Ol' Blue 2010 Turbo Upgrade Thread
Page 21 of 23

Author:  top down@-40 [ Wed Sep 21, 2011 3:15 am ]
Post subject:  Re: Ol' Blue 2010 Turbo Upgrade Thread

ellpee wrote:

Yeeeeessssssss... But engine movement could potentially rub the threads on the side of the bolt where it goes through the manifold plate.
Leads to more movement...leads to loose bolts...just a thought...

In that line of thought, wouldn't the same force be applied to any nut tightened against the manifold, no matter what the threads on the bolt/stud are like?

Author:  Gigapunk [ Wed Sep 21, 2011 11:25 am ]
Post subject:  Re: Ol' Blue 2010 Turbo Upgrade Thread

Just joined teamswift just for your build, and read all 21 pages. Great stuff! What are your wHP goals, and what turbo are you running?

Author:  t3 ragtop [ Wed Sep 21, 2011 4:44 pm ]
Post subject:  Re: Ol' Blue 2010 Turbo Upgrade Thread

i don't know what power i'll get at the wheels, the limitations will be boost pressure under 21 psi and available fuel.

i wasn't happy with the batch fired injection and i'm currently reworking the megasquirt to add a cam angle sensor using a hall effect device that detects a magnet installed into the cam gear and 3 channels to drive the injection for full sequential operation.

the turbo set came from turbine tech. jardamuth sent me a kkk ko4 turbo with a custom manifold and stainless turbo outlet. those parts are beautifully fabricated.

my latest acquisition is a set of 17-4 stainless steel head studs that were custom fabricated. i'll be installing those when i pull the valve cover off so i can do some machining to mount the hall effect device for sequential injection.

welcome to teamswift. :D how did you find my turbo3 project on the internet?

Author:  Gigapunk [ Wed Sep 21, 2011 7:50 pm ]
Post subject:  Re: Ol' Blue 2010 Turbo Upgrade Thread

I found your build by doing an image search for a geo metro 3 cylinder engine. I've been looking for a light weight 4 stroke motor that would work well in a jet ski.

Here's the compressor map for the kkk K04. It looks like it's better suited to a 1.6L motor, or even a smaller engine as long as it makes ~80 hp NA. You are really going to be walking the line on compressor surge, and you'll also always be below 70% efficiency, so your intercooler will count!


The stock XFi motor was supposed to be good for 49-55 hp stock

you can convert the m^3/second to CFM by multiplying by 2119, and then CFM to HP by deviding by 1.62

on the bottom 0.04 corresponds to about 52 hp at zero psi
0.08 = 105 hp at 14.7 psi
0.10 = 131 hp at 22.1 psi

Author:  1987TurboSprint [ Wed Sep 21, 2011 10:14 pm ]
Post subject:  Re: Ol' Blue 2010 Turbo Upgrade Thread

Gigapunk wrote:
The stock XFi motor was supposed to be good for 49-55 hp stock
But he doesn't have an XFi motor. He started with a stock G10T, 1.0l 3 cylinder turbo out of an '89-'91 Sprint/Firefly Turbo =)

Author:  Gigapunk [ Thu Sep 22, 2011 8:02 am ]
Post subject:  Re: Ol' Blue 2010 Turbo Upgrade Thread

1987TurboSprint wrote:
Gigapunk wrote:
The stock XFi motor was supposed to be good for 49-55 hp stock
But he doesn't have an XFi motor. He started with a stock G10T, 1.0l 3 cylinder turbo out of an '89-'91 Sprint/Firefly Turbo =)

I'm pretty sure you're just messn with me... but as you probably already know it doesn't matter when calculating out flow. As a matter of fact the stock G10T fits PERFECTLY on the same verticle green horsepower line. On the Y axis start at 1.5 (the G10T's stock boost of 7 psi), and you cross the horsepwer line at ~0.063 m^3/s which equals about 83 hp. Per wiki the jdm EFI G10T was good for 82 hp.

Don't you just love when math works?

Author:  Gigapunk [ Thu Sep 22, 2011 8:04 am ]
Post subject:  Re: Ol' Blue 2010 Turbo Upgrade Thread

If you started with a G10T block why did you use/need a special adapter for the oil feed. And I thought I remember you adding a new oil return also? no?

