Sunday, June 24, 2012

Frame Build Progress

Digging through the base coat and putty layer on the tank and managed to find these underneath. It seems the emblems have found their rightful place.

Seat frame's done. Pretty sturdy to support my weight.

Underneath the seat frame is the housing for the battery and other electrical. Don't think i can put my MiniMS ECU there but we'll figure out a way later.

The rear fender is finished as well. It's as sturdy as the seat frame but the shape is still awkward for me. Will ask to change it later.

A view under the rear fender. The provisions for harness holder have been made as well.

Saturday, June 23, 2012

Vixion Intake Manifold and CLT Sensor Fitment

With the workshop still slowly progressing the frame build, the engine was taken to a machine shop so that preparations can be done to fit the Vixion intake manifold and the water temperature sensor (CLT). Intake manifold and CLT sensor are the only items that directly attach to the stock engine.

Surprisingly, Vixion and Tiger's manifolds have nearly similar mounting bolt spacing. Here's the manifold side by side.

One of the holes need only enlarging inward by 2-3mm as shown below.

Here's the intake manifold fitted. There' a vacuum line from it that, if i'm not mistaken, on stock form is used to connect to EGR (Exhaust Gas Recirculator). That's nice because i can connect a vacuum gauge there. Or maybe if i want to go forced induction later, it can be used for another MAP sensor.

Next comes the CLT sensor. On Vixion, this goes to radiator water line to sense the coolant temperature which in turn informs the ECU of the engine temperature and allows the ECU to adjust the amount of fuel to inject. The question then will be: where to put this sensor if the engine is air-cooled like mine?

There are many location to choose from to mount the sensor. You just need to make sure that the location gives the best temperature reading and that it's stable and does not get easily cooled by air flow. Some chose to mount on the cooling fins but this is not recommended as the fins should be a few degrees cooler than the engine block. Another option would be to drill the cylinder head and mount the sensor to read the temperature around the valve and camshaft. I prefer to read the engine oil temperature and mount the sensor to the drain bolt.

The drain bolt wall is thin and care must be taken when you drill and tap the hole for the sensor. It's better if you can add some aluminum before to give you more "meat" to hold the sensor.

The sensor will protrude a bit and foul the oil strainer but it's still safe and will only leave minor bent. If you want to be safer, you can put 2-3mm of spacer ring before you bolt the sensor. Don't forget to apply some sealant to avoid oil leak.

Here's the sensor on its final resting place.

CDI Module Mod for Megasquirt

There are two options for the ignition on this EFI conversion: CDI or igniter. Personally, coil & igniter, or inductive ignition, is preferable as it outputs bigger spark energy. Besides, I still have my 4A-GE igniter & coil that i can use after last year i converted to EDIS. However, i'm a bit concerned with the power consumption of inductive system. I'm not sure after i add the fuel pump, oxygen sensor heater, injector and ECU, the alternator will still have enough juice to retain charging and supply the ignition coil and the rest of the electrical. I'm even considering of converting to LED for the lighting just to save power. With the same consideration, i will first try if CDI will work.

If i can somehow successfully trigger the stock CDI with Megasquirt, the alternator will have less load. Unlike DC CDI which is supplied by the battery and alternator, Honda Tiger GL200 uses AC CDI which has a separate charging coil. This means it will even spark without battery and alternator.

There are two problems with using the stock CDI on Megasquirt. First is that the stock CDI is triggered by a pulse generator which is basically a VR sensor inside the left crankcase cover which means that the CDI works with sine wave (analog signal). If i want Megasquirt to trigger the CDI directly, i have to somehow make the CDI digitally triggered. The other problem is that there is already a timing advance curve inside the CDI. How can the ignition timing be controlled with Megasquirt if the CDI is already in control of it? The stock CDI and pulse generator system will run the ignition at 10 deg BTDC during idle (approx. 1400 RPM) and slowly advance the timing until 32 deg BTDC at 5000 RPM. To fully control the ignition with Megasquirt, i need to bypass the timing advance system inside the CDI.

