Harlan's Impossible turbo build.
 

UPDATED 3/16/15
Still going strong at 4psi.
Here's the rundown of my build:
09+ tranny INSTALLED
Agency Power Exhaust and midpipe. INSTALLED
Ebay T35 turbo.
Ebay 38mm wastegate.
PTP turbo manifold.
GM IAT
55lb ev12 injector for preturbo water injection.
Arduino serving double duty as boost controller and WI controller.
IGN-1A coils.



OLD:

I plan on getting the turbo on the car first and setting a mild tune with 4psi then getting all

the software and hardware to play nice.

Yeah, that all sounds completely reasonable. Figuring out the software as I wait for parts to

arrive. Will be doing most of the install in about a month. This is a conglomeration of a lot of things I've tried successfully and otherwise, I'm not expecting everything to work out of the box, but it's the journey that's important to me.

Any encouragement or helpful criticism is great, but please don't turn this into a "It will never work." thread unless you can provide facts to back it up.


NEW:

Well a lot of my plan has changed/been adapted. Skip a few pages in unless you just want to watch me make mistakes. :)

The Mk1.

Mk2

Where did you get the 3.92 diff ratio? (not a 3.90?)

Do you have the correct transmission speedometer gear to match the 3.92 diff?

PS: Theoretical top end speed for a 3.92 diff is awesome.

I believe it's off a miata. Got a good deal on it and it should lower my cruising rpms and give the lower gears longer legs. Top speed should be great.

i see "ebay" , "no intercooler" , "glasspack exhaust"


clicking back button..

Should be fun to watch ;)

Everyone loves a good train wreck :)

Thanks for the support! :) I did not wake up one morning and decide to do a turbo build, but have been researching and planning for almost a year. It might be a colossal failure or it might take months and a lot of money to get right. Either way, sit back and enjoy the show.

for a year and you didn't read how not to use ebay special boost items?

What are you doing for a manifold and why no intercooler? Water injection or meth injection? And why mess with the Arduino when boost controllers and W/M controllers are so cheap?

Grand ideas, eBay parts

Seems a bit silly, but good luck just the same

It's going to be a rear mount turbo where the muffler is now, using fuel injectors for pre turbo water injection. Water is the plan, but may go to water/ethanol or water/methanol. I have a water/methanol controller and I hate it. They did everything right except the application, but I can bitch about certain WI companies later. Fuel injectors let me have complete control over volume, and they have been reliable for now. Manual boost controllers are cheap, digital aren't that bad either, but I like more control and it's darn easy to do with a microcontroler also I should be able to have boost/WI/knock detection on the same screen.

I'm not trying to make a dyno queen or a track star, I'm just trying to make something fun for me. I figure 300hp max with power coming on much lower and staying all the way up.

Also if you were going to inject massive amounts of water pre turbo, would you use a turbo that costs more than $200?

Ah, a rear mount. That presents some additional complications.

If I was going to put a turbo on my lawnmower I would still pay more than $200

And it's more the wastegate that I would be concerned about

Rear mount does have some major advantages: 1) no complicated exhaust plumbing 2) no excess heat under the hood 3) no intercooler blocking airflow to the radiator 4) plenty of room in the space vacated by the stock muffler. A major disadvantage is having to run oil to the back or come up with an independent oiling system with an electrically driven pump.

Glasspacking probably won't last long, steelpacking might be better. An oval sectioned air pipe forward would preserve some ground clearance. I'd be surprised if injecting water upstream of the turbo didn't cause damage to it, but for $200 it's a relatively cheap experiment.

Get an oil-less Comp Turbo

Lag will be a major issue. Rear mount results in a serious volume of exhaust piping to pressurize before the resulting response. You also have the same on the intake side plumbing all the way back to the front. Double lag whammy.

Running a 2" charge pipe so I'm not worried too much on clearance. The pipe itself will give some charge cooling, and the pressure drop over it's length should also cool the charge some while driving the turbo closer to its efficiency peak at high flow. I'm returning the oil with a fuel pump, I'll change over to a normal scavenge pump if it breaks within 6 months. The glasspack blanks I'm getting should be pretty easy to gut and reline with stainless steel wool, which will be my first step with them.

Gotta get back to torturing transistors for science.

harlan--you may want to speak with these people about how to properly set up a water injection system:
http://www.google.com/url?sa=t&rct=j...gujrVoU92IXE8w
These are good people who know how to really set up a great system and they have the parts/experience to get it done.
You will need a pretty big tank.

One thing you may not have considered-- that 3:92 diff ratio could cause a little difficulty in start off's.
Obviously you are going for a cruiser-- in town you may spend more time than you want in 4th gear?
--you may end up shifting a lot?
The glass packs will not last long at all--even with the exhaust cooling you will obviously have in place. The fiberglass will burn out after a little.
Spend a little more money and get one with SS packing--you will be glad you did.
This will be a very interesting build.
I am really interested in how you mount the nozzles and coordinate the injection with the intake valves. Or, are you only going to put nozzles on the primary runners? If so the intake valve transitions on a properly designed system will be difficult to tune for?
Knock sensor developments will also be very interesting to follow.

