How to drive an injector

[The Witzl]

Only proviso to this is that some Toyota ECUs use peak-and-hold injector drivers for low impedance injectors. This could potential be any low impedance ECU that doesn't run an injector resistor pack. I have only ever seen one (the JDM AE86 4AGE ECU) but that's not to say there aren't others, just that I didn't scope the injector outputs and measure the injector impedance.
The idea is that an injector needs a high current to open, but only a lowish current to remain in the open state, so a peak and hold injector driver first provides a current "peak" (1-2ms iirc) to open the injector and then "holds" a lower current. It is possible that a high impedance injector driven by a peak-and-hold driver will open during the peak, but will shut during the hold time as the current may be insufficient to keep it open.

The opening times of the injector are improved with the high current "peak" over conventional injector drivers that just open and close the injector without changing the current during the open state. As has been stated a few times already a high impedance injector, and a low impedance injector with a resistor pack, is effectively the same thing.
A discussion (years ago) about injector opening times ended in the conclusion that for some reason there must be no actual benefit from peak-and-hold drivers over conventional, as toyota (ND) seemed to have moved completely away from peak and hold. What do aftermarket ECUs use?

Thoughts anyone?

Mos.

I can't quote on what most aftermarket ECUs use for their injector drivers, but I will agree that most manufacturers have moved away from the peak-hold injector drivers.

Back to aftermarket ECU's - most are capable of driving either high impedance OR low impedance injectors without the use of ballast resistors, so would that give any indication?

The MEgasquirt-2 has a PWM "flyback circuit" (?) with some TIP125 FETs to handle low impedance injectors. I haven't read far enough into it to know exactly what that means though

[mic*]

I am about to go out and get the reccomended Jaycar bits i need today. As i understand it, this drives the system by creating a signal (for WI inj) from a signal (fuel inj)

I just thought tho... The fuel pressure vaires with manifold pressure whereas my WI system pressure is constant... So the same duty cyle under VAC will inject a lot less fuel than than if it were under 15psi boost... The WI will inject the same vol relative to duty all the time wont it?

Not knowing much about this, or anything really; If i tune it to increase the relative duty of the WI as the actual duty of the fuel injector goes up, will this suffice?

Ben what did you do to make it work?

[jonra23]

The fuel pressure across the fuel injectors is normaaly constant as fuel pressure rises and falls in relation to boost pressure, where your water injection system will not unless you set up with regulator and return line.
Would be a fair bit of work.
Water injection volumes required to stop detonation will most likely not be linear with fuel requirements in most cases anyway.
Have heard of people tuning water injection system volumes on dyno by disconnecting knock sensor from ECU, mnitoring knock sensor for detonation and tuning water that way, have not done it myself would want to be fairly careful.

[o_man_ra23]

The Jaycar module may have non-linear programmability. Either that or you should be able to get a good compromise using the linear programming.

Cheers, Owen

[Ben Wilson]

I just ramped up the duty cycle of the Jaycar controler to compensate for the changes in fuel rail pressure.

[mic*]

I just ramped up the duty cycle of the Jaycar controler to compensate for the changes in fuel rail pressure.

Thanx Ben. Thats what i plan to do too. Was there any other method to your tune than calculating roughly what +/- adjustment the dD would achieve the desired Water:Fuel ratio? For example, on the dyno as above (thanx Jon).

...Well iv finished building the EPA (pulse adjustor) and its the first device like this ive built. Ive just got a couple of q's for anyone who has built one of these. Many thanx.

In regards to assembly; I only partially understand what the switching of each of the LINK's will effect. And, in the schematic diagram, each of the crystals has a link wire bridging it? Its not shown on the wiring diagram, or done in their job photos...?

In regards to injection; I think it was mentioned above that the Gen3 1G uses peak & hold injectors? Given that the Jaycar EPA not like peak hold injectors, I need to use their Peak Hold Injector Adaptor # KC-5395. Hypothetically had we not discussed this, how would you know whether you had a construction or trim problem vs a peak hold injector problem?

[o_man_ra23]

I cant see why your module wont work with a peak n hold setup. The ECU still produces current in its output. Your module should be able to pick up this output and utilise it for the water injection output. I think the P&H adaptor might be for using P&H output for a low resistance injector. I could be wrong on this though, does the manual say that you will require the extra module if your ECU outputs to a P&H??

Also, check for dry solders before using the module. They can be difficult to identify, but often look a dull silver in colour (your solers should be shiny, and not bulge out, but suck into the materials being bonded) or they have a dark ring which looks like a hairline crack around the pin that has been soldered. Also, rubber mount the board to reduce vibrations.

Cheers, Owen

[jonra23]

The Jaycar kits require a Peak Hold adaptor kit as they wont work with Peak hold injector systems due to the larget variations in voltages confusing original kit.

Not sure if Gen 3 1G is Peak hold as haven't played with them.

One problem I have had is that the 25D pin cable to connect the APA and the Hand Controller together has to be strait through and not null wired as most seem to be.

