's going bye bye.

Beautiful work.

Ok, seat and harness time!

*Driver side: Sparco Rev, modified Sparco steel mounts, modified AWR rail, Schroth Hybrid II HANS 6-point harness [wrap around for shoulders, eye-bolts for lap and sub belts]
*Passenger side: Sparco Sprint, modified Sparco steel mounts, AWR rail, Schroth Profi II HANS 6-point harness [same mounting methods]

The Rev and Hybrid II combo is fantastic for my 5'9" 135# build. The Profi sub belt t-bar digs into my inner thighs in longer duration seatings, which is not the most comfortable. The passenger side obviously gets the lower rent and/or more universal accommodations while still abiding to the stricter clubs' HPDE safety rules.

The biggest challenge was to get my seat to sit low enough for me to pass the broomstick test. This was a surprisingly big problem given my average height. I had initially bought Kazespec mounts and they placed my head higher than stock, which is obviously a problem. Doing further test sittings, I needed the seat to be at most 1/4" off the floorpan, if not on the floorpan itself. I took the AWR mounts and cut them apart to place the seat bottom literally on the carpet. I also fabbed 1/8" steel backing plates so the seat mount is sandwiched around the floorpan, creating a much stronger joint.

The sub-belt anchors are sharing hard points with the FM butterfly brace frame rails, using Schroth backing plates.
The lap belt anchors are the factory lap belt hard points with eye-hooks.

All this should be fairly legit in terms of strength and geometry. I strictly abided by Schroth's geometry recommendations for all belt angles and load vectors, as much as the seats would allow.

Bottom line: I pass the broomstick test, and the seat doesn't hit the door because it's perfectly centered in front of the steering wheel.

https://www.miataturbo.net/attachmen...ine=1422849604IMG_5237.jpg by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1422849604IMG_5238.jpg by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1422849604IMG_5253.jpg by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1422849604IMG_5255.jpg by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1422849604IMG_5256.jpg by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1422849604IMG_5262.jpg by Ryephile, on Flickr


https://www.miataturbo.net/attachmen...ine=1422849604IMG_5245.jpg by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1422849604IMG_5242.jpg by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1422849604IMG_5268.jpg by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1422849604IMG_5271.jpg by Ryephile, on Flickr

Cheers,
Ryan

The Sparco's are much more comfortable for ECU tuning than sitting on the floorpan. :bigtu: It's also worth mentioning the Schroth Hybrid belts are suuuper comfy, well worth the moolah over the Profi sub-belt T-bar.

I've spent some time tuning my cranking, ASE, and warmup. I also discovered the magic of batch-fire at cranking and idle advance. Take a look at my not-even-close tune. Only about 10 VE cells are tuned, my ignition map is inferred from my MS1 setup, and I haven't yet touched the AE. It starts pretty good at 10°F and starts solid above freezing. My main problem was idle @ operating temp was oscillating, but the idle advance, even roughed in, has dramatically stabilized idle.

Cheers,
Ryan

A few things got done:
*carbon-fiber spliced these small ABS NACA ducts to the GV lip apertures for the 2.5" brake cooler hose (sorry, no pictures yet, but it's functional, not gorgeous)
*carbon-fiber upper radiator panel made [see below picture] from 3 plies of 3k plain weave. 80 grams is 40 grams lighter than the ugly plastic one the previous owner made, and the new one seals the holes much better too.
*Put the fenders, bumper, and wheels back on.
*Took the car for its maiden drive! :bigtu:

https://www.miataturbo.net/attachmen...ine=1424817613Miata's first drive by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1424817613Miata's first drive by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1424817613Miata's first drive by Ryephile, on Flickr

Overall it was a big success, drove the car for about 20 miles total. I auto-tuned most of the <100KPa and <3.5k RPM VE cells. The coolant thermostat works great per my datalog, the coolant temp cycles around 180°. Rev's base map is heavily biased rich in the boosted cells, to the tune of 30% or so. Even so, the untuned X-Tau is a revelation for throttle response. It seems easy to get a few pounds of boost below 3k RPM, so it's looking good for a nice low RPM boost threshold. I spent very little time in boost on this first time out and my peak TPS was ~50%.

