Nodge68
Well-Known Member
Well, the plot thickens- I have definitely checked rave again - page 124 under EDC CTS.... the chart is the correct.one definitely for the EDC (TD4).. so, how the hell is the i930 reading the bloody CTS (unless of course the RAVE chart is incorrect.....)
Nodge, are you in a position to check your cts sensor resistance in relation to perceived temps ? - this is crazy...
I can test my CTS tonight.
I'll test the CTS resistance and it's actual temperature.
I'll also put my I930 on and give the live data temperature, which should be within a few degrees of that measured with my infrared thermometer.
Joe_H
Well-Known Member
WARNING - Bloody long post - lots of leccy teccy and a bit of math..
---------------------------------------------------------------------------------------------------
Ok, we are getting somewhere - and what a B&stard of a job !
Right, the Rave data is bollox - the numbers are 'correct' but they are not 'resistance'. ! - they are VOLTAGE.......................
This might be heavy stuff -- so if you are not 'leccy techy' orientated skip to the end lol
Ok - first let us try to make sense of what is going on with the limited data.
If we look at a general NTC Thermistor resistance plot for a CTS - (and this is probably what we are actually using - it fully corresponds to a BMW part)
http://www.bosch-motorsport.com/med...sor_NTC_M12_Datasheet_51_en_2782569739pdf.pdf
Now, that is more like it !. the CTS RESISTANCE readings are beginning to look sensible using this plot.
Ok, so let us move on a bit... what RAVE is quoting is the actual VOLTAGE (wrongly labelled as resistance!) that the ecu should be seeing at the sense terminal of the CTS (note - as the CTS is an NTC thermistor - meaning negative temperature coefficient - which basically means that as the temperature increases the resistance drops.) One side of the CTS goes to ground. The other goes to a 'reference' / 'sense' voltage from <> to the ECU. - (you can connect the CTS pins either way around.)
Ok, lets try to make more sense of all this...
Rave is wrong ! - that is fact ... if we look at the table I posted before from rave - and - you replace the heading in the right hand column with 'ECU SENSE VOLTAGE' instead of 'resistance' then it starts to make sense.
We need some other data to reference - so here is a VOLTAGE CAL plot from a BOSCH CTS . - I will link it but copy the relevant data here -
Look at the Temp C plot against VOLTAGE CAL - hmmm. seen this before (well, something VERY similar) eh ? - yup, in the incorrectly headed rave data. !
to refresh - here is the Rave data -
I did say it was a bit 'HEAVY' -----
Now we have some correlation between resistance values of the CTS to temperature (seen i the first image above)
However, that is only half of the story - and - it would seem that both the temp sensors that Radnal has are probably ok.....
So, how does this tie up to voltage and the C versus V table(s) - the incorrectly labelled RAVE one and the correctly labelled BOSCH one...
Right ...
Notice the description in Rave -
Ok, I was presuming that there was a potential divider inside the ecu and that the NTC CTS unit was in parallel with R2 inside the ECU. That would be using the normal descriptives for a potential divider where the 'top' resistor is R1 and the bottom resistor is R2 .
It would appear not so and some reverse engineering is necessary to see how those V figures are achieved
Consider the following scenario - inside the ecu is a 1K resistor from 5V to the 'sense <> ref' pin on the ECU that goes to the CTS.
Consider this as R1 in the potential divider.
Now let us add in the NTC CTS resistance value as R2 of the divider. We now have a complete potential divider consisting of - inside the ecu - R1 = 1K in series with the external CTS resistance R2 which is grounded causing a full divider.
The voltage at the central point of a potential divider with reference to ground - (basically the voltage measured with reference to ground where R1 and R2 join) can be calculated as follows.
Vout = Vin * (R2/(R1+R2))
Substituting values -
Vout for - say 20C is as follow -
We know Vin (5V), we know R2 from first image where 20C = 2.5K ohms (2500 ohm). We know R1 as (I am making an educated guess at it being a 1K resistor.!
