Is the POE thing a matter of aesthetics, e.g., you don't want to have a cable running between the HE and an outlet? I understand the benefit of POE for devices like security cameras, but I don't see the real value of it for devices that typically end up on a shelf, in a rack, etc.
Saves using an outlet on typically crowded power strips. Or if the device is in a location where power isn't available...
I agree that those are two reasons, but the first is easily fixed. I wonder how many HEs fall into the second reason. Seems like it would be a very small percentage.
hub placement is quite important if you only have one. So if you are a slave to receptacle locations then the placement wont always be optimal. Running a data cable is alot easier and cheaper than running line voltage if you already have POE available.
It's not just an issue with Unifi switches. It happens with my Cisco CBS switches as well.
I cannot say I know for sure, but I think the actual issue is probably a current draw problem. When transmitting, the power requirement of the hub jumps significantly for a brief period of time. I suspect that many PoE splitters, or switches, cannot sufficiently satisfy the sudden current demand.
I've tested 3 PoE splitters with the C-8:
- UCTRONICS PoE Splitter USB-C 5V, 2.4A
- UCTRONICS PoE Splitter 5V 4A
- REVODATA Gigabit Type C PoE Splitter 5V,4A
The failures are not all the same. The UCTRONICS units show an immediate problem with transmissions. Easy to see with a sniffer. Z-Wave packets suffer significant delays coming out, and attempting to send a lot of Z-Wave packets in quick succession causes Z-Wave lockups, sometimes lasting for several minutes.
The REVODATA unit does not show immediate problems, however the power level coming out of the hub appears to be reduced. Devices that are within direct reach when using the brick become one hop when using the splitter. Nothing outright fails, it just doesn't work as well as with the brick.
I'm thinking of trying a battery between the PoE splitter and the hub to see if I can find some situation that works as well as the brick.
My C-8 Pro arrived today. I am using the exact same UCTRONICS PoE Splitter that I was using on my C-8 hub, connected to my UniFi Lite 16 PoE switch. Once again, I have experienced no issues whatsoever with any of my Zigbee devices.
And, just to make sure... I paired a GE/Jasco Z-Wave switch as a quick test. I was able to pair it and control it with no issues whatsoever. Perhaps a Z-Wave mesh of just one node is not large enough for the issue to manifest itself? Or perhaps I am just lucky? Or perhaps the UniFi 16 Lite PoE switch is not one of the problematic ones?
I doubt it is an issue with total power output from the UniFi switches, as even their weakest PoE switches can easily handle 15 Watts. The Hubitat hub's average power usage is about 1.5 Watts. Even if we tripled that during radio transmit periods, we're still talking less than 5 Watts.
I 100% believe users have had this issue. I just caution against any blanket statements where Community members are told that a Ubiquitu UniFi PoE Switch + PoE splitter will not work with a C8 or C8 Pro hub.
Which model of UCTRONICS are you using? I know there's a 3A version that I haven't tried.
I think you meant Z-Wave devices...those are the type that are affected by this POE issue.
Oops - just saw that you don't have any Z-Wave devices, forgot that!
Never mind...
I wonder if they changed the hardware on the C-8 Pro to support Thread & Zigbee at the same time. C-8 I think has a chip capable of Thread but couldn't do both at the same time. I still argue that they could have it set to do one or the other so we can pick if we want Thread only vs Zigbee only although there probably isn't enough of a market to implement something like that at this point.
That's definitely not going to be a day-1 capability of the Pro, and since the zigbee guts are the same as plain C8, I wouldn't expect it to happen anytime soon, if ever.
HE staff has previously said they aren't planning to pursue any time-slicing solution or Thread-only option for the zigbee radio.
There's a single 802.15.4 radio (at least as far as I can interpret the FCC filing). So zigbee and Thread cannot be used simultaneously. And, as @hydro311 indicated, Hubitat staff have indicated they will not implement time-sliced use of the radio to support both Thread and zigbee.
Like you, I too would like to use the radio in one C-8 as a zigbee radio and as a Thread radio in my other C-8.
That's a great point, and now that the Pro is out (so there's now a couple flavors of the C8 out there), perhaps staff would be more amenable to a "ZB or Thread" setup option - heck, it could be a good way to tout Hub Mesh and sell more hubs overall.
Okay, thanks. Thatās one I tested. I was hoping it was one that I havenāt, like a 3A.
Have you tried this one, with a metal enclosure? Someone else mentioned this one. I have not tried it, personally.
https://www.amazon.com/gp/product/B09MC4QDYG/ref=ox_sc_saved_title_1?smid=A3B0XDFTVR980O&psc=1
I haven't tied the Waveshare one yet. I saw the post, but I understood that the person was using in conjunction with a Voltaic battery. Unfortunately, it's a long lead time item from Amazon. I have another unit to test arriving on Tuesday. I'll try that one first.
I would like to test a Voltaic, but the smallest one, which is all the power needed, isn't currently available on Amazon. And I would rather avoid a battery backup because it would mean that I have to get on a ladder to do a power pull reset on the hub.
I have not tried any PoE splitters on my hubs but Iāve just been reading along.
Iām wondering if itās more of a conducted interference instead of radiated.
Has anyone tried a clamp on ferrite on the power cable between the splitter and the hub. Hell I would stick a couple on.
Second idea is adding a nice big bulk capacitor on the hub side of the ferrite, but that would be messy without an inline breakout board.
Thoughtful idea. I have a couple around here somewhere...
FWIW, I have put a simple USB meter on a couple of them, and I am seeing voltage fluctuations during heavy transmit. Unfortunately I can't tell much more with the meter I have. There are some nice scope style meters, but the PC software is too sketchy for me.
This whole thing has become a bit of a project.
It's z-wave that tends to have issues. zigbee is unaffected
Which is why I performed the following test as wellā¦