Project OldTree: The 12 Days of Rotormas
04-22-13, 09:48 PM
#352
How to Deal
Mom is in a better place now, tending to an amazing garden, combined with an endlessly-stocked kitchen full of entirely organically raised meats, cheeses, vegetables from said garden and tons more.
Today was Earth Day and what I did was bury a deceased groundhog. I found him on the road just off my place's back porch yesterday. Felt bad for the little guy, so I got the snow shovel and moved him to the brick section by the steps so he wouldn't be flattened by a car. Now he's in Mom's Earthbound Garden.
Two days ago was the first full day I had the house to myself, so I got back in the kitchen and made pan-seared organic Ribeye steak in soy sauce with shells, cheese & tomatoes (Mom's signature dish and my favorite!!!). Turned out really good for not cooking on my own in 7 years. Tonight, I was a bit less ambitious and made Hamburger Helper using the drippings leftover from the Ribeye. Pretty good, but a tad on the salty side since I added a little more soy sauce.
Right before Mom went into the hospital, I found some great deals on ebay for what I need to finish up the intake plumbing. 30ft of 5mm silicone vacuum line for $32 shipped, two 2.5" 90-degree silicone couplers for $11 each and a GM-style 3-bar Map sensor with pigtail for $21 shipped.
The Map sensor is slightly different from the placeholder GM branded 1-bar sensor in that the connector nub is shorter in height and the hose barb is thinner (~7mm at flare & 6mm elsewhere compared to 7.8 & 6.4mm for the GM one). I would assume that these small differences are cost-cutting as well as ease-of-installation measures to preserve crusty old vacuum lines & connectors.
So before dinner, I broke out the wiring gear and re-pinned the sensor's pigtail to match Haltech's color scheme (5v is Orange, Sensor wire is Yellow, Ground is Brown), then extended the wires with leftover wire that matches my color scheme (Gray/Red, Yellow & Black/White respectively). Added some Metripack terminals and dropped it into the TMT subharness while the sensor went right into the recycled GM bracket with no fuss.
Upon closer inspection, the vacuum line I got was actually about 5.5mm ID/11mm OD and fit well on 5mm or 3/16" barbed nipples such as those found on my GM solenoids, but isn't as secure as I personally feel comfortable with on the non-barbed 5mm nipples on the UIM. It did have about the same grippiness/friction as the aquarium line, so I'm probably just being paranoid.
The plus side to this glitch is that it fits the 3-bar sensor securely and can work just as well on the 1/4" nipples on the turbo actuators & CRV/BOV. The downside is that it is too large for the Haltech's internal sensor barb and my boost gauge. Since I have a 3-bar Map sensor now, this isn't a problem and I can figure out something clever for the boost gauge later. As for routing and securing the vacuum lines, I've been talking about this for months with those closest to me and have an ingenious way to do it using what I already have on hand. The next update will detail it, so stay tuned.
Today was Earth Day and what I did was bury a deceased groundhog. I found him on the road just off my place's back porch yesterday. Felt bad for the little guy, so I got the snow shovel and moved him to the brick section by the steps so he wouldn't be flattened by a car. Now he's in Mom's Earthbound Garden.
Two days ago was the first full day I had the house to myself, so I got back in the kitchen and made pan-seared organic Ribeye steak in soy sauce with shells, cheese & tomatoes (Mom's signature dish and my favorite!!!). Turned out really good for not cooking on my own in 7 years. Tonight, I was a bit less ambitious and made Hamburger Helper using the drippings leftover from the Ribeye. Pretty good, but a tad on the salty side since I added a little more soy sauce.
Right before Mom went into the hospital, I found some great deals on ebay for what I need to finish up the intake plumbing. 30ft of 5mm silicone vacuum line for $32 shipped, two 2.5" 90-degree silicone couplers for $11 each and a GM-style 3-bar Map sensor with pigtail for $21 shipped.
The Map sensor is slightly different from the placeholder GM branded 1-bar sensor in that the connector nub is shorter in height and the hose barb is thinner (~7mm at flare & 6mm elsewhere compared to 7.8 & 6.4mm for the GM one). I would assume that these small differences are cost-cutting as well as ease-of-installation measures to preserve crusty old vacuum lines & connectors.
So before dinner, I broke out the wiring gear and re-pinned the sensor's pigtail to match Haltech's color scheme (5v is Orange, Sensor wire is Yellow, Ground is Brown), then extended the wires with leftover wire that matches my color scheme (Gray/Red, Yellow & Black/White respectively). Added some Metripack terminals and dropped it into the TMT subharness while the sensor went right into the recycled GM bracket with no fuss.
Upon closer inspection, the vacuum line I got was actually about 5.5mm ID/11mm OD and fit well on 5mm or 3/16" barbed nipples such as those found on my GM solenoids, but isn't as secure as I personally feel comfortable with on the non-barbed 5mm nipples on the UIM. It did have about the same grippiness/friction as the aquarium line, so I'm probably just being paranoid.
The plus side to this glitch is that it fits the 3-bar sensor securely and can work just as well on the 1/4" nipples on the turbo actuators & CRV/BOV. The downside is that it is too large for the Haltech's internal sensor barb and my boost gauge. Since I have a 3-bar Map sensor now, this isn't a problem and I can figure out something clever for the boost gauge later. As for routing and securing the vacuum lines, I've been talking about this for months with those closest to me and have an ingenious way to do it using what I already have on hand. The next update will detail it, so stay tuned.
05-27-13, 12:34 PM
#353
Vacuum Overhaul Part 1
I apologize for the lack of updates during the past few weeks as it has been quite busy and I have not had much motivation to work on the car due to household issues coming up. Most of all, I need to give a few shoutouts to express my gratitude.
GoodFellaFD3S
RenesisFD
DJSeven
RotaryKytten (on Ohio Rotaries)
The above people offered up their own hard-earned cash to help me help myself out of a bind. I got a better job that can afford the house mom & I were splitting the bills on, but wouldn't start untill May 13th. Everyone above chose to lend me a hand so I could meet the timeframe I initially promised the landlord to have the first month's rent and get a new lease in my name as the old one was in mom's name. With that done, I can keep things going thanks to the above four people. I've never met any of y'all in person, but you've got my deepest thanks for giving me a fighting chance. You guys/gals are incredible!!!
Now, back to the build:
Due to the picture limit per post being set at 12, I am having to split this update into multiple posts.
The $21 3-bar Map Sensor was defective, reading 1.85v at atmospheric pressure (~98.2 KPa Absolute) when it should have put out 1.55v. As soon as I switched the Haltech over to use the 3-bar, the engine flooded! Opened the laptop and found the sensor was reporting ~1.2 psi of BOOST instead of 0.9 InHg of Vacuum. Granted, this could have been corrected by re-scaling the sensor's calibration file, but doing so accurately means running the sensor through its entire range, then logging each reading. A new part should not require such when it is advertised as a plug & play replacement for the official GM 3-Bar Sensor. I suggested to the seller that they go through their remaining stock to test each sensor per the Injector Rehab calibration chart and do an exchange for one that works, but they're just giving me a refund instead. Kind of sucks because if they had some good ones, I'd gladly accept and pass on the good info so they'd have more happy rotorhead customers, but oh well. If/when I find a source of good/cheap GM-type sensors, I'll post it.
The vacuum line was a no-go either, just wouldn't stay secure on the non-barbed 5mm nipples. Thankfully, the seller has been excellent in working with me to correct this numerical glitch. They sent me a sample of their 4mm and 6mm hose for free so I could determine what is the proper size I needed. Just like the FD, 4mm is the one I needed. The actual ID is 4.2mm and it stretches sufficiently to fit onto the GM Map sensor (7.8mm at largest part of the barb). Upon hearing that 4mm was the size I needed, they're exchanging the 5mm stuff for 4mm on their dime. Now, if only they had GM Map Sensors...
One thing I despise about most builds with a large number of vacuum lines is how they are secured in place. Most people use Adel Clamps and that's fine for a few lines, but simply isn't up to my standards when facing 7 solenoids with up to 3 lines each. I'm very picky about making things look presentable/serviceable and stacked Adel Clamps just looked poorly executed when facing this many lines. After taking a close look at how Mazda secured all of the hardlines to the firewall (they just pop into C-shaped clips), I found a similar solution from McMaster-Carr. While clean and straightforward, the problem is that it is made of Nylon and would not survive being in close proximity to the turbos. As I could not find a similar part made of more heat-resistant materials as the OE Mazda solution is (steel coated in vinyl for the stock Map sensor & A/T vacuum switch line clip), I found an answer in the basement.
The last time the car was running, it looked like this:
Over the past month, I have been designing and fabricating a new vacuum system. The basis for this was something my mom & I came up with over the winter. Each clip locks solidly in place, you can mix & match any size of lines at the same location, require no tools for removal/reassembly and the clips stay secure on the line so you won't lose them during servicing. As for securing the brackets, I am using existing holes on the car, both threaded and bare, to attach each one to the car. Again, this goes with my philosophy that drilling more holes isn't necessary when a design can be given a few small adjustments to use what is already provided. Mazda has always been fond of reusing parts where logically appropriate to reduce development costs, so I have done the same and taken it a step further. All of the hardware used here is what I already hand on hand, originally purchased at a local dollar store months ago for my workbench but never used.
