Biggest Intercooler Ever (on the street)

[The Interceptor]

I saw a nice orange RX4 coupe in Perth on the weekend: it had an intercooler
that was not only taller than the radiator, it was so wide that the inside
two of the four headlights were taken out by the intercooler.

I like rotary people, probably because they don't drive front wheel drive
naturally aspirated Lancers. But I reckon this guy will be getting a yellow
sticker sooner rather than later.

Brett

[Fraser Johnston]

"The Interceptor" <thisemailadress@willnotwork.com.au> wrote in message
news:1105445109.209698@quartz.westnet.net.au...
> I saw a nice orange RX4 coupe in Perth on the weekend: it had an
intercooler
> that was not only taller than the radiator, it was so wide that the inside
> two of the four headlights were taken out by the intercooler.
>
> I like rotary people, probably because they don't drive front wheel drive
> naturally aspirated Lancers. But I reckon this guy will be getting a
yellow
> sticker sooner rather than later.
>
> Brett

Must be a truck intercooler. There was a few of them at rotomotion when I
had my RX-7. I hear they are not very good on cars because they are made
for big laggy truck turbos.

Fraser

[Graham W]

Fraser Johnston wrote:
> Must be a truck intercooler. There was a few of them at rotomotion when I
> had my RX-7. I hear they are not very good on cars because they are made
> for big laggy truck turbos.

How would you go about making an intercooler which was not good for a car?

Seems to me there are three main ways:
1) Make it too restrictive for the volume flow rate the car is capable
of drawing. Diesel truck motors draw a lot more air than a small four
cylinder or rotary, but they might also tolerate more restriction. I
doubt this is a problem.

2) Make it too small so that it doesn't actually cool the air enough to
be worth the restriction it causes. I don't see this being a problem.

3) Make it far larger than it needs to be, so you're lugging around more
mass and wind resistance than you need.


The only other possibility I can think of is that a really large
intercooler, like a really large, long pipe, contains a large mass of
air with significant inertia, making lag worse.

Warning: techy back of envelope calculation follows.

================================================== =======================
Air at atmospheric pressure has a density of about 1.2kg/m^3.

Let's estimate an average cross section area for flow within the
intercooler of say 100mm x 100mm, 0.01m^2, which is probably generous.

For each 100 cubic feet per minute flow through the intercooler, (0.047
cubic metres per second) the air velocity increments by 4.7 m/s.

Lets assume you start at 100cfm and stamp on the throttle and measure
lag to 300cfm. At 100cfm, v=4.7, at 300, v=14.1.

The kinetic energy of the moving column of air is 0.5 m v^2, so the
change in kinetic energy of the air column is 0.5 x (177) x m = 88m

I'll take a wild guess at 0.01 cubic metres of air space in the
intercooler, so we're looking at something like one joule.

To impart one joule in say 0.2 second you'll need 5 Newton; over a cross
section area of 0.01m^2 thats 500 Pascals of (transient) restriction.
Atmospheric pressure is 100,000 Pascals.
500 Pascals is about 50mm water pressure.

If the cross section is smaller, velocity squared increases and the area
the pressure works across decreases, making things worse proportional to
one over the cube of the cross sectional area, so a smaller cross
section longer path through the same intercooler is much worse. Halve
the cross section from 0.01m^2 to 0.005m^2 and the transient restriction
would increase by a factor of 8 to 4kPa.
================================================== =======================
Conclusion?

Maybe inertia of the air within the intercooler is significant.
Anyone got any real world experience with this?
If I got keen I could set up a spreadsheet and do more accurate numbers.

