Radiator question, for thermodynamics/heat transfer experts Options

[bondo]

Hopefully I can get some input on this without too much of the inevitable "that won't work" (IMG:http://www.914world.com/bbs2/html/emoticons/smile.gif)

I'm cutting down a large tall radiator into a shorter one.. 1/3 the height to be exact. The 3 layers of radiator will be set up in a cross flow, sequentially. The hottest water will enter the rearmost later, travel across, then across again in the middle layer, and across yet again in the front layer. The idea is the hottest water gets the hottest air, and the coldest water gets the coldest air. This way delta T stays pretty constant.. I'm thinking of it as trying to heat the air as much as possible, which would cool the water as much as possible.

So I realize that cooling capacity will be reduced, as I will be dealing with 1/3 as much air. But then the cooling capacity must be more than 1/3, because it's 3 layers thick instead of 1. My question is what percentage of the original cooling capacity should I be expecting? Would it do me any good to do 5 layers instead of 3? (by adding 2/3 of a second radiator)

The original radiator cooled the same engine just fine, in a heavier car, with an AC condenser preheating the air. I will be running an electric pump, so I will have an advantage at idle.

[Mueller]

I think by the time the air hits the 3rd layer, it's going to be too hot to cool anything and be pretty much useless....

[neo914-6]

(IMG:http://www.914world.com/bbs2/html/emoticons/agree.gif) When I worked on multiple heat exchangers, you'd always plumb in parallel...
You may try to plumb one, record temps and keep adding...Sometimes the best calculations did not match actuals.

[redshift]

Ask Seanery, he knows all about heat transfer.

(IMG:http://www.914world.com/bbs2/html/emoticons/w00t.gif)


M

[skline]

So you are not in a hurry to finish your car and drive it huh? I would just pick up a Renegade setup and be done with it. But, thats just me.

[SpecialK]

"that won't work" (IMG:http://www.914world.com/bbs2/html/emoticons/biggrin.gif)

(IMG:http://www.914world.com/bbs2/html/emoticons/icon_bump.gif)

[ArtechnikA]

i think you're going to have a hard time getting any air through any layers...

you -can- heat up the air really hot by the simple expedient of reducing the airflow, but the object of the game is not to heat the air, it is to cool the water. hotter air does not equate to cooler water. (IOW - you are not as much interested in reducing temperature as rejecting heat. ( where heat == temp * mass ).

i understand your logic on the sequence, but i still think it is backward -- coolers are most efficient at higher delta-T. (still, i see in my mind a conventional coiled-tube lab distillery and we always plumbed those with the coldest inlet water at the exit - but that was not about efficiency, it was about achieving the lowest possible temperature - not the same thing...)

QUOTE

I will be running an electric pump, so I will have an advantage at idle.

will you also be running BIG electric fans to move some air?

QUOTE

So I realize that cooling capacity will be reduced, as I will be dealing with 1/3 as much air.

not just 1/3 as much air, 1/3 as much surface area exposed to the ambient temp freeestream air. i doubt that's a simple multiple (1/3 * 1/3 == 1/9) but i'm thinking it's on that order.

probably this technique has been documented in the HVAC and thermodynamics texts, none of which i have, but it is described in several references i do have on race car and aircraft plumbing, usually on the "don't" page of "do's and don'ts" ...

if you're determined to build such a thing, and if i were you i wouldn't count on my armchair analysis as a gospel, i'd put temperature sensors in front of each layer and behind the last, and in the water stream before and after each layer. and then test it.

i think that if reducing the frontal surface area is really your design goal, you would be better served finding one deep cooler than stacking multiple layers. but i'd take this as a hint: i have never seen any really deep coolers ...

just MHO, since you ask ...

[bondo]

QUOTE (Mueller @ Feb 2 2005, 12:10 AM)

I think by the time the air hits the 3rd layer, it's going to be too hot to cool anything and be pretty much useless....

The third layer gets the hottest water so the air that hits it will have to be cooler than the water going through it. BUT, it may not be enough cooler to do much good.

