Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Automotive ECM/PCM troubles. This is an issue to ask here I think.

Status
Not open for further replies.

fastline

Member
I am bringing a wrecked ford car back to life. Had some wire harness damage but I believe I have it mostly repaired and did so before putting power back to the car.

In a nutshell, the 5V reference voltage for sensors is not right. A pile of codes having to do with low sensor voltages, low input, low output, etc. I tested at the throttle pedal plug, and then the pins right at the PCM for the throttle pedal and MAP. All are showing a consistent 2.1V. I tested voltage to the PCM using the PCM chassis as ground and the correct 12V power pins and see 12V at the PCM.

I know most of the manuals might immediately start pointing to a bad PCM but my gut is not there yet. The car will start and run smooth and nice for about 5sec, then start coughing and die. Being a car guy, I already know that means the PCM is sending good pulsing to the coils, injectors, picking up some sensors, etc. Most all PCMs also so a self check and there are no codes related to a PCM fault, only low voltage at sensors.

I am curious if this sounds like an overloaded 5V rail from a short at a sensor? I do not think I have a wiring issue but possible I suppose. I am trying to determine the easiest way to pin this down. I would think a total loss of 5V reference might point more towards a bad PCM? But 2.1V? It is trying!

Thoughts?
 
The overloaded 5V rail seems like an OK assumption. Generally, the sensors are ratiometric meaning they will be relative to the rail, hence the absolute value of the 5V supply doesn't matter to a degree. The car runs in open loop for a while too, ignoring the sensors.

On a "hunch" disconnect the MAF sensor and try to start the car. Try that first.
 
Well..... It appears I am hosed..... After enough in car testing, I bench tested by just applying power and testing several 5V ref pins. All still test at 2.1V..... I tore it down to the board but I am not sure how one would even find the problem in this maze.


There are some people online repairing these, but only for a failed spark coil issue which I think is exploded transistors. IE, obvious visual repair stuff. There is nothing obvious on mine. Very frustrating to see half the voltage but no clue what might be causing the issue. Usually components either work or they don't so I tend to think something in parallel in the 5V rail is shorted but.... Good luck finding that one!

Any thoughts or ideas? As of now, I cannot even get to the back of the board to see the traces and try to map the path. They decide to use silicone to glue er down. I was able to get the lid off to get at the component side but I am shooting blanks without seeing the back I think.
 
Well, thermal camera, we gots that!

I think since I am otherwise screwed, I will see if I can get the board off. Getting a schematic???? lol Yeah, right! This is Ford Motor Company. Since I cannot find anyone on the net that even repairs them, I would say schematics are not out there. They are getting cheaper though but the biggest issue is then I get to reprogram EVERYTHING to the new ECM. More bucks than just the price of the computer. Plus I still don't know why the 5V is low. Don't want to do this again.


Would you agree that the 5V rail is likely shorted? How do you think a 12V-5V circuit might be made for one of these? Obviously has some protection built in but I am not sure what I am looking for.
 
So.... Your ideas were certainly helpful! With the FLIR, the area of interest lit up like a christmas tree! I have provided files for both the chips in the area of interest. The thermal cannot get close enough to pin down precisely where the heat is but I will do more tests and see if I can get a macro image that might even detect the exact part of the chip.

The question now is this. I cannot unmount the board. I have access to only one side and it looks like there is more stuff on the back but I don't think I will be able to get the board off. It seems glued in a certain area and probably snap it pulling any more.

So, I am wondering if this new info can help pin down the issue more? I am wondering about rolling the dice on a few components to just try it? However, I am not sure if these chips would really be an issue (internal mosfets?) or if this otherwise might mean the components are working but otherwise being over driven from a short elsewhere? However, nothing else is lighting up, only the chip area. I just cannot run the traces from a 5V pin to find this.

I have loaded a close up of the area of interest. Feel free to mark up.

EDIT: by chasing pins, I have determined that at least 3, 5V pins at the connector go to both the VSC and VSD on the L9758 chip, which are 5V outs. Another sensor picks up on solely the VSB. I have verified all these connector pins are seeing 2.1V.

All pin test show 0hms to the chip. Does that seem right? I would think there would at least be series resistors and maybe caps?
I cannot determine if the VSC and VSD pins are really supposed to be bridged or not. I suppose you could for more current but in any case, that is where I am. I think that is the chip getting smoking hot, and the one driving the 5V ref.

EDIT2: I show pins, VPprog 1, 2, 3, boost, ref sel are at 0hms to the ground pin. BUT!! I found that VSA is at 1ohm to ground!!! I am thinking we have our smoking gun! VSA is shorted to ground. Now, if you smarter folks are on board here, how can I determine if this is an internal chip failure or another component? Ironically (or not), I found 2 highly suspect SMD caps that are in parallel and showing continuity through them. They are both in series on the VSA circuit. If I am looking at the diagram right, each of these pins should have a cap from the VS outputs to ground? IE, a shorted cap would short the VS pin to ground?
 

Attachments

  • FLIR0180.jpg
    FLIR0180.jpg
    164.4 KB · Views: 355
  • FLIR0181.jpg
    FLIR0181.jpg
    176.1 KB · Views: 352
  • L9733 chip.pdf
    387.8 KB · Views: 401
  • L9758 chip.pdf
    229.6 KB · Views: 341
  • SDC10080.JPG
    SDC10080.JPG
    519.3 KB · Views: 452
Last edited:
I's look at the signals on PDF page 8 of the power supply and also check using a diode test (powered down) for the OC outputs. e.g. Check for connections. You should discover how the supplies are supposed to be programed and if the error circuits are used and then look at the actual output.