Author:  turbohull [ Thu Sep 22, 2011 12:46 pm ]
Post subject:  Re: Ol' Blue 2010 Turbo Upgrade Thread

oil feed comes from a BPT 1/8 fitting with 28 pitch, NPT threads are 27, that's the "special" part. nothing special really...

you should stick to flow and leave HPs out of your math. displacement is one thing, but you must have forgot about engine speed...

Author:  t3 ragtop [ Thu Sep 22, 2011 4:32 pm ]
Post subject:  Re: Ol' Blue 2010 Turbo Upgrade Thread

this engine's block was a 1998 model g10. i bored and threaded the boss on the block where the turbo3 has the oil return located. i also drilled and tapped the boss between cylinders 2 & 3 for the knock sensor. also, not that it matters, but the engine was re-sleeved and bored for the used mk2 turbo3 pistons that went into it. basically, i assembled the engine from odd spare bits i had laying around.

thanks for posting the compressor curve. i never worried about the selection of the turbo, i have the utmost faith in jardamuth's wisdom when it comes to putting a turbocharger on a suzuki. his work is as much art as it is functional. i just used the parts he sent to me.

when it comes to wind, the g10 seems up to the task. i can hear the turbo spinning up just past 1500 rpm. jess told me that they're easy to come by and fairly inexpensive.

on another note, i sent lizzie to jim's automotive electrical service to pick up a high torque starter i had rebuilt for my new winter turbo3 build. if you ever run across a chance to pick up a gear reduction type starter for the g family suzuki engines, grab it. the typical straight starter, new or refurbished, spins the engine at around 180 rpm. the gear reduction type whips the engine over at 320 rpm. i have one on this blue turbo3 engine and even as tight as it is, it rolls over with authority - ba-bumpa, bumpa, bumpa! :lol:

Author:  Jardamuth [ Thu Sep 22, 2011 8:10 pm ]
Post subject:  Re: Ol' Blue 2010 Turbo Upgrade Thread

HeEeyaAAa folks! Witnesses the famous starter used by Steven Spielberg in his movie "duel".

Don't worry about those nasty truckers. Once you crank it up, nobody can catch you.

Just make sure to disable the clutch switch before fingering any trucker on the highway.

Author:  t3 ragtop [ Thu Sep 22, 2011 8:32 pm ]
Post subject:  Re: Ol' Blue 2010 Turbo Upgrade Thread

i figure it's good for an extra 20 hp. :wink: :lol:

Author:  JamalSpelling [ Thu Sep 22, 2011 9:48 pm ]
Post subject:  Re: Ol' Blue 2010 Turbo Upgrade Thread

The KKK K04 is perfect.
Plenty of headroom, and you'll never run out of wind.
The spike can be tamed with a good EBC.

It does have a bit of lag, but there's ways to improve it.
Exhaust being a key factor, and possibly cam selection.
I think an aluminum flywheel might make a big difference
as well in the low end spool up.
And of course the EBC settings will help also.

I've got a few limitations with the mostly stock G10 internals setup and the
modified VAF, but T3's gonna have almost limitless advantages
and overall complete control with the setup. So much more

Author:  Jardamuth [ Fri Sep 23, 2011 11:24 pm ]
Post subject:  Re: Ol' Blue 2010 Turbo Upgrade Thread

Fuel management is the KEY factor. Proper timing will make a night and day difference in spool up.

Author:  t3 ragtop [ Sat Oct 01, 2011 1:29 pm ]
Post subject:  Re: Ol' Blue 2010 Turbo Upgrade Thread

and to open up the grayed out ve trim maps to tweak the fuel delivery among other things, i made the jump to full sequential operation. now, instead of trying to run a less than perfect "spoofed" operation of a 6 cylinder wasted spark engine i've added the electronics to support 3 channels for injection and true 3 channel cop operation.

i had to add an option board that has the provision for 4 channels of injector driver and 4 channels of ignition driver fets. i only needed 3 channels to support 3 cylinders so i didn't do a full population of the mod board. also, since i had my ignition working pretty well i didn't use the coil drivers. i did route my 3 ignition channels through the mod board so i could use it's db15 connector and a new, dedicated wiring harness to support ignition and injection.