It appears I need to re-bleed the brakes as it feels like there's still some air in there, but the totally new pads and rotors certainly confuse that feel. Oh, and the cold snap we've been having resulted in the [presumably original and heavily clouded] plastic rear window to crack with a *BANG*. So, a new Robbins top is en route.

Tell me more about the Abarth.

:giggle:

Seriously clean work on the Miata sir. Looks fantastic.

Thank you very much.

As a plus, my Robbins top just arrived. Yay!

:hustler: I know you're just kidding, but the Abarth is actually really fun. It's definitely Italian in its quirkiness, but it's been totally reliable, sounds incredible, and is super fun as a DD. I keep test-driving other new cars (e.g. F56 MINI, 86 twins, Mk7 GTI, FoST, etc.) and can't find a satisfactory reason to replace it. For sure all those other cars have better interiors, and most actually handle better, but none of them are nearly as raw fun.

I actually wasn't kidding. The giggle was because of the MT.net tendency to talk more about non-Miata interests. I recently bought a wrecked 2012 500 5-speed just to make a buck. Since I got it fixed and drove it around I've started lusting after an Abarth. They do sound incredible and look like a total riot.

Plus they're dirt cheap used.

Fun stuff.

Did a "rough cut" VE tune this Friday evening as the roads were clean and dry. My friend that helped me is an OEM ECU calibrator, and it's awesome to see a true professional in their element. We did it the correct way; me driving at a designated load and RPM and the tuner focusing on sensor data, map region, and slope calculation with AFR closed loop turned off. The result looks beautiful up to 5psig and redline of 7200 rpm, and is interpolated beyond that on both axes. WOT tuning will happen later. Peak VE appears to be 4500 RPM, which makes sense on the stock cams.

As an aside, the wastegate actuator FM provided was totally wrong, actuating at 15 psig instead of the advertised 6. Obviously that's not going to work on my stock bottom end and intent of EBC.

A few notes:

*clapped-out convertible top is being replaced today with a new Robbins top, yay!
*Discovered my engine setup prefers an injection timing of 300-° in idle region for most stable combustion. More advanced or retarded than that resulted in subtle misfiring which ended up making the idle oscillate through varying AFRs.
*Dialed in the new wastegate, which is a custom Kinugawa with a 0.3bar spring and homemade stainless actuator rod. Figured out base boost, and noted inherent creep, which was minimal from 6k to 7200 rpm. Then dialed in open-loop boost using 4th and 5th gear pulls until we started chasing our tails between percentages. Then clicked-on Closed-loop boost and it was almost magical how well it worked right out of the gate. Dialed in the sensitivity and it's now good to go.

Boost threshold for 180kpa on the road is 3k rpm, which is acceptable. It's a seemingly casual ramp from 2k to 3k, but we also haven't done any ignition timing tuning at this point, it's still inferred from my MS1/Greddy setup. Initial knock datalogs imply zero knock, which makes sense given the conservative ignition map.

Dyna-Pack at MonkeyWrench reserved for April 3rd. That'll be the final tuning where we'll figure out thermal capacity, dial in MBT in the more important regions of the map, keep out of knock, keep the torque below a rational amount, and dial in the XTau (which is already about perfect).

Pics: One of the 1" longer wastegate actuator rod I fabbed out of 303 stainless, and the rest from the new top install today. Shelby Trim did a great job, proving their reputation. This is the Robbins #2905 1-piece with zipperless vinyl window, so it'd be compatible with my roll-bar.

https://www.miataturbo.net/attachmen...ine=1427161603IMG_5408.jpg by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1427161603IMG_5427.jpg by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1427161603IMG_5428.jpg by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1427161603IMG_5434.jpg by Ryephile, on Flickr

As a sanity check I re-re-checked my base timing. Ignition Options / Wheel Decoder settings, Miata 36-2, 0 deg offset, and Falling Edge.

https://www.miataturbo.net/attachmen...ine=1427426159IMG_5438.jpg by Ryephile, on Flickr

*Set MS3 to fixed 10 deg timing
*Observe timing marks on ATI damper and timing belt cover
*Surprise surprise, 10 degrees on ATI is TDC on cover, and 10 degrees on cover is TDC on ATI, spot on.

