LLY Ficm

[2004LB7]

At risk of disclosing TMI, I'm now 100% certain these units are just a stand alone amplifier design. By removing the canbus connection, it verified the control wires do all the heavy lifting related to pulse width and timing. If not, it would have failed to fire. Which should be noted was tested by @1FastBrick starting with key off, and no comms to FICM.

So in that test, FICM had nothing to go by beside a crank signal and injector control inputs. Without a cam reference or some message over canbus, it has no way to know which cylinder is actually up on compression. Conclusion, it doesn't care because it' just amping up exactly what receives off the control lines. It is possible that they do some voltage control or slight adjustment to timing based of that crank signal RPM. But according to Einstein, ya can only slow time, not go backwards... So they can't remove any LAG by adjusting the pulse, only possibly ad delay. And what for? Making it easy to conclude, they followed the KISS protocol. Signal in triggers those big transistors [which if ya check, 8 are directly tied to the command pins] while those coils and caps supply the 48 volts at roughly 15amps.

Only major question I have left is, are they conditioning the signals using the onboard processor, or triggering these R4389 100V 56A MOSFET transistors directly from the control wires?? Not my particular line of expertise, but should be easy enough by tracing out the control lines. Other important part I've noticed, they split the load between two 25Amp fused circuits. Two positive pins, two grounds. Guessing if we check, they followed the firing order to allow dwell time on each coil, and distribute the load equally.

Spent a couple hours reading up on K-Line. Have a better understanding now of how it communicates. And how to verify the pins. K-line Rests at 12V, drops that to zero to wake up the node. Easy enough, noW to communicate..

On the test boards, was thinking maybe you could rob some coils, caps and transistors from a smoked board, and build us a couple soldered breadboard to test the theories right quick. I'm thinking we can have some prototypes boards ready in 30 days if things works as I believe. Just so happens, designing embedded systems, somehow became my primary job since the boats on hold.. :-(

View attachment 117533

Yeah, I have a bit more tracing to do to see how they amped the voltage for the injectors. I have a feeling they used the coils on the injectors as part of the LC circuit along with the two large capacitors. I need to measure the solenoids to see what Henry they are. I think I might still have an LB7 injector around somewhere. Otherwise I'll just have the LLY ones to test. I already know what the Caps measure at. That's printed right on them. I'll do a little more tracing when I get home.

I am guessing the 2 separate power circuits are what get turned off separately when the FICM detects a fault (short/open on injector) and goes into "tractor mode". I would not be surprised if there is one big FET per power supply so the OS can turn off the failed output without taking out a single common power bus.

Which also is evidence that the 8 control wires are directly controlling the injector FETs, because if the CPU was in control, it wouldn't need to disable a whole bank of 4 injectors if one was shorted, it could simply disable that specific injector.

It sure does look like this is a dumb amplifier board with a CPU to monitor and provide some very basic safeties, nothing more.

The two power circuits that supply the 12v to the module where originally the constant power for the "EDC" and the switched ignition. Remember this was likely cobbled together from another design and altered to fit this purpose. If you look at the wiring diagrams it lables them as such which lines up as a stand along ECM would have. They are internally connected together on the fused side and internally on the FICM board. So they probably kept both lines as the pins couldn't handle the full current. Same for the two injectors that have two wires joined together. Not enough large pins to support the current.

There appears to only be one high voltage boost circuit so that would not account for the injector bank disabling. More likely a programming choice. Maybe to keep vibrations on the engine or crank down. Disabling all the cylinders that work in unison.

 

[2004LB7]

If they are just "Dumb" boards, does that also make us "Dumb" for playing with them?

Stupid is as stupid does

 

[TheBac]

I've asked my son to look into this. He has a few engineer friends over at GM that may be able to help with circuitry info or even coding info.

 

[1FastBrick]

I've asked my son to look into this. He has a few engineer friends over at GM that may be able to help with circuitry info or even coding info.

I doubt there is much info on GM side but you never know.

When I was trying to get some info about some controllers and certain tables I was told everything was purged after 10 years the contact I had wasn't even able to look it up.

Was the same when I was trying to get a template to mount something. I ended up finding an enthusiast that had a an original copy of the template and I had to pay him $50 to make a copy of it.

 

[455buick]

Guys, this has to be one of the best technical threads I've read in many years on any of the Duramax related sites. I've been around and lurking since about 09 and I've read many many back to like 03. This one is keeping me refreshing the site quite a few times a day. I just want to say great work, it's just flat out interesting! Just to add, my perspective from what I'm seeing, I would agree it seems to only be an amplifier.

