I'm been thinking that a liquid intercooler would help me over my dsm SMIC. I have an extra DSM around here and was thinking that I could weld some sheet metal around the DSM and flow some water to it. I figure I would use my windshield fuild since its not hooked up, an old fuel pump, and some type of heat exchange from the bone yard. I might take it to a welder to have this done, I dont think my flux welder will do sheet metal.

What do yall think?

 

It could probably work. They say for best results, when converting an air-air unit to air-water, to reverse the flow. As in, make the charge air go through the front of it, and run the water where the charge used to go. I don't know what the dimentions of your IC are, so maybe this is not really feasable. Either way, best of luck.

 

NOt quite sure how feasible that is, considering that the intercooler has a big opening for the air to go thru (2" or so probably)... it would take a big pump to flow water thru that I'd think.

 

It works great. The DSM sidemount is actually one of the easier ones to convert to water. You definately don't want to 'reverse the flow.' You keep the air going where it was designed to go, and flow water across it with a jacket welded around to contain the water. Since it is aluminum, it has to be TIG welded, which can get expensive. I have one in my garage, so I've built a few things like that in the past. My current setup uses a core from a Saab with custom end tanks and water jacket.



-Chris

 

here is my a/w intercooler made from a dms sidemount and my heat exchanger 22x6.5x2

 

looks good

 

anyone else? Can I make DSM a jacket with sheetmetal?

 

I think we should discuss, what pumps, what resovoir, and what kind of radiator, these setups have, kudos to those of you that have done this already.
-PHiZ

 

I was actually pondering about this a little while ago. I came across this link http://www.autospeed.com/cms/article.html?&A=0090. It is a second part of an intercooling guide (first part explained air-air intercooled systems). It explains the parts of an air-water intercooled system, the functions and what not. This should certainly provide you with some insight.

EP

 

here is pic of my resivoir. it will do for now.

 

Todd at Overblownmotorsports has a converted DSM SMIC like Prelude91siT's. He originally fabbed it for his Integra, it's sitting there collecting dust I think. He used to have pictures up on his site, I can't find them now.

There are two popular pump choices, I forget their names. They are the ones used by the Sy/Ty and GN guys... one is stock/stock equivalent that has a shaft driven impeller and is considered inferior. If the shaft seal leaks, coolant/water makes it back to the motor and shorts it out. If a piece of trash gets caught in the impeller and seizes it, the motor burns out. They last a good long while, really, but when dealing with the lifespan of the vehicle and overheating your intake charge, they start looking real sucky.

The other alternative is a sealed motor with a magnetic driven impeller. Motor can't get wet, trash stuck in the impeller can't burn out the motor since it's not mechanically connected, ergonomically terrific.

I forget why the cheap, ubiquitous Shur-Flo pumps weren't liked for this purpose, like they are in water-injection. I think it has to do with the Shur-Flo being a high pressure low volume pump. Look into it yourself if this seems like a route you wish to pursue, I'm just an opinionated jackass with a foggy memory.

 

It's a pity you can't remember the names of the pumps.

 

I do remember http://www.syty.org. Direct your searches there; I'm about to assume the unconscious fetal position for four or five hours.

 

I run a sherwood pump that I got at grainger... My situation may be different than some of you because Im using it on a drag only car... but anyways, if you search grainger, Im sure you could find it...

on a side note, most people say that the pump that im using flows too much for this application, meaning that the water doesnt stay int the IC long enough, but as far as my experience goes, it has worked perfectly... Just load up on ice, and run the thing. the IC is still cold by the time I get back to the pits...

 

i dont know anything about it. my pump is a mayfair marine pump. 500gph, catridge aerator pump. max and draw 2.0. model # 2850. it cam premounted to my resivoir

 

If you use a DSM smic woulden't you still have the air flow restriction of that core? Even if it cooled the charge air better the pressure drop on a higher hp car would be shitty. I would use two of those cores in parallel, since your already fabbing up the water jacket part why not lessen the flow restriction and double the efficency while your at it (space, could be an argument)? or maybe running the water through the tubes, like was mentiond above, would help this problem that I have convinced myself is there

BTW- tig welding is NOT the only way to do it. You can get fine results with a MIG or oxy setup on ALu.

 

The DSM core is a rediculously high flowing piece, it actually flows too well. The only restrictions are the 1.825" inlets, which if you're wholesale modifying the core it no longer matters, things can be changed to suit.

I run the stock inlet on my triple DSM FMIC, three cores welded together, with a custom endtank and 2.5" outlet. For such a ghetto setup it comes scary close to the ideal - the compressor exit immediately 180's (ideal) into the FMIC inlet less than 4" away (ideal) so the small diatemer pipe doesn't act like so much of a restriction... when the charge hits the IC, the pressure drops massively (ideal, adibiatic cooling).

My only criticism is how easy this setup would be to heat soak at high hp. I'd rather run two DSM SMIC joined together, or the turbo Prelude guy's Saab IC is a nice size. If joining the DSM SMIC, be VERY careful cutting the end tanks off as damaging the header panels on those is very easy, and makes for a difficult time welding and getting them to ever seal right.

 

your tripple core setup is in series? I can see how dsm guys go so fast with that tiny little thing since it does flow that well.

 

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by Bailhatch &raquo;</TD></TR><TR><TD CLASS="quote">your tripple core setup is in series?</TD></TR></TABLE>

I hope not. It HAS to be in parallel, else it is a huge restriction.

 

Don't lots of people run IC with the same height & depth as a DSM core but 3 times longer (meaning longer tubes)? I do realize there will be way less pressure drop in parallel but it seems you could do it either way. Especially since JDavis said it flows 'too well'. If it flows more than average than you could get the same efficiency with better packaging (for honda)of a series alignment, again assuming the pressure drop is acceptable.

