No announcement yet.

Burnt up ECU (pics inside)

  • Filter
  • Time
  • Show
Clear All
new posts

  • Burnt up ECU (pics inside)

    Well, I went to install ym Turner 2.5 EVO big green chip and opened up the ECU and much to my surprise I found this.

    Anyone wanna give me a clue what happened here? car was running as early as 2 weeks ago. Don't know when this happened, but it had to be after I did the 2.5 conversion cause it wasn't there before.

    could this be the dreaded ICV/ ECU burn out? Think my ECU is fried? It ran like I said, so who knows.

    The Chuck giveth, and the Chuck taketh away

  • #2
    Re: Burnt up ECU (pics inside)

    Originally posted by Highrever
    Well, I went to install ym Turner 2.5 EVO big green chip and opened up the ECU and much to my surprise I found this.

    Anyone wanna give me a clue what happened here? car was running as early as 2 weeks ago. Don't know when this happened, but it had to be after I did the 2.5 conversion cause it wasn't there before.

    could this be the dreaded ICV/ ECU burn out? Think my ECU is fried? It ran like I said, so who knows.
    That is from the ICV going bad. You need to replace both of those transistors and the ICV valve. Those transistors are like $2 at rad shack.



    • #3
      Any good write ups on how to replace them? I solder about as good as a drunken elephant witha wandering eye

      The Chuck giveth, and the Chuck taketh away


      • #4
        Is it me or are the ECU's on these cars crap. Mine took a crap about a month ago.


        For Aftermarket Engine Management, Visit Electromotive Store:

        Own a shop and need a better website?
        PM me or Visit


        • #5
          If you have a hunting idle and a good/bad ICV,you will be looking for a BDX53A transistor to replace that burnt one. Usually, its neighbour is crisp as well.


          • #6

            Maybe this text is helpfull:


            Read this info from Jose Santana, fellow E30 M3 DME guru.

            Subject: [M3] DME. ICV module. Schematic.
            From: "Jose Santan"a
            Date: Sat, 24 Jun 2000 23:34:33 +0100

            Hello folks,
            I have finished some of my experiments with my DME.
            I have "extracted" the schematic from the "ICV module".
            Note that it's not complete, the ICV module is larger and
            it would take a lot of time and effort to extract it completely.
            The damage caused to the DME by a bad ICV should be
            located at this area, I don't think it can go further. So what
            we have here should be enough to repair a bad ICV module.

            1) How does it work? A brief description.

            1.1) The ICV.

            First of all, sorry for the slight deformation on the lines of the
            schematic. It happened when the picture was reduced from
            440 kb to 34 kb. Print out the schematic and look at it:

            -The ICV is fed by the main relay through the pin #2
            (12-13.8 volts).
            -The coil L1 controls the closing of the "gate".
            -The coil L2 controls the opening of the gate.
            -The longer the coil L1 is connected to ground, the more
            the gate closes.
            -The longer the coil L2 is connected to ground, the more
            the gate opens.
            - The coil L1 is in reality a group is six coils connected in
            series and wound in six different positions (the same is
            applicable to L2 here)
            - The coil L1 is the one that uses to fail.

            1.2) The ICV module.

            As you can see, the "final stage" is quite simple. I expected
            to see something more complicated. The jumpers labeled
            as 27, 40, "pcb jump" and B15 exist on the DME for real. The
            other two jumpers were added by me when I was editing the
            schematic. The tracks that come from the pins #16 and #18
            go on into other modules. However, it has been omitted on the
            schematic. The diodes D562/563 serve as a path to discharge
            the coils and they smooth out the impulsive response across
            the coils. The transistors are working as switchers. When the
            transistor is saturated, the coil finds a low resistance path to
            ground (through the collector-emitter) and it is as if the coil were
            "connected" to ground.On the other hand, when the transistor is
            cut off, the low resistance path becomes a very high resistance
            path, and the coil is then "disconnected" from ground. Who is
            controlling those transistors? Well, it's interesting to note that
            T561 is being controlled by T560, and T560 is being controlled
            by the chip RCA H.719

            When T560 is saturated, the collector-emitter saturation voltage
            (Vcesat) is aprox. 0.8 volts. This voltage is not high enough to
            feed the base of T561 through R561. Final result: when T560 is
            driving, T561 is cutoff, which means that L1 is active and L2 isn't.

