Logging temps

Posted on February 15th, 2004 in Opinion by Julian Edgar

Those of you who have read the series (starts at Using Oscilloscopes on Cars, Part 1 ) will be aware that we’ve recently been exploring the use of digital scopes with cars. As we said in those stories, if you’re doing any modification work that involves input and output signals, the only real way to see what you’re doing is with a scope. Putting my money where my mouth is, I recently bought a digital handheld scope – a Fluke 123 Scopemeter. I got it secondhand but in as-new condition – it’s a product that I have spent nearly two years trying to find at the right price…

In addition to its abilities to display waveforms, it can also be used as a paperless chart recorder. That is, it can plot by means of a line graph the level of a signal over time. Both the time and level parameters auto-scaled, so it doesn’t really matter if you’re logging something for 10 seconds – or 3 days. You can’t download the actual logged numbers from the meter but you can dump the graph itself to a PC. So the logging function of the Scopemeter isn’t as good as you’d get with an adaptor working into a PC, but because of its speed of set-up and ease of use, it’s more likely to be used in everyday measurements.

Click for larger image

An immediate use that I thought of was to log temperature. In many areas of automotive modification, temperatures are vital. From the intake air temp to the temp drop across heat exchangers (radiators, intercoolers, auto trans coolers, oil coolers), seeing what’s happening is important. To quickly test the effectiveness of the Fluke in graphing temp over time, I plugged in a K-type thermocouple adaptor – a device that takes the very small voltage coming from a thermocouple, compensates and amplifies it and then has a known voltage output per degree. For example, the adaptor might have an output of 1 millivolt per degree C – if the meter shows 25mV, the temp is 25 degrees C. (A thermocouple adaptor is shown here, being used with a normal multimeter.)

As a simple test I logged the outlet air temp of my PC. (Using one of the thermostat switches covered at DIY Adjustable Temp Switches, I operate a couple of ancillary fans to keep the insides cool.) I placed the temp probe on the exhaust side of the outlet fan and then set the system going. The results were fascinating – you could clearly see the temp skyrocket when the fan came on (drawing hot air out of the case) and then gradually drop over 5 minutes as the PC was cooled. The fan would then switch off, accompanied by a tiny temp spike, then 7 minutes or so would pass before the fan again kicked into action. It’s not wildly exciting, but without having to do any set-up besides plug the adaptor into the Scopemeter and press a few buttons, it was all pretty easy. And it showed that it was worthwhile pursuing this area further.

The Scopmeter has the facility to take two inputs and graph them simultaneously. That is, input and output temps from (say) an intercooler can be logged. However, there’s then the necessity for two thermocouple adaptors, or an adaptor that can take two thermocouples and spit out two separate signals. I am not aware of any of the latter devices, and running two thermocouple adaptors seemed to be a bit clumsy.

Click for larger image

So I made my own adaptor.

Much of the technique has been covered in our article TempScreen Part 4 and for this adaptor, two of the AD595AQ chips were again used. However, this time I used two 9-volt batteries (giving plus/minus 9V and so allowing temps of less than zero to be measured) and also added a pair of trimming pots. The data sheet for the ICs shows a wealth of info – Monolithic Thermocouple Amplifiers with Cold Junction Compensation. In addition, I made an extension cord with banana plugs at each end so that the adaptor could be mounted remote to the meter. To finish it off I stuck it all in a box and added a flashing LED to show when the device was switched on.

So how well does it work? What’s the information like? Firstly it needs to be said that unlike the previous project, I had some difficulty in getting the two outputs to be identical. There was always a couple of degrees difference in the two temps, and if one pot was adjusted to cancel this difference, at a higher temp it reappeared (ie there was some non-linearity). However, having said that, there was never a difference of more than 2-3 degrees.

And in actual use, where trends are often more important than absolute data, the system works very well.

Click for larger image

This graph shows the logged temps of the airfilter and hot side of the intercooler in my turbo V6 Nissan Maxima. The depicted trip, which takes just under 25 minutes, is the one I do every day – the drive to my post office box. I drive for about 10 minutes, stop for 5 minutes, then drive the 10 minutes home. The top trace shows the hot side of the ‘cooler and the bottom trace, the temp of the airfilter.

