Stop intake air temp heat soak for good on your turbo MX-5

One of the biggest contributing factors to a poor running boosted car is the intake air temperature sensor. This is used to calculate the mass of air going into the cylinder, hotter air is less dense and requires less fuel than colder air.

If we assume that the VE of the engine has been calculated correctly, i.e. the car runs and idles properly and has been ‘mapped’ then it should be able to cope with the changes in intake temperature as the calculation follows the ideal gas law. However, that assumes that all the sensors we are running are perfectly accurate, even a 5% difference in temperature and pressure could result in a hunting idle or a moderate lean condition – making the car feel sluggish off the line on the initial start or in a hot days traffic jam.

It is possible to mitigate this effect by using the IAT modifier common to most aftermarket ECUs, or running the car at a richer AFR at idle, we’re talking about situations where the car is tuned to a reasonably clean AFR.

It’s always been a theory of ours that there is a better way of calculating the fuel requirements of the engine and that off the shelf sensors are highly likely not going to be able to cut it, it had us questioning how fast they respond to changes in temperature and how accurate they really are at each temperature point.

Let’s consider that in reality the engine will heat soak while sat in traffic, heat builds up under the bonnet of the car which is inhaled by the turbocharger – or supercharger, and heated up more – but at idle the intercooler really can’t do its job as effectively – especially when you’re sat still.

So, the engine very much will slowly be taking on a hotter and hotter air mixture – and that’s perfectly OK but we just need to know as quickly as possible and as accurately as possible in order for the fuel requirements to be kept in check.

This all came to a head when we started to look at how we could scientifically, accurately and repeatably test the efficiency of the turbo and intercooler setup on leaf – we know that our new intercooler kit is far more efficient based on the temperatures we’re seeing as well as the overall feel of the car but numbers are key here.

So how do you measure the temperature of the turbo outlet? Well, you’ll need a pretty fast responding sensor, that’s well calibrated and that can handle up to 200c air temps.

The standard offerings of the Bosch / GM style 2 pin simply can’t handle that.

 

 

With a response time of ~10 seconds (yes, 10 seconds!) – let’s talk about that for a minute. What does 10 seconds of airflow look like on an engine?

*Comparing with a similar Open Element sensor GM design, which has a 15 second response time: LINK

Using this calculator: Here

We know that a 1598cc engine, a 1.6  will consume 4,602 litres of air per minute at 7200rpm at the suggested volumetric efficiency of 0.8, or thereabouts. The intercooler itself holds ~ 2.7 litres of air and the intercooler piping is around 2.8 litres for a total air charge of circa 5.6 litres, that’s how much air is in the piping at any one time. Calculated based on the BOFI 230mm Intercooler and Piping Kit – we’ve overestimating here more than under.

With the IAT sensor in the up pip close to the throttle body, there is perhaps 5.4 litres of air in front of the intake sensor itself to be measured.

Now we’ve got that maths out of the way we can look at what 10 seconds of airflow at 7200rpm looks like and then look at a more reasonable 1500 rpm. If we take 4,602 litres of air a minute and divide that by 60 to get litres of air per second we end up with 76.7 litres. In one second the engine will have consumed the total volume of air in the intake system 14 times over! (76.7l / 5.4l = 14.14l) in 10 seconds that’s over 142 times! (76.7l * 10 / 5.4)

At 1500 rpm, say as you’re pulling away, air starts rushing into the engine bay dropping the ambient temperatures and in turn the real intake air temps, blowing over the intercooler helping that charge to cool down further the engine will be consuming 959 liters of air per minute, that’s 15.8 litres a second which is 3x the volume of the air in the intake tract – in 0.3s of time the engine has a whole new set of air, which will likely be at a lower temperature, getting lower and lower as the vehicle speed increases.

A heck of a lot can happen in 10 seconds. That’s not to say that for most of the time when you’re driving the car the temperature fluctuations are so minimal while cruising, even when you’re on track as the whole system is heating up evenly. It doesn’t, however, respond well if your turbo starts to go out of efficiency and creates a surge in the outlet temperatures, it could be 10 seconds before the IAT protection kicks in adding fuel or removing timing to protect against detonation. 

So what’s the solution, can’t the readings be instant? In our searching the question as with most things is a balance between cost and performance. Finding a manufacturer that will provide sensors that even read below 10 seconds brings us into the realms of motorsport. That’s where KA Sensors comes in, who supply sensors to motorsport teams for anything from turbo speed sensors, brake fluid temperature, suspension geometry and far more.

Remember our requirements started from wanting to test the efficiency of the turbo as well as the intercooler itself, so it’s got to stand up to 200c heat and also has to have as fast a response time as possible.

The NTC2F sensors from KA Sensors offer a calibrated sub 2 second response time, while that still sounds long in the context of how much air is being consumed by the engine – its as fast as possible without spending absolute mega-bucks. In reality, the sensor will respond faster the more air is being processed by it.

The other constraint was trying to find a definitive calibration for each individual sensor, now a generic standard is useful for the BOSCH / GM / FAE style sensors but we’re trying to be full Bill Nye science guy about this – it has to be accurate. These sensors are all individually tested against the master calibration ensuring accuracy. On the point of accuracy, we needed to know how accurate the sensors were at each point in the temperature range measured, this varies from 1.4c at the low end to 3.4c at the high end with most sensors. The KA Sensors Intake Air Temp Sensors are calibrated to +/- 1c at all ranges.

On paper, these will be the perfect drop-in replacement for your bosch / gm style sensor, provide faster response and more accurate temperature readings. They are also physically designed to mitigate heat soak to the sensor itself, so that should be far less of a contributing factor.

 

5 comments

  1. Ellis Blackman

    Is this sensor a direct replacement to the standard one?
    Or will I have to do some wiring?

    1. Daniel Marshall

      Hi Ellis,

      This is for use with a standalone ECU i’m afraid, assuming you have one then you’ll need to wire it up with a new connector.

  2. Zoltán Patyi

    Umm, why are you comparing a closed element sensor to an open element one? Of course the open element reacts much faster…. It might still be considerably faster, but it would be a lot more fair of a comparison

    1. Daniel Marshall

      Thanks for your comment, I’ve put in an example of the ‘other’ standard fast response open element sensor which has a thermal constant of 15 seconds.

      1. Zoltán Patyi

        Thanks, now it’s fair 🙂

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Over 300 Brands From

UK, Europe, USA & Japan

85% of Orders

Shipped Within 24 Hours

Worldwide Delivery

to 200+ Countries

Honest Expert Advice

From Real MX-5 Owners