The reason I've come up with this idea is because of the tidbits I've managed to squeeze out of Scott (Renegade) regarding WHY 914 V8s have trouble staying cool. He says the water doesn't spend enough time in the radiator to lose heat. Thus my idea of the multipass design. I have since heard that Renegade's radiator is a multipass design (not 100% sure on this, I just read it here somewhere).

I will have big fans moving the air. They will be at the outlets instead of on the radiator.

If it doesn't work, it would be very easy to flip the whole thing over and do it the other way.. hottest water in the front. That would give a very good delta T for the first layer, but by the third layer the air may be as hot as the water.

The other option is to just run all 3 cores parallel in a conventional 3 core setup. The radiator construction would be easier, but I'm not sure it would be more effective.

As far as just buying Renegade's radiator, there's no way I'm spending $1000 on a radiator. I'm much to big of a CSOB for that (IMG:http://www.914world.com/bbs2/html/emoticons/smile.gif)

[v82go]

Hi
I recently replaced my custom 6 core radiator with a "Howe Racing" aluminum 2 core. My
car now runs aprox. 10 degrees cooler. Don't know if it was the change to 2 cores or the change to aluminum, but she does run cooler. I have also had good results with a Howe in my Monster Miata. In the 914 I am using a 16x27.5 (no fill-chevy) which cost
$190.
Bob

[SLITS]

Rate of water flow thru a radiator is very important. On our water cooled race cars we used a 4 row crossflow aluminum radiator. To slow the water down, we used blanking sleeves in place of a thermostat. The one we ran was a plate with a 1/4" hole in it.

We also experimented with different ratio pulleys to slow the water pump speed down and found this to be the best solution.

That said, we were operating within a specific RPM range on the track (5500 to 8500 RPM). The only time the engine RPM was below 5500 was in the paddock.

I am not a thermodynamic engineer, nor a true HVAC person (spent a lot of time in that industry though). I have never seen heat exchangers plumbed in series...especially a fluid to air heat transfer. I do know that the effeciency is a relationship between the rate of fluid flow, material of heat exchanger construction and the surface area of the fins attached to the liquib bearing tube....break the bond between the fins/tube and effeciency immediately starts dropping radically.

Oh well, it's to early to make any sense

[bondo]

QUOTE (v82go @ Feb 2 2005, 08:50 AM)

Hi I recently replaced my custom 6 core radiator with a "Howe Racing" aluminum 2 core. My car now runs aprox. 10 degrees cooler. Don't know if it was the change to 2 cores or the change to aluminum, but she does run cooler. I have also had good results with a Howe in my Monster Miata. In the 914 I am using a 16x27.5 (no fill-chevy) which cost $190. Bob

Aluminum makes a big difference. The one I'm working on is aluminum, in case anyone was wondering.

[914GT]

I find this to be some interesting stuff and when it comes to 914 V8 conversions there's not a lot of real engineering available on what's an optimum cooling system design. I wish my co-worker mechanical engineer (PhD, specializing in thermal and numerous patents in cooling electronics packaging) was interested but he's not into cars. (IMG:http://www.914world.com/bbs2/html/emoticons/confused24.gif)

I think that for a given engine and radiator setup there is an optimum coolant flow for the system, and this is independent of the power output of the engine. In other words, the flow rate should remain fairly constant and if the engine is dumping more heat into the coolant, then there should be a corresponding increase in the heat loss at the radiator (more air flow). There must be a rate of max. efficiency, if plotted it would peak out at some particular flow rate. Too low (or zero) obviously things would overheat. Increase the flow too much and the same thing happens. Would be interesting to attach a few thermocouples in a system and play around with this.

This is where an electric pump has an advantage as it can keep a constant flow independent of engine RPM and power output. Or its flow can be varied easily according to conditions. I've been very pleased with how mine is performing and I think I basically lucked-out finding a well-matched system. But I would like to understand it better as to why it works so well.