            When T560 is cut off, the current can't flow through the collector-
            emitter, and it takes another path: R561, R563, T561 ground.
            The voltage that appears now across R561/R563 is enough to
            feed T561, which enters in saturation. Now the current can flow
            through L2. Note that although there is a certain amount of current
            flowing through L1, the voltage across it is very low, because R561
            (which has 2200 ohms) is "thieving" that voltage. Therefore L2 is
            active and L1 isn't.

            T560 is being controlled (saturated and cut off) by the signal
            that comes from the chip RCA H.719. Some pins of this chip
            are connected to the microcontroller, and I suppose that the
            real control signal comes from the microcontroller. However I
            didn't pay too much attention to this chip, I just wanted to know
            where the control signal came from, and that's the RCA H.719.

            1.3) The control signal.

            This is what I saw on the oscilloscope. It's a square signal,
            (or a "pulse train" if you prefer) but it's not AC, the voltage
            varies between 0 and 3 volts.
            ______ ______
            | | | |
            | | | |
            ___| |_______| |______...
            Ton Toff

            "High level": 3 volts before R560 / 0.7 volts after R560.
            "Low level": 0 volts
            Frequency: aprox. 77 Hz.

            The voltage and the frequency of this signal never change.
            What changes is the "duty cycle", and that's the way the ICV
            is controlled, by means of variations in the duty cycle (DC):

            -a DC= 50% would keep the gate of the ICV steady, wherever
            it is.
            -a DC > 50% forces the gate to close. The higher the DC is,
            the faster the gate closes.
            -a DC < 50% forces the gate to open. The lower the DC is,
            the faster the gate opens.

            Note that what I'm saying about the DC is valid if we are referring
            to the DC of the control signal or T560 only. For T561, it would be
            the opposite.

            Some examples: when you turn the ignition key to the position II,
            (no matter if the engine is cold or warm) the control signal opens
            the ICV aprox. a 30% and the DC is adjusted to aprox 47.7 %.
            (Ton=6.2ms, Toff=6.8ms). In theory, this DC should be 50%, in
            the practice, I double checked and it's 47.7%. Anyway, since
            it is quite close to 50%, the gate stays steady, although in reality
            it is opening veeeery slowly.
            When you crank the engine, the DC is readjusted to aprox. 35%
            (this opens the gate even more) and as soon as the engine
            catches the DC jumps up to aprox. 68 %. (Ton=8.9 ms Toff=4.1
            ms). As the engine gets warmer, the DC varies slightly, and when
            the engine has reached the operating temperature, the DC is no
            longer modified and is aprox 75% (Ton=9.8 ms Toff=3,2 ms) Note
            that the ICV is not "switched off", it's always active, even when it's
            fully closed and the engine doesn't need it.

            2) Common failures.

            These are the ones I have seen so far:

            2.1) Blown track between the pin #34 and the collector of T560.
            When it happens, the current can't flow into the collector of T560
            nor the base of T561. The final result is that the ICV can't be
            controlled. The position of the gate is unknown, it depends on
            where it was before the failure. The blown track can be replaced
            with a simple jumper wire, but once it's done, only a NEW ICV
            should be used. Otherwise, a new damage to the DME may

            2.2) Blown track between the pin #16 and the emitter of T560.
            This case is not very usual, but it's possible. It's the most
            aggressive of them all, because the return path to ground is
            destroyed, and any other module that uses this path can't take
            the ground. The car here won't start. Replace the broken track
            with a jumper wire and don't even think about using the old ICV.

            2.3) The transistor T560 is melted or bad. It seems to be that
            this case is quite common among you. If the transistor is melted,
            replace it with a new one (see below, step 4). Now, pay attention
            here. The fact of having a transistor whose external appearance
            is normal doesn't mean it's ok. That happened to me. After
            refreshing the soldering points, I determined that my T560 was
            bad, and however the capsule was in perfect shape. Only the
            white plastic sheet that is between the DME box and the PCB
            showed a slight evidence of overheating on this transistor. That
            was the clue. Inspect that sheet carefully!, it could give you some
            information. The symptom of a bad transistor may be an
            intermittent behavior of the ICV: sometimes the gate goes to
            wide open, sometimes it stays closed and sometimes it moves
            between "open" and "closed" very fast, you can even hear a
            funny noise when this happens.

            3) PNP or NPN?