One of the most important projects was building the solenoid rack. What was involved was cutting the stock GM brackets in half, reversing both pieces, adding 1/2" inbetween them and welding everything back together. While the original unmodified brackets would have worked, the orientation had the connectors at the top and I wanted the vacuum lines to be more prominent than the wiring. The downside to this is that I'll have to redo some of the solenoid subharness.
The Front Line Bracket was challenging to make it work suitably, have sufficient density and be presentable. 16 clips from front to back, then another 8 facing the front.
The passenger side needed some plumbing rework to make some room for the vacuum lines to be run. What I had to do was make the BOV's dump hose, 12AN in size, pull in closer to the engine. Add a 1" Pex Elbow and the stock Cosmo rubber elbow and it tucks in nicely. The 12AN hose was a pain to get onto the Pex elbow due to its reinforcement, so I had to grind down the 'lips' and then give it some dish soap to persuade it. Normal 3/4" rubber hose does not have this problem and will slip right on with no fuss.
Also had to use one adel clamp to secure the rear line guide in place at a second point. It is bolted to the body through an existing hole, but the vertical section was moving around too much for my liking. So one 1/4" adel clamp binds it to the brake line and it is nice and solid.
More in Part 2
GoodFellaFD3S
RenesisFD
DJSeven
RotaryKytten (on Ohio Rotaries)
The above people offered up their own hard-earned cash to help me help myself out of a bind. I got a better job that can afford the house mom & I were splitting the bills on, but wouldn't start untill May 13th. Everyone above chose to lend me a hand so I could meet the timeframe I initially promised the landlord to have the first month's rent and get a new lease in my name as the old one was in mom's name. With that done, I can keep things going thanks to the above four people. I've never met any of y'all in person, but you've got my deepest thanks for giving me a fighting chance. You guys/gals are incredible!!!
Now, back to the build:
Due to the picture limit per post being set at 12, I am having to split this update into multiple posts.
The $21 3-bar Map Sensor was defective, reading 1.85v at atmospheric pressure (~98.2 KPa Absolute) when it should have put out 1.55v. As soon as I switched the Haltech over to use the 3-bar, the engine flooded! Opened the laptop and found the sensor was reporting ~1.2 psi of BOOST instead of 0.9 InHg of Vacuum. Granted, this could have been corrected by re-scaling the sensor's calibration file, but doing so accurately means running the sensor through its entire range, then logging each reading. A new part should not require such when it is advertised as a plug & play replacement for the official GM 3-Bar Sensor. I suggested to the seller that they go through their remaining stock to test each sensor per the Injector Rehab calibration chart and do an exchange for one that works, but they're just giving me a refund instead. Kind of sucks because if they had some good ones, I'd gladly accept and pass on the good info so they'd have more happy rotorhead customers, but oh well. If/when I find a source of good/cheap GM-type sensors, I'll post it.
The vacuum line was a no-go either, just wouldn't stay secure on the non-barbed 5mm nipples. Thankfully, the seller has been excellent in working with me to correct this numerical glitch. They sent me a sample of their 4mm and 6mm hose for free so I could determine what is the proper size I needed. Just like the FD, 4mm is the one I needed. The actual ID is 4.2mm and it stretches sufficiently to fit onto the GM Map sensor (7.8mm at largest part of the barb). Upon hearing that 4mm was the size I needed, they're exchanging the 5mm stuff for 4mm on their dime. Now, if only they had GM Map Sensors...
One thing I despise about most builds with a large number of vacuum lines is how they are secured in place. Most people use Adel Clamps and that's fine for a few lines, but simply isn't up to my standards when facing 7 solenoids with up to 3 lines each. I'm very picky about making things look presentable/serviceable and stacked Adel Clamps just looked poorly executed when facing this many lines. After taking a close look at how Mazda secured all of the hardlines to the firewall (they just pop into C-shaped clips), I found a similar solution from McMaster-Carr. While clean and straightforward, the problem is that it is made of Nylon and would not survive being in close proximity to the turbos. As I could not find a similar part made of more heat-resistant materials as the OE Mazda solution is (steel coated in vinyl for the stock Map sensor & A/T vacuum switch line clip), I found an answer in the basement.
The last time the car was running, it looked like this:
Over the past month, I have been designing and fabricating a new vacuum system. The basis for this was something my mom & I came up with over the winter. Each clip locks solidly in place, you can mix & match any size of lines at the same location, require no tools for removal/reassembly and the clips stay secure on the line so you won't lose them during servicing. As for securing the brackets, I am using existing holes on the car, both threaded and bare, to attach each one to the car. Again, this goes with my philosophy that drilling more holes isn't necessary when a design can be given a few small adjustments to use what is already provided. Mazda has always been fond of reusing parts where logically appropriate to reduce development costs, so I have done the same and taken it a step further. All of the hardware used here is what I already hand on hand, originally purchased at a local dollar store months ago for my workbench but never used.
One of the most important projects was building the solenoid rack. What was involved was cutting the stock GM brackets in half, reversing both pieces, adding 1/2" inbetween them and welding everything back together. While the original unmodified brackets would have worked, the orientation had the connectors at the top and I wanted the vacuum lines to be more prominent than the wiring. The downside to this is that I'll have to redo some of the solenoid subharness.
The Front Line Bracket was challenging to make it work suitably, have sufficient density and be presentable. 16 clips from front to back, then another 8 facing the front.
The passenger side needed some plumbing rework to make some room for the vacuum lines to be run. What I had to do was make the BOV's dump hose, 12AN in size, pull in closer to the engine. Add a 1" Pex Elbow and the stock Cosmo rubber elbow and it tucks in nicely. The 12AN hose was a pain to get onto the Pex elbow due to its reinforcement, so I had to grind down the 'lips' and then give it some dish soap to persuade it. Normal 3/4" rubber hose does not have this problem and will slip right on with no fuss.
Also had to use one adel clamp to secure the rear line guide in place at a second point. It is bolted to the body through an existing hole, but the vertical section was moving around too much for my liking. So one 1/4" adel clamp binds it to the brake line and it is nice and solid.
More in Part 2
05-27-13, 01:00 PM
#354
Vacuum Overhaul Part 2
Here is Part 2 of the Vacuum Overhaul.
Last and definitely not least are the Vacuum Manifold and Vacuum Chamber. Nothing too fancy here, but they fit the area well, are out of the way yet still accessible and only required drilling one hole to bolt on the manifold as the chamber uses an existing M6 threaded hole. What I used to glue the bracket onto the chamber is Duco Cement, found at the local dollar store for $1. JB Weld was my first choice, but would have meant another trip to Lowes or Menards and spending $4 more.
Yes, I know the manifold's bolt is a bit long, but there is a logical reason for this. It will also double as a binding point for the future high-pressure A/C line when I have the time & money to give the A/C the attention it deserves.
Before showing the results, let's recap with a picture of how it began.
Here it is at 80% complete as of today, the result of a month of work with a dremel, drill & arc welder:
I began with 30ft of 4mm vacuum hose and ran out early with a few lines remaining. IIRC, they are for the Map Sensor, Boost Gauge, Wastegate Actuator, Turbo Control Actuator (Vacuum and Boost), BOV, Charcoal Canister, OMP injectors and Cruise Actuator.
Those are currently mocked up in the previous blue aquarium hose. Another 20ft of good hose is currently en-route. As for the larger items such as the Brake Booster, 5ft of 6mm hose was plenty. The last bit of work will be illustrated in Part 3, so stay tuned...
Last and definitely not least are the Vacuum Manifold and Vacuum Chamber. Nothing too fancy here, but they fit the area well, are out of the way yet still accessible and only required drilling one hole to bolt on the manifold as the chamber uses an existing M6 threaded hole. What I used to glue the bracket onto the chamber is Duco Cement, found at the local dollar store for $1. JB Weld was my first choice, but would have meant another trip to Lowes or Menards and spending $4 more.
Yes, I know the manifold's bolt is a bit long, but there is a logical reason for this. It will also double as a binding point for the future high-pressure A/C line when I have the time & money to give the A/C the attention it deserves.
Before showing the results, let's recap with a picture of how it began.
Here it is at 80% complete as of today, the result of a month of work with a dremel, drill & arc welder:
I began with 30ft of 4mm vacuum hose and ran out early with a few lines remaining. IIRC, they are for the Map Sensor, Boost Gauge, Wastegate Actuator, Turbo Control Actuator (Vacuum and Boost), BOV, Charcoal Canister, OMP injectors and Cruise Actuator.
Those are currently mocked up in the previous blue aquarium hose. Another 20ft of good hose is currently en-route. As for the larger items such as the Brake Booster, 5ft of 6mm hose was plenty. The last bit of work will be illustrated in Part 3, so stay tuned...
06-10-13, 01:43 AM
#356
Turbos leaking and stuck...
After finally figuring out what is causing low vacuum on my engine, being a vacuum leak at the Throttle Body/UIM Gasket on the rear edge, I discovered my turbos have yet another coolant leak. This time, it's the feed ports and most likely due to lack of a crush washer. Got some more, then jumped in at 7PM to pull the turbos out just enough to get to the feed banjos.