[andys]

QUOTE (v82go @ Feb 2 2005, 07:50 AM)

Hi I recently replaced my custom 6 core radiator with a "Howe Racing" aluminum 2 core. My car now runs aprox. 10 degrees cooler. Don't know if it was the change to 2 cores or the change to aluminum, but she does run cooler. I have also had good results with a Howe in my Monster Miata. In the 914 I am using a 16x27.5 (no fill-chevy) which cost $190. Bob

Bob,

I have the same size AFCO Racing dual pass radiator. Is your's a single pass? Also, what size hoses did you run for the high and low sides?

andy

[davep]

Ah, a topic dear to my heart, and a little within my realm of expertese.

Using counterflow you can get better heat transfer, and that is a good thing. However, by reducing your airflow down to 1/3 of the original design, your cooling will be reduced down to about 1/3 as well. You may be lucky and only lose 50% but I doubt you could get that lucky. Adding counterflow layers is an incremental effect, but having lots of airflow is crucial. The main thing here is that the heat is being dumped into the air. To reduce the airflow by two thirds would require a temperature rise three times greater, and that will not happen in your design. Sorry. It is difficult if not impossible to cheap out on the rad AND the hoses connecting it to the engine. You have to have the correct combination of airflow and waterflow.

Remember that the greater the waterflow, the greater the HP required for the pump, and that reduces the rear wheel HP. The HP used by the pump is transferred as heat to the water coolant, and that adds to the load on the rad. However, if the waterflow is too small, the engine will overheat because the circulation cannot take it away as fast as it is produced. If you want the engine to live, then you have to determine the heat dumped into the coolant by the engine under full load. Then the pump has to move that heat ( so much GPM times deltaT). Finally the rad and airflow must transfer that same amount of heat to the air. It is not trivial.

[bondo]

Ok, it looks like I'm looking at a couple different types of multipass radiator..

Correct me if I'm wrong..

The "typical" multipass radiator has everything in the same layer, but it divides the flow into top and bottom, so it flows one way across the top, and comes back across the bottom.

My idea has layers behind each other, which would obviously be less effective than the top/botom method, but would offer more cooling per frontal square inch.

I need to think more about this and maybe do some testing.

[scotty914]

all this talk about water flow, air temp rise, delta what ever is great but i did it the easy way .... i am using a stock radiator that suby uses with the 2.5 engine. that way i know its is designed right by suby!

minor high jack
the radiator frame is welded up it goes in to the car today, tommarrow is shrouds and reinstall engine. then fun fun stuff starts, hooking up obd 2 wiring (IMG:http://www.914world.com/bbs2/html/emoticons/headbang.gif)

high jack over

[ArtechnikA]

QUOTE (bondo @ Feb 2 2005, 12:34 PM)

My idea has layers behind each other, which ... would offer more cooling per frontal square inch.

"could," i think ...

and yes, multipass heat exchangers change directions in vertical (typically? exclusively?) zones. (almost?) exclusively hot water enters at the bottom and cooled water is extracted from the top. this allows the flow to work with convection and promotes the coalescing of air bubbles at the top rather than remaining trapped in the moving fluid.

testing is good...

i'm with Mike, i think you'll find the cooling will be no better, and possibly worse, with 3 layers rather than two. i think you'll gain more from increased airflow with 2 layers than you may lose from another incremental pass due to the preheat effect.

air is lazy, and at speed, will be happy to bypass the radiator core and go right over the car if internal restsrictions are too high. you can't blow air into a closed box. it's also more efficient to push air than pull it, and one problem with using extraction fans is that without really good sealed ducting, the air would much rather go around the core than through it. tends to preheat the fans too, although there are enough of them in service that the environment doesn't seem to adversely affect their life. (and it does have the plus that it does not shadow the core from incoming freestgream air...)

