PDA

View Full Version : Does Evoscan read accurate coolant temps for the VR4s?



Kryndon
16-10-2020, 10:44 PM
For the life of me I can't get these coolant temps down.

I am using Evoscan 2.7 with the modded Vag-Com cable. I haven't touched any data items or definitions within the software, I just use it for logging.

No matter what I did, temps seem to be between the 97-100C region, especially now after changing the lifters, oil and rebuilding the turbos (thus having to drain most of the coolant and adding it back in again), the temps in my logs are 99-100. Fans kick on but on the slowest setting, they barely move any air.

I have completely flushed the entire cooling system and replaced the thermostat with an OEM Mitsu 82C unit. I am running 3L of good quality green antifreeze and 4L of distilled water. The car is absolutely stock.

Am I simply having stubborn air bubbles in the system, or does Evoscan read a few degrees higher? I can share a log if needed. I've been scouring the net but I can't figure out what temp these motors normally run at.

Confused
17-10-2020, 08:53 PM
Either "Coolant Temp" or "Coolant Temp Scaled" always seems to read high. The other one seems to read better!

The fans are designed to come on slowly, and only ramp up if actual coolant temperature rises.

82510

It could be a bad temperature sensor - often when they fail though they read lower than normal (so the ECU thinks it is in "warmup" mode, and adds extra fuel, and pulls timing/boost).

I've not actually looked too closely at the temperatures, nor compared them to other sensors - but if the gauge on the dash is reading in the middle, then it's probably fine!

Kryndon
17-10-2020, 09:20 PM
I don't trust the dash gauges. On my other Mitsi (using the same coolant temp sensor and sender) I have seen the gauge remain in the same place but one case it's 90C the other is 105C, a very significant difference. Also I have been told especially more modern Mitsus use specific 'filters' on the gauges which keep them more in one place and doesn't let them fluctuate too much. In other words, the gauge might read a wee bit below the mid mark, but it still may not be safe, i.e. it might be 100C+.

I am using the "Coolant Temp" setting. The "scaled" one simply subtracts 40 or 50C from the main value for other reasons so it wouldn't really affect the real value. I'm just wondering if the actual formula is not accurate for a VR-4.

Also, through Evoscan's Actuators menu, I can manually trigger the fans. However, the weird bit is that both Fan (HI) and Fan (LO) both come on at max speed, the loudest setting by far. Yet, when the car is actually running and up to 99C indicated temp, only then do they turn on, and they only run at the lowest speed, which is barely moving any air and you can't hear that distinctive howl.

So the problem is, why are these fans only coming on at 99C, and when they do, why are they only at the lowest speed setting? It is reflected in the logs, too. From 99-100C the temps would drop ever so slightly to 97 then back up again, as if there aren't any fans at all!

I'm thinking there's two possibilities. Either my temps are fine and within the 89-94 range with the fans coming on normally but on the lower setting (50%) as per your data sheet, thus indicating Evoscan itself is showing falsely higher temps than reality.

OR the temp sensor is in fact a goner and reading higher, but then if the sensor is sending high temp values, surely the ECU will be trying to cool the engine down more effectively by triggering the fans on the highest setting?

I will be preemptively changing the sensor tomorrow and checking the logs and monitoring fan behavior. I just don't know how else to check the actual temp beside sticking a cooking thermometer in the rad cap!...... wait :thinking2

Confused
17-10-2020, 10:28 PM
Yes, the dash gauge is damped, as the engine does have a wide "normal" operating range, and it's not until it gets over this that they consider it an issue and to alert the driver by then suddenly moving the needle - you won't notice it if it slowly climbs, but you will notice it suddenly move.

The standard thermostat starts opening at 82 degrees, and is fully open by 95 degrees.

I'm not sure about the HI and LO, as far as I'm aware both fans are driven by a single PWM output, maybe the Evo has a standard 2-speed fan rather than the variable speed like we do? So on both of those, it probably just pulls the output high (which is the equivalent of asking for 100% speed)

Does either temperature read correctly when the engine is cold, and therefore should be at your current outside ambient temperature? If not, then that could explain why it's showing as "hot" to you, because maybe the formula is incorrect for our cars?

Check the resistance of the sensor at 20 degrees, it should be around 2.1-2.7 kΩ at 20 degrees, 0.26-0.36 kΩ at 80 degrees.

Kryndon
18-10-2020, 01:36 PM
Alright so I've been made a fool by my own ignorance and impatience again...

Turns out Garry you was right, the "Coolant Temp" option does in fact read way higher than actual. I ran the "Coolant Temp Scaled" at the same time as the other option and there was a noticeable discrepancy. At ambient temp, both the outside air sensor and temp scaled were reading 18C, whilst the other option was 21. Then I let the car idle for 30ish minutes and the scaled option rose more slowly but gradually, and once it hit 88 exactly, the fans turned on the slow (50%) setting. Meanwhile the "Coolant Temp" option was already reading 95-97.

So, in other words, we should not use the "Coolant Temp" at all for a realistic reading, but rather we should use the "Coolant Temp Scaled" which is further down the items list. The car held 89 throughout its idling period.

Thank you once more for the information and table, Garry. Hopefully this helps somebody later down the line. Here's a few pics to show the difference!

825148251582516

Confused
18-10-2020, 05:09 PM
Excellent, thanks for the feedback and confirmation. I did think that "Scaled" was the correct one to use, but it's been about 3-4 years since I looked closely at a log with coolant temperatures in it, so my memory was a little hazy!

Glad however that everything seems to be spot on and working correctly :)