Don’t let your RB750 go below 32f (0c). 750g seems to handle colder temps fine.
We’ve swapped out 3 of 5 750’s we’ve installed in outdoor cabinets because they can’t take the cold. Have not had a single issue with the 750g, and we are swapping out the remainder of 750s that might see cold weather.
Change your thread subject/title and thanks for the info.
Good to know. It makes sense that the lowest power draw routers have problems with the cold. I guess they don’t generate enough heat to keep working when the temperature drops below the normal IC operating range.
When I had RB112 boards in my network they would randomly fail when the temperature dropped to 0 degrees F or below. I spent many early mornings cursing in below zero weather before I finally changed them all out with RB532 boards.
Tom
either i had lucky test sample or you have unlucky one 
where it was placed and how much traffic did it pass through (how loaded it was), and was it its shiny plastic case?
They are not shiny white like a mac. They are sort of a satin finish white, like an old microsoft keyboard or mouse. I’ve got a few indoors that are working great.
The problem ones have been in metal 1x1x2 foot outdoor cabinets (outdoors of course). Ethernet status stays 10/100 according to the backhaul radio (Alvarion 5.8 VL) going into ether1. The RB750 is unreachable though, disappears off the arp tables, ip neighbors, etc.. The box generally contains UPS, Mikrotik RB, 2 Alvarion POE units and doesn’t generate much heat. Traffic is pretty tame, sporadically getting a few mbit, but usually 500kbps or less. Repowering is a temporary fix. We use 3.30 firmware.
Temps range from 0f to 30f, but can be windy because of being in a place with a good view like on the edge of ocean or lake. We have traditionally used 433 or 433ah for this sort of thing, but we often don’t need the wireless slots, so a simple ethernet router is just the ticket. We have had no trouble with the 750g or the 433/433ah.
One recent changeout required a $200 private boat ride and a half day’s labor to swap. The most recent changeout required me to drive 3 hours in the middle of a 3-day long blizzard, on my day off. When the stakes are like that, spending more $ for the next router up is inconsequential if it works.
2 RB750 just spend few days/nights outside (up to -20 °C) - still working fine, but I will replace them with RB450 as soon as possible.
Reason: I think there is no place for SOHO router in the outside environment - usually it turns out as waste of money (as you yourself pointed out)
we have a special temperature chamber for testing extreme temperatures. We are re-testing the RB750 today, and I am glad to say - it works fine at even -40C (-40F) temperature. It even starts at this temperature if it has been off.
How do you power it? Maybe there is some other component that’s failing, cables, PoE, power adapter?
and with the test what is the max temp for this board as the temps inside my steel boxes go up to about 54C 60C
we are doing that now
thanks would be great
any new ideas for 2010
frying and freezing routers in 2010.
Xezen i would suggest you to shield your box from direct sunlight, so box does not get that hot, i do not think that there are many affordable high power devices around that would survive that without additional help, like shading the box, forced ventilation etc.
final results: with only ping running, the RB750 works from -40℃ to +55℃
By keeping it turned off and cold, the device can start when it’s -30℃
that have a cover built over the box and it still gets so hot we are adding air cooling to try help with the problems
The lack of published temp rating and the initial experiences have created by concern.
Thanks for the testing. They did work in the cold, just not as reliably as needed for us.
Unless you test a dozen of them for a month at a time in the cold, running traffic, I don’t think lab testing will show the assurance I like. I used to do final testing of electronics in a previous job, and know I can expect notable test variation from unit to unit and things will still be within acceptable specs.
I’ll keep buying the G model till winter has passed and the unit has proven/disproven itself in the outdoors.
oh, the test is quite realistic actually. the devices run in those temperatures for quite a while, and we do all kinds of traffic testing to simulate near-reality situations.
just finished testing RB800 and it gives stunning results: -75C to +65C if unit was started at room temperature, and a cold-start at -50C
As for the RB750, I can suggest to put it so, that there is not a lot of air around it, so that it can heat itself. Although our temperature chamber is quite large actually, not sure why you had these issues. Maybe humidity was different? How about powering, you didn’t say?
Sweet; I’ll have to try the rb800 sometime then!
For the 750, I used the standard supplied power supply. I’ll try one close to home with some bubble wrap around the shell sometime.
Does it go offline and that’s it, or did you notice it goes on/off every few minutes?
Normis,
Did you test both the 300MHz and 400MHz RB750 versions? And the new 6 watt power supplies? Both 110v and 220v?
It is great to hear that you are stress testing but I suspect that there are many variables that you are not testing for.
Tom
Normis - did you check the RB750 WITH power supply in the box - or without?
I have just blown a power supply in 42deg hot weather 
it was the factory supplied 24v power supply - and now it delivers 19,5v - without load - and on load 7-8v.