WTF burnt wire content

[kam-325i]

Welcome back Brian !!!

[craigieeb]

Have you checked inside the coil wiring box on the loom Craig? Its the only place left

Yeah iv had a look around in there aswell and it all seems in order,
No split wires etc, it all still looks factory
Sort of stumped tbh.
In a way, I hope it is the ecu just so I know what's caused the problem

[craigieeb]

As far as I can see, you haven't found out why the original driver transistors failed.
Have the replacements gone the same way, or running very hot?
These transistors are being used as a switch in his application, and, if they were a perfect switch would dissipate no heat at all. Example - assume that a coil draws 2A at 12 volts. With the transistor turned off, there will be 12 volts across it, but no current flowing. Heat (watts) = voltage x current = 12 x 0 = 0. With the transistor turned fully on (won't quite happen in practice), there will be 2A flowing through it, but 0 volts across it. 0 x 2 = 0, so again no heat.
If the circuit driving the transistor develops a fault, and only turns the transistor half on, then we now have 6 volts across it and 1 amp flowing through it. 6 x 1 =6, so we now have 6W of heat being produced by the transistor, which it won' like for long!

Your right, I still haven't got to the bottom of why they went in the first place.
When I plugged the ecu back in, after replacing the igbt's , as soon as I switched on the ignition, the case of the ecu just got hot, within say 3 seconds, then it was off with the ignition,

So im not sure if they have blown again or not, chances are they have gone, as these only have a heat range of -40 to 140 deg ish iirc but the case felt damn hot!

Would like to get to the bottom of it asap, just incase the ecu is not at fault

[craigieeb]

Come to think of it..
Im sure I still didn't hear the icv or the fuel pump prime when the ecu was re connected.
So there must be something else away in the ecu,
grrr!!

[Brianmoooore]

I know absolutely nothing about the circuitry used by VEMS, but would expect that all the coil driver transistors should be firmly off if all you did was switch on the ignition, without rotating the crank to provide drive pulses for the coils. The days of ignition systems overheating if the ignition was on without the engine turning should have ended in the early '70's, with the introduction of TAC systems!

[Speedtouch]

It does sound very much like a short somewhere in the ECU. Are you sure you soldered the replacement transistors in correctly, without causing any solder bridges?

Sometimes, all it takes is a loose flake or very thin whisker of solder to cause a short.

Also, how did you mount the heatsinks? Some power transistors require a mica insulating spacer, or you short the collector direct to ground, which would cause such issues.

[craigieeb]

I know absolutely nothing about the circuitry used by VEMS,

That makes 2 of us

i dont protest to know a lot about it, but i do know it makes the car work, sometimes

It does sound very much like a short somewhere in the ECU. Are you sure you soldered the replacement transistors in correctly, without causing any solder bridges?

Sometimes, all it takes is a loose flake or very thin whisker of solder to cause a short.

Also, how did you mount the heatsinks? Some power transistors require a mica insulating spacer, or you short the collector direct to ground, which would cause such issues.

yeah the soldering went straight forward, with no bridges or anything remotely "dodgy looking"

the heat sink works like this..

there is a strip of heat sink erm.. pad of some description, on the inner side of the case,
then i gave all the igbt's a nice plop of heat sink paste,
then there are strips of alloy that go over the top of the igbt's, with a hole in on either end and they hold the igbt's to the side of the case via screws which go in from the outside

ala......




not my board, as mine is under armed guard on it's way to Oxford : )
but it's the same, more or less,

thats the way they come from VEMS i do believe

[Speedtouch]

Well, that all looks very proper.

However, it may be worth fitting insulation pillars to the screws that hold the transistors:



That way, they can't earth via the screws.

[gromgsxr]

In megasquirt the back of the igbt's need to have mika insulation between them and the heat sink, wondering if that's the case hear. If its grounded straight through the heat sink then warmed up and pop ....

[Speedtouch]

Indeed. See this page for a similar setup:

http://wiki.diyefi.co.uk/Fairchild_ISL9 ... ion_Driver

[craigieeb]

I dont think there is a problem with the heat sink,
its more like there is too much voltage coming from somewhere, and thats overloading the igbt's,

now the only place i havnt checked, baring the rest of the ecu board, are the wires running down to the coil plugs, as they are encased in hard plastic, and chances are if i cut into it with a Stanley or scalpel, I would end up cutting the wire,
but
I suppose thats the only place I havnt checked

[Speedtouch]

I would try running the ECU outside of the case for a couple of minutes, just to rule out the heatsink shorting-out theory.

Use a multimeter to measure the resistance of the coil pack, to ensure it's within the maker's spec.

[craigieeb]

I dont think it would last a couple of minutes without blowing up, besides, it's wrapped up and on its merry way to be tested, down south

I'll fish out the old multi meter, find out what there supposed to be and test them

[Speedtouch]

The IGBTs should be able to handle a couple of minutes running without problems - but even just a few seconds of running would be enough to confirm the ECU works outside of its casing.

[craigieeb]

Right, yet another week on...

the ecu has been tested and it was found to be in safe mode, infact here's the explanation from gunni....


So.

I have tested the ecu, first thing I noticed is that it was in safe mode for some reason. Which happens when one does a firmware upgrade but doesn´t work out completely. I re-sent the latest config and proceeded to test the ignition, injector and auxiliary outputs of all worked right.
Also all the inputs.

I then sent the ecu some RPM signal and it worked as well. Easily surpassing 10000rpm.

The charge time for the coils is 3.33ms which will give the best spark at every point but not overheat the coils, what must have happened was that during the config upload when the ecu is supposed to turn all the outputs to OFF, it didn´t happen and it over did something in the system.

As far as I can see everything is just fine now and should not cause you any issues.



So me replacing the igtb's was the right move, but the software needed to be put right by the expert


Anywhoo I got the ecu back today,
put the engine harness back onto the car,
connected the ecu,
connected the battery,
turned the key and there was no icv buzzing and no fuel pump

guess who forgot to put the dme relay back in

so with that in, it started, sounds like its running off on 1 cylinder,
but I can easily sort that............................





IF IT WOULD START AGAIN!!!


i switched it off to take the coil packs off to test them and put some fresh spark plugs in,

now when i turned the ignition back on, there was only 1 click of a relay, from the fuse box i do believe,
but no icv and no fuel pump
and also no ecu display on the laptop again,

now this is saying to me, there is power getting to the fuse box, but not getting out to the ecu for some fekin reason!

im on the verge of just ripping the fuse box out and all the wiring and just replacing the lot,
something's not right somewhere.

[Speedtouch]

Ah, the joys of aftermarket standalone ECUs!

I presume you have a fuse in the power supply to this ECU? I'd be checking that as a first port of call, then all the earths.

[craigieeb]

Put the ecu to one side,
why wouldnt there be power coming out of the fuse box, when all the fuses and relays are fine?

[gromgsxr]

if the m50 loom is set up the same as the m20, the engine loom is separate, the icv injectors and ecu are powered off the main relay, the fuel pump relay is powered by the main relay and runs the pump when the ecu grounds the relay to flick the switch, there is a seporate relay for the lambda sensor, and the coils get an ignition live straight from the ignition switch. there should be the 3 relays out of the main fues box for all that lot.

[bigbird76]

Not sure if it has been mentioned but a dodgy ICV will melt the transistors in a ECU, Happened on my spare M3 ecu soldered in new transistors and that fixed it. I read somewhere that a 5amp fuse into the ICV wiring would stop the ecu getting fried if the ICV goes faulty.

Leigh