to date, guys who have done the sequential thing have used the existing distributor module with the bottom part of the distributor as the cam angle sensor. i never wanted to use the distributor at all. i deleted it and good riddance to a legacy mechanical system. my vision for a solid state cam angle sensor has been to use a hall effect device to sense the cam position via a small magnet on the cam gear. one of the technical problems that i ran into using a variable reluctance pickup on the crank as a sensor is that the signal that a v/r generates is a sawtooth waveshape that needs additional circuits to convert the sawtooth to a square waveshape. a hall effect device doesn't need a circuit to convert the signal, it puts out a nice clean square wave.

my idea for the cam angle sensor was to use a small hall effect device package mounted to the flange on the valve cover behind the cam gear to sense a magnet embedded into the gear itself to index the zero degree point of the cam. to make that happen, the hall effect sensor requires a voltage to power the sensor, a ground, and then it puts out a square wave on a third wire. the megasquirt needs a small optoisolator circuit added to condition the signal before it sends it to an input on it's processor. i built that simple circuit on the proto area on the version 3 board. since i was sending my ignition trigger signals to the gt ignitors via the spare wire pads on the ms board, those 3 wires were freed up when i changed the ignition trigger outputs to the new, separate wiring harness. i used the 3 spare wires to connect the hall effect device for +12 volts, ground, and the return signal.

i'm off to identify the south pole of a tiny 1/16th" x 1/2" cylindrical neodymium magnet, locate and drill a clearance hole in a cam gear's tooth, and set the magnet into the gear with a 2 part epoxy. i'm using a stock cam gear first before i do the expensive adjustable gear. practice makes perfect! :lol:

to be serious, i'm collecting parts to build monster turbo3 version 2 revision 3 and i'll use the practice cam gear on it. :wink:

Author:  t3 ragtop [ Sat Oct 01, 2011 3:03 pm ]
Post subject:  Re: Ol' Blue 2010 Turbo Upgrade Thread

moving right along. here's the miniaturized components to detect the cam angle.

i set up my drill press to make sure that the drill table was square and the quill was plumb to optimize the angle of the magnet to the hall effect sensor's pickup. i located the tooth and the position of the hole in the back side of the cam gear, the tooth preceding the timing notch for zero degrees indexing. then i center punched the spot i want to drill.

just prior to drilling, i re-checked the twist drill diameter to make sure i had selected the right size.

all set to drill. i went 3/4" deep to allow the south pole of the magnet to project 1/4" to increase it's flux.

i used loctite 2 part epoxy to pot the magnet.

here's the magnet set in epoxy.

i used a translucent red spray over a fresh bead blast to approximate the effect of anodized aluminum in hopes that the finish will match my anodized aluminum under drive pulley. also, the adjustable cam gear i'm running on the current engine does have a red anodized aluminum center and i wanted to see if the test gear has a close enough finish to match that. if it does, when i disassemble the adjustable cam gear timing gear i'll bead blast the outer steel gear and spray it with the same paint.

i'll be working on the new wiring harness to connect the megasquirt box to the ignition and injection wiring under the hood using 3 position weatherpack connectors to match what i had set up for the batch fire configuration. it has been cold and raining today. the weather tomorrow is supposed to be more conducive to working outside so i'll pull the cam gear off the engine and mount the hall effect device on the valve cover then. as seen in the first pic in this post, the hall effect sensor has a split peg that pops through a hole, expands, and holds the package in place. it also has a smaller alignment peg so it will require marking a line between the timing notch on the valve cover and the center of the camshaft, locating the mounting holes to align the sensor's head with the arc of the magnet to make sure that the south pole of the magnet intersects with the sensor's position. i'll probably drill a third hole to mount a strain relief to secure the wiring to make sure that it doesn't contact the spinning cam gear and get sawn in half.

it's cool when an idea starts to take shape in reality. :ez_smokin:

Author:  Gigapunk [ Sun Oct 02, 2011 7:25 am ]
Post subject:  Re: Ol' Blue 2010 Turbo Upgrade Thread

A diy cam position sensor is awesome. Is it necessary to have more than one magnet/signal per revolution to increase resolution, especially since the cam moves and half crank speed?