            If you try to look for the reference ON588 in all the transistor
            catalogs out there, chances are you will find nothing (at least,
            I haven't found anything). The prefix "ON" does not appear
            anywhere. However, there some transistors out there (NSP588,
            BD588...) which may lead us to think that they are the right
            substitutes for an ON588, but that idea is just based on that
            "Name" : 588. Here comes the problem. Those transistors are
            PNP, but a PNP does not match my personal idea of switching.
            In my opinion, the ON588 is an NPN, not a PNP.Why? Let's see:

            - Have you ever seen a PNP working as a switcher? I haven't.
            (I'm not saying it's impossible, just unusual)

            - The emitter of T560/T561 is connected to ground. How do I
            know that? If you look at one of the other ON588's inside the
            DME (attached to the heatsink), you will see that the pinout is
            stamped on the PCB. Looking at the transistor as if you were
            reading the capsule (PH ON588 blah blah), the leftmost pin is
            the base, the center pin is the collector and the rightmost pin is
            the emitter. Well, now go to T560 and T561 and you will see that
            both emitters are connected to ground (pin 16). Then, I ask you
            again: Have you ever seen a PNP with the emitter connected to
            ground working as a switcher? I haven't. An emitter connected to
            ground is a typical configuration for a switching NPN.

            - Not being 100% happy with all this, I tested the behavior of the
            ICV module with a PC simulator (Microcap III, yeah a little old).
            The combination T560 NPN -T561NPN is the one that works as
            it's supposed to work. The combination T560 PNP-T561 PNP
            with both emitters to ground simply does not work.

            - Perhaps the most important reason: I replaced an ON588 with
            an NPN three weeks ago and the DME is working again. No
            problems so far.

            So, in my opinion, the ON588 is an NPN. If you don't agree with
            me or you have another theory, let me know.

            4) PH ON588 vs TIP31.

            Yes, a TIP31 C (an NPN) was the transistor I used to replace
            a bad ON588 (T560). Why a TIP31? Because it is a universal
            power switcher, it has the same capsule (TO220), it has the
            same pinout and I had it in my transistors box.

            Once the TIP 31 was installed, the ICV module came back to
            life, the ICV was vibrating and the buzzing was there again.
            But this time I experienced something new : the real cold start
            regulation. When I start the engine now, I get a strong and
            decent idle at 1000 rpm, which lets me put a little more of load
            on the cold engine and it doesn't try to die or stall. That's the
            real difference between an ICV that works 100% and another
            that doesn't. Aprox. 90 seconds later, the idle falls to 900
            rpm for some seconds and then it falls again to 800 rpm,
            where it stabilizes. I like how it works now.

            The TIP 31 is a good substitute, but not necessarily the best
            one. Looking at the Vcesat of both transistors we realize that
            they are not the same. The Vcesat of T561 (ON588) is more
            stable, it's around 0.8 volts all the time . However, the Vcesat
            of T560 (TIP 31) varies from 0.2 to 0.8 with the engine off and
            from 0.2 to a "dangerous" 1.2 volts when the engine is running.
            Anyway, the transistor is doing its work and the ICV is working
            as expected.

            If you want to try with a different NPN power switcher, that would
            be great. We can compare afterwards.

            5) A few comments.

            All the info contained here is based on my experience and
            knowledge (which might be not enough). Although I have
            double checked everything, there might be minor mistakes.
            I you detect something, please let me know.

            David (Yando), the problem with your old DME was the
            lack of spark (if I remember correctly). On the other side,
            you had a melted transistor T560. Note that there isn't any
            relationship between this transistor and the ignition module.
            However, if the ignition module takes the ground through
            the track 16 and this track was damaged when the transistor
            T560 died (due to an excess of current), this might explain
            the problem with your DME. It's just a possibility. Replace the
            T560 and examine the tracks of the bottom PCB (you have
            to remove the remaining metal cover). Regarding the new
            DME, the same thing...have a look at T560 and the track

            Lee, Tim, you both have inspected your DME's and it seems
            to be you don't see any melted tracks nor transistors. All I can
            say is :
            1) Check the obvious (wires, female terminals in the connectors)
            2) Refresh the soldering points of T560.
            3) See if the control signal is reaching the base of T560 (don't
            forget : ignition key in II, ). This last step should be done with the
            oscilloscope. If you don't have it, it's better if you call someone
            with some experience on electronics. Remember that the DME
            must be tested on your car. Otherwise it can't receive all the info
            from the sensors and control the ICV. If the signal is there and
            T560 is not switching (the ICV must be connected!), replace the
            transistor. If after replacing the transistor, your ICV module still
            doesn't work, I bet you are missing something! :-)

            By the way, don't start the engine when the DME is open and
            connected to the wiring harness (when you are watching the
            signals with the o'scope). An open DME is unshielded, and it
            might be disturbed by the EMI created by the ignition system.
            I did it, and the engine stumbled and hesitated as I had never
            seen before.