Pulled all 6 nuts (labeled A-F, M10x1.5, original part number JE10-40-355), all 3 bolts on the Pre-Control Cover (labeled 1-3, M10x1.5, OE# NF01-13-453A), undid the Turbo Control Actuator (bolted to A1 & A2) from its flapper door (precautionary measure, not necessary) and gave it a tug...didn't budge. Pulled again a bit harder and saw the engine move with the turbos...
The lock washers I had placed on each stud were still there and spun freely so they weren't jamming anything up. They're definitely no good anymore, but aren't binding.
Had a chat with ObliqueFD, then Herblenny and opened up the JC Cosmo Parts Catalog to Section 1312A and think I figured out what the issue is: Turbo oil drains. Don't laugh...okay do laugh but it was a perfectly logical reason to leave them in place: They appeared to have enough flex/stretch/give to accommodate dislodging the turbos from the manifold. All of the other fluid lines have enough excess length, so I figured the oil drains would too. Alas, I was wrong and now understand why the Cosmo's front turbo oil return is in two pieces: Directional reasoning
The rear drain is just like on the FD, a bellowed tube with an overall curve of ~90 degrees. This is because it is approximately in line with the rear iron, so the bellows section is enough for installation/removal. OTOH, the front turbo drain is an entirely different story. It has to travel forward first around the turbo actuators, then do a 3D 90 degree turn to meet with the rigid straight section that leads to the front cover. I believe Mazda intended for the twisty part to be bolted onto the turbos on the bench, then mated to the straight section in the car as I did before.
Short Explanation: Pick up a piece of steel tubing and try twisting it. It's difficult, huh? Now try bending it in half, much easier to do thanks to leverage. Pulling the turbos off the manifold was trying to twist the front turbo oil drain, not flex it like the rear one does. The bellows part is along the wrong axis, so the turbos wouldn't move.
Everyone with a FC should know this lesson if they ever had to replace their fuel pump The sad thing is that it was the 3rd thing I ever did to my car... 
Next update will hopefully have less failboat in it...
Pulled all 6 nuts (labeled A-F, M10x1.5, original part number JE10-40-355), all 3 bolts on the Pre-Control Cover (labeled 1-3, M10x1.5, OE# NF01-13-453A), undid the Turbo Control Actuator (bolted to A1 & A2) from its flapper door (precautionary measure, not necessary) and gave it a tug...didn't budge. Pulled again a bit harder and saw the engine move with the turbos...
The lock washers I had placed on each stud were still there and spun freely so they weren't jamming anything up. They're definitely no good anymore, but aren't binding.
Had a chat with ObliqueFD, then Herblenny and opened up the JC Cosmo Parts Catalog to Section 1312A and think I figured out what the issue is: Turbo oil drains. Don't laugh...okay do laugh but it was a perfectly logical reason to leave them in place: They appeared to have enough flex/stretch/give to accommodate dislodging the turbos from the manifold. All of the other fluid lines have enough excess length, so I figured the oil drains would too. Alas, I was wrong and now understand why the Cosmo's front turbo oil return is in two pieces: Directional reasoning
The rear drain is just like on the FD, a bellowed tube with an overall curve of ~90 degrees. This is because it is approximately in line with the rear iron, so the bellows section is enough for installation/removal. OTOH, the front turbo drain is an entirely different story. It has to travel forward first around the turbo actuators, then do a 3D 90 degree turn to meet with the rigid straight section that leads to the front cover. I believe Mazda intended for the twisty part to be bolted onto the turbos on the bench, then mated to the straight section in the car as I did before.
Short Explanation: Pick up a piece of steel tubing and try twisting it. It's difficult, huh? Now try bending it in half, much easier to do thanks to leverage. Pulling the turbos off the manifold was trying to twist the front turbo oil drain, not flex it like the rear one does. The bellows part is along the wrong axis, so the turbos wouldn't move.
Everyone with a FC should know this lesson if they ever had to replace their fuel pump
The sad thing is that it was the 3rd thing I ever did to my car... Next update will hopefully have less failboat in it...
06-14-13, 08:11 PM
#357
Mid-June Update: Remove/Reinstall 20B Turbos in a FC with Banzai Racing like mounts
Turbos were finished up on Tuesday, with the source of the leak being the crush washers on the turbos' coolant feeds. There were 3 out of 4 present, all aluminum and definitely trash. All 3 had seized to the banjo bolts and required some careful prying to extract. 5 minutes later, fresh copper ones went on and reinstall commenced.
As I was dealing with the turbos, I wrote directions on how to do it from start to finish. When I get the current project sorted out, I'll add it here. Also, the vacuum system is finished along with discovering the source of low vacuum at idle: leak at TB-UIM gasket on firewall side.
Okay, now onto the current project. Right before I jumped in on the turbos, a fellow rotorhead on OhioRotaries came by and GAVE me a front swaybar. Was originally gonna pay him $20 for it, but he felt compelled after seeing the car in person. Marcus, thanks again bro!
Now that I have a base S4 swaybar, mounting is what I'm working on. It's still a work in progress. Will post pictures of both the finished mounts and complete vacuum system in the next update.
EDIT:
How to Remove and Reinstall the 20B-REW Turbos in a FC with mounts similar to those from Banzai Racing
I wrote this short How-To as there was no definitive guide covering the 20B-REW's turbochargers. Included are all of the tips & tricks I discovered over the course of 3 days. 3 days is what it took, including learning curve, as I have NEVER pulled the turbos while the engine was in the car and only did so once when the engine was on the workbench in 2011. With this in mind, the average person should be able to do the whole job in 4 hours
Removal:
1: Remove Rear Turbo Intake & Y-pipe (M8x1.25, 5-7 Ft.lbs)
2: Remove Downpipe & O2 sensor (M10x1.5, 28-38 Ft.lbs)
3: Remove Turbo Control Actuator** (M8x1.25, 5-7 Ft.lbs).
Pro Tip for Turbo Control Actuator: This is a real PITA. They're on the exhaust manifold, underneath the front compressor housing/CHRA, between two actuator rods (Wastegate & TCA) and the only way you can see them is with a mirror & flashlight. Its a straight shot in from the side with a 13mm socket, but the working room is cramped and remote so it is easy to drop the nuts when trying to thread them on.
The perfect tool to remove/torque them is a 13mm deepwell 1/4" drive socket, 2" extension and 3/8" drive adapter. Just enough room to slip the 3/8" ratchet on. Omitting the adapter and using a 1/4" ratchet had the square drive too close to the WG actuator rod so you can't get it on. Just using 3/8" for all of it is too long, no room for the ratchet due to the frame rail being so close.
4: Unhook Oil Drains (M6 rear, M8 for front at midway junction, 5-7 Ft.lbs), then Coolant Returns (M14 rear, M16 front, 5-7 Ft.lbs).
5: Remove Heater hose at hardline barb.
6: If using both stock hardlines, unhook Oil Feed line at midway junction above front turbo inlet. If using aftermarket hose, omit this step
*6.5: Disconnect Coolant Feed hardline at 3/8” hose to rear iron barb.
7: Remove all 6 M10 nuts (Turbo to Manifold) & 3 M10 bolts (Pre-Control Cover Plate).
8: Pull turbos up, rest on manifold
*8.5: Remove Coolant Feed Banjo bolts (M14 rear, M16 front, 5-7 Ft.lbs)
*Either 6.5 or 8.5 is fine for undoing the coolant feed, I tried both and found Step 8.5 was easier as the hose was stuck on and I didn't want to cut it.
9: Remove front Oil Drain (M8x1.25, 5-7 Ft.lbs) & Oil Feed hose (M14, 5-7 Ft.lbs) if not using both stock Hardlines
10: Pull turbos up and out from top
Reinstallation:
1: If you did 8.5 above, reinstall Coolant Feed Hardline with NEW/ANNEALED copper crush washers on both banjos (M14 rear, M16 front, 5-7 Ft.lbs)
Pro Tip: Don't use aluminum crush washers here. I did before and ALL of them seized to the banjo bolts. Not fun having to chisel them off. Copper crush washers can be reused in a pinch by annealing with a common butane torch, but getting new ones is easier/cheaper than pulling the turbos a second time. Ask me how much fun it was pulling the turbos and writing this guide because I found leaks here...
2: Align turbos on studs, secure with one M10 nut finger-tight on bottom rear corner stud
3: Reinstall Turbo Control Actuator. Pro Tip: Thread the two M8 nuts on from underneath, by routing your right arm through the downpipe area. Your fingers will naturally curl around the actuator/bracket where the rod feeds through, ring & pinky holding onto the bracket while your thumb, index & middle fingers can put the nuts in their place. Now tighten to spec from above using the combination of socket, extension, adapter & ratchet.
4: Reconnect Oil Drains with new gaskets/RTV. Pro Tip: Use bolts longer than necessary so it pulls the two flanges together as you tighten them. I used 30mm here.