[Chris]

Just reading that made my brain ache! I should have stayed at school longer!
Not even going to try and follow that!
"Graham W" <zebedee@alphalink.commercial.au> wrote in message
news:34jhl6F4a74a7U1@individual.net...
> Fraser Johnston wrote:
> > Must be a truck intercooler. There was a few of them at rotomotion when
I
> > had my RX-7. I hear they are not very good on cars because they are
made
> > for big laggy truck turbos.
>
> How would you go about making an intercooler which was not good for a car?
>
> Seems to me there are three main ways:
> 1) Make it too restrictive for the volume flow rate the car is capable
> of drawing. Diesel truck motors draw a lot more air than a small four
> cylinder or rotary, but they might also tolerate more restriction. I
> doubt this is a problem.
>
> 2) Make it too small so that it doesn't actually cool the air enough to
> be worth the restriction it causes. I don't see this being a problem.
>
> 3) Make it far larger than it needs to be, so you're lugging around more
> mass and wind resistance than you need.
>
>
> The only other possibility I can think of is that a really large
> intercooler, like a really large, long pipe, contains a large mass of
> air with significant inertia, making lag worse.
>
> Warning: techy back of envelope calculation follows.
>
> ================================================== =======================
> Air at atmospheric pressure has a density of about 1.2kg/m^3.
>
> Let's estimate an average cross section area for flow within the
> intercooler of say 100mm x 100mm, 0.01m^2, which is probably generous.
>
> For each 100 cubic feet per minute flow through the intercooler, (0.047
> cubic metres per second) the air velocity increments by 4.7 m/s.
>
> Lets assume you start at 100cfm and stamp on the throttle and measure
> lag to 300cfm. At 100cfm, v=4.7, at 300, v=14.1.
>
> The kinetic energy of the moving column of air is 0.5 m v^2, so the
> change in kinetic energy of the air column is 0.5 x (177) x m = 88m
>
> I'll take a wild guess at 0.01 cubic metres of air space in the
> intercooler, so we're looking at something like one joule.
>
> To impart one joule in say 0.2 second you'll need 5 Newton; over a cross
> section area of 0.01m^2 thats 500 Pascals of (transient) restriction.
> Atmospheric pressure is 100,000 Pascals.
> 500 Pascals is about 50mm water pressure.
>
> If the cross section is smaller, velocity squared increases and the area
> the pressure works across decreases, making things worse proportional to
> one over the cube of the cross sectional area, so a smaller cross
> section longer path through the same intercooler is much worse. Halve
> the cross section from 0.01m^2 to 0.005m^2 and the transient restriction
> would increase by a factor of 8 to 4kPa.
> ================================================== =======================
> Conclusion?
>
> Maybe inertia of the air within the intercooler is significant.
> Anyone got any real world experience with this?
> If I got keen I could set up a spreadsheet and do more accurate numbers.

[atec]

I expect in real term if the drop across the cooler is above a specific
point then it lags like buggery... real world experience tells me
larger pipe dia is very desirable and so is core thickness as opposed to
actual larger surface area gained by height and width if that makes
sense...