For fun, I changed it to the Rising edge and +5 deg, and BAM the timing was 10 degrees retarded. TDC at TDC. Obviously that's not 10 degrees, so that setting combo is wrong for my setup.

I went further, reinstated my timing map and correct trigger settings, and took a terrible potato video of the movement. Video recorders and timing lights aren't too compatible for clean refresh. Nevertheless, the engine timing is perfect, exactly what my timing map indicates.

I'm satisfied that my settings are correct. Obviously yours will vary, since you don't have my hardware configuration.


-->Oh, you can also kinda see the new Kinugawa wastegate actuator, which is very nice and actually works at advertised pressure.

Hey Ryan,

Cool to have a pro working on your tune, are you willing to share a copy?

Also great to see someone playing with injector timing.

Cheers

Hey just wanted to thank you on the awesome documentation. Helped me loads. Subbed for updates & love the car.

Thanks for the compliments. :) As the tune is cleaned up I'll share more information. Here's a sneak peak though. This is the roughed-in VE map in 3D. The car runs within a small percentage of my AFR targets in open loop, so even in closed-loop the EGO correction is never having to adjust much. The units have been left out intentionally.

https://www.miataturbo.net/attachmen...ine=1427940520Screen Shot 2015-04-01 at 9.59.37 PM.jpg by Ryephile, on Flickr


Unfortunately the car just did a cruel April Fools prank on me tonight during some EBC adjustments and the clutch is slipping. It simply can't hold the extra torque anymore. It was problem-free over the weekend, but alas its apparent borrowed time has expired. I'm now shopping for a new clutch that doesn't suck. The dyno tuning will have to wait. :(

That VE map shocks me a bit. I have never seen such a smooth VE map before...

Torque transfer mechanism components ordered. ACT ZM2-HDG6 ordered from Andrew at TSE, ACT Prolite flywheel via Amazon Prime (!), and seals-n-stuff from Arlington Mazda.

FYI: Flyin Miata is currently out of stock on their Level 1 Happy Meal, ETA June-ish.

I'd hazard to say it's because most people just solo road-tune VEAL their table, which usually results in a coarse profile because they're not carefully pin-pointing individual map cells. For sure when I did it that way with my MS1 + VEAL, it was semi-effective but definitely not smooth.

Trans and clutch removed. It was more glazed than a confectionery.

Anyone known exactly what this old clutch is? PP has the Daikin logo stamped in it, and the disc says Exedy.

Old busted clutch & flywheel setup = 17 pounds
ACT HDG6 + Prolite flywheel = 21 pounds


Old Busted
https://www.miataturbo.net/attachmen...ine=1428429816Clutch R&amp;R with ACT HD 6-puck by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1428429816Clutch R&amp;R with ACT HD 6-puck by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1428429816Clutch R&amp;R with ACT HD 6-puck by Ryephile, on Flickr


New Hotness
https://www.miataturbo.net/attachmen...ine=1428429816Clutch R&amp;R with ACT HD 6-puck by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1428429816Clutch R&amp;R with ACT HD 6-puck by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1428429816Clutch R&amp;R with ACT HD 6-puck by Ryephile, on Flickr

^^^Note the balance weight at the bottom of the pressure plate.


Seals for the rear main [which was leaking] and the transmission input and output just arrived. I should make some progress tonight, but I'm also going to remove the rear diff to replace the bushings, so that'll take time.

i like to use a little blow torch action to the eyelets of the diff carrier. Not all that much heat. you can just heat it up enough to hear it sizzle just barely which seems like enough to break the bond. Get a tube/pipe what have you that's the proper OD to line up with the metal sleeve the bushing sits in and give it 2-4 good whacks a 5lb sledge, comes out like buttah.

love the build as always.

yeah, a shitty 1.6L PP with a stock disc isn't going to hold much boost.

*Rear main crank seal replaced
*Trans input and output seals replaced
*New shift fork and rubber booty

Hulked the transmission back onto the car after very lightly lubing the throwout bearing-shift fork contact points along with the input shaft splines and throwout bearing sliding surface. Bolted up the clutch slave and the clutch pedal works great. Medium-Light effort, very nice "knee" point in the middle of the pedal travel.