Author:  t3 ragtop [ Sun Oct 02, 2011 9:15 am ]
Post subject:  Re: Ol' Blue 2010 Turbo Upgrade Thread

the resolution for the count is actually generated at the crankshaft, the only reason for the cam angle sensor is to synchronize the valve timing to the crank speed. one cam pulse works fine in a 2 wheel combination. all the megasquirt has to know is the cam phase to sync the crank count.

you know i have "go to" guys when i want to check myself. sportage4x4, turbohull, and matt cramer at diyautotuning have been great about keeping me straight and they answer all my dumb questions. :lol:

the sun is out today so i'll be under the hood trying to get as much done as i can.

Author:  turbohull [ Sun Oct 02, 2011 9:52 am ]
Post subject:  Re: Ol' Blue 2010 Turbo Upgrade Thread

in fact all the megasquirt WANTS to see is one count per cam rev.

Author:  t3 ragtop [ Sun Oct 02, 2011 1:24 pm ]
Post subject:  Re: Ol' Blue 2010 Turbo Upgrade Thread

turbohull wrote:
in fact all the megasquirt WANTS to see is one count per cam rev.

that's good because that's all i'm gonna' give it. :wink:

the work to mount the hall effect sensor was pretty straight forward. measure twice, drill once. :lol: i had to be careful drilling the hole for the alignment peg as the flat flange is cast in a position that would open into the inside of the valve cover if i drilled too deep. the split peg wanted the flange to be just a little thicker so to keep the package flat against the face i used a little spot of high density adhesive.

drilling and mounting the magnet was just a repeat of yesterday's exercise. i drilled a little deeper and set the magnet flush with the face of the gear on this one to make sure i cleared the plastic sensor pack. the specs say that the gap can be up to 3/8ths of an inch.

the installation was pretty much as i had envisioned it. i'll tuck the wiring into the split loom that loops over the top of the valve cover and the mod should be pretty well covered once i get my belt guard mounted.

Author:  t3 ragtop [ Fri Oct 07, 2011 6:04 pm ]
Post subject:  Re: Ol' Blue 2010 Turbo Upgrade Thread

i finished up the additional wiring harness for the sequential conversion. the mod uses a db15 connector to carry the ignition triggers to the 3 ignitor/ coil sets, 3 injector trigger wires to fire the injectors individually instead of simultaneously, and a couple of additional ground wires to the common ground point.

i have 3 more wires to relocate and connect for the hall effect device, drop in the new harness and plug the connectors together and i'll be ready to run again.

i'm still pushing the development for my next megasquirt/ turbo3 project and i've pretty much decided that the variable reluctor set up i used the first time is a can of worms to dial in. the hall effect devices automatically condition the trigger signal and although the v/r sensors work acceptably once dialed in, the hall sensors are idiot proof and will deliver a perfect square wave up to 19.500 rpm.

here's the new timing gear/ dual trigger wheel set for the next mill.

the changes are that i moved the trigger wheel to the back side of the accessory pulley to make changing the accessory belt easier and i'll use the 1" crank gear from a 95 g10, the 7/8" wide round tooth timing belt, and the late model fixed cam gear. as far as controls go, i'll use hall effect sensors on the crank and cam. i've sourced 1/2" barrel type hall effect sensors so my sensor mounting scheme will change to something really trick and fully adjustable. :wink:

Author:  Phil N Ed [ Fri Oct 07, 2011 8:37 pm ]
Post subject:  Re: Ol' Blue 2010 Turbo Upgrade Thread

Are those 'Weather Pak' fittings
I see

Author:  1987TurboSprint [ Fri Oct 07, 2011 8:55 pm ]
Post subject:  Re: Ol' Blue 2010 Turbo Upgrade Thread

The crank hall effect sensor will have to be different than the one you're using on the cam, right? Otherwise wouldn't you need a magnet in the 30 teeth on the trigger wheel?