Great site OD with lots of great info. Unfortunately I found their site early while searching water injection and that made me throw out the idea of using fuel injectors despite the obvious advantages. After a bit more research and actual testing I've found that newer fuel injectors are designed to run higher ethanol fuels (read extra water) and are made of stainless. The only potential problem is the spring which is regular spring steel. The injectors I have on my desk now are stock rx8 red injectors and have had water running through them several times and were left filled and soaking in water for a week with no ill effects. If these fail I'm gonna go searching for a pair that is rebuild able and then go find a stainless spring.

Start offs may be more difficult, but it's only a 13% change. Also I don't really care about getting off the line. I only plan to be at 2k once after stopping.
May end up upshifting a lot, but the way I figure I already have to shift too much in town, this should give me longer legs on each gear and only hurt 1st.
Repacking the glasspacks with stainless, doing it myself because off the shelf are EXPENSIVE.
A magician can't give away all his secrets, but I do have a trick up my sleeve for injecting in the runners.

Suggest you PM Bumblebee and have a good talk to him . He has made a really good rear mount setup and he may well be able to help you get it right 1st time .

https://www.rx8club.com/series-i-major-horsepower-upgrades-93/rear-mount-turbo-project-204248/


Here is a thread and good luck man . Looking forward to this thread

Thanks guys. I have both Bumblebees and Rotorocks threads bookmarked. Unfortunately my build is closer to Rotorocks build and he hasn't been on in a year. I know I'm the black sheep on this, which is why it's my "impossible" build thread. Without water injection the engine will blow and the turbo will be out of breath at high rpm. Without proper engine management when the water fails, then the engine. If everything works right it should be a blast. I'd just rather try to roll a hard six doing what I've dreamed about for over a year than buy something off the shelf and being like everyone else.

You can still be special if you buy the same kit as everyone else by doing it properly and getting a great result . A very rare feat around here in case you didn't notice!

I wonder what all the little water drops injected pre-turbo will do when they hit the turbine blades spinning at over 130000rpm.

Sounds like an over complicated system that, even if properly implemented, will fail because of the cheap parts used.
Ebay turbos cost 200$ for a reason. It may be smarter to buy an used turbo from a junkyard instead.
A low mount, intercooled gtx family turbo would yield even better results, show way less lag and be reliable.

Why is your dream set up an ebay special?

Dreams are free so why you being cheap?

The Miata ring and pinion fit the RX8 differential; or did you swap the entire differential in the RX8?

The miata ring and pinion doesn't fit, the whole diff. does.

BSE is right, just checked, I have a 7" ring so it won't work without swapping internals or full diff. Oopsie. Need to find a replacement that's actually compatible. Glad I didn't find that out the hard way.

Also for your amusement:
https://www.youtube.com/watch?v=Bwa-...&feature=g-upl

Ok, I need a ring and pinion from a stock 93-97 rx7 automatic. Should be pretty easy to find. What I have is compatible with earlier RX7s but I doubt anyone would want it. Time to sell.

why not just get a 4.1 from an RX8 auto ....sheeesh

I may be missing something but aren't all auto 8s 4.3 rear ends?

up until this thread was posted you seemed like a halfway smart guy ...

Wow Team, that's almost a compliment. Everyone makes a mistake while learning, I just rush into things further beyond the comfort zone than most and can make more mistakes because of it. I'm not ashamed of that, I just keep fallback plans for when I make a mistake. My first engine swap was on my 8. Did it myself with little help, and surprisingly no problems. Had a plan to do twin rear mounted turbo's, got a used pair of garrets, one of them didn't survive rebuild, so I shifted gears and changed plans. Ok, I didn't do enough research and got the wrong differential based on bad information. No big deal. The scope of this build may change rapidly depending on what works or what doesn't, but I have no regrets. I will know more tomorrow than I did yesterday, and again the day after. And if I fail everyone including myself can have a good time through the journey.

Alright I've pretty much resolve the smoking transistor problem with my coil mod. Need to redo the circuit board so it's not such a mess, and need to wire my car, but the hardware issues and most of the coding are now done.
https://www.youtube.com/watch?v=ExK3...Y&feature=plcp

The problem is that you start with a tabula rasa to learn instead of picking up where others left.
So many complicated coil mods, weird turbo setups, lack of some fundamental research (pre-turbo water injection etc), custom boost controller that may or may not work...
There are proven setups that reach your goals in a direct fashion and without many complications. It's like climbing mount everest on your hands while you could walk there instead.

The external parts of a turbo build is well documented and is a straigh forward deal if one knows what he's doing. The challenges are inside our engine. Go with a simple turbo build and have fun with the internals instead....

I really like this :
"I will know more tomorrow than I did yesterday, and again the day after. And if I fail everyone including myself can have a good time through the journey."

Like this one too:
It's like climbing mount everest on your hands while you could walk there instead.

