PoE and the C8 (with extensive testing)

dennypage · February 4, 2024, 5:20pm

In the hope that it will be useful to others, I thought I'd share my testing of the C8 powered via PoE.

There have been many references to issues when powering C7/C8 units using Unifi switches, but it should be understood that these issues are not specific to Unifi. My main switches are Cisco CBS 350s (192W), and while I had no issues powering C7s with inexpensive lightweight PoE splitters, the C8 has been a completely different story. FWIW, I've also tested with a dedicated Trendnet PoE switch (120W) with similar results to my Ciscos.

To be clear: I do not doubt the reports that that there are some switches, in some configurations, that do not have an issue. Nor do I doubt the reports that many Unifi switches have problems. That said, I believe that the majority of the issue lies with the splitters rather than the switches. I also believe that there are people with C8s on splitters that have issues, but do not notice because the hub still performs adequately. Lastly, I do not have any Zigbee devices, so I cannot speak to the reports that Zigbee is not affected. I have only tested with Z-Wave.

I tested 10 different models of USB-C PoE splitters with the C-8. With a few, I also tested multiple copies of the same model to confirm findings. I have made good use of Free Delivery & Free Returns from Amazon.

For those that want to skip the gory detail, the take home point is this: My opinion is that Z-Wave on the C8 appears to be very sensitive to voltage fluctuations.

If you want the gory detail, read on...

How I tested

When I started out I had a simple USB meter available, so I tried it with some of the splitters. I saw voltage fluctuations during heavy transmit, but wasn't able to tell much more with a simple meter. After a bit of soul searching, I found a scope style meter that I was willing to take a chance on. This clarified things a lot.

The test I run looks like this:

Plug in the hub's USB connection. This starts the data logging.
Wait for the green light on the hub.
Wait a brief period of time (down off the ladder and walk back to my office).
Perform a refresh on the problematic device while monitoring the sniffer.
If the first refresh works well, do it again for good measure.
Initiate a shutdown of the hub.
Stop the data logging when the red light appears.

For each graph shown, yellow is the voltage (V) and red is the power (W). The refresh on the problematic device is generally the power spike shortly after 3:00 in the graphs. Note that the software does not allow me to fix the Y axis, so there is some slight variation from graph to graph.

The brick:

For reference, here is a graph of what the test looks like on the included brick:

As you can see, the voltage is relatively stable, and holds or pushes up as the power requirement increases.

I am not an electrical engineer. That said, my opinion is that the splitters that work are ones that have isolated switch-mode power supplies (SMPS) with good voltage regulators. This makes them heavier, and generally more expensive, but they work reliably. I'm sure that there are other splitters of this ilk that can be found, but I've come to the end of my project for now.

YMMV

Now on to the splitters...

dennypage · February 4, 2024, 5:23pm

The failures

UCTRONICS UC-3AF-USBC 2.4A
UCTRONICS UC-3AT-DC 4A
REVODATA TYPEC0503G 3A (SDaPO Type C0503G 3A)
REVODATA TYPEC0504G 4A (SDaPO Type C0504G 4A)
PoE Texas Power for USB-C 3A (SDaPO Type C0503G 3A)
PoE Texas GAT-USB V2
Intellinet Gigabit Ultra PoE (561693)

There seem to be two large manufacturers of what I would categorize as the inexpensive splitters, UCTRONICS, and SDaPO. Both are commonly relabeled by a lot of other "manufacturers." The SDaPO units in particular seem to underlie much of the market, and are sold under a bunch of names--sometimes including text like "C050X" in the part number, but not always.

The failure symptoms for the splitters are not all the same. The UCTRONICS units generally show immediate problems with Z-Wave transmissions, easy to see with a sniffer. Z-Wave packets suffer significant delays coming out, and attempting to send a large number of Z-Wave packets in quick succession can cause Z-Wave lockups, sometimes lasting for a minute or more. But the hub does eventually recover.