Both of the problems i pointed above will require fiddling with the CDI's electronics to solve. This means dissecting the CDI case until i can reach the components. I bought another Honda Tiger CDI for this experimental purpose.

After hours of agony, pain and swearing.. i could only get this far. The manufacturer has really gone all the way in making sure it's a PITA to get to the CDI circuit. That's rock-solid resin covering the components.

Luckily, the bottom resin layer can still be pried off with a knife so that the soldering is accessible.

Before i go explaining what i did to the CDI module, i'd like to share my findings when i was researching about CDI systems especially AC CDI on bikes such as the one on mine. Based on my research, here's a block diagram of AC CDI system.
I won't discuss in detail the function of each block. There's plenty of info for that on the internet. I just want to point out that i need to bypass the Time Delay and Trigger Signal Conditioning so that i can trigger the SCR directly from Megasquirt. The Time Delay is what allows the CDI to control the ignition timing. It's usually done using an RC (Resistor-Capacitor) network which delays the input pulse from reaching its SCR triggering threshold. Trigger Signal Conditioning is sometimes omitted but it's there to get more precise timing control. With Megasquirt, Signal Conditioning is not necessary as the ignition trigger signal is already a good square wave. Keep in mind that my ability to read electronic schematics is barely adequate so take my words with a grain of salt.

So, with that knowledge, i was able to ID and guess some of the parts on the CDI module and made the mod as below.
With just one wire, i made a direct connection from the pulse generator input to the SCR. I utilized the 470 ohm resistor just to be safe and cut off the track going to the caps. If everything goes according to plan, this CDI is now triggerable from Megasquirt using a 5V pullup resistor from LED D17.

Thursday, June 21, 2012

MiniMegasquirt Mod as Bike ECU

The brain of the EFI control will be a prebuilt MiniMegasquirt ECU i won through Ebay auction last year. Quoting from the website, MiniMS is a DIY Fuel Injection Controller based on Megasquirt Ver 2.2 combined with Relays and an automotive connector housed in weatherproof enclosure. MiniMS is hardware compatible and uses Megasquirt Ver 2.2 software, circuit is identical. This design is for those that need a sealed self contained MS.
See those three little black boxes? They are on-board relays. This is what i really like about MiniMS, it allows for a really neat wiring as there will be no relays to wire externally. It even has the fuses inside also.

Since it's based on Ver 2.2 schematic, it inherits some the specs which require me to make some minor changes as they're useless for my particular application on single cylinder motorcycle.

First of all, it uses an on-board MAP sensor MPX4250. This means there should be an access hole for a vacuum line to plug to the sensor. Not really a pretty sight. The Vixion throttle body assy i'm using already has its own MAP sensor that i'm planning to use. Herein lies a problem: the so-called automotive connector is a weatherproof 23 pin AMPseal connector and there are no readily spare pins that i can use to wire a MAP sensor signal line on it. Luckily, there is a solution. On the original design, pin 18 is actually for an "isolated ground from coil". This is meant for the screen wire from ignition coil to minimize interference. Since i'll be using an LM1815 VR conditioner externally, this pin is not necessary. I traced this pin to the PCB and it's not connected to anywhere so it's safe for me to remove the on-board MAP sensor and put a jumper wire from R2 (refer to MS v2.2 schematics) to pin 18 and that's exactly what i did. Keep in mind that for the sensor's +5V reference voltage, you can use the same pin for the TPS (pin 5).