Graham W wrote:
>
> Fraser Johnston wrote:
> > Must be a truck intercooler. There was a few of them at rotomotion when I
> > had my RX-7. I hear they are not very good on cars because they are made
> > for big laggy truck turbos.
>
> How would you go about making an intercooler which was not good for a car?
>
> Seems to me there are three main ways:
> 1) Make it too restrictive for the volume flow rate the car is capable
> of drawing. Diesel truck motors draw a lot more air than a small four
> cylinder or rotary, but they might also tolerate more restriction. I
> doubt this is a problem.
>
> 2) Make it too small so that it doesn't actually cool the air enough to
> be worth the restriction it causes. I don't see this being a problem.
>
> 3) Make it far larger than it needs to be, so you're lugging around more
> mass and wind resistance than you need.
>
> The only other possibility I can think of is that a really large
> intercooler, like a really large, long pipe, contains a large mass of
> air with significant inertia, making lag worse.
>
> Warning: techy back of envelope calculation follows.
>
> ================================================== =======================
> Air at atmospheric pressure has a density of about 1.2kg/m^3.
>
> Let's estimate an average cross section area for flow within the
> intercooler of say 100mm x 100mm, 0.01m^2, which is probably generous.
>
> For each 100 cubic feet per minute flow through the intercooler, (0.047
> cubic metres per second) the air velocity increments by 4.7 m/s.
>
> Lets assume you start at 100cfm and stamp on the throttle and measure
> lag to 300cfm. At 100cfm, v=4.7, at 300, v=14.1.
>
> The kinetic energy of the moving column of air is 0.5 m v^2, so the
> change in kinetic energy of the air column is 0.5 x (177) x m = 88m
>
> I'll take a wild guess at 0.01 cubic metres of air space in the
> intercooler, so we're looking at something like one joule.
>
> To impart one joule in say 0.2 second you'll need 5 Newton; over a cross
> section area of 0.01m^2 thats 500 Pascals of (transient) restriction.
> Atmospheric pressure is 100,000 Pascals.
> 500 Pascals is about 50mm water pressure.
>
> If the cross section is smaller, velocity squared increases and the area
> the pressure works across decreases, making things worse proportional to
> one over the cube of the cross sectional area, so a smaller cross
> section longer path through the same intercooler is much worse. Halve
> the cross section from 0.01m^2 to 0.005m^2 and the transient restriction
> would increase by a factor of 8 to 4kPa.
> ================================================== =======================
> Conclusion?
>
> Maybe inertia of the air within the intercooler is significant.
> Anyone got any real world experience with this?
> If I got keen I could set up a spreadsheet and do more accurate numbers.

[John McKenzie]

I'm certainly not going to say this is for "all" cases, but if you look
at a fair few videos available on the net of really quick rotaries and
even quick high boost 2 litre fours (and even some VLs) they're plenty
laggy. You can actually hear them spooling up before they launch.

If they were less invovled in an ET record, and more interested in match
racing (and lets be honest, that's a whole different kettle of fish,
since the timers only start when you break the staging beam, whereas the
first one across the finish line in a race proper (providing they don't
redlight) is the winner. In other words, in one case you can spool and
load up and launch however you want, and in hte other, a quick
consistent launch is worth far more than the best et.

--
John McKenzie

tosspam@aol.com abuse@aol.com abuse@yahoo.com abuse@hotmail.com
abuse@msn.com abuse@sprint.com abuse@earthlink.com fraudinfo@psinet.com
sweep.day@accc.gov.au uce@ftc.gov admin@loopback $LOGIN@localhost
$LOGNAME@localhost $USER@localhost $USER@$HOST -h1024@localhost
root@mailloop.com president@whitehouse.gov vice.president@whitehouse.gov
abuse@iprimus.com.au abuse@cia.gov abuse@fbi.gov abuse@asio.gov.au
abuse@federalpolice.gov.au

[Graham W]

atec wrote:
> I expect in real term if the drop across the cooler is above a specific
> point then it lags like buggery... real world experience tells me
> larger pipe dia is very desirable

It's not the drag from continuous flow which contributes to turbo lag.
That just saps power all the time. As you rightly say, more pipe
diameter is desirable for steady state pressure drop anyway.

Lag is a dynamic effect, where followingan aburpt change of throttle
opening, the power takes time to rise because of momentum in the
turbocharger elements. Based on the (extremely rough back of an
envelope) numbers I came up with below, I wouldn't have thought that the
kinetic energy of the air within the intercooler would be an issue
unless it was far far larger than it needed to be.

> and so is core thickness as opposed to
> actual larger surface area gained by height and width if that makes
> sense...

As for whether it's better to have more frontal area or just more
layers/thickness, thats also something we could do a back of the
envelope calculation on...

Very very quickly, if the car is doing 30m/s and the intercooler has
area A square metres perpendicular to the airflow at that speed, then it
sees 30xA cubic metres per second.

For a typical intercooler (estimating from what I see on the road) of
say 300mm x 200mm, 0.06 square metres, the amount of air flowing over
the outside at a road speed of 30m/s is about 1.8 cubic metres per second.