)So, we have all the info we need -
so Vout for 20C = 5(v) * (2.5(Kohms) / (1(Kohms) + 2.5(Kohms))
hence = 5 * (2.5 / (1 + 2.5)) = 3.571 V which is virtually identical to the bosch data table above (they have 3.546V) (slightly different to the rave V plot but we are not absolutely certain of the NTC plot - it could be fractionally different - (most Delphi CTS units are 2.795K at 25C for example)
For further proof lets use another calc at a different point on the sensor Temp versus Resistance chart above - 80C the sensor is at 323 Ohms (0.323K)
So for that calc it would be - Vout = 5 * (0.323/(0.323+1)) = 1.220 V
Compare this to the Second image down again - the bosch cal chart - and that says 1.216V - so the math works fine.
After all that - what does this all mean - well - basically as far as I can see it - ithe voltage measured across the plug that goes to the CTS should be 5V or near enough.
However, there should be open circuit effective resistance measured across the plug (ignition OFF!) - or perhaps a very very high resistance - 100K ohms or larger
Any lower measurement and the ECU is captain cooked.
The values to look for at across the sensor are near enough as per the top image. (they could be fractionally different but only fractionally!)
So, that is how the system works and the math ... now all we need to ascertain is if the fault is the ECU - and I do not believe it is the CTS and we now know the resistance values to look for across the cts..... and we know that effectively putting CTS into preset mode by disconnecting it causes the vehicle to function - then unless something else is really magically happening or voodoo - the big finger starts to points at the ecu... however -
The next logical test is to use the data we now have now we understand it - if we back-probe the CTS with the plug attached we can read the voltage across the sender - which should more or less correspond to the column listed here from RAVE wrongly headed resistance.
We can also Double check by measuring the resistance of the sender directly so we have both items checked - a sender with correct functionality if used in a faulty ecu will most likely give an incorrect voltage when connected and back-probed.
I await the results
.... the fact that the 930 appears to be reading the correct temp is a 'bit weird' to say the least. One would expect it to only know the temp by way of the cts ecu circuit.Round and round we go !
At least we can do some definitive testing on the sender and the input circuit now, and even though it appears to be a fairly simple circuit we now know HOW it works and what to expect and how to test it.
Phew, I am beggered now.
Joe
Last edited:
Nodge68
Well-Known Member
That's good as is means I don't have to go outside and get wet to measured my CTS.
Bloody Rave ah!!
I think the ECU map has a hole in it at the temperature it's not working at.
That's the way it's looking to me. Maybe someone has screwed up mapping the ECU in the past?
Bloody Rave ah!!
I think the ECU map has a hole in it at the temperature it's not working at.
That's the way it's looking to me. Maybe someone has screwed up mapping the ECU in the past?
Joe_H
Well-Known Member
I would tend to agree Nodge, that is my current working hypothesis and the only thing that fits the 930 reading correctly.
It is as if the ECU rom is unstable somehow ? - I thought most did a checksum test but perhaps this is too early a unit. It certainly seems as if bypassing the lookup code for the read cts V to 'what to do with it' map is somehow 'captain cooked'
Most strange - Probably why the car was sold originally.
edit - IT WILL STILL be very interesting though to see the actual voltage at the cts sensor on the subject vehicle !. (for Rednal) You can easily make a quick backprobe set out of sewing pins and a length of wire wrapped and solder to them.
Can the 930 look at voltage instead of temp (ie the raw data) ?
Joe
Last edited:
Nodge68
Well-Known Member
This is interesting. Look at this excerpt from the M47R manual. It explains the starting with the CTS open or short circuit.
Joe_H
Well-Known Member
That is VERY interesting Nodge ! - good find.
