Fan cooling design

Hmm the only reference I know is mostly just stories of cooling problems. "Hot air rises and heat sinks." Forced air means keeping the input filter and air path clear. I bought a new water pump for my tractor only to discover my radiator was covered in grass/ weeds. :^)

George H.

Oh I wanted to add it's a book that discusses mistakes that were made, and it that way is somewhat unique. GH

===================

** That book is on sale for US$44 and consists largely of anecdotes and the lessons to be learnt.

FYI: I have been dealing with issues with fan cooling of the last 40 year s.

High powered audio amplifiers became fan *internally* cooled in the late 70 s, previously fan cooling was generally external and only a need to basis. Crown Audio was one of the first to fit a fan to their products plus a new heat sink design that used folded aluminium sandwiched between solid metal plates. Air was forced through the folds exploiting the large surface area .

But what Crown forgot to tell anyone was the idea only worked in an air con ditioned equipment room with no people or carpet anywhere. The narrow fold s soon became clogged with fluff and dust and the whole thing need disassem bly and cleaning. Having a foam filter on the fan inlet also required regular cleaning, a rea l PITA unless you planned ahead for it. It also killed air flow. Some makers followed Crown's lead with mostly disastrous results - especia lly if no thermal shut down was installed.

Others came up with a better solution, using heavy aluminium heatsinks in t he direct blast of one or more fans - with no inlet filter. Fast flowing a ir does not drop dust and fluff everywhere so much.

Most also made the blunder of NOT coating PCB's copper sides, allowing dust build up and failures due to "tracking". One common example ( Perreaux of NZ) had adjacent tracks with 240VDC betwe en that loved to track and eventual cause a major explosion inside the amp.

One issue you might not have thought of is where there are multiple amps in a rack and one of then has an internal PCB fire generating lots of soot. The soot is then immediately sucked in by each of the others coating the P CBs inside them. Massive job to clean up and restore operation.

..... Phil

Yeah, I have that one. It's a fun read, but apart from discussing a mythical thermal simulation program and the misadventures of the narrator's feckless heat-ignoring foil it doesn't have a lot of quantitative info iirc.

There are the 'fan laws' available from Rotron and elsewhere, but that just says that (other things being somehow equal) a deeper fan will take more back-pressure, and a faster fan will move quadratically more air and take cubically more power.

it really seems that people build fan cooling systems like pyramids---keep making the next one a bit steeper until one collapses (the 'bent pyramid') and then standardize.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com

s.

he lessons to be learnt.

s.

70s, previously fan cooling was generally external and only a need to basis . Crown Audio was one of the first to fit a fan to their products plus a ne w heat sink design that used folded aluminium sandwiched between solid meta l plates. Air was forced through the folds exploiting the large surface are a.

onditioned equipment room with no people or carpet anywhere. The narrow fol ds soon became clogged with fluff and dust and the whole thing need disasse mbly and cleaning.

and laptop manufacturers still haven't learned

Right. Air, given the choice, would rather flow around heat sinks and parts. And flow in the opposite direction that it's supposed to go.

I have Steinberg's book "Cooling techniques for electronic equipment" (1980) here at home and probably another textbook or two at work.

Steinberg has a lot of examples of boxes and racks and real situations.

--

John Larkin      Highland Technology, Inc 

The best designs are necessarily accidental.

I have that book somewhere. I recall it being pretty bad, and all in crazy old units.

--

John Larkin      Highland Technology, Inc 

The best designs are necessarily accidental.

----------------------------------- >

------------------------------------------------------

** That's what I'm talking about .....

Can happen in a week where the venue has new carpet installed.

Made worse if the fans are not speed controlled.

..... Phil

I sold a large number of instruments for use in clean rooms. They had tile floors, HEPA filters to circulate air, and the personnel were completely covered in bunny uniforms with face masks, gloves and booties.

I asked how do they recognize anyone in these outfits. The answer was to look in their eyes. These were supposed to be class 100 clean rooms and they appeared to be doing everything right. See

"Cleanrooms are classified according to the number and size of particles permitted per volume of air. Large numbers like "class 100" or "class

1000" refer to FED_STD-209E, and denote the number of particles of size

But whenever I visited these installations, I pulled the filter from my instruments and examined them. They were completely blocked with dirt and lint. I had to make it clear to management that their clean rooms were not as clean as they thought, and they had to pull the filters on my instruments and clean them periodically.

