The Lotus Cars Community banner
1 - 20 of 123 Posts

·
Registered
Joined
·
9,786 Posts
Discussion Starter · #1 ·
I've been trying to get real data on how much air actually goes through the intercooler on an Exige... I'm using a GPS to record position, and speed over time, and a datalogging weather meter to record air flow through the back of the intercooler (as well as temperature, humidity, atmospheric pressure, etc) over time.

I first calibrated my air flow readings by forcing air through the roof scoop and taking readings across the face of the intercooler. While the meter reads in fpm, given this calibration and the area of the intercooler you can then convert fpm to cfm (without getting into the nitty-gritty, multiply the fpm measurements by 0.4606 to get cfm).

I established a fixed "course", and gathered data with both the stock Exige S roof, and the Cup roof... processed all the data generating scatter plots, histograms, and linear regressions.

The bottom line is that there is no significant difference in intercooler airflow between the stock S roof and the Cup roof. How could this be? It seems that the inlet is not the limiting factor. Most of us have suspected the bottleneck where the end of the roof and the beginning of the clam meet is the limiting factor... my next step is to attempt to minimize the bottleneck and compare the results to the baseline.

First: a graph of vehicle speed vs. intercooler air flow speed, for both roofs, showing both the linear regressed results and the average value histograms.

The next two graphs are scatter diagrams of the raw data for each roof.
 

Attachments

·
Registered
Joined
·
9,786 Posts
Discussion Starter · #9 · (Edited)
Thanks for the encouragement everyone...

Yes, there's not much data above 65mph... and it's not clear that the slight increase in flow with the Cup roof is significant above 65... not many data points. I'd love to get data at higher speeds, but as you have guessed, my "course" is on local roads... and I might have problems if I tried to take too much data above 75mph :evil:

EDIT: The "fixed course" is close to a square, with approximately equal legs in each compass direction... and I tried to avoid taking data on windy days, but a sudden gust of wind could always be responsible for data points above or below the regression line.

I wanted to make sure I could reproduce the conditions somewhat reliably so that the data could be compared properly.

I've just started "fettling"... I'll post more results when I've got them...
 

·
Registered
Joined
·
9,786 Posts
Discussion Starter · #11 ·
Great data, even at lower speeds, I'm surprised the Stock S roof flows that well. Now, how do we determine drag of the two types of roof scoops?

But yeah, higher speed data may change the dynamics of that chart considerably. Get your car on a track with a good long straights :)
Yes, I think it's at least possible to say there's no "reverse flow" through the intercooler... I had heard rumors to that effect previously.
 

·
Registered
Joined
·
9,786 Posts
Discussion Starter · #15 · (Edited)
This is excellent! Very good job! Keep the data coming!

Also, can you post a scatterplot with the datapoints color coded according to which roof produced them? And then overlay the regression line against the scatterplot. Would be nice to see how the datapoints group or don't group around or to either side of the regression line.

Thx! Can't wait for more.
How's this?

EDIT: Only data points with vehicle speeds > 5 mph and air flow speeds > 0 fpm were used in the regression analysis...
 

Attachments

·
Registered
Joined
·
9,786 Posts
Discussion Starter · #17 · (Edited)
Great work! I notice a dip in the points of both graphs at 45 mph. This is where I noticed "reverse flow". Do you find any significance to this trough in the data points? Would you consider testing my intercooler ducts that take airflow from the side scoops and funnel it through the intercooler from the back? I'm sure many people on this forum would love to see the comparison data.
I wouldn't be surprised if there are transitions in the flow around the body that affect either flow into the scoop, or pressure in the engine compartment. That could easily lead to dips/peaks around the regression line. I'd love to have a full scale wind tunnel, but none of the F1 teams with those facilities returned my calls ;).

I read your solution with interest... I've been considering a hybrid solution that incorporates your side scoop ducting, but feeds it into the front of the intercooler in conjunction with the flow from the roof scoop. There's not alot of room in there though... My first mod will just be to fettle out the bottleneck and perhaps seal/smooth the flow from the roof scoop.

EDIT: Also, one of my planned tests is to measure air flow at each of the side scoops to get some idea of the potential volume of air available...

EDIT 2: Calibrating the air flow sensor... yeah, OK... it's not NASA or McLaren. But it works.
 

Attachments

·
Registered
Joined
·
9,786 Posts
Discussion Starter · #24 ·
The hybrid solution you speak of: I tried that once and didn't feel there was enough room. But, hey I could supply you with side panels and hoses and you could experiment.

