Suzuki throttle body injection (TBI) system (It's no
Samurai , nor
any MPI , multipoint
1989 to 1995 (Sidekicks and Trackers) 8 valve TBI engines, only. Throttle body injecton, or Single Point Injection SPI = TBI
At wits end?, engine runs bad, and you find the injection pulse widths are all wrong at hot idle? 1.4 mS is normal at idle 1.4/1000 seconds. (and all sensor read correctly?)
To someone, that said to me, wow, my car runs (kinda ok) with the Injector Red wired shorted to the battery Plus terminal. (or B17 hack wired to ground)
Sure it runs, but running and getting 15 MPG is not running correctly is it?
Most tech's discover, the rich running engine via a variety of tests (and symptoms), the simplest being, that the motor never holds closed loop status (OBD1/2 Scan tool discovery)
The Technician (any ASE) then connects his scope to the red injector wire connector, this wire, is the injection rate wire, the yellow is only and enable pulse!
He discovers that the wire is dead. and that the static timed pulse on the yellow wire, is the only means of injection, he knows this is wrong.
He finds that a customer added a hack, an illegal hot wire jumpered from B8, to the battery plus or IG-COIL fuse (etc).
He removes this hack, and the engine dies. (so some one was hidding a bad ECU !)
Now I know why the simple NOID lamp passes, but in fact the injector Drive is 100% wrong ! (keep in mind either injector transistors below, can short and do the same thing (or open and make a dead NOID lamp test)
That is why, we have the advanced NOID test, here, be sure to test both wires for pulsing actions.
The Suzuki Sidekick does not use a full time HOT Red pin A8 pin ! Nor does the Samurai of any kind, but only a Metro/Swift runs a hot 12vdc Resistor block.
The Samurai FSM books have this crazy drawing#1 below, ( you need to ditch this drawing, for Sidekicks ,really !)
Drawing #1 (bogus to the n-th degree) Seen in some Samurai books and is wrong.
This G10 drawing is dead wrong, for Sidekicks, and G16 Trackers of any kind, but is closer to Metro/Swift 1.0L TBI seen below:
The Suzuki G16A TBI , uses this exact method to inject below , this is my reverse engineered drawing. Click the drawing below, to ZOOM it.
Drawing 2: (my drawing !, this is accurate at the pins stated) As you can see it's a DUAL driver and drives ground and power, and both are PULSED, or it will never work correctly.
click to zoom me.
The RED WIRE B8 , LOOKS LIKE THIS ON A SCOPE This is the action of Q101, transistor. A real photo on a 1991 TBI. ( the scope don't lie)
If I accelerate or cause the MAP vacuum to drop the ECU responds on the RED wire with the below ACTIONS !
Note how the injector is cooled with the fast Negative pulses, and that these PW pulses, add up, as you gun the throttle (engine load, MAP changes)
This is the sure sign of working injection. (It is very hard to photograph this , so I made a drawing, to show these actions.
The injector does not close, at the PW fast plus pulses ! (the cooling cycles do not effect injection rates ) (also, the fuel is the other coolant !)
Truly the below facts are all you need. Check it carefully. The whole point of fuel injection, happens , right here !
The RED wire again. B8. (my hand drawing is the combined scope views, fast and slow sweeps. making the scale below wrong, but numbers are correct)
The yellow wire (enable) looks like this. The Low signal is the Enable signal .
The purpose of this signal is to protect the injector if Q101 shorts, or some nincompoop, shorts the Red wire to the battery. Dah?
This one OHM coil Injector burns to a crisp in 1 second at full 12vdc power, full time.
Math: (Ohms law) If you put full power to the injector. here.
P = I² x R or P = 12 x 12 x 1 , given I =12 amps (I=E/R) then 12 squared is 144, or 144 watts, a hopeless overload.
The Yellow wire is the protection signal, if the injector shorts. it protects the $1700 new ECU, too. (not a bad idea, huh?)
The Transverse connected. RED to Yellow wire signal, looks like this, this is what your NOID tests sees.
Sure the car runs, but will run rich and the ECU will go quite nuts, due to having no control over fuel injections.
The ECU may jump in and out of limphome (failsafe) and loose idle controls and run RETARDED, spark advance tables.
Last my injector flow rates. Before and After cleaned. ( we do not have flow rates /versus time spec from Suzuki, that is why I made the drawing above.)
The next drawing is to help , Metro owners that land here by accident. (goggle finds this) The same rules apply, no Spark errors, no CMP errors. (DTC errors) And the Resistor block must be good. (I think the injector sees 1/3 the volts (5v) DC fired , )
The Metro/Swift dead noid can be that resistor block, (named wrong) here as pump.
The Metro injector resistor block is on the top right of the firewall, looks like a heat sink with a yellow clip attaching two white wires coming off of the resistor to a white wire with a blue stripe and a yellow wire.
2.3 to 2.5 ohms was measured across the resistor , by one person.
If the resistor burns up, all injections stop. Infinity readings, is bad in all cases.
The TBI injector coil, on the metro is 0.5 to 1.5 ohms.
The resistor block was used on all TBI metro, and never MPI engines of any kind and used on G10 and G13 engines with TBI, all MPI engines are free of this block.
The Metro MPI injectors read . Resistance should be between 12.0 to 13.0 ohms.
At the Metro, injector yellow wire gets 12vdc key on, not running, and then goes to less than 1/2 that, firing. (only a scope can see that lower voltage durning the 3mS pulse. ( Current flows at near 5 amps, and the 1ohm injector drops about 5 volts, using ohms laws)
The noid lamp still must flash in place of the injector.
The noid must not glow key on. (not running)
If the Yellow wire is not 12vdc keyon, then the resistor is blow, or the main relay is bad, or the "FI" fuse is blown.
Key on both sides of the Resistor must be at 12vdc (not running). Or the problem is obvious.( meter in hand)
ver 4. 4-8-2013 ,