Tonight I'll button up the car and then start putting on some break-in miles so I can finish the EBC tuning.

Brought the car to MOTD'15. The powertrain and the tune did very well. A few notes:

*Despite the R-pack tie-rod ends, the car still has notable bump-steer. This will have to be sorted out.
*The fixed alignment transformed the handling for the better, but there's room for improvement in transitions. The balance was quite neutral; power oversteer is available, trail-braking allows subtle rotation, yet in constant speed invokes mild understeer.

*AMSoil GL-4 for the Miata 5-speed transmission is terrible. Between the Motorcraft unicorn tears, AMSoil GL-4, and Redline MTL, the Redline shifts quickest and the AMSoil is simply incompatible with the syncros. This was painfully evident on the Dragon where you're doing hundreds of shifts per run.
*The ACT 6-puck certainly holds the torque, but it's graceless in 1st and reverse. Once underway however it's transparent and feels good.

*The brake balance needs work. With the Wilwood prop valve full-out, the TSE 11.75 + FM 10.9 still has too much front bias. I'll have to work on this.

*2 miles from home the lower right front caliper bolt fell out, causing the caliper lower outboard bleeder screw to lathe a concentric ring in the wheel. I'll have to investigate whether the TSE bracket is an improper [spongy] alloy choice, poor thread cutting, or an inappropriate bolt choice. To be fair, the Dragon is much more demanding on brakes than any racetrack, save perhaps the Nurburgring. 636 threshold braking zones in 26-ish minutes is many times the cyclic abuse per time unit on most racetracks. I've still never had this kind of failure.

*The GarageStar water pump pulley is very good quality and finally stopped the cold-start belt squeal that the salvage OEM pulley appeared to be causing.

*Actual 29.7 to 30.9 MPG on the I-75 drive to and from the Dragon. This is ~10% better than the same road with the GReddy/MS1 setup and also much better than the bone-stock '95 I had back in the day.

https://www.miataturbo.net/attachmen...ine=1430772404MOTD&#x27;15! by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1430772404MOTD&#x27;15! by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1430772404MOTD&#x27;15! by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1430772404MOTD&#x27;15! by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1430772404MOTD&#x27;15! by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1430772404MOTD&#x27;15! by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1430772404MOTD&#x27;15! by Ryephile, on Flickr

Are the other seven bolts used in the exact same environment loose as well? Could you have left the one bolt loose? It's worth mentioning that in 5+ years, we've never had anyone else complain about bolts working themselves loose from the brackets.

Full-out is max brakes to the front. If you want max rear brakes, you need to turn the valve full-clockwise (full-in).

I'll check the other bolt torques tonight, however during assembly I triple-checked torque, since it's a safety-critical item. I don't bet my life on shoddy mechanic work. The Dragon is not easy on parts, so I'm thankful it survived at all. The fact the bolt backed out completely implies an unintended elasticity to either the bracket or bolt material.


That would explain the prop valve then. I'll have to remember it as looser = less [rear] instead of righty-tighty. Here are the directions I should've read.

Just checked the current torques of the remaining 7 bolts:
*All 4 knuckle-to-bracket maintained torque
*Upper 2 caliper-to-bracket 1/4 turn loose
*Remaining lower caliper-to-bracket 1 turn loose, not even finger tight.

This is very interesting. It means that for some reason the caliper-to-bracket bolts aren't holding their torque, and the leading edge [lower] bolts are loosening significantly compared to the uppers.

Based on this symmetrical torque failure, it seems logical that either the brackets and/or bolts I received from TSE are unique in their failure mode or there is a design problem here. Surely more to come.

What did you torque them to? Like I said, we never, ever have this problem, and neither does anyone else. This is unique to you, not our kit.

60 LbFt, which is on the high side for Grade 8 of that size. You did not provide any recommended torque settings in the included printed instructions, so I went off good old charts.

The bolts you provided are different between the caliper and knuckle mountings, so I'm not ruling out that the knuckle bolts are acceptable [given they're holding torque] but the caliper bolts are faulty [since they don't hold torque]. It can be deduced the threads cut into the bracket are acceptable, assuming we're not running into a differential thermal expansion coefficient problem.

