Has anyone some recommendation for a 57V PSU with at least 2A-2.5A for the RB5009? Since I would like to use more power than the supplied supplied 48V PSU can deliver, I am searching a higher psu.

Has anyone some recommendation for a 57V PSU with at least 2A-2.5A for the RB5009? Since I would like to use more power than the supplied supplied 48V PSU can deliver, I am searching a higher psu.

I had a UBNT Unifi7 Pro Max that would reboot when under load because it couldn't pull enough watts out the POE port on my RB5009. I wanted to use the 2.5Gb port with POE but I couldn't find a 57v high enough amp PSU so I finally just gave us and put in a TRENDnet 2.5G PoE+ Injector I got from Amazon for $40. It's worked really well so far, but it's another brick in my cabinet...

I know it's not the answer to your question, but it was what I resorted to.

also have a look at the used market for old Cisco PSU's used for Access-points (for those you couldn't or didn't want to use PoE). They are rock solid and cheap to get. I'm talking about really old stuff like for the AP1242AG series...

I wouldn't go to the max voltage (57V). It won't bring you any benefit over 48V. Better to get a beefy 48V PSU than a flimsy 57V. The router's internal electronics (DC/DC converter) will step down the voltage anyway. It's never a good idea to operate a device at its absolute maximum ratings if you want to keep it running for long.

I wouldn't go to the max voltage (57V). It won't bring you any benefit over 48V. Better to get a beefy 48V PSU than a flimsy 57V. The router's internal electronics (DC/DC converter) will step down the voltage anyway. It's never a good idea to operate a device at its absolute maximum ratings if you want to keep it running for long.

An issue with the RB5009 (and many other Mikrotik products featuring PoE out) is a (arbitrary) hard limit on currrent, in the case of RB5009UPr+S+IN that is:
Max out per port output (input 30-57 V) 440 mA

Now, in theory
48V*0.44A=21,12W
and
57V*0.44A= 25,08W
so, if the device attached on a port needs (say) 24W, you have no choice but use a higher voltage.

I completely agree with jaclaz. Additionally higher voltage can help mitigate voltage drop on longer PoE runs. And the output PoE voltage is NOT stepped down by the RB5009 when using a 56V PSU.

I've had success with the PoE Texas PS-56V120W unit.

Alternatively Meanwell has a wide variety of indoor/outdoor PSUs that can output at 56/57V and those also work well with the RB5009s.

Re higher voltage than needed power supplies.

If you have a PoE device that that has a voltage range for possible voltage input to power that device , I suggest to use a power supply on the lower end of the voltage input range ( when possible ).

Example , if you have a Mikrotik which comes with a 24-Volt DC power supply and the input voltage range is 24 to 57 volts , always use a 24-Volt power supply.
If/when you supply 48 or 57 Volts to the PoE input or power supply jack , then the device will voltage regulate the voltage down to 24 Volts. The problem here is that most internal DC power supplies use Linier power supplies ( not switching power supplies ). When a Linier power supply regulates 48 Volts down to 24 Volts , there is a 24 Volt loss that results in heat and the Linier power supply becomes very inefficient ( heat ).

A linier power supply is similar to a house-hold wall switch light dimmer. As you turn down the brightness , you drop more voltage through the variable resistor in the light dimmer. That voltage drop causes heat. *** Note - almost all space heaters are simply a huge resistor across a power supply load ( this is where the heat comes from ).

Most electronic devices will last longer is they do not run hot.

If you have a long long ethernet cable that is dropping voltage while powering a PoE device , keep in mind you are injecting heat into that long long PoE ethernet cable.

You are normally better off fixing or replacing the long long PoE Ethernet cable ( example cat5 to cat6 Ethernet cable ).

North Idaho Tom Jones

As often happens It depends, while having the lowest acceptable voltage may be a good thing for the internal power supply, It Is not such a good thing for the cable (and connectors).

If we have an hypothetical powered device that accepts *any* voltage and needs the same 24W of the previous example, powering It with 1A@24V or with 0.5A@48V has different effects.

The pins on a RJ45 connector and the crimping on the wires of the thin "blades" are more prone to issues when you have higher current going through them.

Besides, voltage losses on the cable will be much more.

On a normal cat5e cable 50 meters in length (a long cable but not incredibile long), at 24V there will be a power loss of around 7 W, at 48V that will be reduced to around 1W.
It Is a lot of difference.
BTW, to actually get 24V at the device you need to inject around 30V, while to have 48V at the device you need to inject 50V.

A cat6 cable Is better, but not that much better, at 24V there will be still around 5W loss.

A good method to measure heat on a 24-Volt to 48+Volt board with a temperature sensor and a common on-board linier power supply.

Power the device at 24 Volts, wait for the temperature to become stable , then query the temperature sensor.
- next -
Power the device at 48 Volts, wait for the temperature to become stable , then query the temperature sensor.


FYI , in electronics , all devices are actually current devices - not voltage devices.
Example; when you use an analog or digital multi-meter and measure voltage , the meter is actually measuring current ( not voltage ).
An analog meter needs a certain current to perform a full swing of the meter. When you change the settings on the meter , you are actually adding or subtracting resistors that are in series with the meter.
A Digital multer-meter basically does the same thing.
When you measure current , you are adding a resistance wheatstone bridge in series and the meter then is reading current in the middle of the 4-resistors.
When you measure voltage , you are placing in-series resistors, the lower the value of the resistors , the more the meter swings.
When you measure resistance , you are placing a voltage through the external device and the meter in series with the battery to the device and you measure current across the meter's internal wheatstone bridge resistors.

The problem here is that most internal DC power supplies use Linier power supplies ( not switching power supplies ).
North Idaho Tom Jones

Where did you come up with that one? Almost nothing uses linear power supplies any more. Switching regulator chips are easier to come by and in many cases are cheaper than linear regulators.

Re Linear power supplies ..

I am not talking about external power supplies - good external power supplies are indeed normally switching power supplies.
I am talking about simple single IC voltage regulators on circuit boards - for low current , they are normally linier voltage regulators.