5: If using both hardlines, reconnect Oil Feed at midway junction above front turbo inlet (M14, 5-7 Ft.lbs)
6: If you did 8.5 above, reconnect Coolant Feed hose at hardline
7: Reinstall Heater hose at Hardline Barb
8: Torque all 6 Turbo-Manifold nuts/3 Pre-Control Cover bolts (M10x1.5, 32-42 Ft.lbs) in stagger pattern
Pro Tip: The Top Middle and Front Nuts are difficult to access by hand, especially the front one. Long, slender fingers and patience is all I can suggest to thread these on, followed by a 3/8” drive socket, 1/4” drive adapter, 3” extension and 1/4” ratchet to snug it down. Since even I had difficulty with these, it might be better to replace these nuts with Coupling Nuts, the kind used to join two bolts together.
9: Reinstall Downpipe (M10x1.5, 28-38 Ft.lbs). Bottom nut was too crowded to get my HF 1/2” drive torque wrench onto, so I played it by ear here.
10: Reconnect Coolant Return lines (M14 rear, M16 front, 5-7 Ft.lbs)
11: Reinstall Rear Turbo Intake, Y-pipe and associated intercooler plumbing (5-7 Ft.lbs)
As I was dealing with the turbos, I wrote directions on how to do it from start to finish. When I get the current project sorted out, I'll add it here. Also, the vacuum system is finished along with discovering the source of low vacuum at idle: leak at TB-UIM gasket on firewall side.
Okay, now onto the current project. Right before I jumped in on the turbos, a fellow rotorhead on OhioRotaries came by and GAVE me a front swaybar. Was originally gonna pay him $20 for it, but he felt compelled after seeing the car in person. Marcus, thanks again bro!
Now that I have a base S4 swaybar, mounting is what I'm working on. It's still a work in progress. Will post pictures of both the finished mounts and complete vacuum system in the next update.
EDIT:
How to Remove and Reinstall the 20B-REW Turbos in a FC with mounts similar to those from Banzai Racing
I wrote this short How-To as there was no definitive guide covering the 20B-REW's turbochargers. Included are all of the tips & tricks I discovered over the course of 3 days. 3 days is what it took, including learning curve, as I have NEVER pulled the turbos while the engine was in the car and only did so once when the engine was on the workbench in 2011. With this in mind, the average person should be able to do the whole job in 4 hours
Removal:
1: Remove Rear Turbo Intake & Y-pipe (M8x1.25, 5-7 Ft.lbs)
2: Remove Downpipe & O2 sensor (M10x1.5, 28-38 Ft.lbs)
3: Remove Turbo Control Actuator** (M8x1.25, 5-7 Ft.lbs).
Pro Tip for Turbo Control Actuator: This is a real PITA. They're on the exhaust manifold, underneath the front compressor housing/CHRA, between two actuator rods (Wastegate & TCA) and the only way you can see them is with a mirror & flashlight. Its a straight shot in from the side with a 13mm socket, but the working room is cramped and remote so it is easy to drop the nuts when trying to thread them on.
The perfect tool to remove/torque them is a 13mm deepwell 1/4" drive socket, 2" extension and 3/8" drive adapter. Just enough room to slip the 3/8" ratchet on. Omitting the adapter and using a 1/4" ratchet had the square drive too close to the WG actuator rod so you can't get it on. Just using 3/8" for all of it is too long, no room for the ratchet due to the frame rail being so close.
4: Unhook Oil Drains (M6 rear, M8 for front at midway junction, 5-7 Ft.lbs), then Coolant Returns (M14 rear, M16 front, 5-7 Ft.lbs).
5: Remove Heater hose at hardline barb.
6: If using both stock hardlines, unhook Oil Feed line at midway junction above front turbo inlet. If using aftermarket hose, omit this step
*6.5: Disconnect Coolant Feed hardline at 3/8” hose to rear iron barb.
7: Remove all 6 M10 nuts (Turbo to Manifold) & 3 M10 bolts (Pre-Control Cover Plate).
8: Pull turbos up, rest on manifold
*8.5: Remove Coolant Feed Banjo bolts (M14 rear, M16 front, 5-7 Ft.lbs)
*Either 6.5 or 8.5 is fine for undoing the coolant feed, I tried both and found Step 8.5 was easier as the hose was stuck on and I didn't want to cut it.
9: Remove front Oil Drain (M8x1.25, 5-7 Ft.lbs) & Oil Feed hose (M14, 5-7 Ft.lbs) if not using both stock Hardlines
10: Pull turbos up and out from top
Reinstallation:
1: If you did 8.5 above, reinstall Coolant Feed Hardline with NEW/ANNEALED copper crush washers on both banjos (M14 rear, M16 front, 5-7 Ft.lbs)
Pro Tip: Don't use aluminum crush washers here. I did before and ALL of them seized to the banjo bolts. Not fun having to chisel them off. Copper crush washers can be reused in a pinch by annealing with a common butane torch, but getting new ones is easier/cheaper than pulling the turbos a second time. Ask me how much fun it was pulling the turbos and writing this guide because I found leaks here...
2: Align turbos on studs, secure with one M10 nut finger-tight on bottom rear corner stud
3: Reinstall Turbo Control Actuator. Pro Tip: Thread the two M8 nuts on from underneath, by routing your right arm through the downpipe area. Your fingers will naturally curl around the actuator/bracket where the rod feeds through, ring & pinky holding onto the bracket while your thumb, index & middle fingers can put the nuts in their place. Now tighten to spec from above using the combination of socket, extension, adapter & ratchet.
4: Reconnect Oil Drains with new gaskets/RTV. Pro Tip: Use bolts longer than necessary so it pulls the two flanges together as you tighten them. I used 30mm here.
5: If using both hardlines, reconnect Oil Feed at midway junction above front turbo inlet (M14, 5-7 Ft.lbs)
6: If you did 8.5 above, reconnect Coolant Feed hose at hardline
7: Reinstall Heater hose at Hardline Barb
8: Torque all 6 Turbo-Manifold nuts/3 Pre-Control Cover bolts (M10x1.5, 32-42 Ft.lbs) in stagger pattern
Pro Tip: The Top Middle and Front Nuts are difficult to access by hand, especially the front one. Long, slender fingers and patience is all I can suggest to thread these on, followed by a 3/8” drive socket, 1/4” drive adapter, 3” extension and 1/4” ratchet to snug it down. Since even I had difficulty with these, it might be better to replace these nuts with Coupling Nuts, the kind used to join two bolts together.
9: Reinstall Downpipe (M10x1.5, 28-38 Ft.lbs). Bottom nut was too crowded to get my HF 1/2” drive torque wrench onto, so I played it by ear here.
10: Reconnect Coolant Return lines (M14 rear, M16 front, 5-7 Ft.lbs)
11: Reinstall Rear Turbo Intake, Y-pipe and associated intercooler plumbing (5-7 Ft.lbs)
Last edited by Akagis_white_comet; 06-22-13 at 04:11 PM. Reason: How-To guide for remove/reinstall Cosmo 20B-REW Turbos in FC
06-22-13, 04:35 PM
#358
Swing and Sway the Elephant Way #1 of 2
Update is split due to pictures.
After the turbo pulling drama, I found that it still wants to leak somewhere on the passenger side. Checked all the fittings, everything is torqued to spec and dry as a bone...
So after reinstalling the Y-pipe, I snapped a pic of the complete 20B-REW Turbo Control System
Also today, I pulled the throttle body and found something disturbing, especially since it was a manufacturing error in this case.
This is the 20B UIM to throttle body flange mating area with its gasket on top
Here is a detail shot of it
And the Throttle Body side with the gasket on top as well.
See the areas where air can go around the gasket? We call these Vacuum Leaks and it explains why the vacuum nipple on the Throttle Body near the BAC Valve was showing lower vacuum than everything else. It got a dose of red RTV on both surfaces today, should be nice and happy now.
Now, onto the main event. As previously stated, I got a base S4 swaybar and attempted to use it. Short story is that it wouldn't work. If you flip the bar over so the 'horns' face up, the middle section is easily 1.5" higher than the ends. This and the 'horns' kept hitting the oil pan, steering rack or oil cooler no matter how I tried it mount it. So another brainstorming session ended up with me getting this:
Look close and you can see that the FC bar isn't straight/flat, while the other one is...
The ends are shaped differently as well. FC one was hitting the tires on full lock & full droop.
New one clears just peachy:
But there's a slight issue here:
And on the underside:
Driver's side is affected as well
Out comes the dremel, armed with the almighty #420 Cut-Off Wheels and a solution is made
Lather, rinse, repeat for the driver's side:
After the turbo pulling drama, I found that it still wants to leak somewhere on the passenger side. Checked all the fittings, everything is torqued to spec and dry as a bone...
So after reinstalling the Y-pipe, I snapped a pic of the complete 20B-REW Turbo Control System
Also today, I pulled the throttle body and found something disturbing, especially since it was a manufacturing error in this case.
This is the 20B UIM to throttle body flange mating area with its gasket on top
Here is a detail shot of it
And the Throttle Body side with the gasket on top as well.
See the areas where air can go around the gasket? We call these Vacuum Leaks and it explains why the vacuum nipple on the Throttle Body near the BAC Valve was showing lower vacuum than everything else. It got a dose of red RTV on both surfaces today, should be nice and happy now.
Now, onto the main event. As previously stated, I got a base S4 swaybar and attempted to use it. Short story is that it wouldn't work. If you flip the bar over so the 'horns' face up, the middle section is easily 1.5" higher than the ends. This and the 'horns' kept hitting the oil pan, steering rack or oil cooler no matter how I tried it mount it. So another brainstorming session ended up with me getting this:
Look close and you can see that the FC bar isn't straight/flat, while the other one is...
