Coolant Temp Sensor Incorrectly Displayed

[bachig24u]

Hi Matt,
I have two Type 1 Boards.
1x VLT base image.
1x CR31 Base image.
When using a VLT base image in a 3EA NA ECU the Coolant temp sensor is reporting incorrectly.
The temp slowly rises from 21 degrees to 80 degrees with no coolant temp connected.
The temp slowly rises from 21 degrees to 80 with short circuit coolant temp connected.
both scenarios with no engine running.
After adding various resistance to the ECU terminals in order to simulate a CTS some erratic behaviour on the display gauges is noticed.

Using the same ECU, with the CR31 base image board installed, the CTS operates as per normal showing correct temp correlating to Coolant Temp sensor terminal resistance.
added : I also used both type boards in an ECCS J7100 CR31 ECU with similar results.

Any idea whats happening here?

Kind Regards
Sime

[Eric]

it's probably nomal.

most ECU's have a CTS emulator, meaning the CTS is "estimated" when the sensor gives false values or is missing completely.
The estimated temperature value slowly rises up to 80C depending on the time the ECU is powered up.

[bachig24u]

Eric,
It's isn't a function of the ECU. it's a function of the ROM version or bug in the ENT file.

[Matt]

Not a bug in the ENT file since VLT/R31 use the same coolant sensor memory address. I just tested it then (also run VLT in my car too...) and on the bench it matches what the sensor is set to

Coolant temp address is 0013

If you put the RAM trace on address 0000 with length 20 and monitor 0013 you should see it match the temperature. If that works then the read register is screwed in the consult table for your board, otherwise potentially corrupt ECU image??

Can check your board out here if you wanted to post it to me?

[bachig24u]

I've just re-installed the binary from a recently downloaded rom file pack, same results.
I will try and grab the other VLT base image type 1 board from the spare ECU and test again.
This was one of two purchased/delivered at the same time.

[Matt]

getting the maps from the rom pack wont fix this one... it only uploads the maps (not the whole BIN file)

not sure why its having issues without seeing in on the bench (you can check the memory address if you wish)

[bachig24u]

Ok, I've tested all my VLT & CR31 base imaged type-1 boards in VLNA(3EA-3EM) & VLT(3TA-3TM) & R31 ECCS-NA ECU's.
Only the VLT base image has this behavior with an open circuit or short circuit CTS in any chosen ECU.
With a resistance value similar to a working cold/warm CTS it seems to operate as per normal.

This leads me to the conclusion that some fail safe programming is hidding in the VLT rom image to prevent a warm Turbo charged engine from getting cold run time spark advance if the CTS fails.

An open or short circuit CTS will swing the temp from -40 to 205.
It's has been a quick test to check wiring in the VLNA's with the CR31 base image.
I guess not in the VLT's .. Arghhhh

[Matt]

Guess that makes sense. The code modifications between R31/VL and VLT are very minor, but that must be one of those changes (as well as secondary warmup table)

[bachig24u]

I'll never understand normal as the reference always changes in this world of technology.
I did prove you were correct although I'm amazed and still bewildered.

When ever the ecu receives a valid signal it displays as it should, if during that time it gets a bad signal it reverts back to the previous predetermined temp and continues counting up to 80C.

it's probably nomal.

most ECU's have a CTS emulator, meaning the CTS is "estimated" when the sensor gives false values or is missing completely.
The estimated temperature value slowly rises up to 80C depending on the time the ECU is powered up.

[bachig24u]

I just couldn't imagine so much forethought in a 1986 Holden but then again it was developed in Japan no doubt.

Guess that makes sense. The code modifications between R31/VL and VLT are very minor, but that must be one of those changes (as well as secondary warmup table)

[Matt]

Those ECUS are JECS (Japanese Electronic Control Systems) so definately code done in Japan. However I have a VL turbo prototype ECU here used by Holden. It has a full size second board on it with knock sensor circuitry and 4 x 4KB EPROM chips. The top EPROM chip had is cover removed as it was the one used for remapping development. Japanese writing on the lid with knock sensor frequency. Would have been a development unit provided by JECS to Holden for tuning