Jump to content


Photo

Mpi Water Temperature Gauge Not Reading Correctly


  • Please log in to reply
12 replies to this topic

#1 BlueFunk83

BlueFunk83

    Learner Driver

  • Noobies
  • PipPip
  • 12 posts
  • Location: Miami, FL

Posted 15 March 2015 - 03:06 AM

Cheers all, 
Recently obtained my first Classic Mini; a late model MPi. Been sorting common things to get it to perfect running order, but one issue that continues to elude me is the water temperature gauge. During normal operation, the needle never moves past C or the blue field on the gauge, yet the aux fan kicks in as needed so I know the engine is running warm. 

What I've done to date:

-Replaced temperature sender connector (prior owner used Mickey Mouse connectors)
-Replaced temperature sender (aftermarket common replacement part)

-Replaced thermostat (88C/190F)

Yes, the thermostat is between the sandwich plate and thermostat housing and not in the head.

So after all this and running around town today, the gauge never really moved beyond C.

Now here is the fly in the ointment...

 

The prior owner had the MEMs and MPi removed, so now the car runs on an Electromotive Direct Ignition system with Weber DGV...

Runs well I must say, however, whenever surgery such as this takes place, there are always niggly bits that don't get sorted... like this temperature gauge.

So I ask you, because that sender spoke to the ECU as well as the gauge, by not having the ECU in the circuit, is there a reason why the gauge would read colder than normal? 


Edited by BlueFunk83, 15 March 2015 - 03:07 AM.


#2 FlyingScot

FlyingScot

    Up Into Fourth

  • TMF Team
  • PipPipPipPipPipPipPipPipPip
  • 2,664 posts
  • Location: Inverclyde Scotland
  • Local Club: Mini Cooper Register

Posted 15 March 2015 - 08:21 AM

The sender didn't connect to the ECU as well as the gauge on the MEMS set up.
It connects to the ECU only and the ECU itself drives the gauge from an output circuit.
Is it wired directly to the gauge from the sensor now?

FS

#3 Dan

Dan

    On Sabbatical

  • TMF+ Member
  • PipPipPipPipPipPipPipPipPipPipPip
  • 21,354 posts

Posted 15 March 2015 - 08:51 AM

On most current OE injection systems the gauges are fitted simply because drivers want them, they don't actually do much. They are idiot gauges and are set to show everything as normal until something goes way out of normal spec. It may be that this aftermarket system is setup to show the same, and it won't show small fluctuations or the temperature climbing high enough for the fan to run but it will show a rise if the temperature continues to climb once the fan is running.

#4 BlueFunk83

BlueFunk83

    Learner Driver

  • Noobies
  • PipPip
  • 12 posts
  • Location: Miami, FL

Posted 15 March 2015 - 01:16 PM

The sender didn't connect to the ECU as well as the gauge on the MEMS set up.
It connects to the ECU only and the ECU itself drives the gauge from an output circuit.
Is it wired directly to the gauge from the sensor now?
FS

Thanks for the clarification, FS. While I have not climbed behind the dash to trace the wires exactly, yes it would appear that the gauge is wired directly.

I do know that without the sender connected the gauge reads null, so yeah.

Could it be that the resistance of the sender itself is too weak to operate the gauge and that it would have received a stronger signal from the ECU?

Edited by BlueFunk83, 15 March 2015 - 01:26 PM.


#5 BlueFunk83

BlueFunk83

    Learner Driver

  • Noobies
  • PipPip
  • 12 posts
  • Location: Miami, FL

Posted 15 March 2015 - 01:23 PM

On most current OE injection systems the gauges are fitted simply because drivers want them, they don't actually do much. They are idiot gauges and are set to show everything as normal until something goes way out of normal spec. It may be that this aftermarket system is setup to show the same, and it won't show small fluctuations or the temperature climbing high enough for the fan to run but it will show a rise if the temperature continues to climb once the fan is running.


While that may be true Dan in most instances, the Electromotive system here is only an ignition system and has no ECU. The only connections out of it are a pulse output to go to the tach.