And I notice you are in San Diego? I would love to brainstorm with you. Just spent the last couple days perfecting a filter that saves Chrysler minivan transmissions from self destructing. Let me know if I can supply anything you need.
Yes, my Lotus is in San Diego, and as of now, so am I. I would be interested in trying out the panels and hoses... it might be a couple of weeks before I'm ready to get to that stage... I'd also like to see your mod in person too.

check out this pic of an Exige S whose rear clam was fettled... Looks to me you guys can go even further to open it up... maybe where i circled and maybe on the inside of the opening there's a few more mm to grind away?
It's hard to see the way the ductwork is attached to the clam, but there's not too much you can grind away in the area you indicated. Here are pics from the "fettling" so far:

1) Before fettling... rubber snorkel is disconnected and a light is shining in the gap.

2) Beginning of fettling... you can see how much material I'm removing

3) After fettling (for now). Road test tomorrow...

EDIT: Note that the rubber gasket on the clam's inlet has been removed... the Cup roof has a gasket on the mating surface already, the stock S roof does not. I moved the gasket from the clam to the stock roof for testing purposes...
 

Attachments

·
Registered
Joined
·
9,786 Posts
Discussion Starter · #31 ·
APK - anything I can do to help, let me know. I will supply anything I have to save you time/trouble/expense. I am interested in data points that show conclusively which configuration would cool the best whatever that may be.
Thanks... do you have your car (or a car) set up with your ductwork? I can fairly easily log data from any car with a standard sized intercooler. All the equipment is portable and easily configured. Maybe we could do a baseline with your configuration
 

·
Registered
Joined
·
9,786 Posts
Discussion Starter · #32 · (Edited)
After the "First Fettle"

One of the things I had briefly mentioned was the need to calibrate the air flow meter... I set up a blower to force air into the roof scoop, and then took readings from the back face of the intercooler. Dividing the face into 15 segments (5 wide by 3 high), I took flow readings (in fpm) at all 15 segments, then a flow reading from the location that the meter would occupy during the rolling test. By averaging the readings, I could estimate the volumetric flow being forced through the intercooler by the blower by multiplying the average fpm by the area of the intercooler (0.6076 sqft.). Then the ratio of the average flow to the meter location flow would give me a correction factor for the logged data, i.e. if the average was 283.5 fpm, the meter read 374 fpm in its logging location, I'd have to apply (283.5 / 375) * 0.6076 to the logged fpm readings to approximate cfm through the intercooler.

Below are the calibration matrices for the pre- and post- fettle calibrations. Notice that after fettling that the average flow is higher, and better distributed across the face of the intercooler. I also tried two positions for the meter during the post-fettle calibration... the original pre-fettle position, and an alternate position that gave the same static flow as the pre-fettle position

First (left) image: Calibration matrix before fettling
Second (right) image: Calibration matrix after fettling (with alternate sensor position)
 

Attachments

·
Registered
Joined
·
9,786 Posts
Discussion Starter · #34 · (Edited)
And now some results...

First, it was not a good day for data logging... it was gusty enough to cause a lot of noise in the data. Second, I tried two different locations for the air flow sensor, as mentioned above. The first location was identical to the pre-fettle sensor location. I actually got slightly lower flow, but that was anticipated because of the changed calibration matrix. The second location was chosen to give the same flow during the static calibration post-fettle as the first location gave pre-fettle. That gave higher flows. But once both locations were corrected by the calibration factors, the results were close to identical (as they should be).

Basically the amount of fettling so far has increased flow over stock by about 20%. I'm not particularly happy with today's data though... so more investigation is in order. On the other hand, 10 minutes with a dremel tool and you get at least 20% more air through your intercooler. Not bad, but I'm not done yet. Also, I noticed during the static tests that some air was escaping from under the roof/clam interface and along the clam's rain gutter. More room for improvement.

Note: The vertical axis on the previous graphs were in fpm... those below are in "corrected" cfm (as per discussion above)

Graph 1: pre- and post-fettle data with original sensor location post-fettle
Graph 2: pre- and post-fettle data with alternate sensor location post-fettle
 

Attachments

·
Registered
Joined
·
9,786 Posts
Discussion Starter · #36 ·
apk the graphs are fantastic! Really clear data. Very, very cool to see someone going through with real data collection. You're a true Mythbuster!

Any chance of throwing in the non-cup roof (Exige S) into the data? Do you still have one available?

Thx for the great reporting!
Thx! I've been getting better/faster at swapping roofs... but I think that I'm going to wait until I've done more fettling, etc. before I swap the stock roof... but that is in the plan... (and yes, I have access to both the Cup roof and the stock Exige S roof)
 

·
Registered
Joined
·
9,786 Posts
Discussion Starter · #43 · (Edited)
...i'm interested to see comparative data using side scoop intakes...i've wondered how cool the air really is coming into the exige roof scoop by comparison to other points along the body - i know that when i put my hand out the roof of my elise at speed, i can feel all that radiator heat in the laminar flow licking up and over the roof...
The meter I'm using is logging temperature (and a few other parameters) in addition to air flow. There's some interesting data there, but since it's downstream of the intercooler it won't show the temperature of the inlet air... it's a complex combination of air heated by the intercooler and convected/radiant heat from the engine. I might yet try to extract some information from this data as well.

