Caveats: To quote Kyle Reese, I didn't build the f'ing thing. I bought it used. It's a v1.01 that has (appears to have) been updated to the v2.2 flyback setup. Not the flyback board. It is currently using the original 2001 firmware, meaning: no fuel pump prime, and no PWM disable while cranking.

--- Background info ---

I had this ECU set up on my GSL-SE engine, then my 12APP, and now my 13B un-turbo. Every time, the same fuel injectors (GSL-SE 680cc, note low impedance) and a PWM setting of 75%.

When I first put the fuel injection setup in, I found that 60% PWM wouldn't cut it in a 60 degree shop. I tried all sorts of things with fuelling and nothing would allow it to start, until a lightbulb went off and I set it to 75%. Started right up

I did have to do some unusual things for cold start to get it to start in freezing temps in the morning, I recall my "cold" cranking pulsewidth was 20ms and my hot was 8. I just figured this was because I was running two injectors and not two pairs of staged injectors that would fire all at once when cranking. (This engine never saw over 80% duty cycle)

For the P-port, which was only ever used in the summer, I had to set it to 25ms and 25ms for it to start at all. It really, really wanted fuel when cranking. After a while I learned to cycle the key a few times to get extra prime pulses. (This engine never saw over about 65% duty cycle, it was really choked down!)

Now the 13B un turbo always had a bit of trouble when cranking. (It's a hee-yuuge street port, the exhausts are bigger than my PP's and the intakes close later, too) It requires a lot more fuel than the stock -SE engine did at idle and it even needs a little more than the 12APP did. (This engine sees up to, er, 102% duty cycle, but only when it really *really* gets wound out, which isn't often)

Maybe it's just me or maybe I'm just paranoid, and the cold front that came through made things worse, but it stopped wanting to start on its own recently. I'd either have to give it a snort of starting fluid, or bump the starter to get the fuel pump going, cycle the key five or six times, bump starter again, etc. For up to fifteen minutes... but it would be fine starting hot.

Finally it left me stranded and late for work. Cold and raining, natch. I happened to think about the PWM. I dug out my laptop, set PWM to 85%, and it lit immediately. Drove it to work, careful not to let revs over 5000rpm. (Two low impedance injectors on the #1 circuit, mind you! I was worried about burnout)

--- Background over, crux of question ---

This kinda *sounds* like flyback failure is gonna happen soon... but, on the other hand, did I just have the PWM settings too low to begin with? Or is trying to run two small-ish injectors without PWM disable simply too much to ask?

I checked the MegaManual. Half of the things they said to look at didn't seem to exist, but nothing looks/smells burnt. (I have a Digifant VW, familiar with electronic meltdown) I *do* have a new MS-I chip that just came in today, all nice and preloaded with the 3.0 firmware, and also with a bootloader so I can go to -extra goodness and run quad 480s like I really should be doing... but should I worry about the flyback circuit? I'd really rather not have to run resistors, it defeats the purpose of low impedance injectors IMO, and I need the car running by the 6th for a National in Detroit.

 

I have those same injectors and have quantified the opening time difference between PWM/flyback and using resistors (5 ohm 20w). My scoped and physically observed difference? Zero, no difference. Opening time of 1.0 ms worked exactly the same both on the scope (looking for the current "bump") and watching them under pressure with varying opening times. YMMV! My motivation for this was to place smaller primaries (hi-Z) on my developing 12A turbo project for more precise metering at low loads and cruise and to eliminate possible failure points.

-Mike

 

How can I tell if the flyback stuff is damaged? Or is >75% PWM in order to crank reliably normal.

 

Update...

I have the V3.0-firmware chip installed, and let's just say that having a fuel pump prime and no PWM while cranking has made starting the car almost as good as an OEM. And that only because I haven't done much tuning, just set cranking pulsewidths down to something more reasonable (15ms cold/5ms hot).

BUT.

It's lean everywhere. Does the v3 firmware calculate things different from the original firmware or is this another sign of issues? I had to add about 10% more fuel at idle, haven't touched the rest yet because I haven't had time to tune but I HAD to keep it from stalling all the time.

 

I have 10W 10 Ohm resistors on my low impedance siemens 870cc secondaries. Do you think that will make them too slow?

 

It's all relative - at 10 ohms your opening time may be a bit longer than a peak-and-hold or 5 ohm constant resistance scheme would give. However, the opening time would be consistent and I doubt that 870 secondaries would be getting anywhere near 85%+ duty cycle given your app. There's a lot of marketing hype and general BS that is just not correct. Were I you, I would determine an accurate opening time for your install (there are a couple of different methods) and see what duty cycle you observe while carefully tuning into the higher output end of things. If you get much past 85% or so then you need to possibly reconfigure.

-Mike

 

The highest duty I've seen is 55% and that was with a lower fuel pressure. Talking full throttle NA at nearly 9000 rpm.

 

I wouldn't worry about using resistors instead of PWM/peak and hold - there's a lot of market hype on the opening time/advantage but I'll be damned if I've been able to quantify it objectively or subjectively. Resistors should be fine. Remember that many OEMs use resistor packs for low impedance injectors, including Mazda. At 55% you have lots of headroom - for example, a .3ms increased opening time (I haven't observed this though I am absolutely not the end-all, be all expert on this!) would affect your DC very little.

-Mike