The ends are shaped differently as well. FC one was hitting the tires on full lock & full droop.
New one clears just peachy:
But there's a slight issue here:
And on the underside:
Driver's side is affected as well
Out comes the dremel, armed with the almighty #420 Cut-Off Wheels and a solution is made
Lather, rinse, repeat for the driver's side:
06-22-13, 04:52 PM
#359
Swing and Sway the Elephant Way #2 of 2: Front Swaybar for 20B FCs
Continuing where we left off...
Now it clears perfectly
But what about the endlinks?? FC ones are 25.4mm thick, while the new bar's original ones are 36mm...
So I built new ones from scratch on the cheap. At their shortest, the new links are 15mm longer than the previous ones
And the necessary mounting blocks. The top chunk of square tube wasn't needed, so I used shorter bolts to tidy it up. Was annoying to thread them in, but worth the effort.
As for their location...
You guessed it...the original bolt holes but 3" lower. I used the rear holes since the front ones were in poor to unusable shape due to broken bolts.
And the extra length in the endlinks paid off:
It's mounted RIGHT BEHIND the oil cooler, fits like it was meant to be there, clears through the entire stroke of the suspension and was an off-the-shelf part I found for $40.
It's a FD Swaybar using the FC bushing brackets and will get new bushings from Racing Beat since their FC swaybar is the same diameter, 1.125" or 28mm.
Total cost: $65
Now it clears perfectly
But what about the endlinks?? FC ones are 25.4mm thick, while the new bar's original ones are 36mm...
So I built new ones from scratch on the cheap. At their shortest, the new links are 15mm longer than the previous ones
And the necessary mounting blocks. The top chunk of square tube wasn't needed, so I used shorter bolts to tidy it up. Was annoying to thread them in, but worth the effort.
As for their location...
You guessed it...the original bolt holes but 3" lower. I used the rear holes since the front ones were in poor to unusable shape due to broken bolts.
And the extra length in the endlinks paid off:
It's mounted RIGHT BEHIND the oil cooler, fits like it was meant to be there, clears through the entire stroke of the suspension and was an off-the-shelf part I found for $40.
It's a FD Swaybar using the FC bushing brackets and will get new bushings from Racing Beat since their FC swaybar is the same diameter, 1.125" or 28mm.
Total cost: $65
06-23-13, 03:07 PM
#360
666rpm, 13.5 inHg, an oil leak and still-leaking turbos...
Idling at 666rpm on its own, pulling 13.5 inHg of Vacuum. A little low on both, but much closer to where it should be. 750-850rpm with a proper mixture should put it right around 15-16 inHg as it should be.
I'd like to mention that the throttle body gasket mentioned in my previous update is a FD one, not a 20B one. Was labeled/bagged as a Cosmo one by Mazda though...Red RTV did the trick in sealing up the vacuum leak though.
Front turbo coolant feed is still leaking/steaming with NEW crush washers on it...
I'd like to mention that the throttle body gasket mentioned in my previous update is a FD one, not a 20B one. Was labeled/bagged as a Cosmo one by Mazda though...Red RTV did the trick in sealing up the vacuum leak though.
Front turbo coolant feed is still leaking/steaming with NEW crush washers on it...
07-14-13, 07:57 PM
#363
The Chuck Norris of Updates: You'll **** bricks when you see it...
Time for another installment of the 12 Days of Rotormas, my 20B-REW Saga. Since the previous update, some cockstain thought it'd be cool to smash my driver's side window when there was nothing of value in the car. In the manner defined by Chris Rock, I hate n****hs. On the plus side, the ******* couldn't get the hood open thanks to salt country having done a number on the hood's underside...right at the hood latch! Giving the ****** tetanus made me feel much better. Rust is a good thing 
. One of the local junkyards claimed to have a window, but what I found out upon arriving is that it was for a Vert. Apparently, Car-Part.com--Used Auto Parts Market does not differentiate between coupe and verts, nor are their participating junkyards inclined to upload such a difference...
***EDIT: Apparently since I visited the aforementioned junkyard, car-part.com has corrected this issue.***
So after all of this and pulling a ludicrous 70 hour week at work (40 of which was Overtime!). work on the car is resuming.
Higgi, I took a close look at the Z32 swaybar and found that it has the same issue as the FC one in that it's not a straight bar. While it may work fine on your build, I needed a perfectly straight swaybar to clear the oil cooler (mounted ~1/8" in front of the swaybar), steering rack (directly above the swaybar) and oil pan (occupies where the original bar mounts to the frame rails). Any bends/offsets in the middle section would interfere with one or more of these as I found out with the FC swaybar. On an interesting sidenote, an RX-8 front swaybar could have worked as well. It's got a straight middle section like the FD one, but the ends are shaped like the FC one. If my front tires were spaced out a little further than 20mm, it would have fit quite nicely. But since the FC one (and the RX-8 one would too) was touching the tires at full lock & full droop, the FD swaybar got the nod.
Right before hell week, I pulled the turbos again, replaced all of the crush washers (Banjo bolts torqued to 30 Ft.lbs per recommendation by Dale Clark) and did a small mod to the coolant supply hose since I couldn't pull it loose from either barb. Now it has 06AN fittings in the middle for easier removal. During reinstall, I did the aforementioned coupling nut mod for the two difficult studs, finding it was significantly easier to thread them back on from above. Also made new gaskets for the oil returns with rubber-fiber gasket sheet and RTV. Front turbo oil return still has a leak at both the midway and front cover flanges (haven't touched this one, EVER!), I'm not exactly sure why though. Coupled with the still-persistent oil leak at the oil pan flange, these would necessitate sidelining the car for an engine pull to be dealt with properly. At this time, I can't afford to be without a car any longer so I'll just watch the oil level like a hawk. Plus, I'm sick of dealing with oil leaks, being on my back underneath the car, working in a small, cramped space at the whim of the weather and tracking in mud as well. Dealing with all of this was pushing me dangerously close to hitting burnout again, so time to change gears and do something different.
ECU is mounted on the stock ECU bracket with plenty of space leftover for other electronic goodies. One original mount stud needed shortened by ~3mm, but it fits and only required drilling one hole but I drilled two to make wiring access easier due to the bulk of wiring. Bracket is secured by 3 of the 4 nuts/bolts and the covers fits well enough that I'm happy with it. Also implemented the last wiring hack for the eventual manual override switches. Cut the trigger wires for the Turbo Control Relay and BOV Solenoid, then routed them to a 3rd connector which houses the gauge interface wiring as well. With this in place, I can manually force the car to run Non-Sequential or even N/A. Picture now:
A few months ago, I got a S4T2 gauge subharness. When pulling my gauges out, I found my car had a FB01 subharness in it. The only difference between the basic FB01 and the upgraded FB44 one are the two boost gauge wires and single cooling fan wire (for cars with ABS). On a small side note, I'd like to mention that one does NOT need to remove the speedometer cable from the transmission to pull the gauge cluster. Just reach under and pop it out of the two clips along the driver's side frame rail, then the single clip under the hood to get enough slack to undo it from the cluster. Now the FB44 subharness is modified with an additional connector with terminals spliced into the Ignition-switched 12v, Illumination and Ground wires, Metripack of course. Vacuum line for boost gauge is threaded through the dashboard in a clean yet reasonably accessible route. Wideband gauge wiring along with a few other useful items got the same attention via a 4-conductor cable strung up through the dash to the accessory gauge connector in the modded FB44 subharness.
Modified the boost gauge pod by sistering it up with another homebrew one for the Air/Fuel gauge. Nothing too fancy here, just a 1.5" ABS pipe coupler & endcap reamed out to 2" ID, glued together with Duco Cement & a hole drilled for the wires to pass through. Drill a matching hole in the boost gauge pod, then run a bolt through the wire hole and glue both pods together. Add some extra reinforcement on the underside made from storage box dividers and it'll do just fine. Last, a hole in the rear of each pod to mount the gauges solidly with one M5 bolt in each.
The gauge-to-pod mounting isn't fancy, but works well. Since most gauges have two mounting holes and use annoying small hardware, what I did was cut up the C-shaped bracket from my boost gauge into three flat sections. Then I drilled holes to match the mount holes on the gauge (M3 for the Boost Gauge, M4 for the Wideband), then drill a 3rd hole halfway between them. Run a M5 bolt of suitable length through the middle hole facing backwards & secure with a nut. Now pop the gauge in the pod so the bolt goes through the pod's rear hole and thread a nut on it. Doing it this way makes removal a breeze too
Boost gauge is modified with a 2-pin connector left from my R/C car days (JST Connector, used for 90s-era receiver battery packs). This was necessary to make install/removal not a nightmare. In all seriousness, EVERY aftermarket gauge should have a connector on it. Something that's been bugging me for awhile in regard to gauge wiring is that it seems like every aftermarket gauge uses tiny wires in the neighborhood of 22-24awg, with only select ones having a connector at the gauge itself. This makes no logical sense to me for two reasons:
1. Wires this small don't belong in a car to deliver such vital information due to their frailty.
2. Lack of a connector at the gauge makes it significantly more difficult to route the wiring intelligently
While I am reasonably adept with wiring overall, this is one area that the average person may have difficulty due to the excessively small wires. It's not just on my cheap Harbor Freight boost gauge, the Innovate Wideband gauge is the same deal. Due to the size, the wires for the boost gauge connector needed to be soldered together, a skill I've yet to be sufficiently adept at. While soldering these together, my brother Dan commented that in tinning the wires, the insulation was melting off much sooner than he normally expected. For being one well-versed in building guitar effects pedals, and well practiced with soldering components of this scale, making these at least one size larger would have been a good move. Big thanks to my brother Dan for doing this for me. As for the other end, I managed to crimp on 16awg Metripack terminals and 18awg seals without too much difficulty, making the whole gauge pod entire plug & play.