I suggest you check your torques. It's your safety, and that's more important than your pride. Just because nobody else has reported it doesn't mean it hasn't happened and they fixed it with band-aids like over-torquing, loctite red, or a real fix like adding a lock-nut.

News to me. It should have been 8 identical bolts. Email me some photos of the hardware you have?

Some cool pictures from Killboy.com from last weekend.

https://www.miataturbo.net/attachmen...ine=1431056315737780.jpg by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1431056315737779.jpg by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1431056315737783.jpg by Ryephile, on Flickr

Busy day.

Went here:

https://www.miataturbo.net/attachmen...ine=1432090010RYE_0193.jpg by Ryephile, on Flickr

232 wHP and 208 LbFt at 10 psi on 93 octane. Perfect for keeping the con-rods and 5-speed in the car in functioning condition. Obviously the boost threshold isn't accurately demonstrated in the plot; on the street in 5th gear target boost occurs at 3.0k RPM. The dip at 4300 on the plot is an artifact of the current boost controller setup, not a VE dip.

Discovered some interesting stuff about injection timing, picked up 10% torque for a notable region of part throttle. Perfected the VE table, did some ignition timing investigation, spot-checked the X-Tau [working perfectly], and did some thermal capacity checks [quite impressive].

Discovered the SwainTech White Lightning is so opaque you can't see the manifold glow red even during sustained peak torque. The closest proximity item to the manifold is the heater hose, and that showed a post-pull IR reading of 170F, which isn't even coolant temp. Impressive.

https://www.miataturbo.net/attachmen...ine=1432090010RYE_0225.jpg by Ryephile, on Flickr


Then went to Autoblog's Miata ND reveal on Woodward. A surprising number of Miatae showed up. I was one of three boosted; the other two were both NC's. The demographic was as-expected; hipsters being supes fab and fat old white men pretending they still fit in the car.

https://www.miataturbo.net/attachmen...ine=1432090010RYE_0195.jpg by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1432090010RYE_0213.jpg by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1432090010RYE_0196.jpg by Ryephile, on Flickr

Panel fitment not so hot
https://www.miataturbo.net/attachmen...ine=1432090010RYE_0204.jpg by Ryephile, on Flickr

Learned from a Mazda engineer that the Skyactiv runs Atkinson cycle only for some of part-load, then uses it's 70 degrees of intake cam phasing to run Otto at WOT. He alleged they were able to run de-throttled during Atkinson, which sounds fishy to me without any valve-lift capability.

https://www.miataturbo.net/attachmen...ine=1432090010RYE_0199.jpg by Ryephile, on Flickr

Here's a fun data excerpt from doing fixed RPM load-step testing. You can see the transitions are clean and precise. It's also a good test of thermal capacity, depending on how long you actually hold the higher loads. We were trying to maximize the Dyna-Pack's capabilities at this point while also verifying both the base VE map and the X-Tau settings made sense versus the AFR target.

https://www.miataturbo.net/attachmen...ine=1432151215graph.jpg by Ryephile, on Flickr

So keen to see your injector timing, VE and ignition maps! :)

Doubt he'll share them, since he went through all the effort of removing actual numbers from the map he posted earlier.

But can you or your tuner PLEASE chime in on at least how you went about dialing in injector timing?

This seems to be a mysterious topic around these parts, pretty much no one seems to know or comment on it.

I attempted to question the difference in injector timing maps that came with my verious megasquirts, since they all seemed very different with no rhyme or reason to the differences, and got nowhere with that.

I'll post the 3-D shapes soon, but it's important to look at the bigger picture. The numbers we settled on aren't that important unless you plan on building the exact same hardware setup. If that's the case, PM me and I'm happy to help.

The final VE map looks very similar to the one posted prior in terms of shaping. The slope through load was adjusted slightly, and also the curve shape after VE peak. This is where spending the time on a fixed-speed dyno and doing load sweeps at set RPM intervals is very helpful in pin-pointing exactly how your VE table is functioning, leaving O2 feedback and AE out of the equation is important, not trivial.

