Starlink dishes go into “thermal shutdown” once they hit 50°C

[Centurius]

15 minutes ago, Forbidden Wafer said:

Now imagine a roof with asphalt shingles (which are dark and I assume heat up quite a lot more than other materials such as clay).

I mean, just touch the asphalt on a road on a very hot day. You'll immediately feel just how much hotter it gets than most other materials. Even in places like Scandinavia those can get really hot in summer.

[Forbidden Wafer]

21 minutes ago, Centurius said:

I mean, just touch the asphalt on a road on a very hot day. 

Not completely sure on the shingles because they're not sold at all over here, so never had first-hand experience. 

 

Here, on a hot day, the asphalt from the road kind of melts. Not crazy enough to touch it xD

[Kisai]

17 hours ago, Murasaki said:

Maybe it can be fixed if the dishes get coated with some reflective layer.

Then it would melt the receiving parts. In a traditional dish, the curved antenna itself focuses the signal on the LNB/LNBF. I looked up the Starlink dish and... it doesn't appear to be arranged this way.

Photo's show it as being parabolic (but that's just hiding the motor assembly), but it's not that large, it's not that different from a LSD (Little Silly Dish, aka pizza dish) DTH (Direct to Home) unit. So I suppose I have to find a teardown... oh look one exists.

 

 

So I skimmed over it, and it looks like they could design a model that could cycle either a liquid coolant(or warming fluid for the north) but they could also just put fins on the back of the metal to make the heat go somewhere in the first place.

 

The array itself is completely different from how a conventional dish works, which makes it have some redundancy I suppose. However it it got hot enough I think the motors would fail as well.

 

[Eaglerino]

8 hours ago, Spindel said:

My guess on most sun heated surfaces, like a satellite dish 

So you're making it up

[AlwaysFSX]

38 minutes ago, Eaglerino said:

So you're making it up

Have you ever measured the temperature of something outside in the sun? Ever? It gets hotter than ambient.

 

[trag1c]

9 hours ago, Zodiark1593 said:

CPUs, yes. The supporting components such as capacitors and MOSFETs, not so much. 

You're right about aluminum electrolytic capacitors...at least partially since many of them can survive 125C but their lifetimes are typically rated at about 1000-2000 hours at that temperature. However, there are plenty of capacitor technologies that are much more resilient with some notable examples would be MLCC (Multilayer Ceramic Capacitors) and Tantalum Polymer caps that have ratings of at least 100C with many more going up to 125C and even some ceramics being rated at 260C. You have to perform calculations to derate the life of the cap, capacitance and voltage (depends on technology used) but they hold up for quite some time even at high temperatures if you put the work into your design.

 

As @cj09beirapointed out the vast majority of MOSFETS can reach at least a 100C with even more ranging between 125C-150C (TJ - Thermal Junction) .There are also tons of them available (especially automotive rated) that will go up to 175C (TJ). 

 

IC's such as microcontrollers and microprocessors actually tend to be on the low end of the spectrum for operating temperatures.

[wanderingfool2]

2 hours ago, Kisai said:

So I skimmed over it, and it looks like they could design a model that could cycle either a liquid coolant(or warming fluid for the north) but they could also just put fins on the back of the metal to make the heat go somewhere in the first place.

 

The array itself is completely different from how a conventional dish works, which makes it have some redundancy I suppose. However it it got hot enough I think the motors would fail as well.

I wonder whether coolant would create an issue with the radio waves.  Given it's high frequency, and if the signal could potentially get distorted by having a semi-large body of water running behind the dish.  Although I guess it works in rain, but I think super close range and having more density might have more of an effect (but that's just an uneducated guess...I could be wildly wrong).

 

Still, my guess is that they will just increase the limit with software and in revision 2 they will have it built better.

[Taf the Ghost]

@Kisaithe normal trick is to shove some heatsinks on the back of things for a temporary fix. But they're going to need to run out a new model for non-temperate environments.

 

125C isn't hard to hit in full UV exposure for an entire day with a high ambient temp.

[cj09beira]

and i was planning to try to jump on this the moment they offered it in africa, guess not

[dalekphalm]

We do not know how wide spread this issue is. Some users are saying it's only affected a single individual.

 

Either it's fairly low-spread at the moment, or that specific person has a defective antenna.

 

In either case, I really don't think this is that big of a deal. Remember that Starlink is in Beta still. That inevitably means the hardware you get now may not even be the public release hardware when they exit Beta and go live with a full release.

 

I think most likely they will issue new hardware designed with these temperature exposures in mind. They may also update the software on existing hardware to provide some kind of mitigation.

 

Does anyone know if SpaceX has made any statements on this yet, beyond confirming operating ranges?

[Kisai]

9 hours ago, Taf the Ghost said:

@Kisaithe normal trick is to shove some heatsinks on the back of things for a temporary fix. But they're going to need to run out a new model for non-temperate environments.

 

125C isn't hard to hit in full UV exposure for an entire day with a high ambient temp.

I think if someone modified their unit to put some heatsinks on the back, it would probably solve it, particularly if the heatsink covers the back of the SoC part.

 

However a re-engineering of it would probably involve a Peltier cooler, which is what would have been used in orbit for the same reason. Though I imagine that would increase the cost.

 

Likwise, I haven't heard of it, but anyone using it in a snow region would have the same problem, where the snow would accumulate on the dish antenna. 

 

I wonder how well Starlink fares with solar-fade.

 

[Bombastinator]

On 6/16/2021 at 3:40 AM, Murasaki said:

Maybe it can be fixed if the dishes get coated with some reflective layer.