            Jose Santana
            '83 316
            '88 ///M3 Macaoblau 2191116
            Canary Islands, Spain.




            • #7
              That Jose Santana is one awesome eletrical engineer. I've followed his posts many times in awe. Much respect...

              Jake Larsen


              • #8
                what a write up!

                i wish i had paid more attention in school lol
                "You must drive this car before you die, it really is that good." BMWCar Magazine re:the e30m3; Oct. 2002


                • #9
                  Anyone know someone to solder on those 2 new chips? Sensative electronics is one area I will not "experiment on" especially when the part is 1100 dollars new from the dealer. Or if anyone has done this resolder, PM me, I'll send it out to you and pay you whatever fee you wish.

                  P.S. Would replacing the ICV (I did already) and running it with just this one burnt chip affect anything? or would it just do the same thing it did before I discovered this? My Idle was always really rough, and that's cause I thought the cams did it, but I guess not.
                  Last edited by Highrever; 06-11-2004, 09:02 AM.

                  The Chuck giveth, and the Chuck taketh away


                  • #10
                    I've done this, it isnt THAT bad.
                    Charley Terhune
                    89 Zinno M3- stock 90k miles
                    89 Zinno M3- 2.5 etc. etc.


                    • #11
                      I havent done it, but I have worked on many circuit boards including those very small surface mount ones. The key to a good removal/install is a heat CONTROLLED iron such as a Weller WTCPT or similar. The second is to be very mindful of where you are putting the iron and how long you leave it there. Hot temperatures are not a PC boards best friend. You can buy a solder wick from Radio Shack which is a wire braid that will soak up the solder. You place this on top of the joint and then place the iron down. Once the solder is mostly removed, I cut the part leads as close as you can to the component. This will leave you with several long pieces of wire that you can grab with a pair of small pliers. Apply the heat again while pulling at the same time from the top. The little wires should come right out leaving you with a hole in the board. Of there is still solder in there, soak it completely up with the wick. Place your new component (making sure you have the polarity correct) into the holes, turn over the board and apply the heat, then solder. I try to move away from the board, sliding up the wire with the iron when I am done so that the puddles of solder leave those factory tear drops. Cut the excess remaining wire form the back and install. Its nice to have one of those helping hands to hold everything since it usually requires more hands than you can give. Take your time and it wont be that hard. I think that every shop addict should have a nice soldering station hanging around. Most of the ones you see that just plug into the wall are way too hot for anything but large wire. I picked up mine from Ebay for $35 used. Its only a transformer and a few misc. parts for the heat control so there is nothing really that can break with age or use.


                      • #12
                        Just how common is this "burn out" if the ICV goes bad? I can't believe that somebody would design something so poorly that a circuit would be fried from a malfunction or failed ICV...then again, nothing suprises me. Just how common is this fried ECU from an improperly functioning ICV?


                        • #13
                          damn good info, sorry we missed that earlier, hard to search every possible word combination
                          A friend will come bail you out of jail, but a TRUE friend will be sitting next to you in the jail cell saying, "Dude, that was focking awesome!"


                          • #14
                            paragraph I found on "turdfly"

                            good info if you dig far enough

                            Testing ICV

                            located near the fuel rail, closer to the front side of the motor. Unplug the electrical connection, and measure the resistance between the center pin and the two outside pins (one at a time). The resistance should be close to 20 ohms each. You can also measure the resistance between the two outside pins, and that should equal about 40 ohms. Apparently if your center-to-outside pin reading is around 10 ohms or less, your ICV has a short and is fried......replace it ASAP before you fry your ECU (or a transistor in your ECU).

                            I don't have the exact author for this piece, but most of my info saved and stored in the glovebox is from "John 88 ///M3 Cinnabar"
                            Last edited by autox320; 01-03-2005, 08:05 PM.


                            • #15
                              ICV transistors...

                              Mine burnt out about a year ago, thanks to Jose's write-up and help, I was able to replace the ICV transistor.

                              He's one HELL of a guy to work with.