You surprised me with a FIR camera.
 
Update: I pulled the suspected caps going to ground. They are not bad. So I decided to VERY carefully pull the suspected VSA pin up from the board so I could isolate that from the rest o the circuit. It seems VSA is reading right at 1.0 OHMs to ground. Unless someone can find another reason in mind that VSA would have 1.0 OHMS to ground, I believe the chip is toast.

The next question is, what is the likelihood there is some level of programming done on the chip that would render a replacement useless? My bro seems to think if there is anything, it is likely put in a certain mode by something else in the board.

One thing is for sure, all 3 of the programming channels are at ground reference so I don't believe any 'in circuit' programming was done. However, I am curious if ST would have possibly did any in-house programming before delivering the chips?
 
5.4
Tracking regulators
Four low drop-out tracking regulators (VSA, VSB, VSC and VSD) are supplied by the output
of the switching pre-regulator. They track the output voltage of the VDD5 linear regulator
with the accuracy as specified in the electrical characteristic table.
The VSA regulator also tracks an external voltage reference (TRACK_REF pin) and the
tracking voltage is selected by the REF_SEL pin.
If REF_SEL is tied High (5V) then VTRK_REF is tracked. If REF_SEL is left open then VDD5 is
tracked. There is an internal pull-down on REF_SEL.
The tracking supplies are intended to drive loads that are external to the ECU so they are
short circuit protected with the current limited. The outputs of the tracking regulators also
withstand short circuit to the battery.
A short circuit to GND, continuous or intermittent on one tracking supply will not affect any
other supply, including the preregulator output voltage VB. In addition to these requirements,
all sensor supplies shall be capable of operating with up to a 15μF load on the supply line.
This load may be present during initial startup, or be applied after the supply has been
powered up. In either case, the application of this load shall not cause the tracking
regulators to be permanently disabled.
VSB, VSC and VSD regulators can be disabled with VS_EN pin.


Vdd5 is a 5V regulator and can be tracked as VSA. VSA is designed for external circuits.

I doubt that the regulators were programed prior, but it would depend on a lot of stuff. I suppose and I think you suppose that Vsa is supposed to be 5V.

==

When you remove part, cut the pins. Chip quick has some solder that will lower the melting point and it;s used for removal.

You may be able to get a stencil from www.proto-advantage.com.

Otherwise flux, Solder the corners. and solder upside down. For some reason it works. e,g, 63/37 solder wiped across the terminals and the component faces in the direction that it would fall off.
 
I have sourced where the power goes from VSA. It seems to power the fuel tank pressure sensor, a source voltage for an AC pressure switch, and the IAT sensor. I could not immediately find any faults with that wiring.

Here is where I am. I have powered the ECU again with VSA pin completely disabled. Again, as expected, it is pulling the same current and heating the chip to roasting hot. VSA shows a short to ground. Now I realize the VSA tracking reg talks to some other items BUT, if you look at the block diagram, the output of that reg goes directly to the VSA pin. So can I assume that just seeing the low resistance from VSA to ground, there is an internal short somewhere?

This brings up another question - The chip seems to be VERY well protected. Current limits, short protection, even 12V protection. The question comes to mind, who in the world would you kill it? There is really nothing in this system that exceeds 15V. Even if the chip is bad and I can replace it, I am not sure I really have the source of the problem found. SOmething killed the chip.


VSA is connected differently in the block diagram. I am trying to understand if there is any other way possible to see the low resistance condition at VSA to ground?
 

Attachments

  • L9758 chip block diag.pdf
    29.7 KB · Views: 332
Low resistances to ground are a BAD thing, but look at it with both ohms/ohms reversed, diode and diode reverse function of your meter and compare it to another similar regulator.

In a automobile 12 V is anything but

the OP said:
The question comes to mind, who in the world would you kill it? There is really nothing in this system that exceeds 15V. Even if the chip is bad and I can replace it, I am not sure I really have the source of the problem found. SOmething killed the chip.

This is a really BAD assumption.

See https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=3&cad=rja&uact=8&ved=0CC8QFjACahUKEwiz-fzX277IAhULVD4KHaGYDas&url=https://www.littelfuse.com/~/media/electronics/application_notes/tvs_diodes/littelfuse_tvs_diode_automotive_circuit_protection_using_automotive_tvs_diodes_application_note.pdf.pdf&usg=AFQjCNGWrLRFmeSukPCrbzIOdAnsOUdg2A&sig2=27Hmcz0Sbg9zqmbCqUD0JQ


Suppose one of the protection mechanisms was a reverse-biased diode and it was overstressed (spike) you would have a short. So, it seems as if your supposed to have a current limited reference at the supply voltage for ratiometric sensors.

A jump took out a digital clock on a car of mine. That was a non-permanent break. i.e. fixable with a little known trick.

Loose battery terminals also play havoc.
 
Well, the chip is finally replaced! Took hours as it is TIGHT in there. Obvious that chip was never intended to be touched. Finally got it in and seated the heat sink with Chip quick as the melting point is still higher than the thermal shutdown limit of the chip so I should be ok.

I tested the ECM on the bench and now pulling .38A instead of 2.6A and seeing 5V on all rails! Looks promising so far. Going to go drop it in the vehicle shortly. I will know what I have very soon.
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top