The full results of the Injection timing experiment I won't share at this point, as we discovered what appears to be a bug in the firmware, so the implementation is not as ideal as preferred. Basically, the MS3 has an issue when doing certain injection timing transitions. For now, we're running 300° BTDC for the entire map, which is not ideal for certain loads, but avoiding the firmware transition error is more important than sticking to ideals. In the end, the car has to drive perfect, so there are compromises you make to achieve that smoothness.

To elaborate further on the test itself, we chose 500 RPM intervals and held specific throttle angles, then swept injection timing and observed measured torque output. This isn't perfect as you also have to be mindful of MAP changes with torque output. At higher loads and RPMs, there was no change [WOT, for example], but at lower loads and RPMs, we were able to measure up to a 10% change in torque output at the same MAP, which is significantly more than we anticipated. Basically, an idealized injection timing map will improve lower load and RPM BSFC.



If I had 4 or 5 more engines to dispose of, then we would push the envelope further on things like AFR, ignition timing, and aggressive knock sensor tuning and feedback. As it stands now, we feel the setup will keep the reliability high, which is the primary goal.

Thanks for the explanation, and no worries about not posting numbers, I know they're next to useless on other cars anyway just like you said.

My car runs flawlessly, but is still on the injector timing map that DIYautotune shipped me the ecu with, and that is the only part of my tune that I have no idea what it does/should do, hence my curiosity.

I'm really liking the approach you guys are taking with the tune (good job on the dyno numbers btw) so I'm "paying attention" over here :)

Interesting info Re: 10% torque increase as well as bugs. In for resolution.

Also makes sense that Inj timing would make less difference at high loads, especially for those of us that will be running close to 100% open at max load / RPM.

I definitely appreciate the compliment. For arghx7 it's been a fun diversion from his OEM work, and I'm thankful he's willing to share his incredible wisdom. We wrote out a specific game-plan prior to going to the dyno, so our tests were methodical and organized. The tune foundation was sorted first [fuel], then ignition timing, then injection timing, then X-Tau. All the data-logs and tune revisions are labeled so we can look back on them and understand what they're for.


I'll say that if you're happy with how the car drives, that's the most important facet. It's quite difficult to do any sort of injection timing tuning on the street, and in the big picture it's port injection, so it's a fairly small adjustment anyway.


In a few weeks I'll have one more major update for the car! :party:

A few brief comments:

Only loading dynos actually measure torque. So you can't hold speed and load (sort-of, by throttle angle) and sweep parameters if you are using a basic dynojet, as they measure acceleration during transient pulls and back calculate torque and power.

The injection timing we are using is fixed end-of-injection, which can be set under the sequential fuel injection settings. That means the start of injection is constantly changing according to the duty cycle.

Remember that duty cycle = 720 crank angle degrees - injection duration in crank angle degrees. If I'm at 50% duty, my injection duration should be ~360 degrees, with start of injection back calculated from the end of injection time. That's assuming steady state operation with no X-tau etc (that stuff complicates it a lot). If I set a fixed value to an entire map, the injection timing is still changing all the time. If I have an EOI of 300 degrees BTDC firing and 20% duty, my duration is 144 degrees, so my SOI is 444 degrees BTDC.

300 deg btdc results in the injection event ending while the intake valve is still opening ("open valve injection"). You can sweep EOI earlier and later and see what happens. Without an emissions bench (to look at mixing according to HC ppm, CO%, and O2%) or combustion indication (to look at IMEP, combustion stability, etc), you can can only go on subjective assessement (did the engine run smoother at idle) or measured torque on a loading dyno.

The shape of the VE table is important, and in a way more important than the numbers themselves. If you are looking horizontally across the map (as rpm changes), it should roughly follow the VE curve according to intake valve closing timing. Later intake valve closing timing = VE peak later. With a fixed cam engine like this it's easy. With variable valve timing (cam phasing or lift change) the complexity increases drastically.

If you are looking vertically across the table, the shape of the curve is proportional to the airflow addded with boost (1st derivative of the curve), with enrichment baked into the rate of change between cells (2nd derivative of the curve).

As for spark timing, the closer you are to where it needs to be, the less effect the changes will make. If you sweep timing a few degrees and only see a couple hp difference, within test-to-test variation, then you're timing is in the right range. It's probably not adviseable to advance it just to squeeze a few hp out of it.