That is more detailed than Rave - far more - ! - that explains something that was really bugging me - Rave says it defaults to 80C - when I read that - I thought WTF ??? what happens on startup (for example if cold !) the engine wouldnt like to think it was at 80C ..... -
SO rave cocks it again as the above say a value -10c is used on start up and 80 WHEN running. .- so when warm, it tries to use a -10 map to start until it is running. . That is a good piece in the Jigsaw and definite further investigation is needed. A lookup or corrupt area would do this presuming that is AFTER the read data to the instrument cluster and 930.
The terms 'interrupted or corrupted' are informative also
That is more detailed than Rave - far more - ! - that explains something that was really bugging me
- Rave says it defaults to 80C - when I read that - I thought WTF ??? what happens on startup (for example if cold !) the engine wouldnt like to think it was at 80C ..... -SO rave cocks it again as the above say a value -10c is used on start up and 80 WHEN running. .- so when warm, it tries to use a -10 map to start until it is running. . That is a good piece in the Jigsaw and definite further investigation is needed. A lookup or corrupt area would do this presuming that is AFTER the read data to the instrument cluster and 930.
The terms 'interrupted or corrupted' are informative also
Last edited:
Joe_H
Well-Known Member
I still want to see what actual voltage he is getting on the CTS circuit - presumably 'normal' if the 930 is reading accurately.
Also, what the temp gauge does during these issues and after he finally gets it running .....
How accessible is the ecu on the TD4 ? can you get to it with freeze spray easily ?
Nodge68
Well-Known Member
ECU on the TD4 is easy to get at. It's easy to back probe the connections too. The cover of the "E" box needs removing but that's only a few 5mm Allen bolts. Once the lid is off the "E" box, all the ECU connections are right on top for easy access.
First of all I cannot thank you enough for the time and effort you have dedicated to this for me.
I will put it back together and go for a run. I think the needle works normally settling in middle once up to temperature. Yes the ecu is very accessible it sits within plastic box behind battery. I did use freeze spray on that case previously with no effect.
If I am reading this correctly it seems that part of the ECU data relating to the coolant temperature has somehow become corrupt. Can this occur through normal circumstances or does it mean it has probably been remapped and this error was inadvertently introduced ? Going forward what are the most cost efficient means to resolve this ? I would ideally like to keep the car as its low mileage nice condition and apart from this nightmare starting issue it runs very nicely. It is however a 2005 model and cannot justify throwing loads of money at it.
I will put it back together and go for a run. I think the needle works normally settling in middle once up to temperature. Yes the ecu is very accessible it sits within plastic box behind battery. I did use freeze spray on that case previously with no effect.
If I am reading this correctly it seems that part of the ECU data relating to the coolant temperature has somehow become corrupt. Can this occur through normal circumstances or does it mean it has probably been remapped and this error was inadvertently introduced ? Going forward what are the most cost efficient means to resolve this ? I would ideally like to keep the car as its low mileage nice condition and apart from this nightmare starting issue it runs very nicely. It is however a 2005 model and cannot justify throwing loads of money at it.
Skinny Mike
Well-Known Member
First thing to do, is remove all the connectors off the ecu and reseat them.
BTW, isn't there a temperature sensor in the ecu box?
Mike
BTW, isn't there a temperature sensor in the ecu box?
Mike
Joe_H
Well-Known Member
Hi Rednal Well, I had nothing else to do today
Firstly - you need to hook up the 930 and monitor coolant temps from start to stop - I have the Lynx, not the 930 so am unaware of its logging features. Can you actually log the data and look back ?
When you say you tried freezing spray - was it directly on the ecu case ? (the metal case) ?
Do you have any resistors ? - for example 2.2K ?
If we can get a clean - no drop out temp log from the 930 - at all times from a cold start to full temp and be able to look back and check that for continuity (in the context of a valid and plausible reading being maintained at all times) - and effectively completely rules out cts external ecu circuitry - then unless someone can pluck out of the air a valid and reasonable theory why the backup mode works when unplugging it but the vehicle struggles without then it is down to the ECU - there is nothing else. ECU's are dirt cheap I could get you one for between 10 and 20 quid. What you need is someone to be able to program it to your vehicle - I presume this would need a T4 or similar ?