I'm planning on taking airflow/temperature data from the side scoops as well. This scope of the project is beginning to grow...
 

·
Registered
Joined
·
9,786 Posts
Discussion Starter · #44 · (Edited)
More interesting info... for those of you that want to see the temperature of the air that's passed through the intercooler. Note that this is not charge temperature (although I'd love to log that as well... maybe in the future).

Here's a graph of downstream intercooler air temperature and vehicle speed over time. There are some very interesting things happening here... but this is just raw data and I haven't done any analysis...

EDIT: Outside air temperature was about 86 degrees F
 

Attachments

·
Registered
Joined
·
9,786 Posts
Discussion Starter · #46 · (Edited)
Now that's a very interesting dataset. It appears that once you're at roughly 40mph or more that the post-cooler can get to and stay in the range of 102 plus or minus a couple degrees. So 15 degrees warmer than the presumably ambient pre-cooler air.

Except that presuming the pre-cooler air is ambient is probably a mistake, what with all the warm air coming off the radiator and flowing right up the windshield into the scoop.

xtn
That's true... plus the temp sensor is only a few inches away from the valve cover... note how the temperature shoots up at the end of the dataset when I've shut the engine off in the garage...

I'm thinking of rigging up something to allow me to put the airspeed/temp sensor in the Cup roof scoop just behind the inlet grill... that would generate flow/temp readings vs. car speed without the other problems... and should show whether the radiator air flow is affecting inlet temperature in a significant way
 

·
Registered
Joined
·
9,786 Posts
Discussion Starter · #48 · (Edited)
time to decrease temp across speed per temp drop would make for some interesting comparisons. Looks like you have the data for it already (that was about a 40 min drive, if I read correctly). Plenty of stops/gos in there too. Those kind of comparisons would be good benchmarks for comparing differences in heatsoak/dissipation for other intercoolers.

At the end, was that just sitting at idle, where the speed=0 and the temp continues to rise?

Might I recommend you connect with Thomasio from rls? It would be great to throw his rls intercooler into the mix of your data collection. I think a lot of people would benefit from the knowledge of how changing the intercooler does/does not impact heat soak and charge temp.

And also, of course, to test Shoco's side-intake cooling kit. What a difference that would make for folks to see the end results of these kits.
One of the things that's hard to see in the graph is engine load... obviously when I'm accelerating the supercharger is working harder, increasing the flow and heat of the charge... I think you can clearly see some of that heat being transferred to the cooling air. At the end the engine was off, not idling... you can see the effect of the latent heat in the engine block without coolant circulation. What's displayed is about 15 minutes of data... the samples are spaced 2 seconds apart.

There are so many ways to take this study at this point... I'd like to exhaust the "fettling" first, since that's the simplest and cheapest mod to increase intercooler efficiency. After that, side scoop ducting, different intercoolers, etc. would be interesting routes to follow.

What I'd also like to do is get my datalogging OBD-II reader in the mix. Then I can log coolant temps, engine load, MAF sensor readings, etc. Lots more work to do...
 

·
Registered
Joined
·
9,786 Posts
Discussion Starter · #50 · (Edited)
I have a datalog for the trip from las vegas to los angeles, i keep meaning to post it up, most of it is just freeway constant but there are some traffic and hill climbs, let me find it and post it.
Cool... what parameters did you log?

By the way... it seems someone else has attempted to combine roof and sidescoop air into the intercooler: Elise.ch :: Connexion

 

·
Registered
Joined
·
9,786 Posts
Discussion Starter · #52 ·

·
Registered
Joined
·
9,786 Posts
Discussion Starter · #58 ·
Andy are you going to order the RLS intercooler? Just wondering if you did and if so, it would be interesting to run the same tests on it.
I'm not ordering it at this point... first I'd like to push things as far as possible using the stock equipment... and right now I'm not acquiring data that could be used to compare the efficiency of the intercoolers (to do that properly I should be logging charge temperature, both before and after the intercooler).
 

·
Registered
Joined
·
9,786 Posts
Discussion Starter · #60 ·
Do you have any data with the roof removed completely? (i presume this will be difficult though as the air flow and turbulance over the gap where the roof would be would make it hard to gather accurate readings) or even with the canvas roof from the elise fitted?
I haven't yet but I'd like to... I don't have a soft top (or the grommets for the support ribs), so it would just be a "topless" test. I think I'd still be able to calibrate the sensor for that configuration, so I don't think additional turbulence would affect things too badly...
 
1 - 20 of 123 Posts
Top