 

[kidturbo]

I'm no trained electrical engineer, just a kid who busted open every electronic toy my parents bought me, only to short the pins and see what would happen. Today we have the internet, to help us all figure out which pins to short, to get the best sparks..

Watched movie Blackberry this week. A group of computer geeks built a military grade encryption cell phone out of toys and parts from Radioshack. Then I realized, these FICM's were probably designed that same year, using many of same chips as the Blackberry.. Of course I'm game to crack one open.. lol

 

[1FastBrick]

So we need to start looking for and buying old black berries???

 

[kidturbo]

So we need to start looking for and buying old black berries???

No, but we might need to go to RadioShack..

Sent from my SM-S901U using Tapatalk

 

[DAVe3283]

I've still got a Radio Shack gift card in a desk drawer. Starting to think it is aging more like milk than a fine wine

Sent from my FlashScan V2 using Tapatalk

 

[kidturbo]

At least they were around slightly longer than the Blackberry.

Further examination of the Kodiak donor harness reveals more broken wires than previously noted. Also missing the harness piece that goes through firewall.

So, believe we will sacrifice this one for Science.. Just easier to start from the ECM and pick the wire we need to make it run on the bench. Apologies to all the LB7 Kodiak owners reading this in 2033...

Sent from my SM-S901U using Tapatalk

 

[2004LB7]

Spent some time one the LLY FICM JoshH sent me to look at. Traced out the main power, boost and switching circuits. Looks to have a pretty typical boost converter (need to double check the mosfets line location as that doesn't match the normal boost converter). On the drawing I did, the boost converter that take the vehicle 12 volts and pumps it up to 48 volts is the top portion of the circuit. The lower circuit is the injector control. I need to trace out where the ECM control signals are going. It they drive the mosfets directly or do they go to one of the chips that then control the mosfets?

Anyways here it is


I'll try and add more to it as I prob into it more. Then I'll draw it up on the computer so it's nice and neat. Then we can run some simulations and see how it works. See what happens if we increase or decrease the capacitance. Different injectors like the LB7, etc

 

[kidturbo]

Looking good. I checked all the pins with this lb7 powered up, and found an extra 5v and a low reference not pinned out in harness. Also a 12v on 52 I believe is our K-Line. Besides that, all other pins were dead air..

Sent from my SM-S901U using Tapatalk

 

[kidturbo]

Tested the couple pins i thought was K-line with a Bluetooth ELM chip and android apps, then my V3, using every known k-line ecm option, then J1939 options. Got it to respond to a single J1939 request, which it said SN = 0. That was it. So broke out the Bosch MDI , still no comms and then it blue screened my Toughbook..

Calling it a day, done for now. Off to go enjoy the weekend.

Sent from my SM-S901U using Tapatalk

 

[DAVe3283]

I need to trace out where the ECM control signals are going. It they drive the mosfets directly or do they go to one of the chips that then control the mosfets?

If it is like you show and the FETs are on the high side of the injectors, they have to be used with a driver chip. The gate needs driven to a higher voltage than the source, and we know the control pins are only 5V, but the gate/source are ~48V.

Edit: haven't had my caffeine yet, your diagram clearly shows low side switching. So they could be directly driven by the ECU, but I'd still expect a driver/buffer chip to condition the control signal to the FETs.

Sent from my FlashScan V2 using Tapatalk

 

[2004LB7]

here's a more properly drawn up schematic of the LLY FICM. fixed the error of the boost converter mosfet too

 

[2004LB7]

It also looks like this chip is the one receiving the control signals from the ECM and turns on the injector mosfets

 

[2004LB7]

From the diagram. You can see that there are two BUK9222 mosfets, one for each injector supply bank. Very likely the means that the FICM shuts off injector banks in the event of issues. I wonder if the chip above that turns on and off the injectors is still trying to operate them even when one bank is off. If so you can probably jump those BUK9222 mosfets out, or put a switch on them to turn a bank back on. One bad injector. Unplug it, flip the switch and be back on the road

 

[1FastBrick]

But the real question? Did you get your smog done so you can play with your new toy???

 

[2004LB7]

But the real question? Did you get your smog done so you can play with your new toy???

Yes!

Eating lunch right now. Afterwards I go outside and play

 

[1FastBrick]

Yes!

Eating lunch right now. Afterwards I go outside and play

Nice!