This engine is now more akin to an SPi converted to carbs or a carb model than an MPi.

#6 BlueFunk83

BlueFunk83

    Learner Driver

  • Noobies
  • PipPip
  • 12 posts
  • Location: Miami, FL

Posted 15 March 2015 - 05:26 PM

Ok, discussed this with my dad, and he mentioned how the MGB and early Minis had a voltage stabilizer on the electrical temperature gauge... Could that be the missing piece here?

#7 BlueFunk83

BlueFunk83

    Learner Driver

  • Noobies
  • PipPip
  • 12 posts
  • Location: Miami, FL

Posted 17 March 2015 - 03:53 AM

Scratch voltage stabilizer as I've read that the Nippon Seiki gauges have that built in to the cluster. Current theory now is that the resistance for the MPi sender is different from others as it was feeding info to the ECU, and then the ECU would migrate that info to the gauge. Now with no ECU, sender resistance may not match the gauge, resulting in faulty readings.

So what about fitting a sender from a carbureted model, as the gauges should be common between the two, but the senders are not. 



#8 Dan

Dan

    On Sabbatical

  • TMF+ Member
  • PipPipPipPipPipPipPipPipPipPipPip
  • 21,354 posts

Posted 17 March 2015 - 07:58 PM

The gauges are not common to anything else, people have similar problems when fitting an MPi dash to an older car. The gauge expects a signal from the ECU.

#9 FlyingScot

FlyingScot

    Up Into Fourth

  • TMF Team
  • PipPipPipPipPipPipPipPipPip
  • 2,664 posts
  • Location: Inverclyde Scotland
  • Local Club: Mini Cooper Register

Posted 18 March 2015 - 12:06 AM

The gauge gets a PWM signal from MEMs so you are stuck with what you have...

FS

#10 BlueFunk83

BlueFunk83

    Learner Driver

  • Noobies
  • PipPip
  • 12 posts
  • Location: Miami, FL

Posted 18 March 2015 - 01:43 AM

So even if I fit the early sender, the gauge won't read it? I've been reading about the SPi guys converting to carbs, and others have said to tap the head and install the GTR101 sender and the temperature gauge will work. 

 

The temperature gauge in the MPi cars is different from the SPi cars?



#11 BlueFunk83

BlueFunk83

    Learner Driver

  • Noobies
  • PipPip
  • 12 posts
  • Location: Miami, FL

Posted 18 March 2015 - 04:58 AM

Just read a post on another forum where an MPi owner transferred the leads from the back of the MPi temp gauge (with stock ECU and wiring) and connected them to a Smiths accessory temperature gauge and stated the gauge was working fine.

So in my brain, that tells me the ohms coming from the ECU to the MPi temp gauge are equivalent to the ohms of the old style sender to a Smiths gauge...

Please correct me if I'm wrong. I don't feel like throwing money at fruitless efforts. Thanks.

#12 FlyingScot

FlyingScot

    Up Into Fourth

  • TMF Team
  • PipPipPipPipPipPipPipPipPip
  • 2,664 posts
  • Location: Inverclyde Scotland
  • Local Club: Mini Cooper Register

Posted 18 March 2015 - 08:14 AM

The system works by the MEMs unit measuring the resistance of the CTS by way of supplying 5V from the internal voltage supply.
Ignoring any other features the ECU in turn creates a proportional signal used to drive the gauge by way of the driver circuit in MEMs.

Your problem is to re create the relationship between the changing resistance of the CTS and the gauge position which it would seem is not in the same range as you have already worked out.
In non injection cars the sensor is designed to work with the battery voltage (regulated) rather than the 5V the MEMs supplied, so sounds like using this could be the way forward.

FS

#13 BlueFunk83

BlueFunk83

    Learner Driver

  • Noobies
  • PipPip
  • 12 posts
  • Location: Miami, FL

Posted 15 October 2015 - 09:27 PM

So I solved this by using an adapter allowing me to mount the old style sender where the heater take off is. Wired using old lead that goes straight to gauge, now gauge works.






1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users