MS v2.2 is not meant to control ignition without some mods. As such, i had to find another pin i can use for the ignition output as i want to avoid having a wire coming out of nowhere from the case. I chose to use the 2nd Injector Bank input on pin 23. I'll only be using one injector on this conversion so a 2nd Injector Bank is not necessary. To achieve this, i had to remove Q7 (IRFZ34 driver) and put a jumper wire from LED D17's cathode to pin 23. Since i'm planning to use my 4A-GE igniter module, i had to come up with a 5V pullup using 1K resistor. The 5V source i'm using is from C3 which is supplied to pin3 of the on-board MAP sensor. Then i wire R17 directly to ground to avoid floating the MC34151's pin 5 output. I might also be able to use the stock AC CDI module this way by wiring this ignition output to the pickup coil's input pin, triggering the internal SCR with a 5V pulse but i need to remove the timing capacitor first. I don't know if this is possible yet, so i'll update you on this later. Also, since pin 9 will no longer be used to supply 12v to the injector, i can use this for the O2 sensor's heater and LM1815 VR conditioner power supply.

Last thing i did is to put an SPST momentary switch for the bootloader jumper pins and mount the switch to the case. This is not really necessary especially if you're not flashing the CPU many times but knowing this project is still experimental, i figured it will be useful.

Below you can see the jumpers and the switch mounted to the case. Note that Q7 is still there and the ignition output wire is not yet wired to the AMPseal connector. You can also see the the 1K pullup resistor and the jumper for MAP sensor signal from pin 18.

One note about using the LM1815 VR conditioner externally to the MiniMS. Unlike the suggested method of wiring the conditioner's output to pin 6 of the 4N25 optoisolator (base of the internal NPN transistor), i'll be wiring the output to MiniMS's pin 4, the coil (-) input of the AMPseal connector where eventually this will go to pin 1 of the 4N25 (anode of the internal LED) via R10 (390 ohms) while the Wing (D5) and John (D8) diodes are jumpered. This has been done on my 4A-GE before i switched to EDIS so it should work. Best of all, this method retains the opto-isolation as i'm not wiring the input directly to the CPU so it's safer.

To summarize the changes, i made this diagram below.

With all the changes done, here is the final pinouts of the AMPseal.
1. Injector 1 +12v
2. Fuel pump output +12v
3. DB9 pin 2  RS-232 data T1 out
4. Coil – terminal ---> INPUT from LM1815N
5. TPS +5v ---> Shared with MAP Sensos
6. Idle output +12v
7. TPS signal
8. Coolant temp sensor
9. Injector 2  +12v ---> Power supply to LM1815N and Oxygen Sensor Heater
10. DB9 pin 3  RS-232 data R1 in
11. DB9 pin 1  +5v
12. Ground
13. Ground
14. Oxygen sensor
15. Manifold temp sensor
16. +12v in battery
17. Ignition switch +12v
18. Isolated ground from coil ---> MAP Sensor Signal Input
19. Ground
20. Ground
21. Ground
22. Injector 1 -
23. Injector 2 - ---> Ignition Output to Igniter or CDI

Wednesday, June 20, 2012

Poke On A Budget

Before we get all the haters rolling in, understand that i did this just for the fun of it. I'm basically going to sell my Axis OG and need something to fill the fenders until i collect enough money for better wheels. That's why cost was priority for this.

It all started with these steelies lying on my frontyard.

One set was from The Dog when i first got her and the other set was from my KF40 truck. They're all are 14x5jj. Mom had been bugging me to get rid of them so i thought instead of selling them for peanuts, why not make something out of them?

The idea was to make a set of widened steelies by combining both sets. One set will be the barrels and the other will be the outer lips. Running tubes is never a good idea for me so the end result must still be able to run tubeless.

To make the outer lips, start by machining the inner lip off.

Afterwards, grind off the welds that hold the center to the rim and get that center off with a BFH. It surprised me how very little welds are needed to make a steel wheel. There're basically only tack welds at four corners of the center. The press-fit fabrication must have really helped on keeping a steel wheel strong.

With the center off, you get the outer lips..

Making the barrels are much easier as you retain the centers. You need only to machine off the outer lip.

To ensure the barrels and lips stay centered when they're combined, the edges are actually made into steps, ensuring both edges match when they're combined. It's a bit difficult but below you can still see the step edge.