The boost gauge has 3 wires for its LED backlight, Orange, White & Black. Orange is Power (for recall of user-selected backlight color), White is Illumination and Black is Ground. In the interest of reason as my light-changing button is broken, Orange and White got merged together, wired into the Illumination circuit.
One thing that really annoys me is that the defroster/defogger/demister vents for the driver/passenger windows seem to be the missing from every FC I know of. If it is anything like the S4 stereo trim, I would assume because they are always breaking. and at $20 each from Mazdatrix, this didn't sound like much fun. So I made a template of the area and made one out of steel, adding provisions for the gauge mount while retaining the functionality of a defroster vent for the driver's window. Painted it black, threw on the gauge pod and installed it. Two gauges, one mount to dashboard, all wires hidden and under $10 to make. Why? Because FC!
I know that wasn't the best pic, so I got another one:
By the time I did all of this, I couldn't help but power it up...and get a pic. Yes I know the boost gauge lights up green. Green is the default color that it goes to on initial power-up. The reason for this is because I accidentally broke the light button when making the boost gauge pod 2 years ago.
The Innovate LC1 Wideband Controller is wired up & configured for WB use on both outputs. Decided to do things a tad differently by splitting off the second signal output and calibration wire into a separate connector as their use would only be intermittent and I wanted to have the calibration button & LED on a plug & play pigtail. As always, it got the metripack treatment. Had to drag out my previous 900Mhz laptop to reprogram the LC1 as my current one lacks a serial port, I don't have the appropriate adapter yet and my desktop (running Ubuntu) didn't want to behave properly in its XP Virtual Machine.
Here's a useful tidbit about the LC1 that I logged:
Default Calibration Data:
Output #1, Yellow Wire (Simulated Narrowband by default)
1.099v: 0.958 Lambda (14.08 AFR)
0.100v: 1.022 Lambda (15.01 AFR)
Output #2, Brown Wire (Wideband by default)
0.000v: 0.500 Lambda (7.35 AFR)
5.000v: 1.523 Lambda (22.39 AFR)
Just for fun, I mocked up the stereo to see how it would look. Looks good, but didn't want to sit perfectly flush. Turns out there's a small lip at the back of the radio slot along the bottom of the opening and my switch panel's 'legs' were hitting it. I'm picky and made the switch panel the same length as the stereo. Some dremel work to nudge everything forward 1/4" got it closer but not sufficiently flush to my satisfaction. It needs another 20mm taken off, 8mm to be flush and another 12mm to not mangle the wiring upon final installation. Total hole slotting will be 1".
Last, but far from least, I have an announcement for the ENTIRE rotary community. There's no clever way to put it, so I'm just gonna say it:
I have the Eunos Cosmo Factory Service Manual, both Volumes 1 and 2.
With some linguistic assistance from one of my Japanese coworkers during a lunch break, the vacuum diagram is now in English
Let me repeat that:
The Eunos Cosmo 20B-REW Vacuum Diagram is now in ENGLISH and color enhanced for clarity.
It is not 100% done yet though. While adding color to each line via Gimp & Kolourpaint, I noticed a bit of strange plumbing in regard to the Wastegate & Pre-Control Actuators & Solenoids' lines. Thinking it was just a fluke on the 20B diagram, I doublechecked it on the 13B one and found the exact same issue. As such, Herblenny has volunteered to verify everything on his personal Cosmo to ensure accuracy for the English-speaking community.
Now you can **** bricks
. One of the local junkyards claimed to have a window, but what I found out upon arriving is that it was for a Vert. Apparently, Car-Part.com--Used Auto Parts Market does not differentiate between coupe and verts, nor are their participating junkyards inclined to upload such a difference...***EDIT: Apparently since I visited the aforementioned junkyard, car-part.com has corrected this issue.***
So after all of this and pulling a ludicrous 70 hour week at work (40 of which was Overtime!). work on the car is resuming.
Higgi, I took a close look at the Z32 swaybar and found that it has the same issue as the FC one in that it's not a straight bar. While it may work fine on your build, I needed a perfectly straight swaybar to clear the oil cooler (mounted ~1/8" in front of the swaybar), steering rack (directly above the swaybar) and oil pan (occupies where the original bar mounts to the frame rails). Any bends/offsets in the middle section would interfere with one or more of these as I found out with the FC swaybar. On an interesting sidenote, an RX-8 front swaybar could have worked as well. It's got a straight middle section like the FD one, but the ends are shaped like the FC one. If my front tires were spaced out a little further than 20mm, it would have fit quite nicely. But since the FC one (and the RX-8 one would too) was touching the tires at full lock & full droop, the FD swaybar got the nod.
Right before hell week, I pulled the turbos again, replaced all of the crush washers (Banjo bolts torqued to 30 Ft.lbs per recommendation by Dale Clark) and did a small mod to the coolant supply hose since I couldn't pull it loose from either barb. Now it has 06AN fittings in the middle for easier removal. During reinstall, I did the aforementioned coupling nut mod for the two difficult studs, finding it was significantly easier to thread them back on from above. Also made new gaskets for the oil returns with rubber-fiber gasket sheet and RTV. Front turbo oil return still has a leak at both the midway and front cover flanges (haven't touched this one, EVER!), I'm not exactly sure why though. Coupled with the still-persistent oil leak at the oil pan flange, these would necessitate sidelining the car for an engine pull to be dealt with properly. At this time, I can't afford to be without a car any longer so I'll just watch the oil level like a hawk. Plus, I'm sick of dealing with oil leaks, being on my back underneath the car, working in a small, cramped space at the whim of the weather and tracking in mud as well. Dealing with all of this was pushing me dangerously close to hitting burnout again, so time to change gears and do something different.
ECU is mounted on the stock ECU bracket with plenty of space leftover for other electronic goodies. One original mount stud needed shortened by ~3mm, but it fits and only required drilling one hole but I drilled two to make wiring access easier due to the bulk of wiring. Bracket is secured by 3 of the 4 nuts/bolts and the covers fits well enough that I'm happy with it. Also implemented the last wiring hack for the eventual manual override switches. Cut the trigger wires for the Turbo Control Relay and BOV Solenoid, then routed them to a 3rd connector which houses the gauge interface wiring as well. With this in place, I can manually force the car to run Non-Sequential or even N/A. Picture now:
A few months ago, I got a S4T2 gauge subharness. When pulling my gauges out, I found my car had a FB01 subharness in it. The only difference between the basic FB01 and the upgraded FB44 one are the two boost gauge wires and single cooling fan wire (for cars with ABS). On a small side note, I'd like to mention that one does NOT need to remove the speedometer cable from the transmission to pull the gauge cluster. Just reach under and pop it out of the two clips along the driver's side frame rail, then the single clip under the hood to get enough slack to undo it from the cluster. Now the FB44 subharness is modified with an additional connector with terminals spliced into the Ignition-switched 12v, Illumination and Ground wires, Metripack of course. Vacuum line for boost gauge is threaded through the dashboard in a clean yet reasonably accessible route. Wideband gauge wiring along with a few other useful items got the same attention via a 4-conductor cable strung up through the dash to the accessory gauge connector in the modded FB44 subharness.
Modified the boost gauge pod by sistering it up with another homebrew one for the Air/Fuel gauge. Nothing too fancy here, just a 1.5" ABS pipe coupler & endcap reamed out to 2" ID, glued together with Duco Cement & a hole drilled for the wires to pass through. Drill a matching hole in the boost gauge pod, then run a bolt through the wire hole and glue both pods together. Add some extra reinforcement on the underside made from storage box dividers and it'll do just fine. Last, a hole in the rear of each pod to mount the gauges solidly with one M5 bolt in each.
