I have PoE device which in particular requires PoE 802.3af variant A (+DC on pin 1 and 2, -DC on pin 3 and 6), and want to know, if it will work with this switch (that device is Ubiquiti UniFi Mesh UAP-AC-M).
Model: GSM4328PA|M4300-28G-PoE+ - 24x1G PoE+ Stackable Managed Switch with 2x10GBASE-T and 2xSFP+ (550W PSU)
I believe this access point supports both Passive 24V ou Active 802.3af PoE modes and any standard PoE switch will work just fine in standard 802.3af mode. As such, the M4300 switches should be OK.
Thank You for a quick reply, but I still have doubts.
This device (UniFi AP) supports passive PoE, as stated in datasheet:
24V Passive PoE (Pairs 4, 5+; 7, 8 Return)
802.3af Alternative A (Pairs 1, 2+; 3, 6 Return)
So if switch will power device on pins 4,5;7,8 (variant B), may not detect it properly, or even damage by putting 50V, where 24V is expected.
If switch puts power on pins 1,2;3,6, then all should be fine. That's why is so important for me, which variant is implemented.
As a last resort solution is possible to connect device with only 1,2,3,6 pins, but then connection speed will be down to 100Mbit, what i want to avoid.
I think the access point manufacturer datasheet is leading to some confusion, and I can attest you will find many reports and feedback on other forums with this access point working fine with perfectly standard 802.3af switches.
So if switch will power device on pins 4,5;7,8 (variant B), may not detect it properly, or even damage by putting 50V, where 24V is expected.... That's why is so important for me, which variant is implemented.
This problem isn't a problem, absolutley not important, and not applicable at all as long as using IEEE standards PoE standards: Different from most of these passive (dumb, cheap, ...) proprietary schemes, a 802.3af or 802.3at PoE source does not blindly supply any power without a proper negotiation to the PD (powered device).
So if switch will power device on pins 4,5;7,8 (variant B), may not detect it properly, or even damage by putting 50V, where 24V is expected.... That's why is so important for me, which variant is implemented.
This problem isn't a problem, absolutley not important, and not applicable at all as long as using IEEE standards PoE standards: Different from most of these passive (dumb, cheap, ...) proprietary schemes, a 802.3af or 802.3at PoE source does not blindly supply any power without a proper negotiation to the PD (powered device).
I'm perfectly aware about that, and using non-standard device because it cheap, and know the risk.
As far as I know, standard compliant device can suply power with variant A or B, or even both. Just wanna know, how that standard is implemented in my particular switch. Unfortunately, nobody can tell so far.
@WojtekC wrote: As far as I know, standard compliant device can suply power with variant A or B, or even both.
Again: Unless there is a standard compliant PoE device plugged, NO power will be supplied by the switch. There is an (amazingly complex!) handshaking happen when a PoE device is plugged - before power will be supplied.
Don't waste your time about a problem which does not exist on 802.11af and 802.11at swiches - they will power your UniFi without blackening them.
As we talk of PoE for Gigabit Ethernet where all four pairs are used for data ... the power is provided "in-band" anyway and the subject question obsolete.
I have 2 UAP-AC-M-PRO connected to a M4300-28G-PoE+, yes this is the bigger model that only supports 802.3af but I can say that this particular model does work as expected.
Other than that I agree with schumaku concerning the sensing that happens with PoE. It gets even more complex with PoE+... 😉
"On a switch with greater than 16 ports, a PD on Port 1 powers on but a PD on ports 2, 3 & 4 does not power on. Similarly, a PD on port 5 powers on but a PD on ports 6, 7 & 8 does not power on."
------
After switch reboot, all devices works O.K., but if I unplug device and plug it back, it is not powered. Only one from fourth devices is powered, and if I unplug that one, another one (just one) starts being powered and works.
Only workaround I can find is to connect device with patchcord with only pins 1,2,3,6 connected. Device is then instantly recognized and powered. You can also set port to 100Mbit full-duplex, with no auto negotiation, so device gets connected much quicker.
As article suggests i contacted support, but with no success so far...
What is odd is not the switch here. The switch is fully compliant with 802.3af and 802.3at standards (hardware handshake and/or software handshake). One could blame proprietary third-party PD devices that are far from the standard, here?
We will need tech-support file out of the switch when the problem occurs with standard patch cables in first place, I believe, to understand when the access point is deviating from the standard.
As I said before, I am aware of all consequences from incompatibility of my devices, and prepared for all sort of problems.
Just found the mentioned article, and hoped, that if similar problem is already known, some solution also exists. I like to know that solution, even if problem mentioned in article is caused by standard-compliant devices, and some other switch.
I'll be happy to provide support file to my support ticked, if somebody ask.