I was thinking a merely sunshade of some sort because I don’t know a lot about radiant solar heat, which I think has to have been the problem but that’s one option.  It is a beta test after all.  I suspect this is one of the kinds of data they were looking for. 

[Taf the Ghost]

6 hours ago, Kisai said:

I think if someone modified their unit to put some heatsinks on the back, it would probably solve it, particularly if the heatsink covers the back of the SoC part.

 

However a re-engineering of it would probably involve a Peltier cooler, which is what would have been used in orbit for the same reason. Though I imagine that would increase the cost.

 

Likwise, I haven't heard of it, but anyone using it in a snow region would have the same problem, where the snow would accumulate on the dish antenna. 

 

I wonder how well Starlink fares with solar-fade.

 

My assumption is that they'll end up rolling out 3 models for different regions. Cold, Moderate and High Heat. I'd also expect specific mounting/location requirements for each. The current operating temperatures look very much like it was tested in California's coastal region, lol.

[Bombastinator]

21 minutes ago, Taf the Ghost said:

My assumption is that they'll end up rolling out 3 models for different regions. Cold, Moderate and High Heat. I'd also expect specific mounting/location requirements for each. The current operating temperatures look very much like it was tested in California's coastal region, lol.

The preferred locations for car testing in the US is rugby ND specifically because it has some of the most severe weather in North America.  Not just very cold winters but also hot summers:   There is apparently there is an area that is even worse in Russia.  Both are very near continental centers which makes me think that may have something to do with it.

[Bombastinator]

On 6/16/2021 at 8:26 AM, suicidalfranco said:

Why is he at home watching YouTube in such a perfect day?

Should be at the beach or mountain

Have you ever lived in a place where it is 110f (43-44c) in the shade? It saps at you.  Walking in the sun is the last thing you want to be doing.  It can actually be dangerous.  Sweating helps.  Humans having sweat glands on every surface of their skin is quite unusual in tha animal kingdom and makes them better at temperature regulation than almost any animal.  Still it’s bad.

[akio123008]

I think this'll be fixed in a software update. They expected it not to go above 50C, it did. Now they'll set the limit to 55 or 60. The electronics will be fine.

 

I know y'all are more interested in water cooling systems, but a water cooled antenna system makes no (financial) sense at this power level. Would make for a fun youtube video, but that's about it.

[StDragon]

8 hours ago, Kisai said:

I think if someone modified their unit to put some heatsinks on the back, it would probably solve it, particularly if the heatsink covers the back of the SoC part.

 

However a re-engineering of it would probably involve a Peltier cooler, which is what would have been used in orbit for the same reason. Though I imagine that would increase the cost.

No, nothing so complicated.

 

The issue here is the SOC on the PCB is a layer directly behind the dish side where it's exposed to sunlight. That right there is a design failure. The ICs, or rather the SOC itself should have been isolated away from direct exposure. Maybe put the logic in a box at the base of the mounting shaft or something.

[Bombastinator]

3 hours ago, akio123008 said:

I think this'll be fixed in a software update. They expected it not to go above 50C, it did. Now they'll set the limit to 55 or 60. The electronics will be fine.

 

I know y'all are more interested in water cooling systems, but a water cooled antenna system makes no (financial) sense at this power level. Would make for a fun youtube video, but that's about it.

Electronics should be.  Plastic and glues may be another matter.  I don’t know what the case is made of or how it is put together.

[Arika]

I'll see how mine goes in full australian summer. i installed mine at the end of summer so it would already have been over the hottest period.

 

I've got ways to cool it down where it is if i need to anyway.

[mr moose]

8 minutes ago, Arika S said:

I'll see how mine goes in full australian summer. i installed mine at the end of summer so it would already have been over the hottest period.

 

I've got ways to cool it down where it is if i need to anyway.

How is it going so far?  better than fixed wireless?

 

[Arika]

Just now, mr moose said:

How is it going so far?  better than fixed wireless?

 

i couldn't even get fixed wireless where i am...so yes XD i went from adsl to starlink. though i have had to keep my adsl for working from home since SL is not 100% stable yet. i've only had 2 minutes of downtime in the last 24 hours and i'm pretty sure it's while i was asleep.

[akio123008]

6 hours ago, Bombastinator said:

 I don’t know what the case is made of or how it is put together.

Good point. I hope it doesn't melt at 60C haha. That would surely indicate some poor build quality.

[Guest]

5 hours ago, Arika S said:

i couldn't even get fixed wireless where i am...so yes XD i went from adsl to starlink. though i have had to keep my adsl for working from home since SL is not 100% stable yet. i've only had 2 minutes of downtime in the last 24 hours and i'm pretty sure it's while i was asleep.

Goes good for me. I was in the same boat - Spark (Telecom) only offered ADSL, so we switched to Starlink. Sometimes it says it's obstructed (its on the roof but we have a few trees here and there), but i haven't noticed any drops. Not too fussed with the ping, seeing as its the same, or better. Usually get about 60-100 ping, 100 down, 20-30 up. The speed varies a lot, but its always faster than ADSL (70-90 ping, 5-6 down, 0.8 up). 

Would love to do 4G, which is in full reception here, but Spark don't seem to think so, even though we live right by their cell tower.  

[Spindel]

1 minute ago, RorzNZ said:

...faster than ADSL (70-90 ping, 5-6 down, 0.8 up). 

Damn I'm happy I live where I live from this regard. 

 

Now I have fiber, but even with ADSL I would get waaaaaaaaaay better than those numbers (except that I have cut all incoming copper to my house  )

[Arika]

8 minutes ago, RorzNZ said:

Usually get about 60-100 ping, 100 down, 20-30 up

thats such a difference just across the pond

 

40 ping

250 down

40 up

 

My adsl was about the same though