I have recently bought a RB5009UPr+S+IN and trying to get my head around how POE works. My internet connection is delivered by an ISP (SFR France) provided “box” that takes a Coax in and ethernet out, which I connect to my Mikrotik router. I would like to power the "box via passive POE, but have not managed to find any power specifications, the adapter is 12 volt from what I gathered from websites that sells replacement adapters. I have a POE splitter and when I connect the box with POE set to auto on the port, the box powers up OK, but power resets itself after 30-60 seconds. In the RouterOS logs I have the event
detected poe-out status: current_too_low
Does this mean that the power output from the RB5009UPr+S+IN is not sufficient to power the internet box and is there anything I can do to make this work ? The RB5009UPr+S+IN is powered by the 48v brick that comes with it, of there are options to how its powered
Possibly I’m using the wrong terminology here with passive POE, but I am powering a TPLInk Deco WiFi access point from it as well, by using POE splitter, so I was hoping to do the same with the box. However, if it’s not possible, I’ll have to find other options
Post the exact model of that ISP box, maybe specs can be found.
Loosely, if the replacement power supply is @12V, it seems like it wants 12V (or maybe 24V) while the Mikrotik will output PoE at the highest available power source (in your case only the default power brick @48V).
The PoE splitter that you have likely is a “plain” splitter, so it gets 48V from the RB5009 and outputs the same 48V.
There are splitters including converter, that take the 48V in from PoE and outputs 5V or 12V or 24V.
Be very careful when looking for one of those as many are 100 Mbit compatible only (I presume that the connection between your ISP box and the RB5009 is 1 Gb).
The RB5009 has a limitation on the amount of current it can provide over PoE out on a single port, 440 mA, so @48V the max power is 0.4448=21.12 W, usually those ISP boxes are low power devices, so the available power/current should be more than enough, but cannot say without more information.
Without any loss, 21.12W would be 21.12/12=1.76 A, but you have to take into account the conversion loss (and depending on the length of the ethernet cable also the cable loss), I wouldn’t be surprised if there was a 25% loss, so 1.760.75=1.32 A.
If the replacement adapter you have found is 1A@12V or 1.2A@12V (or less) the RB5009 should be able to power the ISP box without issues through the splitter/converter, if it is 1.5A@12V maybe, if it is 2A@12V forget about it.
RB5009 (Connected to a UPS) --2M Cat6–Basement Patch Panel–est 10-15M Cat6 – Upstairs Patch Panel-- 1M Cat6 → Internet Box
The box must be upstairs as the incoming coax is there, so the reason I want to POE power it is that the RB5009 in my little data centre downstairs is UPS powered.
It looks a lot like an Altice UBC1326 (AA00) made by UBEE in Vietnam.
I thought that our good guys at italian TIM/Telecom were good at re-branding/re-naming routers in the most confusing manner possible , but it seems like the french SFR people are even better at it .
The AC adapter (don’t mind the US plug) can be seen on page 4/5, it is an APD WA-36N12FU, labeled 12V 3A.
That makes it a 36W power supply, no way you can power it from a RB5009.
You will need a 802.3bt PoE power supply. (the next step after 802.3af and 802.3at, capable of up to 60W), besides the splitter/converter, and splitter/converters of that power are not common (please read as “expensive”), they common ones are usually 1A or 2A @12V.
The box is indeed the one you identified. To close the loop, I got a bit disappointed with the power requirements you identified so I got the clever idea to disable the WiFi on the box to bring down the power draw, which (touch wood) seem to have been sufficient, the box has been on for about 10 hours now without any power resets or log entries.
But your splitter is one with the conversion at 12V?
It is entirely possible that the “default” power supply is oversized ( and by turning off the radio surely you reduced the power needed).
As well it is possible that a device provided with a 12V supply actually accepts a much wider range, like (say) 8-30V, but it is rare that it can actually receive 48V (and needs a splitter).
The 8.9 W (183 mA@48.9V) that you see on the RB5009 is what I would have expected for a device that is essentially a media converter/modem, but even if you add to it the wireless (and that is enough to have the Mikrotik freak out) the added draw for the radio(s) must be in the order of magnitude of 440-183=257 mA, let’s say 300mA*48V=14.4W which is a “high” but “reasonable” number for an access point (“modern” ones are usually within the 802.3af specs. i.e. 12.95W, but older ones often wanted more “juice”).
So the device (with radio on) should be around 9+14=23 W, let’s say 24.
Even taking into account power spikes when booting, using a 36W power supply for it is definitely overkill, but maybe the device used for the FCC tests was intentionally given a beefier power supply, to be on the safe side.
Usually power supplies are 10-20% over provisioned, not more, so that they have some reserve for the boot spikes but otherwise work normally at 75-80% of their target current draw which is usually the more efficient point, 50% is unusual.
, but it seems like the french SFR people are even better at it 
.