[Ben Wilson]

Thanx Ben. Thats what i plan to do too. Was there any other method to your tune than calculating roughly what +/- adjustment the dD would achieve the desired Water:Fuel ratio? For example, on the dyno as above (thanx Jon).

I built up a PWM generator*, then measured the actual water injector output at different input duty cycles, it wasn't that hard to come up with a rough estimate based on the rate of pressure increase of the malapassi and the change of injector duty cycle through the rev range.

*There's another Jaycar kit which comes with a duty cycle reader and a generator.

[o_man_ra23]

The Jaycar kits require a Peak Hold adaptor kit as they wont work with Peak hold injector systems due to the larget variations in voltages confusing original kit.

If you use an op-amp based input buffer outputting to CMOS level logic, you wont have a problem. I dont know why Silicone chip wouldnt have thought of this when making the kit. Its all digital logic, so it doesnt matter what level you work at, so long as you work at the same level throughout, then you drive a 13.8 volt MOSFET with the CMOS output to run the auxilliary injector. Pretty simple stuff(*). I should maybe make my own and submit it to SC.

Cheers, Owen

(* to people who arent Electronic Engineers its probably not all that simple, but its not a complex problem, and the solution has been readily available for 30 years)

[mic*]

I could be off base here Owen, as i am not famillar with all the electronics components you have mentioned, and im not an electronics engineer. But i think its to do with the massive voltage spikes at the start and end of the "hold" phase.

Ill post up some stuff later. All very new and interesting to me.

Is it easy to stuff up solder joints? I reckon ive done ok... A few are not completely 360degrees around the pin. But all have good contact to pin & board.

The instructions vs magazine photo's for soldering the dual header pins to the LCD (on the hand controller) are contradictory. Photo's show the solder on the LCD side of its PCB, whereas instruc. say DO NOT solder to the LCD side of the PCB... I have soldered to the bottom as per instruc. This was the hardest soldering job, esp since my tip had burnt away a bit... Oh which brings me to ask, it appears by photos & instruc. that the header pins are not soldered to the base PCB? They just push through? Sounds wrong. Any thoughts?

[Ben Wilson]

The header pins should be soldered to the board.

I'd suggest getting the book - It has colour photos and it makes life a lot easier.

[o_man_ra23]

voltage spikes are easy to fix with spike arresters (backward facing zener diode, and a forward facing regular diode). The spikes are due to inductive intertia, same as running a DC motor. These problems were overcome before dinosaurs became extinct, so im confused as to why they werent addressed in this kit. Possibly the unpopularity of P&H injection and the extra cost of adding a few more components and a bit more board into the kit.

Ive managed to stuff up solder joints many times with an iron that isnt hot enough, or solder that is too thick. If you waited for your iron to heat up completely, and used the thin solder that comes in the pack, and kept everything clean, you shouldnt have much of a worry. One thing to look for is tiny bits of solder that find their way between tracks (use a magnifying glass and a decent light). This will find any inadvertant shorts (took me a week to find one once). I would go by the instructions, as they will be the most up to date version, the photos could be out-dated. The header pins will need to be soldered if they are not a special solderless type (you can tell, they would have a screw or other fixing device on them). I would say solder them anyway. We had a problem with a camera at work due to a terminal header with dry solders.

It can be hard to get 360 degrees around a join sometimes. You are better to leave it if its a reasonable join than to risk overheating the board and lifting the track trying to fix it (and you probably wont fix it anyway).

[The Witzl]

If you use an op-amp based input buffer outputting to CMOS level logic, you wont have a problem. I dont know why Silicone chip wouldnt have thought of this when making the kit. Its all digital logic, so it doesnt matter what level you work at, so long as you work at the same level throughout, then you drive a 13.8 volt MOSFET with the CMOS output to run the auxilliary injector. Pretty simple stuff(*). I should maybe make my own and submit it to SC.

Cheers, Owen

(* to people who arent Electronic Engineers its probably not all that simple, but its not a complex problem, and the solution has been readily available for 30 years)

The problem here is that the "designers" at Silicon Chip have little forsight. They do not seem to be able to design a project with flexibility of applications.
..... but lets not get into that....


Owen is pretty much spot on with his construction tips.

I just built the IEBC (same PCB as DPA, just different PIC microcontroller chips), but havent constructed the hand controller for a long time now, so cant remember the method.

Anyhoo... some REAL pics of the IEBC completely constructed

NOTES:

• take note of the soldering on the back of the PCB, see how each solder pad is completely covered with a nice "cone" of solder (except one which is half covererd, BAD ME!). Dont use too much solder!!
• take care when laying out your components. Bend the legs and test fit them FIRST, then solder once its in the right place.
• do the lowest profile parts first (resistors, wire links, diodes etc), and work your way up in component heights. This avoids problems with getting little parts into the PCB right next to a massive heatsink or D25 connectors etc.
• Crystals - there is no need to do the wire linky thing that the kit instructions say to do. This method is only done to hold the crystal in place and stop it vibrating (and becomeing damaged). A better option is to use some hot glue or double sided tape as a vibration dampener to hold the crystal in place.
• take your time!!

[Ben Wilson]

And once you think you've finished all the bloody links go back and look again, you will have missed at least one....