Then comes the combining and welding. The welds were ground smooth and rewelded a couple off times, while carefully checking for pores that could leak air.

Once done, i went to check for the fitment on The Dog. Here's how the front looks.

Crazy poke, huh? Keep in mind that i had to use about 30mm of spacer to fit them as my RX7 caliper would foul the wheels without the spacers.

A few hours and two cans of spray paint, they look easier on the eyes. They're 9.5JJ rears and 9JJ fronts. 9.5 inch is the maximum width you can get with two sets of 5JJ since getting the lips off would reduce some of the width.

Since budget is priority, i decided to get 185/60 R14 used tires, cost me only 50 bucks. Now here's the fun part. Stretch tire fitment has always been a challenge, especially here in Lampung where not many guys have done it. I tried first using the brake fluid method but it didn't work. Perhaps i didn't use enough brake fluid. Luckily, the machine shop has this awesome "air blaster" made, as i like to call it.

The final result, 185/60 R14 on 9JJ.

I let a fellow Daihatsu Charmant Community member and friend of mine, Jefry, borrow these wheels. He's the fellow who's going to buy my Axis OG. He just sold his 14 inch alloys and with the money, bought a set of new 185/55 15 inch tires to fit my Axis OG. We'll later exchange the wheels after he pays me. I chose to keep my tires as i'm planning to use them for the new wheels later. His Charmant is stock, at least around the brake and suspension area. So the next pics will show you how these steelies will look under stock Charmant


Tuesday, June 12, 2012

Yamaha V-ixion EFI Parts Specs & Pinouts

Just got all the Yamaha V-ixion (or Vixion or FZ150) parts i need for the EFI conversion. Knowing the specs and pinouts of the parts are mandatory as i need them for the wiring later and also to map them to Megasquirt so i searched the web for the manual and also measured the parts myself.

Here's the specs i managed to gather.

Throttle Body Assy
Vixion's TB assembly is made by Mikuni and on it you can find the following:
- Intake Absolute Pressure sensor (IAP)
- Intake Air Temperature sensor (IAT)
- Throttle Position Sensor (TPS)
- Fast Idle Solenoid Valve (FID)

The TB valve diameter is 28mm which is good news since my Honda Tiger GL200 carb is only 26mm, ensuring that the engine won't choke. Judging from the amount of "meat" on the TB inlet, i'd say you can ream out the TB to 30mm safely.
FID is on separate housing from the other three, which is the upper part from the pic above. TPS, IAT and MAP can be found on the lower part and form what Yamaha calls Modulated Air Quantity Sensor (MAQS). Never take off the lower part unless you know what you're doing as you may need to adjust the TPS position. Click here if you want to see the TB without MAQS and FID.

Fast Idle Solenoid Valve
There are two pins from FID and it doesn't matter how you connect them to the ECU. I can't really tell exactly for now, but i'm willing to bet that it's just a simple on-off valve, not the PWM type.
You can also find the idle air screw on the other side of the TB. My initial plan for now is to only use FID for warm up enrichment, perhaps via a simple manual switch or use the B&G On-Off setting. That way, just like a carb choke, i'll only activate it on cold starts. Let's see if that works.
On normal condition, you'll measure around 31.5-38.5 ohms between the pins. Mine is 35.7 ohms so i'm safe.

Intake Air Pressure Sensor
Intake air pressure sensor or MAP (manifold absolute presure) sensor can be found inside MAQS module. MAQS module has five pins labelled V, T, G, M and A as shown below.
V is the 5 volts reference input for TPS and MAP while G is the ground pin for all the sensors. MAP sensor output is on pin M. When supplied with 5 volts between V and G, pin M will output approximately 0.789 volts@20kPa and 4.000 volts@101.32 kPa. That's already two mapped values so i won't have any problem making Megasquirt's kpafactor file there.

Below here is the internal diagram for MAQS module.

Throttle Position Sensor
TPS output is on pin T which will output approximately 0.63-0.73 volts at closed throttle.