The initial part of the drive involves climbing a steep hill – on this 20-degree C day you can see that intercooler temp rapidly rises to about 45 degrees. The temp spikes that follow are as I negotiate a series of intersections – the Maxima’s turbo comes up on boost very fast and the spikes show this. Meanwhile, the airfilter is just gradually rising in temp – note that the scale on the lower trace is only 15 degrees from top to bottom, unlike the intercooler temp trace which is scaled 0 – 75 degrees.

I park the car at the 10 minute point and switch it off. And here’s one of the most important things that can be seen – it’s something we’ve commented on many times in the past, but seeing is believing. With the car switched off, both temps start an inexorable climb – the airfilter rises from 25 to 28 degrees, and the underbonnet intercooler from about 25 to 35 degrees. Then when the car is restarted, the airfilter (even though it’s sealed to a bonnet vent!) jumps nearly 5 degrees, while this and the heatsoaked ‘cooler cause the next use of boost to give the greatest temp measured – 53.3 degrees.

Note how the peak filter temp was recorded at 11:16 am (nearly 11:17 in fact) and the peak intercooler temp was just over 1.5 minutes later. Basically, even after driving for only 10 minutes and then switching off the engine for just 5 minutes, the affect of heatsoak was sufficient to increase the intercooler temps when the car was again moving.

Trade an ambient of 20 for 40 degrees C, and swap having the car switched off for 5 minutes for having it stationary in city traffic with the engine running, and you can see why we’ve often said that the effects of heatsoak are critical to the actual intake air temp recorded on turbo cars.

Two more interesting points: on this little drive, the average filter temp was 23 degrees and the average of the hot side of the intercooler, 30; and there were eight temp spikes (ie uses of boost) on the way to the post office, and six on the way home.

Click for larger image

So that was the intercooler, but what about something else under the bonnet? This graph shows the inlet and outlet temps of the auto transmission cooler. When we did the tranny cooler project on the Maxima (Cooling the Trans) I did some temp measurements, but the single probe tended to show that the trans cooler was hot when the car was stopped and cooler when it was going. But the temp drop across the cooler was hard to quantify – it seemed to be all over the place.

So what was happening? The two graphs are on different scales – the top trace, which is the trans cooler outlet temp, is on a 10 – 40 degree C scale and the bottom trace, the inlet temp, is on a 0 – 75 degree scale.

The scale differences help explain the spikiness of the upper trace, but you can see that if both traces were plotted on the one scale, the traces would mimic each other fairly well. However, the max temp on the inlet pipe was 61 degrees C, and the max temp on the outlet pipe was 44 degrees – a 17degree drop across the core. However, even more interesting is the Scopemeter’s calculated averages – the average outlet temp was 32 and the average inlet temp 45 degrees – a 13 degree drop.

In fact, given that we know the day temp was about 20 degrees C, we can calculate the average efficiency of the auto trans cooler. To decrease the average intake temp of 45 degrees to the ambient of 20 degrees would need a 100 per cent efficient core, so a drop of 45 to 32 indicates an efficiency of 52 per cent (13/25 = 0.52). That’s a pretty good for a trans cooler than cost under five bucks!

Click for larger image

So what’s the radiator like? I went for a drive around a few hilly blocks, measuring inlet and outlet temps. The max inlet temp was 83 degrees C and the max outlet temp, 60 degrees – a 23 degree drop across the radiator. The averages showed 78 degrees and 52 degrees respectively – a 26 degree change. Again with an ambient temp of 20 degrees, this represents an average efficiency of 45 per cent.

One of the advantages of the Scopemeter over a PC logging adaptor is that the Fluke can handle anything thrown at it – you can’t easily blow it up by inputting too high a voltage, for example. In this sort of application it’s also very quick and easy to set up. And the input signals don’t have to be temperature, of course. It’s just as easy to graph the voltage coming out of a boost pressure sensor (perhaps with intercooler temp on the other trace) or the oxygen sensor signal.

Comments are closed.