Sometimes the journey is more important than the destination. Thats my excuse anyway. lol.

Pre turbo water injection can be done and has been done. It is much different than the usual set up.
We supercharged guys have pre injected for a long time. Some precautions do have to be followed.
Got to admit all--this is going to be interesting.
__________________

My 2 cents would be to go 1 step at a time and have better focus. If your hell bent on the rear mount then get the turbo setup without fancy stuff to start off with (water injection, boost controller, rear diff....) and before running water injection there are some small water to air intercoolers that would do the job very well. Were the rear mount will get you is it's real on or off power not really a smooth transition, I've seen drag cars made like this. I would get the basic setup first then go from there, fab should be easy so thats a bonus, and I do agree with others get a used turbo off a car T28's and T3 are easy to come by for $150 - $300 (check out 240SX sites they typically like to upgrade the stock SR turbo's and sell off the old ones pretty cheap) .

Good luck!! Cheers

there's no better way to start off an experiment than by purchasing parts of questionable specs and tolerances :rolleyes:

Mini Update:
Almost all the components are here now. Wastegate looks beautiful. Turbo is mechanically sound and only needs some minor modifications. 3Bar Map sensor works great and tracks linearly with voltage right up to 28lbs. Boost control solenoid working well with PWM. And I have one of the glasspack repacked with SS wool waiting to get a new acetylene regulator to weld it back together. I'll post more when I put everything together for the boost controller and test it with compressed air.

Also I've found that Howard Coleman on the RX7 forum is using bosch fuel injectors for his water injection. Wish that info had been around 2 years ago, I would have skipped a lot of experimentation and went right to what I'm trying now. I'm gonna uncap a yellow injector and that should let me flow about 1000cc, and if the spray pastern is nice enough I may do the same for my existing injectors instead of going bigger.

NYET ! on the air filter setup PM Bumblebbee why that is a badski idea .....................

Not my air filter setup, just placing the components I have. Will be adding the maf tube, water injector and shield to surround the filter. Sent Bumblebee a msg.

/\ nothing wrong with learning from what others have done . A lot of the mistakes you are about to make have already been made and overcome .... :)

Thanks for the advice Brettus. I'm willing to take suggestions from others, and really do appreciate the input. I'm just tired of hearing "OMG you'll blow up your engine." or "You need to have a ball bearing Garrett chosen by professionals or your build will suck."

I've msged oltman about stock knock detection but he seems to have disappeared. I'm gonna try feeding the stock ecu some tones and figure out if I really need to do my own knock retard scheme or can just use an amplifier or frequency multiplier and let the ecu do the job.

Yellow injector is uncapped. Took about 30 seconds when I did it the easy way. (The hard way was still going after an hour.) Will test spray pasterns today.

Also testing higher spark frequencies and putting the boost controller together.

Edit: tested uncapped spray paterns/flow
Capped: 620cc/min (water flow, not at stock fuel pressure.)
Uncapped: 890cc/min (again, water flow.)
About a 43% increase in water flow which should correspond to a similar increase in fuel flow. Pattern isn't bad. Stock has 4 holes which flow pretty straight. Uncapped flows in a circular pattern which focuses down and crosses itself then flows straight. I personally wouldn't use either for Primaries since they directly effect fuel economy. But as P2/S they should do just fine.
Also corresponds to the increase here: KGParts.com - Modified RX8 Fuel Injectors

Well... can't find injectors much larger than stock that have more than 4 holes. Guess it's just hard to flow that much finely. Oh well, back to plan A.

Tried to experiment/learn about FFTs on the arduino for knock detection. Boy does that stuff go deep fast. Hopefully I'll have something to work with soon that will be better than my original plan.

Interested in this WI idea using injectors . Wonder about the tendency to clog ?

Clogging isn't the biggest problem. The spring isn't stainless so the injectors themselves will be wear items. I already have plans to use a check routine. The pump will stop, the system will wait for a minute for a low pressure switch to come in, then the injector will fire for a predetermined amount of time and the pressure switch will have a band of time it is allowed to come in at. Should both check for leaks and injector flow rate. If it fails the self check it will lock boost at 4psi and retard timing.

How long do you expect them to last ?

I don't know. Howard Coleman on the RX7club is using bosch injectors and seems to like them. If I have get a new injectors yearly but it gives me the performance I want then it's a good deal. Just started a thread on the 7club. Hopefully someone there doing similar will chime in.

Distilled water with a 1micron filter will be Ok for the clogging... corrosion may be problem though

true--even the water meth nozzles will "clog" after a certain period of time. Deposits from the fittings/plastic tank/pump,lines all come into play. Water is pretty corrosive stuff.
harlan so the injectors are going to control the volume not the pump? Correct?now here is a vid showing the spray pattern for the rx8 yellow FI ( capped)

Mazda RX8 fuel injector 450cc flow rate Denso 4450 - YouTube

here is a vid with different w/m nozzles
http://www.google.com/url?sa=t&rct=j...C97tyVqw1eq5DQ

It would seem to me that you really want the better atomization spray?