Here are more examples:

Ventblockers

Bring out your dead: Reg readers reveal filthy, filthy PCs

Whenever you use fans for cooling, you have to provide some way to clean the heat sinks and filters or they will eventually get plugged up.

--
The best designs occur in the theta state. - sw

One big national lab has the world's largest clean room. They sent us back some of our fiberoptic gadgets that weren't working right. The photodiodes and lasers were full of dust bunnies.

--

John Larkin      Highland Technology, Inc 

The best designs are necessarily accidental.

Remember how refrigerators used to have big coils on the back for cooling?

I just bought two new refrigerators. One is a Samsung, the other is a Haier. Neither have any external coils for cooling, and there is no hump at the bottom where the compressor used to go. I looked for it, but I have no clue where they put it.

However, the sides get warm to the touch when they are first turned on, and after they have cooled down, the temperature rise is not noticeable.

Also, even though they are two difference brands, neither make the slightest sound. Both are set to 34F for the fridge section, and 0F for the freezer. I monitor the temperature with ordinary fridge gauges, and it is astonishing how well they maintain the correct temperatures.

I also monitor the humidity in the fridge section, and both maintain the same relative humidity as in the room, around 30%. If you open the door briefly to get something, the humidity skyrockets to 50%, but quickly settles back to normal.

These are not expensive units. Both were around CAD$1,500.

There is a lesson here for cooling electronic equipment. Make the units as efficient as possible to minimize the amount of waste heat, seal the units to keep dust out, don't use fans, spread the heat over a large area, and use convection cooling.

--
The best designs occur in the theta state. - sw

s.

he lessons to be learnt.

s.

70s, previously fan cooling was generally external and only a need to basis . Crown Audio was one of the first to fit a fan to their products plus a ne w heat sink design that used folded aluminium sandwiched between solid meta l plates. Air was forced through the folds exploiting the large surface are a.

onditioned equipment room with no people or carpet anywhere. The narrow fol ds soon became clogged with fluff and dust and the whole thing need disasse mbly and cleaning.

eal PITA unless you planned ahead for it. It also killed air flow.

ally if no thermal shut down was installed.

the direct blast of one or more fans - with no inlet filter. Fast flowing air does not drop dust and fluff everywhere so much.

st build up and failures due to "tracking".

een that loved to track and eventual cause a major explosion inside the amp .

in a rack and one of then has an internal PCB fire generating lots of soot. The soot is then immediately sucked in by each of the others coating the P CBs inside them.

Right, As Phil H writes below, it's a fun read, but short on useful informa tion. One fan cooling story (of mine) involves an AC power conditioning unit. Th e pass transistors (TO-3) were mounted in a heat sink tunnel, with fan at one end. One morning the u nit wasn't working. Examining the unit it was found that the pass transistors on the top of the heat sink tunnel had unsoldered themselves and fallen out of the pcb. Solder as thermal shut-down :^)

George H.

Our fridge has the coils on the bottom. And we have a cat.

--

John Larkin      Highland Technology, Inc 

The best designs are necessarily accidental.

Design?

You mean limiting noise levels?

Fan cooling design - what a laugh!

RL

Google 'Forced Convection Cooling'.

Lots of stuff from the 1990s.

RL

** That story is very unlikely. TO3s cannot fall out, there are two bolts holding them in.

If they somehow came loose, there is no current flow.

..... Phil

Motorola diodes in 194-04 plastic puck bodies tended to desolder themselves, when forced air cooling failed.

TO3 thermal and electrical paths are supposed to be independent, but many mistakes were made in combining the main case electrical contact/insulation with the mechanical mountng hardware.

You used to be able to buy TO3 plastic 'caps' that softened under mounting hardware and low temperature nylon insulator bushings. Even involving FR4 boards in the mounting hardware could be problem.

Obviously anything that softened in the mounting hardware's development of mounting pressure could result in thermal runaway - plastic eventually melting below the melting point of both internal and external solder bonds.

Melted solder conducts just fine.

RL

=======================

** Huh ??????????

Why post this load of irrelevant s**te ?

Got SFA to do with my post above.

..... Phil

Forced air cooling?

Desoldering events?

No current flow if 'loose'?

It's not your post, it's the OPs.

Suck it up.

RL