Here's the beef: I've been trying to watch my compressor outlet temps on my twin turbo setup on a HSR to make sure i stay out of any detonation. As of now, it is non-intercooled and running ~9psi. The compressor outlets are simply y'ed together and connect to the throttle body. Probably less than 2 feet of pipe from each turbo to the TB.

Problem is, i never see the temps move more than 10-15*. I would think transitioning from light throttle cruise on the highway to a sustained, WOT full boost blast would show significant swings in temperatures.

But it doesnt.

Heres a snippet from a datalog i took tonight while driving on the freeway. Its only reaching 5-6psi for reasons im going to investigate tomorrow, and yes, i know its getting DANGEROUSLY lean as revs increase, also to be addressed tomorrow.



The sensor i'm using is the stock brass encased piece. I've read a lot about the "open element" sensor with a plastic cage that will react faster, but it seems like the brass one should change somewhat under boost. Im suspicous theres somethign else going on, what could cause this?

 

I pretty sure I'am having the same problem with my sensor to (brass one) seems to me if I can remember correctly it only moves a degree at the most with my setup. To bad my laptop is not working right now otherwise I would check my logs to make sure.

 

That factory sensor is about worthless for IAT reading on a boosted engine, especially when it is mounted in the bottom rear of the plenum. It will usually increase in temp very slowly, and decrease even slower because it absorbes heat from the plenum. I mount mine in the IC piping just before the TB, but it still was way too slow to be meaningfull. When I needed to know IAT readings for both my TT L98 and on my turbo Z24, I used my DVOM temp sender and just run it thru a connecting hose in the IC piping (I put the sensor in the pipe and then slide the hose back over the wire leads and clamp it just hard enough so it doesn't come apart under boost). This thing responds very quickly, it even goes up and down when as you go on and off boost. Plus, if you have an IC, you can easily test temperature at the compressor outlet and IC outlet to verify it's effeciency.

Here is a quick pic to show you how I mount the sensor:

 

Hey I like that idea!! what make is you DVOM?

 

My DVOM is a PDI. It is not a name brand like Fluke, but it was very well optioned for what it cost. Actually, I got mine free in a Ford/AAA auto diagnostics contest from when I was in school. I've heard other techs who really like them as well. The temp lead is very useful, as the sensor is even smaller than the wire leads- it fits anywhere.

 

I got my PDI from the Ford/AAA Contest too! I love it.


Zac

 

Well, I did some research on this and read a bunch of posts in the DIY PROM board and decided there were 2 options. One was to use a v6 (i think) sensor that has an open plastic cage instead of a brass cap. The other was to cut away the cap from the stock sensor and use the bare thermistor. I did the latter and made some fairly decent improvements.

(I did not come up with this myself, but read about it in several threads on the DIY PROM board some time ago which a simple search should turn up)

I cut the stock sensor just above the threads so i would expose as much of the sensor as possible.



There was a small "paper towel tube" kind of cardboard spacer between the sensor and the brass plug, and the sensor itself was glopped up with what most likely would be a thermal paste.



I removed all of the paste that i could using paper towels, toothpicks etc. and was left with this bare thermistor. The two legs go right to the pins of the connector, its that simple.



Everyone who wrote about this cautioned that engine vibrations probably would break the solder connections if left as shown above, and send the thermistor on through the engine and through the turbos. Not good. The solution is to fill in around the sensor with silicone (i used hi-temp gasket maker). I'm sure i could have covered the thermistor itself and gotten pretty accurate readings, but i tried to keep it completely clean of rtv for the best results. This is after meticulous cleaning of the sensor with a toothpick.



I let the RTV set up and then did some comparison testing between the modified sensor and a stock one (it was actually the CTS since I had already cut the IAT up, but they are the same).

I tested the sensors using a hairdryer with an output of around 160F. I made sure each sensor was at ambient temperature, then hit it with the hairdryer on high about 1" away from the sensor for 30 seconds. I then switched the hairdryer to "cool" (which turned out to be a little warmer than ambient) for another 30 seconds to keep air flowing. I logged data from each test and compiled it here. Apologies for the crappy screen captures. I should have used the same sensor for both tests, and the timing was a little off, but it gives a pretty good picture of the improvements.



You can see the that the stock sensor is slower to respond, and did not even reach a stable temperature in 30 seconds, where the bare sensor did. Most boost situations probably will be much less than 30 seconds, so this fast response is critical.

Now, this test was done in open air. I used a box end wrench to hold the sensor end up in the air stream. It doesnt take heat soak into account whatsoever, mostly because installing the IAT requires removal of the plenum (stealth ram).

I'll have to get some driving results as soon as i get it all back together, and get some new hose connectors. The crappy home depot rubber pvc connectors didn't like that kind of heat and pressure and split clear open (this is why I wasnt making much boost in the first datalog ).

 

I like this idea- great info here!

BTW, I had the same problem on my turbo cavalier- tried to save money by buying the cheap PCV pipe rubber connectors- two of them blew apart within one day of hard driving. Ended up costing me more in the long run because I still had to upgrade to real turbo hose in the end. All in the name of R&D I guess.