Homebrew EFI -- the throttle bodies

The throttle bodies I'm using are two of the four used on a Kawasaki GPz1100B2, I think this is the Aussie name for what is elsewhere called the Z1100B2, and I think it's the same system as on the ZX1100A as well. It's a bank of four 34mm throttle bodies, each with a single butterfly and injector. [ (photos, scroll down)(diagram) ]

Fitting to the XLV

The XLV is a narrow-angle (45°) V-twin, normally fitted with a pair of downdraught CV carbs between the cylinders. The frame above the engine forms the airbox, with the airfilter installed in it. Thankfully, the carbs are quite bulky, so there's plenty of room for the throttle bodies despite my clumsy engineering! The headroom is a bit restricted though -- it's not going to be easy. The carbs are 33mm venturi, but their outlet stubs are 42mm OD. In contrast, the GPz1100 throttle bodies [PDF] are 34mm throughout, with an OD of 40mm. This means, sadly, that they won't squash neatly into the XLV carb mounting rubbers after all :-(.

To get around this I plan to machine up aluminium 'adaptor rings', which will match the ID and OD of the original carbs at one end and push/glue onto the throttle bodies at the other. Here's an approximation [PDF] Similar ones will be needed at the top of the TBs to to match the rubber horns from the airbox.

Being a narrow V, nothing lines up with anything else, including the carbs. The cylinders are at 45° to each other, but the inlet ports are angled 37.5° from the plane of the head. A bit of judicious protractoring suggests that the TBs will be at 30° to each other, and their common shaft will be 73mm above the port surface (their outlet is 43mm below the shaft, so that's 30mm to be made up with the mounting rubbers and the adaptor rings). They're offset 80mm sideways between centers, slightly further apart than the stock GPZ ones. I should draw up a diagram of this too!

GPz TBs:

Inlet Rings:

The usual row-of-carbs-joined-by-linkages needs a twist in the middle, so the first step is to make a couple of brackets to hold the two TBs properly aligned and separated. I made up some prototype brackets out of steel to test this idea, and they seem to fit together okay. I'll have to make up a little piece to connect the two shafts though.

The prototypes are drilled slightly oversize and bent by hand, so they're not terribly accurate -- the bodies end up incorrectly aligned. I'll have to be much more accurate for the final version.


On the Engine:

It looks like I'll be able to use the 'pull cable' setup from the GPZ more-or-less unmodified, although I may have to reverse it since the carb is now laying down.

The GPZ idle setup doesn't really translate to the V-twin. Instead, a small block of aluminium bolted to the base of the supports an idle speed setting screw with a knurled knob. I'm not sure how to handle fast-idle yet ... perhaps a solenoid-controlled 'leak' into the MAP sensor ports?

FAR LEFT: The TB and throttle wheel seen from the underside of the carb, with the bracket connecting the TBs visible edge-on. A block of aluminium (green) screwed to the TB base with extended M6 screws (red) has a hole (blue) for the idle set screw, which pushes one of the little spurs on the throttle wheel.
LEFT: the first stage of idle set block, before I'd tapped the thread for the adjuster. It's not really made of paper, it's just overexposed.

Throttle Position Sensor (TPS)

The TPS is just a sealed 5Kohm variable resistor, mounted on the end of one of the throttle shafts. It's wired back to the ECU (via screened cable?) It seems to fit easily enough, and I like the cheesy "DFI" logo on the cover! If it turns out to be shot I'll buy a Farnell part instead, but keep the DFI cover :-)


Yeah, I know some of the photos are crap (the yellowish ones, generally). I'm only just working out how this camera thinks, I'll put new ones up as I go ...

The injectors came with the bodies, they have "JECS", "E" (or maybe "F"), "MADE IN JAPAN" and "A46-00" moulded into their tails and "C146" and "3516A" stamped into the top. They're a low-impedance, pintle-type injector.

Apparently they're a 180cc/min, used by the Nissan 200SX/LZ20B motor and the Nissan 280ZX/L28 motor, and need 2A peak and 500mA hold (thanks Tony).

From Arttu:

These values seems to be correct, according my tests.

I measured flow rate and it was about 190-200cc/min. Fuel pressure was roughly 2.7bar with GPZ1100 regulator.

In my installation maximum injector current is limited to about 2A by external resistors and currently I use 68% PWM limit, which should produce less than 1A hold current. So also current values seems to be in right magnitude.

I almost forget to mention that I also measured opening time of these injectors. It was 1.2ms at 12V voltage.

One nice thing about them is they come with fuel line 'stubs' on their inlets, rather than the more usual automotive practice of sealing to a fuel rail with O-rings. This should make it easy to connect them to the pump and regulator with EFI-rated T-pieces and fuel line.

Now, the XLV should be making in the order of 45 kW. The fuel consumption rule-of-thumb seems to be "half a pound per horse per hour", which would work out to 3.87 cc/min/kW, I think (better check that). That'd be 174 cc/min, or about 48% duty cycle for two injectors. Max duty cycle is generally ~ 80%, so that should hopefully be okay and leave some room for power increase ... up to 75 kW (not that that's likely out of the poor old beast :-) )

Of course, all this depends on fuel pressures, etc, so the first thing I'll have to do once I get the drivers sorted is to flow test them ...

homepage / motorcycles / Homebrew EFI