 

That's NOT the same thing.

If you are going to merge three cores together, you NEED to do it in parallel. If you do it in series, there will be a huge amount of turbulence between the cores (since you can't weld the individual tubes together) which will increase pressure drop.

Not to mention that if you have 3 cores in series, the total pressure drop will INCREASE compared to one core (the air has to flow through the same area, only three times as far). On the other hand, if you put the three cores in parallel, then you will increase efficiency, increase heat sink mass, and most importantly decrease pressure drop.



Modified by kpt4321 at 6:14 PM 2/10/2004

 

Ok, ya there would be some turbulance between the cores, depending on how you did it.

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by kpt4321 &raquo;</TD></TR><TR><TD CLASS="quote">
On the other hand, if you put the three cores in series, then you will increase efficiency, increase heat sink mass, and most importantly decrease pressure drop.
</TD></TR></TABLE>
You must mean "put three cores in parallel". I don't think efficiency changes if all other thing are equal (turbulance etc..)


Are you positive the efficiency is not independent of pressure drop?
In Corky Bells book he has a pic showing two ICs both with the same size cores but with the tubes orientated as I have them below.

"The top and bottom cores have the same frontal area, heat transfer area. and efficiency, but the bottom core has much greater internal flow area...and, therefore, lower pressure drop" (that last part is what made me question your word choice in your quote above)

picture sucks but you can see the bottom one has 12 tubes and the top has 4. frontal area would be the same.
---------------------------
---------------------------
---------------------------
---------------------------

l l l l l l l l l l l l
l l l l l l l l l l l l
l l l l l l l l l l l l

The bottom tube configuration is superior because it has the same efficiency but much less pressure drop. That is how I see it. If i'm wrong please explain how and why.

 

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by kpt4321 &raquo;</TD></TR><TR><TD CLASS="quote">

I hope not. It HAS to be in parallel, else it is a huge restriction.
</TD></TR></TABLE>

No, Mister T, you canny DSM ECU hacker you, it is not in parallel and it is not a huge restriction. Have you looked at how HUGE the SMIC's cross section is? The core is 4" thick, it barely fits behind a CRX bumper after you cut most of the bumper out.

As far as ideal ICs go, pressure drop is not exactly a bad thing, within reason. Despite what flow numbers would indicate at first glance, you require some friction with surface area to help transfer heat. Also take into account adiabatic cooling infers a certain pressure drop.

We could argue this for days. I think I'm looking forward to it

 

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by Bailhatch &raquo;</TD></TR><TR><TD CLASS="quote">Ok, ya there would be some turbulance between the cores, depending on how you did it.</TD></TR></TABLE>

Regardles of how you did it.

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote &raquo;</TD></TR><TR><TD CLASS="quote">You must mean "put three cores in parallel". I don't think efficiency changes if all other thing are equal (turbulance etc..)</TD></TR></TABLE>

Yeah, I did. I changed it just now, thanks for catching that.

If everything is the same, yes, efficiency will not change. But everything isn't the same.


<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote &raquo;</TD></TR><TR><TD CLASS="quote">Are you positive the efficiency is not independent of pressure drop?
In Corky Bells book he has a pic showing two ICs both with the same size cores but with the tubes orientated as I have them below.</TD></TR></TABLE>

I didn't say that it was not independant of pressure drop. Depending on the situation, it may be or it may not be.

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote &raquo;</TD></TR><TR><TD CLASS="quote">"The top and bottom cores have the same frontal area, heat transfer area. and efficiency, but the bottom core has much greater internal flow area...and, therefore, lower pressure drop" (that last part is what made me question your word choice in your quote above)

picture sucks but you can see the bottom one has 12 tubes and the top has 4. frontal area would be the same.
---------------------------
---------------------------
---------------------------
---------------------------

l l l l l l l l l l l l
l l l l l l l l l l l l
l l l l l l l l l l l l

The bottom tube configuration is superior because it has the same efficiency but much less pressure drop. That is how I see it. If i'm wrong please explain how and why.
</TD></TR></TABLE>

I'm not sure how this made you question me....this is exactly what I was saying.

I said that mounting them in parallel, which would make the core look like the bottom picture, which has lower pressure drop (and likely higher efficiency).

The top picture isn't as good, and that's magnified when we're talking about merging three SMIC because there is extra turbulence between the cores...extra pressure drop.

 

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by J. Davis &raquo;</TD></TR><TR><TD CLASS="quote">
Have you looked at how HUGE the SMIC's cross section is? The core is 4" thick, it barely fits behind a CRX bumper after you cut most of the bumper out.</TD></TR></TABLE>


Well, I just might own a DSM, so I just might know how big the bastards are.

The issue is not the flow of one individual core, which is pretty big. The issue is that merging three cores together creates turbulence and pressure drop at each connection.

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote &raquo;</TD></TR><TR><TD CLASS="quote">As far as ideal ICs go, pressure drop is not exactly a bad thing, within reason. Despite what flow numbers would indicate at first glance, you require some friction with surface area to help transfer heat. Also take into account adiabatic cooling infers a certain pressure drop.</TD></TR></TABLE>

Agreed. You need friction/turbulence, at least a little, within the heat transfer areas, and also I agree that cooling with naturally cause a pressure drop.

However, the pressure drop from turbulence BETWEEN sections of intercooler does neither of these. It's extra turbulence, but it doesn't do much to cool the air since it is in an area where heat transfer isn't happening (behind the weld between the cores). Since there isn't any cooling happening there, the pressure drop sure isn't because of that.

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote &raquo;</TD></TR><TR><TD CLASS="quote">We could argue this for days. I think I'm looking forward to it </TD></TR></TABLE>

Uh-oh. I don't need another argument.