Remembering that 100 cubic feet per minute is equal to 0.047 cubic
metres per second, the flow inside the intercooler is of the order of
0.1 to 0.2 cubic metres per second.

That gives an outside to inside flow ratio of the order of 9:1, which in
turn would mean the temperature change ratio, outside to inside, would
be 1:9.

Even allowing for the fact that perhaps only half the air goes through
the intercooler rather than around it, that means the air on the outside
doesn't get very much hotter, so whilst more layers instead of increased
area isn't ideal, it's not much of a penalty either. I would think that
twice as thick probably works as well as 50% larger area, with less drag
and probably a similar weight.

Again, this is really rough numbers, but enough to check the concepts.
If the ratio of air flow on inside and outside had been 1:1 then it
would be clear that more area was needed to get more cooling.

[Graham W]

Chris wrote:
> Just reading that made my brain ache! I should have stayed at school longer!
> Not even going to try and follow that!

Hey, I gave fair warning!

> Graham W wrote:
>>Warning: techy back of envelope calculation follows.

[Graham W]

John McKenzie wrote:
> I'm certainly not going to say this is for "all" cases, but if you look
> at a fair few videos available on the net of really quick rotaries and
> even quick high boost 2 litre fours (and even some VLs) they're plenty
> laggy. You can actually hear them spooling up before they launch.
>
> If they were less invovled in an ET record, and more interested in match
> racing (and lets be honest, that's a whole different kettle of fish,
> since the timers only start when you break the staging beam, whereas the
> first one across the finish line in a race proper (providing they don't
> redlight) is the winner. In other words, in one case you can spool and
> load up and launch however you want, and in hte other, a quick
> consistent launch is worth far more than the best et.

Absolutely, there are applications where lag is a complete non issue,
the really obvious ones being turbo diesel generator sets and really
large trucks, where nothing changes very suddenly anyway, whereas at the
opposite extreme it would be really nasty on a road ridden sports bike,
where the instant power on tap is used to affect the handling.

[fulliautomatix]

aburpt

well, it would eventually...after the lag

[Graham W]

fulliautomatix wrote:

> aburpt
> well, it would eventually...after the lag

COuld you perhaps type in english and leave sufficient of the previous
post to give us some clues to what you are trying to say?

[The Interceptor]

> Absolutely, there are applications where lag is a complete non issue,
> the really obvious ones being turbo diesel generator sets and really
> large trucks, where nothing changes very suddenly anyway, whereas at the
> opposite extreme it would be really nasty on a road ridden sports bike,
> where the instant power on tap is used to affect the handling.

I'd argue that power generation is one place where you *can't* cop a big
amount of turbo lag, especially if you are talking about off-the-grid
generation where you can get large impact loads, such as large electric
motor starts and so on. However, in power gen the diesel gensets *usually*
are normally operating well into the boosted zone, so it's a bit of a moot
point in that regard. I could give you some examples of EMD generating sets
having issues with turbo lag when copping big impact loads when running
around 50% load, but it's a long and complicated story.

Brett

[Graham W]

>>Absolutely, there are applications where lag is a complete non issue,
>>the really obvious ones being turbo diesel generator sets and really
>>large trucks, where nothing changes very suddenly anyway, whereas at the
>>opposite extreme it would be really nasty on a road ridden sports bike,
>>where the instant power on tap is used to affect the handling.

The Interceptor wrote:
> I'd argue that power generation is one place where you *can't* cop a big
> amount of turbo lag, especially if you are talking about off-the-grid
> generation where you can get large impact loads, such as large electric
> motor starts and so on.

I thought about that, and you're right, off grid they do need to have
stable speed, but the inertia of the generating components helps with
that too.

> However, in power gen the diesel gensets *usually*
> are normally operating well into the boosted zone, so it's a bit of a moot
> point in that regard. I could give you some examples of EMD generating sets
> having issues with turbo lag when copping big impact loads when running
> around 50% load, but it's a long and complicated story.