The gauge-to-pod mounting isn't fancy, but works well. Since most gauges have two mounting holes and use annoying small hardware, what I did was cut up the C-shaped bracket from my boost gauge into three flat sections. Then I drilled holes to match the mount holes on the gauge (M3 for the Boost Gauge, M4 for the Wideband), then drill a 3rd hole halfway between them. Run a M5 bolt of suitable length through the middle hole facing backwards & secure with a nut. Now pop the gauge in the pod so the bolt goes through the pod's rear hole and thread a nut on it. Doing it this way makes removal a breeze too
Boost gauge is modified with a 2-pin connector left from my R/C car days (JST Connector, used for 90s-era receiver battery packs). This was necessary to make install/removal not a nightmare. In all seriousness, EVERY aftermarket gauge should have a connector on it. Something that's been bugging me for awhile in regard to gauge wiring is that it seems like every aftermarket gauge uses tiny wires in the neighborhood of 22-24awg, with only select ones having a connector at the gauge itself. This makes no logical sense to me for two reasons:
1. Wires this small don't belong in a car to deliver such vital information due to their frailty.
2. Lack of a connector at the gauge makes it significantly more difficult to route the wiring intelligently
While I am reasonably adept with wiring overall, this is one area that the average person may have difficulty due to the excessively small wires. It's not just on my cheap Harbor Freight boost gauge, the Innovate Wideband gauge is the same deal. Due to the size, the wires for the boost gauge connector needed to be soldered together, a skill I've yet to be sufficiently adept at. While soldering these together, my brother Dan commented that in tinning the wires, the insulation was melting off much sooner than he normally expected. For being one well-versed in building guitar effects pedals, and well practiced with soldering components of this scale, making these at least one size larger would have been a good move. Big thanks to my brother Dan for doing this for me. As for the other end, I managed to crimp on 16awg Metripack terminals and 18awg seals without too much difficulty, making the whole gauge pod entire plug & play.
The boost gauge has 3 wires for its LED backlight, Orange, White & Black. Orange is Power (for recall of user-selected backlight color), White is Illumination and Black is Ground. In the interest of reason as my light-changing button is broken, Orange and White got merged together, wired into the Illumination circuit.
One thing that really annoys me is that the defroster/defogger/demister vents for the driver/passenger windows seem to be the missing from every FC I know of. If it is anything like the S4 stereo trim, I would assume because they are always breaking. and at $20 each from Mazdatrix, this didn't sound like much fun. So I made a template of the area and made one out of steel, adding provisions for the gauge mount while retaining the functionality of a defroster vent for the driver's window. Painted it black, threw on the gauge pod and installed it. Two gauges, one mount to dashboard, all wires hidden and under $10 to make. Why? Because FC!
I know that wasn't the best pic, so I got another one:
By the time I did all of this, I couldn't help but power it up...and get a pic. Yes I know the boost gauge lights up green. Green is the default color that it goes to on initial power-up. The reason for this is because I accidentally broke the light button when making the boost gauge pod 2 years ago.
The Innovate LC1 Wideband Controller is wired up & configured for WB use on both outputs. Decided to do things a tad differently by splitting off the second signal output and calibration wire into a separate connector as their use would only be intermittent and I wanted to have the calibration button & LED on a plug & play pigtail. As always, it got the metripack treatment. Had to drag out my previous 900Mhz laptop to reprogram the LC1 as my current one lacks a serial port, I don't have the appropriate adapter yet and my desktop (running Ubuntu) didn't want to behave properly in its XP Virtual Machine.
Here's a useful tidbit about the LC1 that I logged:
Default Calibration Data:
Output #1, Yellow Wire (Simulated Narrowband by default)
1.099v: 0.958 Lambda (14.08 AFR)
0.100v: 1.022 Lambda (15.01 AFR)
Output #2, Brown Wire (Wideband by default)
0.000v: 0.500 Lambda (7.35 AFR)
5.000v: 1.523 Lambda (22.39 AFR)
Just for fun, I mocked up the stereo to see how it would look. Looks good, but didn't want to sit perfectly flush. Turns out there's a small lip at the back of the radio slot along the bottom of the opening and my switch panel's 'legs' were hitting it. I'm picky and made the switch panel the same length as the stereo. Some dremel work to nudge everything forward 1/4" got it closer but not sufficiently flush to my satisfaction. It needs another 20mm taken off, 8mm to be flush and another 12mm to not mangle the wiring upon final installation. Total hole slotting will be 1".
Last, but far from least, I have an announcement for the ENTIRE rotary community. There's no clever way to put it, so I'm just gonna say it:
I have the Eunos Cosmo Factory Service Manual, both Volumes 1 and 2.
With some linguistic assistance from one of my Japanese coworkers during a lunch break, the vacuum diagram is now in English
Let me repeat that:
The Eunos Cosmo 20B-REW Vacuum Diagram is now in ENGLISH and color enhanced for clarity.
It is not 100% done yet though. While adding color to each line via Gimp & Kolourpaint, I noticed a bit of strange plumbing in regard to the Wastegate & Pre-Control Actuators & Solenoids' lines. Thinking it was just a fluke on the 20B diagram, I doublechecked it on the 13B one and found the exact same issue. As such, Herblenny has volunteered to verify everything on his personal Cosmo to ensure accuracy for the English-speaking community.
Now you can **** bricks
Last edited by Akagis_white_comet; 07-15-13 at 04:28 AM. Reason: Car-Part.com Correction
07-15-13, 09:08 AM
#364
Very nice! I cant wait to see this thing running. I was tempted at one point to try FD turbos on my FC but figured it would of been a really big headache. Glad to see you stuck through all of this (with an extra rotor too!)
P.S...I think its been more than 12 days of rotormas.
P.S...I think its been more than 12 days of rotormas.
07-16-13, 03:04 PM
#365
She moves
Today was the big day.
After trying a M7x1.0 speedbleeder on the clutch slave cylinder and, not fitting and having to run again to Jegs last night for a M8x1.0 one, I bled the clutch this morning. Very odd that my spare SC is M7 while the one on the car is M8...
Still had a LOT of dead play in the pedal, so it got readjusted to make happy. Fired her up and...
Stalled, then the breaker tripped, didn't want to restart for a bit. 3 minutes later, she puttered around the front of my house and got a flat tire.
After trying a M7x1.0 speedbleeder on the clutch slave cylinder and, not fitting and having to run again to Jegs last night for a M8x1.0 one, I bled the clutch this morning. Very odd that my spare SC is M7 while the one on the car is M8...
Still had a LOT of dead play in the pedal, so it got readjusted to make happy. Fired her up and...
Stalled, then the breaker tripped, didn't want to restart for a bit. 3 minutes later, she puttered around the front of my house and got a flat tire.
07-18-13, 08:22 PM
#367
0-6250 rpm datalog, need some input on it
Today was kind of sucky but balanced with happy goodness.
The Bad: Gauge pod broke in half where the glue was. It's re-glued with extra reinforcement now.
The Good: Engine revs to 6000rpm now
After a conversation with Stevesimon about the oddly-low rev ceiling, he suggested I check the ignition settings. Turns out I had left Ignition Timing Lock set to Enabled
:
Fixed that and put everything back in the car to give it a go. Doesn't idle as smooth or stable as before, runs a bit rough and takes some throttle to keep it alive thanks to the AFRs being in the 11-12 range overall. Pulls 15inHg at 800rpm, so it's a step in the right direction. Had a brainstorm and took a datalog as I revved it in neutral to its new artificial ceiling of 6250 RPM. The log includes RPM, manifold pressure, battery voltage, ignition advance, injector duty cycle, Air Temperature, Coolant Temperature, TPS & Wideband O2 readings. You'll need Haltech's ECU Manager software to view the log. Also useful to have virtual gauges for each channel too.
Could a wisened fellow Haltech user please take a look at the attached log and give it a critique? I don't believe it is running too rich at 6000rpm at the moment, but a second opinion on why it hesitates is always appreciated.
PS2K 0-6250 log.csv.zip
The Bad: Gauge pod broke in half where the glue was. It's re-glued with extra reinforcement now.
The Good: Engine revs to 6000rpm now
After a conversation with Stevesimon about the oddly-low rev ceiling, he suggested I check the ignition settings. Turns out I had left Ignition Timing Lock set to Enabled
:Fixed that and put everything back in the car to give it a go. Doesn't idle as smooth or stable as before, runs a bit rough and takes some throttle to keep it alive thanks to the AFRs being in the 11-12 range overall. Pulls 15inHg at 800rpm, so it's a step in the right direction. Had a brainstorm and took a datalog as I revved it in neutral to its new artificial ceiling of 6250 RPM. The log includes RPM, manifold pressure, battery voltage, ignition advance, injector duty cycle, Air Temperature, Coolant Temperature, TPS & Wideband O2 readings. You'll need Haltech's ECU Manager software to view the log. Also useful to have virtual gauges for each channel too.
Could a wisened fellow Haltech user please take a look at the attached log and give it a critique? I don't believe it is running too rich at 6000rpm at the moment, but a second opinion on why it hesitates is always appreciated.
PS2K 0-6250 log.csv.zip
07-19-13, 03:56 PM
#368
Without actual map used its hard to say whats the problem.
I would suggest logging much more values as well, there is no reason not to - everything displayed at any of the tabs in ECU manager will be logged.
Turn on home, trigger and missed counter windows to see if CAS readings are OK.
I would suggest logging much more values as well, there is no reason not to - everything displayed at any of the tabs in ECU manager will be logged.
Turn on home, trigger and missed counter windows to see if CAS readings are OK.
07-19-13, 08:56 PM
#370
What's great for a snack, fits on your back, it's LOG LOG LOG!
This was begging for a Ren & Stimpy reference!
Thanks for the critique on the datalog, I now understand how to read it properly. Unfortunately, I did not log both primary and secondary injectors' duty cycle/time, only primary DC. As for fuel pressure, I am using the stock 20B pressure regulator. However this brings an interesting point in that a decrease in fuel pressure once the secondaries hit a certain Duty Cycle would logically point to the fuel pump being unable to keep up. As I'm currently running a FD pump, it would imply that it can't flow enough to feed six 550cc injectors adequately, hence it leans out. The exact numbers can be derived mathematically, but would require more effort.
