C. Ludwig

iTrader: (27)

Here's the way I did it. There are a 100 ways to skin a cat and it doesn't need to be this complex but I wanted a somewhat automated system with a low fuel fail safe. Thought I would pass along what I've done. The ECU used is a Haltech E11v2.

I used a Hella solid state relay for the high speed switching mechanism. http://store.summitracing.com/partde...5&autoview=sku It's rated at 32A and can switch at around 60 hz. The speed and load works well with the common Surflo pump. Since the relay is designed for PWM control it makes no noise and runs cool. In short, when driving the pump through this relay with the PWM output of the ECU everything works as it should and you have very good pump speed (pressure) control.

The relay itself presented a problem that had me scratching my head for the better part of an hour. I was getting odd voltage feedback through the system and the relay just wasn't working right. Turns out the internal circuitry of the relay was the culprit. After spending some time on the phone I found out they just act this way. On a normal relay you need to feed power to pin 86 and ground 85 to close the circuit. With Hella solid state relay it seems that pin 86 pulls positive power from pin 30. What happens when the relay is in an open state is that you get positive current to pin 86 and that was playing havoc. Easy solution is to simply unplug pin 86 and remove it from the system.

Next problem was that when the ECU is powered down it dumps the PWM outputs to ground. Since pin 30 of the solid state relay was wired directly to the battery positive through a fuse when the key was turned off the relay went to 100% duty cycle and the pump came on! So one more relay had to be added to cut power to pin 30 when the key is in the off position.

The low fuel fail safe works like this. When the arming switch is activated it sends a signal to the ECU to switch to the high boost map. This changes the fuel, ignition, and boost control maps. The relay is normally closed at this point and the signal goes from the arming switch to the ECU. When the in-tank float falls with the fuel level it closes a circuit to ground which causes the relay to change state. The signal circuit to the ECU goes open which causes the ECU to revert back to the low boost map. At the same time the relay closes a circuit for the low fuel warning light mounted in the dash to alert the driver of the situation.

Haven't had a chance to actually do a live test yet. The weather has been crap the past couple days. I have done some static testing of the pump output while manually manipulating pump duty cycle with the ECU.

stylEmon

iTrader: (2)

thanks for the write up. This has been on my mind for months.

C. Ludwig

iTrader: (27)

Finally had a chance to test the system tonight. Best I can say is that it worked like it's supposed to.

I decided I wanted full spray at 20 psi and up and somewhat arbitrarily picked 12 psi as the start point. So all load sites from 20 psi and up were set to 100% duty. Picked the load site just below 12 psi and linearized from there to 20 psi. So the system comes on at 12 psi and builds progressive pressure to 20 psi.

I figured my 700cc nozzle would replace about 22% of my fuel so I started by removing 20% from the load points from 20 psi and up. I then went back to the same load site I chose to start the pump and linearized from there to 20 psi to progressively remove fuel as the methanol comes on.

The results were that it was almost spot on right from the start. It's pretty rich in transition and about a 1/2 point fat overall. Will be easy enough to tune it out.

slo

iTrader: (1)

keep in mind that the frequency is static unlike the frequency of the fuel injectors which is chaning with RPM.

I'm sure you get what I'm saying, but to make it more clear for others whom may be reading, if the halltech is set to inject fuel at 5ms, then at 4K rpm at that setting it will inject exactly twice as much fuel as it would at 2K at the same setting.

For the generic pwm map on the other hand X% is X% regardless of rpm, so the duty cycle should be ajusted based on rpm an boost to avoid having the amount of fuel substituted go down as the engine spins up.

C. Ludwig

iTrader: (27)

This is not correct. An on time of 5ms produces the same amount of fuel whether you're opening the injector at 0 rpm or 20k rpm. While a 50% duty cycle at 2k (15ms rotary or 30ms piston sequential) rpm will produce twice the fuel that a 50% duty cycle will at 4k rpm (7.5ms rotary or 15ms piston sequential). Maybe that's what you meant?



You're correct. However, under full load the PWM output is 100%. Any fraction of that is only available in boost transitions. Speaking strictly of full load tuning, with the 700cc nozzle I am using, I'm getting roughly 700cc of methanol regardless of RPM. The torque curve of this particular engine (Mitsubishi G54B, Starion) peaks as soon as full boost is seen at 3700 rpm and rolls off until the engine is done in around 6000 rpm, the total fuel demands lessen as RPM rise. Since my methanol volume is constant it actually becomes a greater percentage of the overall volume of fuel as RPM rise.

With a different engine that sees the torque output rise after peak boost is achieved the volume of methanol versus the total volume of fuel would fall as the engine reaches peak torque and begin to rise again as torque, and overall fuel demands fall. In this case you could program the pump output to provide a linear amount of methanol regardless of engine speed or load is desired.

