6
\$\begingroup\$

We are working on a product which is extremely cost sensitive. It generates a lot of heat so we are planning to use a heat sink. For those 'very rare edge causes' where heat sink alone cannot cool down the system, we have also equipped the system with a cooling fan.

Due to cost constraints, we are planning to go with cheap 12V DC cooling Fans (basically used parts, the ones which don't come with a guaranty). So now we are focusing to judiciously use the fan ONLY when absolutely required (The tests show heat sink alone can keep the system cool for ~75% of scenarios). Our system can tolerate a max. temperature of 85 °C.

To maximize the fan's life, we'll have to reduce it's wear and tear. And to do that, we have to decrease it's usage. For this, we have come up with two different solutions (there might be even more possible, we are open for suggestions!):

1. We control the fan's rotation.

We switch it on at a (relatively lower) 60 °C but at very low speed. Basically, the fan's rotation is directly proportional with temperature (like laptops). But because cooling fans rapidly lowers down the temperature, so even if running at 35% speed, the system will cool down rapidly and the fan can be switched off (say at 50 °C).

2. We raise the trigger point to ~70 °C.

The fan is switched on at full speed at 70 °C and switched off at, say ~60 °C. Fan speed is not controlled.


Which approach will work better? You may assume that our design is perfect in terms of heat transfer. The device is vertically mounted and hot air can easily escape from upper face (while cool air enters from bottom face). Assume no hot air is trapped inside the device.

Approach #1 tries to minimize power consumption (thus wear and tear), but it does not minimizes fan usage time. Approach #2 tries to minimize fan usage time (at the expense of power consumption).

So more Power Consumption for less time or less power consumption for more time, which will work better?


We have tried to weigh both of these approaches on following factors, but due to limited experience, we have been unable to choose which one is better:

1. Amount of dust deposited

More dust means more resistance and thus more power consumption leading to more wear and tear.

2. Air Temperature

Hotter air moving across the fan means more heating of fan coil, thus causing more damage and reducing fan life.

Approach #1 looses at Factor #1 but wins at Factor #2. Approach #2 looses at Factor #2 but wins at Factor #1.


Please we are looking forward for an expert advise on this matter. We have already looked into this, this, and this but are still stuck. Please forgive us if it is plain obvious, we have very little experience in dealing with such issues. We do not intend to ask a subjective question but to discover and learn about the best industry approach for such cases (if there is any?).


EDIT 1 Providing some more technical details.

Expected lifetime: min 5 years. Environment temperatures vary between 0 °C to 50 °C. Device can tolerate maximum of 85 °C.

What is the average / max power consumption?: Fan consumes ~1.32 Watts at full speed. System consumes constant ~10 KVa.

How big is the fan? Fan diameter is 73 mm.

How much dust is expected in the environment? If there is any, how do you prevent it from interfering with the passive cooling?: It is going to be a consumer product and thus we have to expect a dusty environment. Being vertically mounted, theoretically, only some dust should stick inside the heat sink's grooves (reducing it's surface area and heat conduction). Currently we have nothing to prevent this. Any suggestions for the same? i'll have to contact the the pollution department to get the exact data. i will get the data and revert back.

How many start / stop cycles of the fan are to be expected?: This depends upon the kind of approach we choose. But the way i see it, it depends how often we cross the 'critical' temperature without any fan (in the current prototype the fan is always running). For that we'll have to conduct some tests. We will publish the results in a day or so.

\$\endgroup\$
  • 5
    \$\begingroup\$ Side note, the bearings are more likely to fail than the actual fan motors. Orientation of the fan could have an impact on bearing life; I'm not sure which direction is optimal. Make sure the fan is on the 'cold' side of the device, so it's not sitting in hot air. \$\endgroup\$ – Someone Somewhere May 20 '16 at 9:00
  • \$\begingroup\$ Can you give us some more information? What is the expected lifetime of the object? What is the average / max power consumption? How big is the fan? How much dust is expeced in the environment? If there is any, how do you prevent it from interfering with the passive cooling? How many start / stop cycles of the fan are to be expected? \$\endgroup\$ – Posipiet May 20 '16 at 9:05
  • \$\begingroup\$ 10 kVA as a permanently running consumer device does not compute. Please give Voltage and Current. 10 kVA would consume in the order of 86.000 kWh per year. In europe, this would cost like 17 k EUR / year in electricity. \$\endgroup\$ – Posipiet May 20 '16 at 16:34
  • 2
    \$\begingroup\$ In a well-designed thermal system, often times you can get by with just the slightest draft. It's much like an analogy to an insulated attic where insulation thickness isn't the only factor if the draft is non-zero. If you run all the time at 10% of full RPM, does the system ever even reach 85C? I would design for the case where your system tries to exceed that 85C and find the PWM that will keep it below that and then always run it there. Lower RPMs will extend bearing life as well. \$\endgroup\$ – Joel Wigton May 24 '16 at 16:00
  • 1
    \$\begingroup\$ Don't Start up a fan on low speed setting (i.e power/duty cycle). Always give it a push, then wind it down again to the desired set point. The reason is that starting up requires the most power – a good computer fan I tried required 50mA to get going (reliably), yet maintained rotation with only 7mA from there on. \$\endgroup\$ – sekdiy May 30 '16 at 5:47
3
\$\begingroup\$

This is just my two cents but it would seem that you wouldn't have to worry about heating up the fan coil so much if you had the entire casing using fans in a positive pressure system. I'm assuming that the fan(s) you intend to use are brushless and you are putting a filter in front of the fan(s). This would make having to worry about normal dust as well as carbon dust from the brushes a non factor.

As for running the motor either at a higher set point or controlling the fan speed I would opt for the latter. Strangely enough I would as much as I could try to keep the fan moving when possible if you are going for longevity. Letting anti-friction bearings sit is possibly one of the worst things you can do for them. After a amount of time they will get flat spots and make metal to metal contact as their lack of movement would mean no wedge of lubricant to ride upon. That would be the sound you hear on old computer fans if their bearings are going out.

It might seem counter intuitive, but you could always have a way to plug the fan motor(reverse it's direction) if you have a filter built in ahead of it. This would allow it to automatically clean the filter a bit if the rpm of the fan motor dipped below a certain amount for the voltage being supplied to it.

This all depends though, I would look at the MTBF of the fan(s) in question before I made any decisions. Hope this is helpful. -J.G.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.