For boost control, the mechanical characteristics of the system need to be right. You have to resolve spring pressure, solenoid plumbing, preload of internal wastegate, etc. You need that good foundation before you set duty cycle tables, controller gains, and (hopefully if MS adds the code) air temp and baro compensation to the feedforward portion.

^^^Thanks for chiming in!


In other news, Autoblog posted a little video from the ND reveal on Woodward, and at 3:06 they did a shot of my Miata's engine bay. Sweet. :)

One of the last "install" updates, and it's a big one. Anze Suspension bespoke Penske setup. These guys are 1st class to deal with. Their wisdom and experience is abundant and their professionalism is top-notch. While I've been familiar with Anze for well over a decade, it was my first actual project with them and I'm very pleased to be part of their family now.

https://www.miataturbo.net/attachmen...ine=1433910402RYE_0235.jpg by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1433910402RYE_0244.jpg by Ryephile, on Flickr

The dampers are based on the 7500 DA's but with proprietary Anze valves and valving. Stroke and positioning is to my spec. Total stroke is more than sufficient for the new spring rates, which rely on Anze and V2 Motorsports's huge knowledge database. Upper mounting hats are billet aluminum with serviceable spherical bearings.

Even the mounting hardware is jewelry:
https://www.miataturbo.net/attachmen...ine=1433910402RYE_0246.jpg by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1433910402IMG_5743.jpg by Ryephile, on Flickr

https://www.miataturbo.net/attachmen...ine=1433910402IMG_5755.jpg by Ryephile, on Flickr

The car is roughed in for corner balance at the moment, but it's clear the Penske's are worth every penny. They ride smooth, have fantastic quick weight transfer, absorb classic Detroit "roads", and stay composed at all times just like a legit quality damper should. It'll get setup fine tuned next week, and then finally it'll be track time! :)

https://www.miataturbo.net/attachmen...ine=1433910402IMG_5792.jpg by Ryephile, on Flickr

:burncash::burncash::burncash::burncash:












:burncash::burncash::burncash:


Nice dampers

The car spent a few days at the race-shop spa.

*ES poly bushings in all suspension arms
*ES poly diff bushings, the hard way. Pressing out the stock bushings fractured the housing, so I now have a very pretty brand new diff housing in the car.
*Corner balanced and aligned. Going with a hybrid concept, blending V2's SM experience with my half-street half-track no rule-book application. We'll see how the first track day goes, lots of trailblazing to sort through yet.

I was surprised and/or relieved to see the car isn't a total armored tank on the scales. 2,218 lbs without me but with >1/2 tank of fuel. That's only 2 pounds more than the advertised weight from the 1992 Miata brochure. Not bad for 10 pounds of shit in a 5 pound bag.

I've been missing out on this thread. Nice work!

I'm putting my 1990 Silverstone back together now. Awesome color. ;)

Cool project man. Nice job.

Can I ask the reason your using x-Tau over EAE?

I can't tune either for shit, I was just wondering why you went for x-tau. I was under the -possibly misguided - impression that EAE is a superior wall wetting algorithm.

It appears the MS guys wrote EAE without fully understanding why it's used [originally as a hydrocarbon emission countermeasure] and it ends up being a tip-in band-aid with a steady-state calculation, whereas looking at their x-tau presentation, the implementation is much closer to a model based delta MAP.

In any case, whatever ends up working for your setup is what's important. X-Tau works brilliant in my application.

If you can't tune either, it's very likely then your VE map isn't correctly tuned. Without that correct no amount of AE tuning will ever work right. You should be able to drive the car with just the VE map and no AFR feedback or AE turned on and it should drive perfect except for bigger tip-ins and tip-outs. Casual tip-in's don't usually need AE.

I will chime in and say that when I look at EAE and X-Tau and compare it to OEM level code I have used or been exposed to, EAE logic is very much like the hydrocarbon countermeasures. You can tell by the way the maps are speed load based, rather than with a time component for transient operation.

When you're trying to pass emissions on a port injected engine you typically use an unburned fuel model, which is more like the EAE. The transient fuel calculation has components based on time. That's an oversimplified answer to a complicated matter. I've got some diagrams I can dig up that might help a little.