PSU cooling fans  

Re: PSU cooling fans

Can you feel any air coming out from the bottom? I'd think that putting both fans facing the other side (with open slots) would cause most of the air to just blow across the very top of the case.

Re: PSU cooling fans

The fans aid natural convection by pulling warm air out of the top. You can definitely feel the breeze they create. The cover plate helps create a channel where the air enters at the bottom, flows up along the circuit board and exits at the top.

Re: PSU cooling fans

I'm not convinced it's heat related. My power supply is hand warm ( I can keep my hand on it). In addition, the PSU case is not designed to be active cooled (like Meanwell).
Without seeing an IR image and thermic distribution is hard to tell what's happening with/without fans. Most likely you'r just distributing the heat more evenly.

Re: PSU cooling fans

I wired them in series and connected them to the MK3 at the Einsy.

Robert, how did you wire them to the Einsy?

Re: PSU cooling fans

I wired them in series and connected them to the MK3 at the Einsy.

Robert, how did you wire them to the Einsy?

2x 12v fan in series to the DC connectors.

Im also not convinced that the PSU failure is related to the heat, i suspect it's caused some part due to to power overdraw
My 320Watt Meanwell measures 295W power draw from the wall for a short time during every cold start heat up phase, and the Prusa PSU is rated to 240Watt max.

PC builders always leave some headroom (PC components draw lets say 400 watt, then go with a 650Watt PSU, for overclocking go 750Watt PSU), and the LED PSUs we use are especially prone to failing if you use them at max power for a while.

Considering the both heaters are almost direct short circuits, these poor PSUs get shorted for a a good while during warmup, and then get PWM pulsed shorts after that.
Heater cartridge is 14 ohms, drawing 41 watts of power, heated bed is 3 ohms drawing 192 watts of power. Einsy electronics + steppers draw an additional 15 watt. Rest comes from other smaller stuff.

Fans will help for sure, and help even better if there is a optimal airflow (blow in from the bottom, suck out from the top), but the PSUs will fail regardless over time.
Faster than on the MK2, the MK2 PSU had some headroom to spare.

Re: PSU cooling fans

devilhunter - the failure you describe is exactly due to heat. If an electronics device doesn't immediately fail or suffer an exceptional event like a lightening strike, short circuit or dog attack, it will gradually degrade over time at a rate that is heavily temperature dependent. That is the basis of electronics life predictions and accelerated life testing. It's why your computer slows down when its CPU gets too warm. Most processes follow an Arrhenius relationship with an activation energy that yields about a doubling of rate for every 10 degree increase in temperature.

Yes, everything fails over time. The practical problem is to slow the degradation to the level where the device time to failure is long enough that you don't care.

I agree that the PSU is poorly engineered. I think its poor practice to spec a device to operate at its limit in normal use.
A switching type power supply is 80%-90% efficient so 295 watts at the input is not a 295 watt output, but point taken - the Prusa PSU is too wimpy for its application.

nikolai.r - The fact that the case is only warm doesn't mean that there is not a critical component that is hot.

The fans don't just stir the air in the PSU, they help draw in cooler ambient air at the bottom, flow it over the board and exhaust warmed air at the top. The transfer of heat from a device to the air is heavily dependent on air flow rate. Forced convection rates are far higher than those of thermally driven convective flows. Again, that's why your computer CPU has a fan on its heat sink.

clayton.r - I put spade connectors on the fan power wire ends and attached them to a pair of power leads in the terminal strip in the bottom of the Einsy. I snaked the power wires into the existing cable bundle.

The fans cost $7.79 for the pair delivered. I hope that the small help they provide keeps this poorly designed PSU alive while I'm interested in this printer. If not, I'll replace the PSU with a quality, properly sized one with a brace and the power sense added.

Re: PSU cooling fans

nikolai.r - The fact that the case is only warm doesn't mean that there is not a critical component that is hot.

Two thing.
1. Because the PSU is relatively cold, your cooling efficiency is low. Difference between ambient air and inside PSU is probable around 10C and we are still far away of rated 50C.
2. Because PSU wasn't designed for active cooling, adding fans might generate exactly the opposite. You might create turbulence and heat stall inside PSU.

My thoughts are equal good or bad as your fan solution. Without IR camera, I wouldn't do this adjustments. Because you're changing something without knowing/checking if it's getting better or worse.