Another thing I noticed is that when it begins to lean out, ignition timing starts to roll back a bit as well. Could this be related to it leaning out or is it just a coincidence? Just thought I'd ask to have the correct understanding (still a Haltech Tuning N00b).
Good news: I have a new window in the car now, thanks to GXL87Turbo. He not only pulled it from his S4 partout that I found on craigslist, but delivered it from Marysville for $20 total. People like him are what make the rotary community so awesome!
90 minutes to vacuum all the shards out from inside the door with a shop-vac (NOT fun at all) and install the new window including learning curve. And I didn't break the mirror triangle!!
Next time, I'll log more channels and post it up.
Thanks for the critique on the datalog, I now understand how to read it properly. Unfortunately, I did not log both primary and secondary injectors' duty cycle/time, only primary DC. As for fuel pressure, I am using the stock 20B pressure regulator. However this brings an interesting point in that a decrease in fuel pressure once the secondaries hit a certain Duty Cycle would logically point to the fuel pump being unable to keep up. As I'm currently running a FD pump, it would imply that it can't flow enough to feed six 550cc injectors adequately, hence it leans out. The exact numbers can be derived mathematically, but would require more effort.
Another thing I noticed is that when it begins to lean out, ignition timing starts to roll back a bit as well. Could this be related to it leaning out or is it just a coincidence? Just thought I'd ask to have the correct understanding (still a Haltech Tuning N00b).
Good news: I have a new window in the car now, thanks to GXL87Turbo. He not only pulled it from his S4 partout that I found on craigslist, but delivered it from Marysville for $20 total. People like him are what make the rotary community so awesome!
90 minutes to vacuum all the shards out from inside the door with a shop-vac (NOT fun at all) and install the new window including learning curve. And I didn't break the mirror triangle!!
Next time, I'll log more channels and post it up.
07-20-13, 11:32 AM
#371
Moderator
iTrader: (3)
Join Date: Mar 2001
Location: https://www2.mazda.com/en/100th/
Posts: 30,841
Received 2,604 Likes
on
1,847 Posts
congrats on the first drive, that is cool!
looking at the data log, if you put injector duty right on the o2 trace, they follow each other, which makes no sense, unless it was misfiring. its possible it was too rich, so it misfired which will throw off the o2 sensor.
you can see throttle goes open, fuel increases, timing goes down, but afrs go lean, so either its too rich or not rich enough.
other things, idle seems a little slow and rich, not problems, but maybe idle quality could be better. it also seems like it jumps around a little, maybe cell to cell? the rpm, and timing move around a bit more than i like, although this is IMO!
looking at the data log, if you put injector duty right on the o2 trace, they follow each other, which makes no sense, unless it was misfiring. its possible it was too rich, so it misfired which will throw off the o2 sensor.
you can see throttle goes open, fuel increases, timing goes down, but afrs go lean, so either its too rich or not rich enough.
other things, idle seems a little slow and rich, not problems, but maybe idle quality could be better. it also seems like it jumps around a little, maybe cell to cell? the rpm, and timing move around a bit more than i like, although this is IMO!
07-20-13, 03:25 PM
#372
Log, Internacionale!
Took another log and it's quite different. Between the two logs, I have made no changes at all, but it revs slightly easier and went to 6500 RPM this time. Several more data channels this time, ignition advance (leading & trailing), Injector Duty Cycle & Time (Primary & Secondary), CAS counters and the OMP position & feedback.
The meat of the log begins at 3:47 as I wanted to give it some time to warm up before letting it loose. Coolant temperature is logged here as well.
Overall, it was still rather stumbly due to the rich AFR, but I didn't notice any panic-worthy leaning out behavior this time, peaking at 13.4:1@5025rpm (same AFR on decel at same rpm) and bottoming out at 10.6:1@1200rpm
Primary injector DC peaked at 56.8% while the secondaries were completely silent the entire time. I believe this may be due to not setting them up properly in regard to the staging bar (again, a n00b error on my part). J9, could you please advise me on the correct settings here?
The meat of the log begins at 3:47 as I wanted to give it some time to warm up before letting it loose. Coolant temperature is logged here as well.
Overall, it was still rather stumbly due to the rich AFR, but I didn't notice any panic-worthy leaning out behavior this time, peaking at 13.4:1@5025rpm (same AFR on decel at same rpm) and bottoming out at 10.6:1@1200rpm
Primary injector DC peaked at 56.8% while the secondaries were completely silent the entire time. I believe this may be due to not setting them up properly in regard to the staging bar (again, a n00b error on my part). J9, could you please advise me on the correct settings here?
07-20-13, 05:19 PM
#373
Moderator
iTrader: (3)
Join Date: Mar 2001
Location: https://www2.mazda.com/en/100th/
Posts: 30,841
Received 2,604 Likes
on
1,847 Posts
yes, actually datalog as much as you can, you can always filter out extraneous stuff.
re injector staging: Mazda turns the secondary injectors on @3200rpm OR 1psi of boost,
the haltech does load only, which is fine. i try to start out with a map that is as close to Mazda's as possible, because the OE map works. so i'll set the staging bar to 1psi or less, usually -1psi.
the OMP seems to work, the feedback follows the stepper.
the trigger miss count is 0, so that means the ECU hasn't missed a CAS tooth, you'll want to check this here and there.
so the log, i switched it to PSI, cause KPA sucks IMO, and the thing never actually sees boost, peak pressure is -6.2psi, so i'm not sure where your staging bar is, but mine wouldn't have turned the secondaries on yet.
the 13.4:1 actually happens after you let off the throttle, so nothing to worry about.
re injector staging: Mazda turns the secondary injectors on @3200rpm OR 1psi of boost,
the haltech does load only, which is fine. i try to start out with a map that is as close to Mazda's as possible, because the OE map works. so i'll set the staging bar to 1psi or less, usually -1psi.
the OMP seems to work, the feedback follows the stepper.
the trigger miss count is 0, so that means the ECU hasn't missed a CAS tooth, you'll want to check this here and there.
so the log, i switched it to PSI, cause KPA sucks IMO, and the thing never actually sees boost, peak pressure is -6.2psi, so i'm not sure where your staging bar is, but mine wouldn't have turned the secondaries on yet.
the 13.4:1 actually happens after you let off the throttle, so nothing to worry about.
07-20-13, 05:43 PM
#374
yes, actually datalog as much as you can, you can always filter out extraneous stuff.
re injector staging: Mazda turns the secondary injectors on @3200rpm OR 1psi of boost,
the haltech does load only, which is fine. i try to start out with a map that is as close to Mazda's as possible, because the OE map works. so i'll set the staging bar to 1psi or less, usually -1psi.
the OMP seems to work, the feedback follows the stepper.
the trigger miss count is 0, so that means the ECU hasn't missed a CAS tooth, you'll want to check this here and there.
so the log, i switched it to PSI, cause KPA sucks IMO, and the thing never actually sees boost, peak pressure is -6.2psi, so i'm not sure where your staging bar is, but mine wouldn't have turned the secondaries on yet.
the 13.4:1 actually happens after you let off the throttle, so nothing to worry about.
re injector staging: Mazda turns the secondary injectors on @3200rpm OR 1psi of boost,
the haltech does load only, which is fine. i try to start out with a map that is as close to Mazda's as possible, because the OE map works. so i'll set the staging bar to 1psi or less, usually -1psi.
the OMP seems to work, the feedback follows the stepper.
the trigger miss count is 0, so that means the ECU hasn't missed a CAS tooth, you'll want to check this here and there.
so the log, i switched it to PSI, cause KPA sucks IMO, and the thing never actually sees boost, peak pressure is -6.2psi, so i'm not sure where your staging bar is, but mine wouldn't have turned the secondaries on yet.
the 13.4:1 actually happens after you let off the throttle, so nothing to worry about.
My current understanding is that the fuel map can be leaned out across the board quite a bit from its current settings and it would help curb the hesitation. Once the turbos are plumbed in, it would need to be re-enriched accordingly, correct?
07-20-13, 07:19 PM
#375
Moderator
iTrader: (3)
Join Date: Mar 2001
Location: https://www2.mazda.com/en/100th/
Posts: 30,841
Received 2,604 Likes
on
1,847 Posts
Would this be Load-Fuel, Load-Ignition or Delta Load? Forgive the n00b question, but I'm a big believer in the philosophy of better safe than sorry. Currently, mine is set for 0.9 psi. By your reasoning, the secondaries wouldn't be activated untill positive pressure is introduced into the intake manifold.
Also, since the Cosmo's turbos hit full boost at 1800rpm, wouldn't this make it run richer than if they came online at 3200?
the literal answer to your question is yes, assuming that full boost gives the same mixture across the rpm band, it'll be richer if the turbo comes on earlier
My current understanding is that the fuel map can be leaned out across the board quite a bit from its current settings and it would help curb the hesitation. Once the turbos are plumbed in, it would need to be re-enriched accordingly, correct?
off idle the turbos being plumbed in should make a big difference, although its hard to say how much