Author:  t3 ragtop [ Fri Oct 07, 2011 10:59 pm ]
Post subject:  Re: Ol' Blue 2010 Turbo Upgrade Thread

yeah, i used weatherpak connectors on several circuits under the hood including the injectors and ignition to make sure i didn't have moisture or corrosion related electrical issues. on the previous wiring iteration running batch fire injection, i used the 3 position weatherpak connectors looking ahead towards the sequential wiring to make the change over easier. :wink:

as to the hall effect sensors for the new design, i spec'd a 13/16th" barrel type that's compatible with any ferrous metal toothed wheel. it doesn't have to see a magnet per se, it can discriminate the presence of a ferrous metal tooth if it's moving and change it's electrical state going high to low - high in a gap, low on a tooth. the same barrel type hall effect sensor also changes state when it detects the south pole of a magnet. the magnetic trigger doesn't require as small of an air gap. when it's triggered by a ferrous metal tooth, the air gap is more critical.

in part, that's why i moved the trigger wheel to the rear side of the accessory pulley, to improve run out on the teeth as they spin. the id of the trigger wheel has been machined to a very tight concentric fit to the od of the pulley's hub which maintains concentricity on the trigger wheel. there's only a couple thousandths run out at the od of the teeth - well within the .010" permissable wobble.

the new hall effect sensors aren't as cheap (by a factor of more than 10) as the flat pack type i used on the first rendition but they're going to be much easier to use than the variable reluctance sensor from an electrical/ electronic standpoint. i spent more time trying to dial in the v/r signal on the first version than i thought i should have. i farted around with polarity, sensitivity, and hysteresis trying to get a suitable trigger signal, and i don't like engineering that is that finicky.

another gripe i have with the v/r sensor is that the faster the crank spins, the larger the signal amplitude gets. the signal is in the mud at idle, at a half volt, and around 70 volts at 7000 rpm. the signal response from the hall effect device is constant at 12 volts up to 19,500 rpm from it's output lead. then the 12 volts is used to fire an led in an optoisloator package which in turn switches on a phototransistor with it's biasing set to fire a 5 volt trigger signal suitable for the megasquirt's microprocessor. the processor always sees a solid 5 volt trigger no matter what the rotational speed is and all of the signal conditioning is taken care of automatically between the hall effect device and the optoisolator. no fiddledy farting around trying to dial it in. :wink:

my next megasquirt build is going to be streamlined, modular, and slick. now that i have a handle on what the control and wiring set should be, it'll be much neater. i'll build the fet coil ignitors into the ms case along with the fet injector drivers, pare down the number of separate gauges by integrating oil pressure and temperature into the megasquirt, and using the megasquirt's pid boost controller. i will also start out building the megasquirt with a better handle on which components i can delete from the board during assembly.

there's a heck of a learning curve but the megasquirt is a true hobbyist's delight. it is so feature rich that you need to have a game plan together before you ever heat up a soldering iron. i have a way better handle on what i need from the ms now (and what i don't need.)

Author:  t3 ragtop [ Tue Oct 11, 2011 4:35 pm ]
Post subject:  Re: Ol' Blue 2010 Turbo Upgrade Thread

well, ol' blue is officially running sequential operation. i have some sync issues to work out, it seems to have a little trouble around 3500 rpm. i caught the little sync indicator flash red and it sounds like the cam makes another rev before the engine straightens up and the ms sync indicator turns green again. i have my thinking cap on now trying to puzzle out if i should play with a cap value or twiddle the mechanics of the hall effect sensor. i'm leaning towards the electronics because the mechanical end seems pretty solid.

i pulled the controller back out. i found a post at the ms forum that pertains to the usb serial converter board i'm using. apparently, if the usb isn't connected, the early manufacture ms2 daughter boards decide to go into virtual boot loader mode. that explains why my controller clears it's firmware when i try to start the car without my netbook connected. the fix is adding a 29 cent transistor. :lol:

i just received another turbo3 valve cover and i'll be building another hall effect device enabled cam angle sensor rig, version 2.0 :-P version 1.0 seems to work okay but the new design will allow for adjusting the air gap between the magnet and the pickup. it will be more industrial looking with not so much effort to keep things small.

as usual, practice makes perfect. i think my next megasquirted g10t will be a simpler, yet more refined, unit.

Author:  turbohull [ Wed Oct 12, 2011 11:05 am ]
Post subject:  Re: Ol' Blue 2010 Turbo Upgrade Thread

that's great news! I knew the usb converter had something to do with it :P

now for the sync loss. you can log it and in the log there should be a sync loss reason, this can lead you to your problem.

I remember reading that putting the cam sensor at TDC is not recommended. since your pulley has 2 positions you could try it on the other position and see if the sync loss is still there.

Page 21 of 23 All times are UTC - 5 hours
Powered by phpBB® Forum Software © phpBB Group