MBE defaulting to 90C coolant temp... coolant calibration map to compare

[Trevor Little (Trevturtle) - Treasurer]

Hi I hope someone can help me?

 

Some background first...

Winter checks on the car highlighted that the coolant temp in the ECU (a MBE 967) was defaulting to 90c.

I checked the following:-

The ECU has coolant temp calibration maps, (See attached)

The upper/lower and default (90c) coolant temp limits are set. (see attached)

The coolant temp plug connector at the engine are 5v and ground and the wire does connect back to the correct pin (pin 2) on the ECU.

I've checked the coolant temp sender fitted to the engine in a pan on water and I can see the resistance change..so it works however the read  at 8C is 5M (yes 5M!) ohms..and sends the ECU to the default of 90c

 

I have also connected a variable resistor and confirmed the ECU does read a value ., not just the default 90c. ..so the Ecu is working

I also connecting a spare coolant sensor (1k ohms) gives a 111c reading on the ECU with a 8C room temp  

 

I've started a conversation with Ian Oddified to see if it might be the ECU, but wondered if anyone on here had any experience/guidance ( have read all the other WSCC threads already..no luck there:)

 

My only thoughts so far are:-

1. An incorrectly matched coolant temp sensor...but how I find the right one I'm not sure  

2. And incorrect calibration map in the ECU.  (See attached)

 

The previous owner Mark Hoy swapped the 1.8 zetec for a 2.0 Zetec engine... so I don't know if could be a factor.. unless the  coolant temp sensor fitted to the 1.8 engine was used ..things maybe be adrift now.  (I assume there's no guarantee the old 1.8 engine had a ford 1.8 temp sensor..so i can't just buy one of those.)

 

My questions: 

Does anyone have examples of coolant temp calibration maps I can compare mine to .

Any guidance on part number for temp sensors (matched to a calibration map)

 

Sorry for the long text, but I thought I would give as much background as possible first.

 

Thanks 

Trevor

 

Coolant temp map attachment

 

 

I can send the files if it helps..the image was the only way I could find to upload my file..it not perfect 

 

[Welly Jen]

Hello Trevor,

Can you tell us what the x and y axes are on the calibration graph? The resolution of the picture is too low to make them out. I had a similar problem with my car when the plug to the coolant sensor was bad. The ECU sensed infinite resistance (open circuit), giving a temperature from the calibration curve of over 100C. This meant that the car was extremely difficult to start from cold with no enrichment being added. Once up to temperature it was fine. Mine is a Microsquirt ECU, which is a bit simpler than your MBE in this respect.

 

Your ECU isn't receiving a valid signal, so is defaulting to 90C. The fact that you are seeing 5V at the plug suggests that one of the two wires is connected to the ECU. Exactly the voltages you see will depend on the internal circuitry of the ECU as it is using the resistance of the sender to derive a voltage, so it isn't a good test. However your connection of a variable resistor and the ECU seeing a variation in temperature suggests you have good wiring.

 

Most probably the sender isn't matched to the calibration, or possibly the ECU. There may be a range of allowable resistances that will work with the circuitry inside the ECU. that your sender exceeds. This is why I was asking about the axes on the calibration graph.

 

Can you get the resistance of the sender at 0C (ice/water mix) and 100C (boiling water), plus an intermediate known temperature? These senders are thermistors and they don't have a linear temperature to resistance relationship. You need a minimum of three temperatures to calibrate them properly.

 

Jen

 

Edited to add:

The resistance from your coolant sensor is orders of magnitude higher than mine. I use a Sierra Cosworth sender which is 6000Ohms at 10C, 1600 at 40C and 380 at 80C. If the MBE ECU is using an internal resistance in combination with the sender to create a voltage to read, then too high a resistance sender will lead to errors. The fact that you get sensible values from the ECU with a resistor in the kOhm range suggests that you have the wrong sender for the ECU.

[Trevor Little (Trevturtle) - Treasurer]

Hi Jen

thanks for the quick response.  An update following a discussion with Ian at Oddified.

 

  It seems I have Westfield MBE ECU that has modified resistors to match ford sensors.

Ian said it would state this on the ECU..it does it's written on the front in pen

 

A few tests Ian suggested and Ian's interpretation of the results confirmed I have a modified ECU and the problem is likely to be a mismatched sender.

A trip to the breakers this afternoon and  I have returned with a handful of Ford senders and found one that works   and reads the correct temperature.

 

Ian's email 

Your testing makes sense and the ecu is ok.

 

I would be fairly confident in saying that the car has a genuine westfield mbe ecu (normally they have 'Ford Sensors' printed on the front) which has the resistor mod inside specifically for rescaling the water temp input for the original ford sensor, and changing the engine has changed the sensor/resistance.The original Westfield ecu's have a 10k resistor fitted in place of a 1k to change the overall scaling. It's basically a divider chain, if the ecu had the normal 1k resistor inside and the sensor was at a temperature that gave 1k then the ecu input would be at 2v5 and around 42c. I

f however yours does have the 10k fitted inside, with a 1k sensor then it'll be around 0v5 input. Looking at a standard Westfield 1.8 zetec calibration map, 0v5 is around 115c (which is coincidently more or less what you measured...).

 

From that i'd be confident in saying you have the specific westfield ecu and the wrong temperature sender.

 

You have a few options;-

Change the calibration map for the current set-up. It doesn't really work though because you'll find that you'll have no resolution and run out of scaling.

Find a new sender with the correct resistance range that fits the 2.0 zetec. The spare sensor you have isn't the right one, it'll need to be for the 1.8 to be sure it's correct, but i don't know if they're interchangeable.

Change the resistor on the pcb, or i can do it for you, and then do a bit of fine tuning on the calibration map to make it 100% right (pan of hot water and thermometer, tweak the map as it cools down...simples!).

 

If the sensor fitted at the moment reads several M ohms then it's probably duff any way....it might actually be the correct one if it worked.

 

The remaining problem, which is much simpler to fix is: car has 1/8 NPT fitting and the sender from the breakers has what looks like M10 or M12.  I think I'll retap the manifold, as it already has a spare 1/8 NPT port.

 

Thanks again for the quick response...hopefully Ian's guidance may help someone else in the future.

 

Trevor

[Welly Jen]

Hi Trevor,

That is good news. Looks like the MBE ECU measures temperature the same way as Megasquirt ones. With Megasquirt you have the option of changing the resistor on the PCB to rescale it for a sender with different resistance ranges.

 

In AsciiCad

 

  | +5V

  |

  -

 | | Sender variable resistance.

 | |

 | |

  -

  |------------ Voltage measured. Varies with sender temperature.

  -

 | | Internal ECU resistor. Similar value to sender resistance.

 | |

 | |

  -

  |

  | Earth 0V