Intake Air Temperature Sensor
Between pin G and A, you'll find the IAT sensor's resistance. Unfortunately, the manual only says that it should measure around 5.7 - 6.3 kOhms without any further info on the temperature (WTF Yamaha?!). Funny that i got only 1.78 kOhms at room temperature.. I guess i'll have to measure myself to find the values for Easytherm input. It'll be a bit challenging as i can't take off the IAT sensor from the TB. Perhaps 30 minutes in a refrigerator and a thermometer would help. I have managed, however, to find Yamaha XF50 scooter manual that says 5.7 - 6.3 kOhms at 20 deg Celcius while another Yamaha bike manual, YP250, says 2.21 - 2.69 kOhms at 20 deg Celcius. If i have to make a choice, i'd follow YP250's at it's closer to my own measurement.

Worth mentioning is that the Yamaha Vixion manual i'm referring to may be outdated. I downloaded it from our local Yamaha website where i got a Vixion part catalog as well. When i ordered the EFI parts using the part numbers i found on the catalog, i was informed that some of the numbers have changed. TB assy is one of the parts that changed so perhaps the IAT sensor changed as well.

I googled "2.21-2.69 Air Sensor" as keyword and found more values for Easytherm input!
> -20 deg C = 13.6 - 18.4 kOhm
> 20 deg C = 2.21 - 2.69 kOhm
> 60 deg C = 0.49 - 0.67 kOhm
> 80 deg C = 0.29 - 0.35 kOhm
It seems that i'm on the right track. This post about Yamaha WR25X mentions that it should be around 10 kOhm at 0 deg C and it falls to the correct range on the values above.

Coolant Temperature Sensor
The CLT sensor looks quite similar to the one on my 4-GE engine.
On the body, you'll find 3P20 and 179700-0480 markings.

It has only two pins and it won't matter how you connect them to the ECU. The manual says it should be 310-326 ohms at 80 deg Celcius.

Using the markings on the body, i managed to find more specs for this Denso sensor from the Yamaha YP125 manual.
> 2.32 - 2.59 kOhms at 20 deg
> 0.310 - 0.326 kOhms at 80 deg
> 0.1399 - 0.1435 kOhms at 100 deg Celcius
Sweet, three mapped values, good for Easytherm input. When i measured at room temperature, i got 1.54 kOhms so i guess i'm good.

Fuel Injector

This is a 6-hole injector with 3C11 marking on it. Unfortunately, i can only inform you that it measures 12.5 ohm on the pins, meaning that it's saturated type. You can connect the pins anyway you like.

I'm still looking for the exact flow rate but i'm guessing it's around 100 cc/min. If i plug that value to this calculator, it will be enough for up to 15 hp which is just about the same value as Vixion's advertised hp. It's barely enough for my Honda Tiger as it outputs 16.2 hp. Any engine mod done later will ensure the need to upgrade the injector.

Fuel Pump
This is an in-tank fuel pump with integrated fuel pressure regulator (FPR) which will keep the fuel pressure at 250 kPa (36.5 psi), a bit lower than the typical car engine. The FPR is kept in place with the black plastic clip. The fuel level sender is also integrated to the fuel pump assembly.

There're four pins found, designated as the following:

The fuel float is attached to a variable resistor that measures 7.4 ohms at full and 95.1 ohms empty on the pins. Well within the manual specs of 4.0 - 10.0 @ 20deg C full and 90.0 - 100.0 @ 20deg C empty.

You can pull the FPR unit out after you unhook the black plastic clip that holds it in. This is the FPR unit. It has A3H1I6 markings on it. By the looks of the fuel outlet, it seems inapplicable for external use.

This is where the FPR plugs in. You can see two rubber 0-rings sealing the FPR.

The fuel pump itself seems applicable for external use, with the input on one end and the output on the other. It has the following markings:
11082906 C1
Made in China

Last but not least, here's how the fuel flows on the pump assembly.