Netonix Forums

So has anyone verified that the NEW AFX radios support 48V or 48VH?

Also did someone say if the model supports 48V it indicates so inside the Ethernet cover?

If it does support 48V does it also support 24V and 24VH still as well?

Things we need to know:
Did you power it with 48V or 48VH?
Does it still accept 24V and 24VH?
How exactly can you tell the AFX radio you have will support 48V?

What are new AF5X? We have some bought end of last year which all run 24VH.

I run an AFx5 on 24v with 10m of cable. It uses 7-11 Watts.
I ve read somewhere that early models are not supporting 48v and the change can be seen from MAC address.
Since mine works ok and it is an AC site i didnt want to try raise the voltage.

I didn't try it but that would be nice to power those radio without using VH ports. If Ubnt could give use the supported radio with mac address, that would be nice but I don't expect that from Ubnt.

I have one site where I have a AF5X and AF3X on 48VH. Drawing between 5-13 watts watching it the last few minutes. I found that URL that says which ones dont support 48v power and it appears we got one of the original AF5X's that do NOT support it. All of our others do, I just havent tried them on 48VH yet.

Here's the URL on UBNT's wiki that details which MACs do/dont support the 48V power on 5X. Note that 2X and 3X all support it.
https://help.ubnt.com/hc/en-us/articles ... ut-voltage

CLICK IMAGE BELOW TO VIEW FULL SIZE

I have not put any of my X's on 48V , I did find that my 2X link was less stable when powered at 24VH on the switch vs running on the supplied brick. I did observe occasional blips as high as 24 watts on the 2X . When I have expedient access to mountaintops again I can deploy some better backup and experiment then.

We have run 4x AF-5X @ 48VH all newer than the serial number break above.

For what it's worth, 1 died, but could have been a manufacturing defect - rattling parts in radio case. Unit was RMA'd with no complaints.....

rockhead wrote:I have not put any of my X's on 48V , I did find that my 2X link was less stable when powered at 24VH on the switch vs running on the supplied brick. I did observe occasional blips as high as 24 watts on the 2X . When I have expedient access to mountaintops again I can deploy some better backup and experiment then.

Define more stable as I find no logic to your claim they run better on the supplied POE brick verses 24VH supplied by our switch as there is no magical electrons injected into the wire from the POE brick?

In fact our switch produces a much cleaner power, meaning less ripple, than the cheap power supply found in the POE brick. Our AC switch models produces DC power with a ripple of less than 600mV peak to peak and our DC switches produce less than 400mV of ripple peak to peak whereas the POE brick is more than 600 mV peak to peak.

Furthermore the POE brick UBNT uses is manufactured by Gospower or Gospell in China which is who also makes our power supplies. So the DC "power" is all generated by the same companies technology.

I have all my AFX radios (24 units) running on either 24V or 24VH and have no issues at all????

Our 24VH is capable of delivering 1.5A or 36 watts which far more than the bricks can produce.

Our normal 24V is rated at .75A which is 18 watts but that is at 55C and constant draw (more at cooler temperatures), in reality those ports can handle spikes in demand up to 36 watts.

By inserting a POE brink between our switch and the AFX radio you have created 2 more connection points which will significantly increase the chances of Ethernet errors as each connection inserts signal loss (roughly 1.5dB) and you now have one more component to fail and no way to reboot the device remotely.

More stable as in observed greater stability of data rates. the brick not only gives me more failure points but also removes my ability to power up/down the device from the comfort of my office but on this profanely long 2X link it seems much happier on the brick.