Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

cooling conundrum

Status
Not open for further replies.

unclejed613

Well-Known Member
Most Helpful Member
i'm trying to put together a cooling system for a PC. problem is: the PC will be running unattended for extended periods of time, so the cooling system should have zero maintenance requirements. i was thinking of water cooling, except that usually requires pumps. cooling with water convection would be preferable, but i'm not sure if i can get any substantial flow rates. vapor phase (unless it's a closed system) would require maintenance of a water supply. the goal here is to eliminate moving parts (fans, pumps, etc...). the system should be workable in harsh environments (high, but not extreme heat to cold, but not neccesarily freezing temperatures). there are also power limitations, and that's why i've ruled out thermoelectric cooling.

it's for an unattended 802.11 repeater for someone who wants their DSL connection available at their workshop on the other end of their land, and there's a ridge between the house and workshop. i tried a passive repeater (two cantennas wired together) but there's not enough signal for it to work reliably.
 
Last edited:
I don't have a complete answer but I can say that I've seen vapor phase used - called heat pipes - for some applications. Convection will occur in a liquid system (was thinking of glycol/water mix) as a result of temperature differential. You'd have to experiment with it. Absorption refrigerators such as those used in RVs are powered, so to speak, by heat - you might adapt those principles to your situation.
 
I take it you consider fans too unreliable. But a well designed fan can run for several years continuously.

You might Google "fanless computer cooling". There are numerous links on that.
 
Fans with ball bearings are designed for a long life. They are more expensive, though.

Another improvement is to use a temp sensor to power the fan up only when required. A further refinement is to use fan speed control and use an oversized fan to run it at slow speed most of the time, and at normal speed in only those extreme situations that do require it.

These techniques are employed in server and telecomm power supplies, which operate 24/7 for years and years.
 
I am running a DSL repeater system as well.
All I have is a pair of wireless routers running at the repeater station. One runs a pair of directional antenee pointing towards the source and the other router points towards me.
The total power consumption is under 4 watts for the pair. The wireless routers I have both run off of 12 volt 300 ma power packs.
I have electrical power at the repeater staton but being so low powered a good deep cycle battery and a 20 watt solar panel would easily carry it and keep the battery charged.

As far as the source computer a old stripped down bare bones one with a small processor and a oversized heatsink on the CPU can run passively.
Strip the software programing down to the absolute minimum. Just running a DSL modem card and doing signal handling is very low power.
You probibly could run just fine off a windows 98 machine with a 150 MHz processor, 5 gig hard drive, 64 meg ram, and still be under 10% of the processors load capacity at peak.

I would recomend that you put a UPS system on it. Pick up one that has around 400 -750 VA capacity and rework it to run the battery leads out side of it. Then stick a big deep cycle battery on it as well. A small computer load on 100 Ah 12 volt deep cycle battery could run for 6 to 8 hours before needing a recharge. This would give you tremendous reserve time!
 
Last edited:
You could use peltier elements for cooling very effectively. They generate a max temperature difference between both plates of max 67 deg/C.

E.g. ambient temperature +35deg/C - resulting cooling temperature -32deg/C.

PWM-controlled you can design the cooling device to stay well above freezing point to avoid water condensation. (+15 to +20deg/C)

There are different types and sizes peltier elements on the market, ranging from 30x30x4.7mm/8.6V/3.9A (33.54W) to 40x40x3.3mm/15.4V/8.5A (130.9W). The temperature range of all is the same, just the amount cooling power varies.

If you are not going to use a fan, plan for a large heatsink for the hot side of the peltier, which should be placed outside the enclosure.

Boncuk
 
I think the investigation and research you are doing is good. The problem - anything other than the very common methods (such as a fan) will take some time and effort to develop to a point where you can really call them reliable. As already suggested, fans are fairly reliable - possibly more so than your first effort at alternatives. Things you might do are add backup fan and controls or possibly controls to shut the repeater down to cool if needed. I know it isn't exactly what you requested.

All of this presumes you can't just use an grossly oversized heat sink with natural convection.
 
The most reliable systems have redundancy as a feature.

If you have a stream nearby, you could use that as a continual water source and release the water futher downstream. That could be pretty low maintenance. Maybe a huge rain barrel calculated based on annual rainfall. :)
 
Last edited:
this isn't the first time i've been asked to put something like this together. tnx for the responses. looks like i've got a bit of prototypring to do. i can probably scrounge up a couple of 150W transistors to use as a heat source (and use their measured Vbe as a temperature guage). if i set them up to run as constant current sources, i can control the current through them with a pot and run them hot. if i clamp them to a heat exchanger i can measure the efficiency of the cooling system almost directly. the Vbe gives me the die temp of the transistors, and the water temperature can be monitored with a diode as the sensor. silicon has a -2.2mV/deg C forward voltage coefficient so it's a very linear relationship to the die temp. i'll also be able to test thermally conductive rubber sheets and see if it's as efficient as it sems to be. i've seen a lot of it in use in DVD players as an interface between processor chips and heatsinks, as well as in class D amps between the output chips and a heatsink.
 
I would seriously look at simple wireless routers as the repeater devices. You just plug them in back to back. The system I have has three of the four units mounted outside in PVC boxes. There are no fans or cooling to worry about and they have lasted for over a year in all forms of weather from -40 to +110 in the sun. ;)

And as I said before if a computer is needed just for the conversion from land line to wireless, very little power and processing speed is needed. A old cheap laptop would even work! :)
Having helped design and set up a system that transmits at 85 meg while shooting a half mile while going through a repeater and has ran smoothly for over a year I have found out for a fact over complicating it does not help! :)

Just a thought!
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top