[bondo]

QUOTE (ArtechnikA @ Feb 2 2005, 10:56 AM)

QUOTE (bondo @ Feb 2 2005, 12:34 PM) My idea has layers behind each other, which ... would offer more cooling per frontal square inch. "could," i think ... and yes, multipass heat exchangers change directions in vertical (typically? exclusively?) zones. (almost?) exclusively hot water enters at the bottom and cooled water is extracted from the top. this allows the flow to work with convection and promotes the coalescing of air bubbles at the top rather than remaining trapped in the moving fluid. testing is good... i'm with Mike, i think you'll find the cooling will be no better, and possibly worse, with 3 layers rather than two. i think you'll gain more from increased airflow with 2 layers than you may lose from another incremental pass due to the preheat effect. air is lazy, and at speed, will be happy to bypass the radiator core and go right over the car if internal restsrictions are too high. you can't blow air into a closed box. it's also more efficient to push air than pull it, and one problem with using extraction fans is that without really good sealed ducting, the air would much rather go around the core than through it. tends to preheat the fans too, although there are enough of them in service that the environment doesn't seem to adversely affect their life. (and it does have the plus that it does not shadow the core from incoming freestgream air...)

Good point. I was thinking about it from a standpoint of: for every square inch of radiator, there is some peprcentage of open space for air to flow. Say that percentage is 50%.. So I was thinking of it like 3 plates with a 50% sized hole in them, would have roughly the same airflow as 1 plate with a hole that size. I didn't think about the turbulence and such of all those little fins. I was thinking about this from the standpoint of "more can't possibly hurt", but that only works up to the point where the additional restriction of airflow outweighs the increase in surface area.

Man, this is getting deep (IMG:http://www.914world.com/bbs2/html/emoticons/smile.gif)

[ArtechnikA]

QUOTE (bondo @ Feb 2 2005, 01:03 PM)

Man, this is getting deep

welcome to engineering...

the essence of engineering is one word: "enough." how much strength is enough? how much cooling is enough? how much power is enough?

like Dave said -- how much cooling is enough? it's enough when your worst-case combination of power output (usually high) and airflow (sometimes but not always low) creates a temperature no higher than the most sensitive component (often, the boiling point of the coolant at operating pressure) can take.

you can make some reasonable assumptions on some things (engine efficiency i.e. - how much waste heat is produced per HP) but sometimes the variable interact in unexpected ways (like a cooling system airflow restriction that works perfectly well at 60 mph but stalls at 80) and then, God help us, there are the nonlinearities ... where 2X of one thing gives you (or requires) 4X, or X**4, of something else...

we're not trying to poke holes in your design for the fun of it - there are innovations made every day.

we are trying to help you think through the design issues so your car can be (close to) working right, right out of the box.

[bondo]

Thanks for all the info and suggestions.. what a great group! (IMG:http://www.914world.com/bbs2/html/emoticons/smile.gif)

While I'm thinking about it, Scott had mentioned the other problem with cooling was thermostat delay/runaway. Apparently by the time the termostat opens, and the hot water runs up to the radiator, cools, and comes back to the engine, it can already be too late.

My idea to solve this problem (and the problem of firewall interference) is to place the entire LT1 water pump assembly in the front trunk, with an electric conversion on it. For those unfamiliar with the LT1 water pump, it has a 2 part thermostat valve. A typical thermostat blocks the cooling water, with a small bypass for the heater. The LT1 thermostat is mounted on the water pump and continiuously circulates the water throulgh the engine. Once it gets warm the thermostat opens and it adds the radiator to the loop. The way I plan to set it up will have water continuously flowing to the front of the car and back.

Question 1. Am I correct in thinking that the return hose (from the thermostat to the engine) should be as small as practical so the cool water gets to the engine quickly?

Question 2. Where should I put the reservior/filler? I'm thinking it should be on the cool side of the radiator, so it can store a quantity of cooler water that will be immediately available when the thermostat opens. There are basically 6 choices. Either end of the radiator, just before or after the thermostat/pump, or just before or after the engine.

I'm just full of crazy ideas! The fun part is that if my cooling system doesn't work, i won't have the foggiest idea which part isn't right... hooray for too many variables!