Sounds like you know more about gensets than I do...

[Fraser Johnston]

"Graham W" <zebedee@alphalink.commercial.au> wrote in message
news:34jhl6F4a74a7U1@individual.net...
> Fraser Johnston wrote:
>> Must be a truck intercooler. There was a few of them at rotomotion when
>> I
>> had my RX-7. I hear they are not very good on cars because they are made
>> for big laggy truck turbos.
>
> How would you go about making an intercooler which was not good for a car?
>
> Seems to me there are three main ways:
> 1) Make it too restrictive for the volume flow rate the car is capable of
> drawing. Diesel truck motors draw a lot more air than a small four
> cylinder or rotary, but they might also tolerate more restriction. I doubt
> this is a problem.
>
> 2) Make it too small so that it doesn't actually cool the air enough to be
> worth the restriction it causes. I don't see this being a problem.
>
> 3) Make it far larger than it needs to be, so you're lugging around more
> mass and wind resistance than you need.
>
>
> The only other possibility I can think of is that a really large
> intercooler, like a really large, long pipe, contains a large mass of air
> with significant inertia, making lag worse.
>
> Warning: techy back of envelope calculation follows.
>
> ================================================== =======================
> Air at atmospheric pressure has a density of about 1.2kg/m^3.
>
> Let's estimate an average cross section area for flow within the
> intercooler of say 100mm x 100mm, 0.01m^2, which is probably generous.
>
> For each 100 cubic feet per minute flow through the intercooler, (0.047
> cubic metres per second) the air velocity increments by 4.7 m/s.
>
> Lets assume you start at 100cfm and stamp on the throttle and measure lag
> to 300cfm. At 100cfm, v=4.7, at 300, v=14.1.
>
> The kinetic energy of the moving column of air is 0.5 m v^2, so the change
> in kinetic energy of the air column is 0.5 x (177) x m = 88m
>
> I'll take a wild guess at 0.01 cubic metres of air space in the
> intercooler, so we're looking at something like one joule.
>
> To impart one joule in say 0.2 second you'll need 5 Newton; over a cross
> section area of 0.01m^2 thats 500 Pascals of (transient) restriction.
> Atmospheric pressure is 100,000 Pascals.
> 500 Pascals is about 50mm water pressure.
>
> If the cross section is smaller, velocity squared increases and the area
> the pressure works across decreases, making things worse proportional to
> one over the cube of the cross sectional area, so a smaller cross section
> longer path through the same intercooler is much worse. Halve the cross
> section from 0.01m^2 to 0.005m^2 and the transient restriction would
> increase by a factor of 8 to 4kPa.
> ================================================== =======================
> Conclusion?
>
> Maybe inertia of the air within the intercooler is significant.
> Anyone got any real world experience with this?
> If I got keen I could set up a spreadsheet and do more accurate numbers.

Mainly pressure drop across an inefficient core and lag by the increased
distance of the inlet tract from the turbo.

Fraser

[Fraser Johnston]

"John McKenzie" <jmac@alphalink.com.au> wrote in message
news:41E4B931.191A@alphalink.com.au...
> I'm certainly not going to say this is for "all" cases, but if you look
> at a fair few videos available on the net of really quick rotaries and
> even quick high boost 2 litre fours (and even some VLs) they're plenty
> laggy. You can actually hear them spooling up before they launch.
>
> If they were less invovled in an ET record, and more interested in match
> racing (and lets be honest, that's a whole different kettle of fish,
> since the timers only start when you break the staging beam, whereas the
> first one across the finish line in a race proper (providing they don't
> redlight) is the winner. In other words, in one case you can spool and
> load up and launch however you want, and in hte other, a quick
> consistent launch is worth far more than the best et.
>
With launch control this is much less of an issue than it used to be.

Fraser