Bernardd

Can you run this setup using a spare injector driver? Would it damage or hurt the driver?

limbar85

iTrader: (21)

Do you know if the E8 (which is what ecu I'm looking at) is able to switch between maps based on your low level sensor as well? I plan on using 2 700cc injectors so all I need is 2 pwm outputs?

C. Ludwig

iTrader: (27)

The injector driver isn't able to sink the 10+A that the pump will pull so for this particular system it's not an option to directly run the pump off the ECU. That's what made the solid state relay a necessity. If you wanted to run a single solenoid, ala FJO, you could as the current draw should be low enough. The problem with that then becomes independent control of the AI versus the gasoline injection. That's why I'm using the PWM output. It's programmable independently of the base injection maps.

C. Ludwig

iTrader: (27)

The E8 will do the same thing. The system described above uses one PWM output and one Spare A/D input.

Again, this system controls pump speed, not an injection solenoid. So if you are saying you want to run two nozzles with two separate solenoids the issue becomes more complicated. You can run one directly from the ECU in the same manner as I'm controlling the pump but two solenoids would draw too much current off a single driver. There is also no way to independently control two separate solenoids if that's what you want to do.

Bernardd

That's what I wanted to know. I have two spare injector drivers I can use with the relay.

slo

iTrader: (1)

I'm not sure what your saying here, what I said is correct, for a given pulse width in MS, as RPM changes the volume or mass of fuel injected changes. That’s why with rotary's or other engines with very flat VE curves you can inject about the same pulse width of fuel at 3K for a given load (say 10 PSI) as you do at 6k or 9k, and yet you make much more power as the rpm climbs, with the additional fuel. This isn't because the engine is becoming more thermally efficient.

Duty cycle is percentage of on time, there is no other definition of this, it doesn't change with rpm, 50% duty cycle in theory should produce the same amount of fuel at any rpm if the ecu is accounting for the time it takes the injector to open and close.

edit: this is really besides the point of this discussion but this is alot easier to do with an injector or PWM valve.

C. Ludwig

iTrader: (27)

No. That is just not true at all.

Power, horsepower, is a function of torque and rpm. hp = (torque x rpm)/5252. So, if our example engine makes 200 lb/ft from 2000 to 6000 rpm it will make 76 hp at 2000 rpm and 228 hp at 6000 rpm. All the while the fuel requirement (in pulse width, lb/hr, cc/min, however you want to figure it) is the same because the engine is moving the same amount of air per cycle at 2000 rpm as it is at 6000 rpm. You are NOT burning more fuel per cycle or moving any more air per cycle. The additional horsepower is simply created, really mathematically, by the additional rpm.

slo

iTrader: (1)

Right, you have the answer right here, but this sentence is wrong: "All the while the fuel requirement (in pulse width, lb/hr, cc/min, however you want to figure it) is the same because the engine is moving the same amount of air per cycle at 2000 rpm as it is at 6000 rpm. "

Pulse width cannot be used to determine the actual amount of fuel injected without the aditional data of RPM added, and its not equivelent to the other 2 ( lb/hr, cc/min) and cannot be converter into an actual measure of fuel without aditional data.


The fuel requirment does changes with rpm because:

The same amount of fuel per cycle (this is the key per cycle), but you don't measure volume or mass of fuel per cycle you measure volume or mass over time such as liters per minute or pounds per hour. (the key being per hour or per minute).

So if you have twice as many cycles (in a given time) then you have twice as much fuel which is how it relates to the original discussion.

So the actual fuel requirment does go up with the number of cycles, (RPM).

I think we are both saying the same thing, just in different terms, here is another way to say it. If you have a pulse width of 1MS, and a 1CYL rotory or 2 stroke engine operating at 100 RPM there will be one injection event per revolution for a total of 100 1MS injection events.

If you then double the rpm to 200 rpm but keep the pulse width at 1MS there will then be 200 1MS injection events for that same minute, you will have put twice the volume or mass of fuel into the engine with the same pulse width.

Really what you have done in your previous posts is explain why pulse width doesn't change (with rpm), and why the actual amount of fuel (LPM or pounds per hour etc) does (with RPM), which is my point.

This is all besides the original point which you understood probably before I made my first post which was that since frequency isn't tied to RPM (with the generic PWM map), the amount of fuel you will be replacing will be proportional to RPM, so in an ideal world to keep the ration of meth to gas at a steady 20% (or whatever) the map has to adjust the dutycycle by RPM and not just boost.

C. Ludwig

iTrader: (27)

So you're saying that with a given pulse width, say 5ms, your using more fuel at 6000 rpm than you would at 2000 rpm when measured in volume/unit time? Absolutely. I agree 100%. But that same pulse width will provide the same amount of fuel per cycle at 2000 or 6000. Your simply using more pulses per unit of time to create a larger volume/unit time.



Agreed. Again, if the torque curve of this particular engine were different this would be of more consequence. However, with this engine things work out rather simply when looking at this aspect of fuel delivery.