Water or Water/Methanol Injection DIY

[06YarisRS]

DISCLAIMER: The following only represents my current understanding and, as such, may not be fully factually accurate. I welcome correction or elaboration as appropriate.

I have been experimenting with adjustments to my part throttle, partial boost AFRs. Unlike a standalone ECU, the AEM F/IC does not have any true closed loop fueling capability. More on that in a minute. The car's factory ECU will fight to bring the AFRs to around stoich (14.7:1) largely for fuel economy. The F/IC will try provide a more rich AFR to compensate for the increased heat and cylinder pressures associated with ramming in more air and fuel - bigger explosions - and this is done by increasing fuel percentages in the cells of the fuel map. Once the throttle is depressed far enough, probably 90%+, the car will transition into open loop, no longer taking into consideration wideband O2 inputs, revert to pre-defined fueling maps, and allow the F/IC control over fueling. In this state, the AFRs become rich - in the area of 11.5:1 and the engine is in a safer operating mode. It seems counter intuitive; more power output, yet safer for the engine.



Although my car could be considered a low boost application, there still exists, however slight, a chance that knock could occur on, say, a really hot day, high engine temps after hard driving and increased cylinder pressures under 'high' boost, low rpms where there is the greatest stress on connecting rods, bearings, pistons etc. The 2ZR-FE engine has a compression ratio of 10:1 which is high to start with in a forced induction application (many engines built for FI are in the 8.0:1 - 9.0:1 range. So, in the interest of safety, I have been pursuing AFR enrichment in the part throttle, partial boost zones.

I have a couple of options:

Option 1: Attempt to "spoof" the O2 sensor signal in the F/IC tune map. The F/IC has the capability to basically trick the car's ECU into thinking that different AFR's - in this case richer ones - are stoich AFRs. The O2 sensor might pick up a ratio of 12.5:1 or 13.0:1, for example, but the F/IC will intercept the voltage signal and send the message that the O2 sensor is seeing 14.7. This works fine, except that the ECU will eventually get wise and start to trim out the extra fuel. There is a process to collect fuel trim data which is then put into the fuel map as percentage increases/decreases. This apparently completes the fueling charade and the ECU remains happy. Below is what I have so far. I've also included a table of what typical wideband voltages should per AFR. You will see that my numbers do not coincide, but I have managed (at least for the time being) to improve AFRs significantly in the partial boost areas. Now I am getting approximately 13.5:1 (2 - 4 psi boost), 12.5:1 (4 - 6 psi boost) and 11.5:1 (7 - 8 psi boost, when throttle position, I believe, puts the car into open loop).

My current O2 map:



Typical voltages from O2 sensor for different AFRs:



My current issue is this: I have not been able to accurately establish AFRs associated with each voltage adjustment as my Wideband AFR gauge fluctuates a bit during experimental voltage inputs. If I can't iron down steady values, I won't be able to put accurate fuel percentages in the fuel map cells. Should this approach not yield stable and predictable results, long-term, there is Option 2. But so far, it's working very well. Actually, I may go with both options.

Option 2: Install a water, or water/methanol injection system. I am leaning toward water only injection and here's why. If I can't tame the AFRs via the O2 map (reconcile the fueling in a way that the ECU doesn't get wise), the ECU would trim out any additional fuel (methanol) along with the gasoline. The O2 sensor and ECU will not detect an AFR enrichment with water only as there is no fuel component. Since, I'm not looking for power increases, but rather just greater safety (in the form of increased "effective octane rating", which water will supply), water seems the way to go. Theoretically, there would not be a power loss associated with injecting water as water expands forcefully as it converts to steam (steam engine effect). It simply dramatically drops cylinder temps, due to latent heat of water vaporization. Theoretically, a cooler more dense charge also permits more air and therefore, more fuel, but I doubt this will come into play as I'm injecting post MAF sensor.

If I go for option 2, water injection, I will be building my own system based upon what others have done and presented online. It will be a 'single stage' system, boost pressure triggered. I will experiment, but expect it to start injecting around 2 - 3 psi. I have already started to collect parts.

This is a seemingly very well built pump. It looks identical to the FlowJet pump in every way (rebadged) at $30.00 vs $95.00. I made a contact with an alternative energy company and they directed me to quality Chinese made pumps, vs the $20.00 eBay pumps. This has a completely sealed motor, unlike the cheaper open motor units on eBay. The pump is remarkably solid and surprisingly heavy and much quieter than I had expected. It is adjustable from 17 - 100 psi. It also has a built-in check valve and methanol compatible diaphragm and seals - in case I do decide to do a 50/50 water/methanol mix.

I would mount the pump on the passenger side front crumple zone frame section. It may require the fabrication of a mounting bracket - likely a flat aluminum plate.





Some fittings from McMaster-Carr. I bought three different nozzles for experimentation purposes. I also picked up a check valve which I'd placed close to the injector. I don't want water drawn into the intake under vacuum conditions. These misting nozzles meet the same specs as the very expensive 'made for water/methanol systems' nozzles that retail for between $60.00 - $75.00 CAD with shipping.



EDIT: It occurred to me that the check valve I ordered won't work. It cannot stop water from being sucked into the engine under vacuum. The check valves that are used in W/M injection systems are specially designed with a spring that has a pressure rating of around 16 -20 psi. So, I picked up a water solenoid.



I still need to source many other parts including a low-pressure (2 - 10 psi) Hobbs switch, tubing, fittings, a relay, fuse and fuse holder (may have one on the shelf in the garage already) a tap adapter solution for my windshield water tank as I'll be drawing off that. A few advantages of that are: it already have a level sensor, built in, it takes up no additional space and I don't want to give up hatch space, run wires and tubing through the car etc. If I go with water/methanol, many run winter windshield washer fluid, though there's much debate about that. There is a local agricultural business that will sell me straight methanol at $5.00/gal.

There is another option for triggering the system on. The F/IC has a switched 12V auxiliary power option based upon variables such as manifold pressure, TPS postion and rpm. In theory, I would just have to set TPS and RPM at low values and MAP at 17.7 PSIA (14.7 + 3 boost). I will experiment with this before searching for a Hobbs switch.

Hobbs Pressure Switch



Under Construction....