Cooling SE Inverter with external fans

[sorka]

We're having a pretty hot week in the Central Valley right now. I know the my SE was designed to work in these temperatures but seeing internal temps from the API of 170F during the height of production has me worried about longevity. Yes the warranty is long but if I lose the inverter in the middle of the summer, I could be without solar for a few months given how long it takes for Tesla to get the ball moving on replacements.

My first attempt was a few 120mm cooling fans blowing down on the top from an angle and then just set flat on the top. Both of these had the same effect....about a 10F drop.

I figured I needed a better seal on the fans around the top of the fins so I tinkercadded some shrouds up and printed them up:

The result is a 20F drop from 170F to 150F. Here you can see where I've turned the fans on and off. Also, having them in place and not running them doesn't seem to result in higher temperatures from theoretically blocking the top of the fins. In this first graph, the shrouds and fans were placed put on top at 11:30 am without turning them on which did not result in a bump.

Yesterday I waited until it hit 170F the first time and then turned the fans on. Turning them off at 5:30 pm shows less of a difference because the device temperature was closer to ambient temperature and the cooling effect is going to be a percentage of the difference between air temp and cooling fin temp.

 

[power.saver]

We're having a pretty hot week in the Central Valley right now. I know the my SE was designed to work in these temperatures but seeing internal temps from the API of 170F during the height of production has me worried about longevity. Yes the warranty is long but if I lose the inverter in the middle of the summer, I could be without solar for a few months given how long it takes for Tesla to get the ball moving on replacements.

My first attempt was a few 120mm cooling fans blowing down on the top from an angle and then just set flat on the top. Both of these had the same effect....about a 10F drop.

I figured I needed a better seal on the fans around the top of the fins so I tinkercadded some shrouds up and printed them up:

The result is a 20F drop from 170F to 150F. Here you can see where I've turned the fans on and off. Also, having them in place and not running them doesn't seem to result in higher temperatures from theoretically blocking the top of the fins. In this first graph, the shrouds and fans were placed put on top at 11:30 am without turning them on which did not result in a bump.

View attachment 675580

Yesterday I waited until it hit 170F the first time and then turned the fans on. Turning them off at 5:30 pm shows less of a difference because the device temperature was closer to ambient temperature and the cooling effect is going to be a percentage of the difference between air temp and cooling fin temp.


View attachment 675582

Nice. I did a similar thing with one of my SMA inverters that doesn't have internal fans. I can't see where you ran your wires, but in my case I used a 12V DC computer fan, and mounted a small solar panel above the inverter. The fan runs when the sun is shining and shuts off at night. I get a similar 10-20 degree temperature drop.

 

[jrweiss98020]

Why not have them blow upwards, and let natural convection aid in the air movement?

 

[sorka]

Why not have them blow upwards, and let natural convection aid in the air movement?

I've been experimenting with both. So far no difference in temp so they are currently blowing up even though the picture was taken with them blowing down.

 

[gati2]

Did you cut into the top case with the shroud or is the only gap in the area in the back? If you didn't cut into the top case to allow the fan to blow unimpeded into the unit, I'm wondering if you're generating a bunch of turbulence that is hampering smooth airflow in/out. If you are blowing the air out, I'd also wonder if you'd get better results if you covered the other vent area in the back so as to avoid air getting sucked in from the sides of the fan.

 

[sorka]

Did you cut into the top case with the shroud or is the only gap in the area in the back? If you didn't cut into the top case to allow the fan to blow unimpeded into the unit, I'm wondering if you're generating a bunch of turbulence that is hampering smooth airflow in/out. If you are blowing the air out, I'd also wonder if you'd get better results if you covered the other vent area in the back so as to avoid air getting sucked in from the sides of the fan.

It's open on the back and I didn't cut into anything. The vents don't share air between them. They are completely isolated except for whatever air can flow out the back with the fins shoved against the wall. I'm going to modify the design to make it a single unit with a single fan push air through all fins. I've been playing with some mock ups and 2 fans vs one doesn't seem to matter.

Also surprising is that if I completely cover the top of the fins with something to block all airflow, it still doesn't seem to make a difference in the temperature. It seems there's enough leakage on the sides and in the back that natural convection still does it's thing.

 

[gati2]

I'd agree that moving the air across all the fins would make the most sense. Do you have an IR camera? I'd wonder if there was an area/fin hotter than the rest.

...A somewhat off the wall idea - if you were looking to decrease the temperature more, have you thought about using a Peltier cooling block?

 

[sorka]

I'd agree that moving the air across all the fins would make the most sense. Do you have an IR camera? I'd wonder if there was an area/fin hotter than the rest.

...A somewhat off the wall idea - if you were looking to decrease the temperature more, have you thought about using a Peltier cooling block?

I'd have to somehow modify the inverter itself and remotely locate the cooling fins of that type of cooler. Cooling air with one and then blowing it over the existing fins like I'm doing now would take a huge amount of power and be way less efficient than a standard heat pump.