Re: PSU cooling fans

OK nikolai.r - don't make one 🙂

A few comments so others aren't led astray:

Can you provide any example where the forced convection heat transfer coefficient is less than that of free convection?

Turbulence is your friend in convective heat transfer. The heat transfer rate increases significantly when you transition from laminar flow to turbulent flow.

What is "heat stall"?

Re: PSU cooling fans

Hi Robert,

I don't think anybody is led astray. Your solution looks straight forward for many people. Adding fans => colder power supply. This is something everybody understands.
But analyzing and understanding the air flow inside a case, testing the heat dissipation for each part in different conditions. This is something I'm missing here. All this things were done in the test labs and compliance checks. I'm not saying your solution doesn't work, I'm just raising concerns because it's untested. In best case you increased cooling efficiency by 5,10,20,30%. In worst case some parts are getting hotter now.

By "heat stall" I meant when the air got stuck somewhere, is not flowing anymore.

Re: PSU cooling fans

Chiming in, my PSU gets too hot to hold my finger on the top metal. The idea of adding a fan to it that pulls hot air out (although that might be hard on the fan running so much heat through it, rather than pushing air into the PSU) is how I ended up on this thread, with the hopes of splitting off the heatsink fan (which would sync it with heatsink cooling) are where I'm going on this experimental journey.

So yea, off of 110VAC, my PSU too hot to touch, a fan *IS* required, period.

EDIT: After looking closely at PSU layout, the design is actually using the outer heavy metal case as a heatsink, connected at two points, which explains why the case itself gets so hot, so it seems to me that the smartest way to cool it would be to NOT push or pull air at all through the PSU, but rather, cool the metal itself, possibly using those two holes on the large back side which has the least venting, and possibly wall off what vents are there from the air flow of the extra cooling, which means a clever part design, and printed in ABS. Either that or just add heatsinking to the case.

Re: PSU cooling fans

I see this is a relatively old thread, but I'll contribute anyways.

I mounted one 92mm Sharkoon fan on the top of the MK3 PSU, similar as the OP did, using silicone fan mounting pegs. I made a small PCB that switches the fan on above 30°C, and a thermistor is taped to the top side of the PSU case. It keeps the fan running only when the printer is actually printing something, to prevent unneeded dust accumulation.

It surely does its job, since without the fan the temperature of the PSU case reaches above 40°C after half an hour of printing ABS in the enclosure, but with the fan it keeps below 34°C all the time.

Currently the fan blows air in, but I'll also test the effect when the fan blows the air out.

Re: PSU cooling fans

I had a PSU fail. I may have to try this.

Re: PSU cooling fans

Looking forward to the MK4, I would think that having the PSU mounted beneath the heat bed would be smart. It could be heatsinked to the frame/chassis of the printer itself, thus reducing the load on the bed heater. It would be a smart use of power and make the entire device more efficient.

Re: PSU cooling fans

I think the MK4 will be a welcome addition.

Re: PSU cooling fans

i was thinking of putting a big noctua fan and wiring it directly to the power supply. any thoughts?

Re: PSU cooling fans

i was thinking of putting a big noctua fan and wiring it directly to the power supply. any thoughts?

If you figure how, please post it.

Re: PSU cooling fans

i was thinking of putting a big noctua fan and wiring it directly to the power supply. any thoughts?

If you figure how, please post it.

i will

RE: PSU cooling fans

I live in the tropics and I've had prints just stop for no reason. I thought the PSU might be overheating as it certainly feels hot, so about 6 months ago I rigged up a cooling fan (independently powered) directed at the PSU. Since then I haven't had a single "no reason" stoppage. I would say the PSU definitely needs a cooling fan if your ambient might exceed 25 C. I've ordered a second Mk3S and the first thing I want to do is add some forced cooling and do it 'properly' so it comes on when the printer is switched on. I've seen some threads with links to Thingiverse so I will take a look. 

RE: PSU cooling fans

I live in the tropics and I've had prints just stop for no reason. I thought the PSU might be overheating as it certainly feels hot, so about 6 months ago I rigged up a cooling fan (independently powered) directed at the PSU. Since then I haven't had a single "no reason" stoppage. I would say the PSU definitely needs a cooling fan if your ambient might exceed 25 C. I've ordered a second Mk3S and the first thing I want to do is add some forced cooling and do it 'properly' so it comes on when the printer is switched on. I've seen some threads with links to Thingiverse so I will take a look.