IDEA for a passive rad - will it work?

[iakovl]

my idea is to take along and big round copper pipe and use force to shape it into something like this ( | ) and put ALu fins on both sides.

will it work?

[jman1310]

Originally Posted by iakovl

my idea is to take along and big round copper pipe and use force to shape it into something like this ( | ) and put ALu fins on both sides.

will it work?

do you mean something like this?
it knocked a degree or 2 off my temps but nothing impressive
note that it is only a smooth bore, no trubulators

if you want it, it's for sale (cost of the copper fittings and shipping ~$30)

[redleader]

I've never liked the fin on pipe idea. Much easier to just make a lot of smaller pipes for the same effect. Nice work though on those fins. You soldered those on by hand? If so its pretty impressive.

[snowwie]

looks like my hot water heater

edit

and i doubt the alu fins are soldered, just a tight fit

[jman1310]

it's a bit of extra tube from my parent's house's radiant heat system
yep tight fit

[iakovl]

its more like in this pic:

[Long Haired Git]

Calculate the surface area and compare.
By my calcs (I'd love someone to confirm) a Big Arse heater core (Toyota Camry 87) which is approx 220 x 140 fin area, has a surface area of approx 735,000 mm2.

So, take a copper pipe and squish it into a shape that's, say, 1 inch on the long side and a quarter inch on the short side, and it will need to be 13 meters long (40ft) to get the same surface area as the BA heater core.

Assuming you want to keep the sam max distance of fin tip to coolant, then you'd want to have 4mm wide fins, and lets say you're going to hang the pipe in mid air and hence can solder said fin on all four sides. Now your tube only needs to be a mere 7.6m long.

The problem with making a better radiator than a heater core is mostly that, comparitively, they're cheap to buy and easy to fit.

Volenti puts it best: Surface area is king.

[maxSaleen]

You could try making several 1m lenghts of 1/4" pipe with Cu fins . Run about seven of them in parallel and you might see adequate performance. I say Cu fins because they can be directly soldered to the Cu pipe and they will dissipate heat better than their Al bretheren. I assume that weight is not an issue here.

[Long Haired Git]

maxSaleen: did you do the maths on your suggestion?
Assume 1/4" pipe means a pipe with 1/4" OD, and with seven I get 140mm2 of SA per mm of length. Add in 3.8mm wide fins, again four per pipe, and that adds 106mm2 of SA per mm of length. Total SA per mm length = 246mm2, hence length per pipe to equal 120.2 heater core = 3m.

[redleader]

Originally Posted by Long Haired Git

Calculate the surface area and compare.
By my calcs (I'd love someone to confirm) a Big Arse heater core (Toyota Camry 87) which is approx 220 x 140 fin area, has a surface area of approx 735,000 mm2.

So, take a copper pipe and squish it into a shape that's, say, 1 inch on the long side and a quarter inch on the short side, and it will need to be 13 meters long (40ft) to get the same surface area as the BA heater core.

In a pipe design, all surface area is equally distant from the coolant. In a fin design, most of it is not, therefore you can't compare the surface areas directly. The fine design makes much more effcient use of its surface area, therefore comparing raw numbers will be misleading.

[iakovl]

but....
what about zelman's passive rad ? it's not too big but it still works
and there are other passive alu rads around

mine is copper+alu fins + 120 fan thas will increas the air flow

[jman1310]

the zalman rad probally has fins on the inside (i haven't seen it but i think i remember a pic of the internals)
makes a big difference in amount of heat transferred

my rad above is designed for a temperature difference of 40-60 degrees (room is about 70 rad is almost to hot to touch, which i would estimate at 130-140)

[Long Haired Git]

I confirm red-leader, hence why I made the fins the same max distance from the tube: 3.8mm for a BA heater core. Particularly for a passive rad, the fins make a big difference.

I used to run dual 130x150 HCs in vertical orientation (not horizontal) with shrouds and fans fitted (should have been removed) and still ran just 20 degrees Celsius above ambient measured at the CPU.

For a reasonable CFM of air situation, heater cores rule because their compact shape directs air flow to do most good. For passive situation, heater cores rule because they condense so much surface area so close to the coolant. I'm trying because there are things about HCs which are not ideal, but so far any design I come up with to improve on them is either impractical or prohibitively expensive (even for us extremists).

Reasonably confident a horizontally mounted HC of good size (120.2) in free air will outperform most home-brew-copper-pipe solutions that also rely on air (bury in ground doesn't cut it) unless its outrageously huge.

[Marci]

Passive Cooled TEC System that works - Medusa - http://www.over-clock.com/ivb/index....17&#entry35917

[maxSaleen]

Add in 3.8mm wide fins, again four per pipe, and that adds 106mm2 of SA per mm of length.

I was thinking more along the lines of 50 fins per pipe. I probably should have mentioned this as it increases the surface area substantially. I know that this would be a pain to hand solder but I would imagine the performance benifit would be worth it.

[Long Haired Git]

You can safely ignore radiation as a method to cool your radiator, which is ironic.
Instead, you will be using convection to the air next to the radiator, and then relying on the old "hot air rises" principal to replace that now hot, useless air with cool air.
Now cut to looking at your arm, which I assume is hairy like mine, and think about why we get goose-pimples and why they keep us warm.

So, 50 fins only works if there's air flow, and won't work if all it does is keep the hot air nice and close.

This is why the reserator is vertical with vertical fins. This way as the air heats and rises it moves over the fins and leaves the unit - being replaced by cool air from the bottom.

Similarly why I say a HC should be mounted horizontally and not vertically.

Also, you'll need to provide dimensions, as fitting 50 fins around a 1/4" tube (yes, not you who suggested 1/4" I know) will be impossible to solder.

Reminder of the challenge: 7 sections of 1/2" pipe with 12 fins each 3.8mm wide needs to be 1.2m long to match a 120.2 HC surface area.

Also, you'll need to squish the pipe because a normal 1/2" tube has a lot of water in the middle and not near the edges whilst a HC has very thin tubes (1.8mm wide by my ruler) and so coolant is, at most, 0.8mm from the sides (assumes 0.1mm wall)

[Les]

Originally Posted by Long Haired Git

......Assuming you want to keep the sam max distance of fin tip to coolant, ........

A rather bold assumption.
Would suggest that 6x would be more appropriate for purely passive cooling where h~ 10-20 w/m^2*c
Would expect about the same finning efficiency as forced convection with h~ 100-200 w/m^2*c

Some aspects of the theory for passive radiators were discussed, from post 5 onwards,here

[Long Haired Git]

Wow - nice link Les.
I've been designed a better forced convection radiator (or trying to) and so didn't see this bad assumption. Passive rads will result in a higher coolant temp and hence the fins can be bigger which means that the whole thing can be smaller.

Eg: 8 x 20mm fins on 7 x 1/2" pipes now only 0.5m tall.

That linked sites gots lots of nice calcs....