SO, what logging facilities have you on the 930 ? - and did you try freeze spray on the ecu metal box.? - It s still so weird that the thing always starts in the morning .
Any fault in the circuit should put it into default mode so it sounds like the ecu is somehow reading a totally incorrect value yet it sees it as plausible (otherwise it would go default - same as you unplugging it - which it does not seem to do ! - if it DID see corrupt data or an o/c fault or whatever, it would still be ok as it would operate the same as when unplugged - hence it comes back again to some form of thermal related data corruption that at the point of map comparison only) the ONLY thing I can see between morning and later is thermal- what else could it be ? - it is a seriously weird issue. I think you need to get a cost for adapting a s/h ecu to your motor.
I hate to call ECU but yet, that is the situation. Parts are for nothing - I know a guy with a stack of them ! - just need to get someone to match it.
Joe
Firstly - you need to hook up the 930 and monitor coolant temps from start to stop - I have the Lynx, not the 930 so am unaware of its logging features. Can you actually log the data and look back ?
When you say you tried freezing spray - was it directly on the ecu case ? (the metal case) ?
Do you have any resistors ? - for example 2.2K ?
If we can get a clean - no drop out temp log from the 930 - at all times from a cold start to full temp and be able to look back and check that for continuity (in the context of a valid and plausible reading being maintained at all times) - and effectively completely rules out cts external ecu circuitry - then unless someone can pluck out of the air a valid and reasonable theory why the backup mode works when unplugging it but the vehicle struggles without then it is down to the ECU - there is nothing else. ECU's are dirt cheap I could get you one for between 10 and 20 quid. What you need is someone to be able to program it to your vehicle - I presume this would need a T4 or similar ?
SO, what logging facilities have you on the 930 ? - and did you try freeze spray on the ecu metal box.? - It s still so weird that the thing always starts in the morning .
Any fault in the circuit should put it into default mode so it sounds like the ecu is somehow reading a totally incorrect value yet it sees it as plausible (otherwise it would go default - same as you unplugging it - which it does not seem to do ! - if it DID see corrupt data or an o/c fault or whatever, it would still be ok as it would operate the same as when unplugged - hence it comes back again to some form of thermal related data corruption that at the point of map comparison only) the ONLY thing I can see between morning and later is thermal- what else could it be ? - it is a seriously weird issue. I think you need to get a cost for adapting a s/h ecu to your motor.
I hate to call ECU but yet, that is the situation. Parts are for nothing - I know a guy with a stack of them ! - just need to get someone to match it.
Joe
Skinny Mike
Well-Known Member
Joe_H
Well-Known Member
The only benefit in removing the ecu (or unbolting it) is to allow you to freeze spray the case.
Which temp sensor in the 'ecu box' are you referring to ?
Joe_H
Well-Known Member
It would be a natural occurrence I believe - if it was continuous and always repeatable it would be be possibly CAUSED by alterations, but it is not continuous - hence again I can only see it being thermal.
Again - what logging facilities has the 930 got ? - that would be most useful although I think at this stage it is academic - I would like to see a full log of cts data from the 930 and also the results of freeze spray when it has trouble starting - the freeze spray should NOT be sprayed into connectors ! it is a case of chilling the ecu case down and leaving it for a few minutes then trying again. It would just be nice to be able to 'influence' the operation and bring about or remove the fault prior to replacing. (not necessary - but still nice)
Skinny Mike
Well-Known Member
The ecu box on the td4 is temperature controlled, has pipes running from the cooling/heating system to the rear of the box.
Mike
Joe_H
Well-Known Member
Hmmm, I have seem those pipes on images I wondered what they were ! - I what else is in the box besides the ecu ? - is there any LIQUID pipes running through this 'box' ? - Can the ecu be placed outside as a jury rig to get it away from the pipes. That again opens up a new aspect. Tell me more about this 'heating' to the ecu box....