The SDaPO units do not show immediate problems, however the Z-Wave power of the hub appears reduced. Some devices that are on the edge of direct reach when using the brick become one hop when using the splitter. Nothing outright fails, it just doesn't work nearly as well as with the brick. And again, pounding out a large number of packets exacerbates the problem. But they probably work well enough for most folk. Honestly, if I didn't have some temperamental 300 series devices, or spend so much time looking at Z-Wave strength and routes, I would probably not know that they are lacking.

The UCTRONICS UC-3AT-DC 4A and the REVODATA TYPEC0504G had been returned prior to the scope multimeter arriving, so I do not have graphs of those. Just the gross statement of "I see a lot of voltage fluctuation" using the dumb meter.

Here is a graph of what the test looks like on the UCTRONICS UC-3AF-USBC:

The voltage is a bit unstable and suffers drops when the power draw spikes.

Here is a graph of what the test looks like on the SDaPO / REVODATA / PoE Texas Type C0503G:

While the base voltage is higher, similar to the UCTRONICS unit the voltage is unstable and suffers drops when the power draw spikes.

The Texas PoE GAT-USBC V2 was a spectacular failure. Here's what the graph looks like:

Graph doesn't look that bad, right? The hub even functioned okay, about as well as the REVODATA units. The key problem is that the PoE device actually requires some form of USB power negotiation, and will not work without it. So the splitter works okay as long as the multimeter is in-line because the multimeter provides the negotiation, but if the multimeter isn't there then the splitter will not power the hub. Total fail.

The Intellinet Gigabit Ultra PoE (561693) was also a spectacular failure. I had high hopes for this unit, but it simply would not power the hub, even with the meter in-line. I tried connecting an iPad to it for testing, and it just cycled between power and no power every couple of seconds. It's possible that the unit was simply defective.

dennypage · February 4, 2024, 5:26pm

The successes

I really didn't expect that there would be so few splitters with performance to match the brick. I also did not expect that they would be from one company.

Here is a graph of what the test looks like on the Procet PT-PTC-AT:

While the base voltage is slightly lower, it is extremely stable. The C8 likes this. This thing is built like a tank. Heavy, with a metal shell.

While I have not tested it, expect that the more expensive 60 watt PT-PTC-BT (802.3bt) will perform the same as the 25 watt PT-PTC-AT (802.3at) I tested.

Here is a graph of what the test looks like on the Procet PT-ATC-5V:

While there is more fluctuation than with the previous unit, however it's still performs well enough to make the hub happy. It comes with handy cables, though they aren't great. Unfortunately it is bright white, which makes it hard to hide. Interestingly, this unit appears to consume ~350mW more power than the PT-PTC-AT.

Here is a graph of what the test looks like on the Procet 100Mb PTC-AF-5V:

This is the only 100Mb splitter I tested. Given that the hub is only 100Mb, it's not currently an issue but it doesn’t feel future proof. I honestly didn't expect it to work, because size wise it looks about like the other inexpensive splitters I tested. I took a chance on it because the other Procet splitters I tested both worked. However, when it arrived, I was surprised at its weight. While not nearly as heavy as the PT-PTC-AT, it's still heavy for its size.

It has the advantage of being inexpensive, having built in cables, and comes with a USB-C to Micro USB adapter which could be handy if you aren't sure if you will be using a C7 or C8. It consumes ~400mW less than the PT-PTC-AT. I just wish it wasn't white.

Caution: At least one user has reported that the PTC-AF-5V unit did not function properly with a C8 Pro. I will update if more information becomes available.

dennypage · February 4, 2024, 5:28pm

An almost ran

There was one other splitter that I really wanted to test, and had high hopes for, which is the Waveshare POE-SPLITTER-TYPE-C. Someone in an earlier thread had commented that they were successfully using one.

By the time I really got rolling on this, they had begun disappearing from Amazon. I was kind of stuck on the thing, so I found the original manufacturer and was about to order some shipped over. Just before completing the purchase, I noticed that there was a slight change in the specs from the units listed on Amazon.

One of the things that had liked about the Waveshare was that they advertised it as having a full SMPS, and that it used the SiLabs SI3404 switching power regulator. I like stuff I can find/read specs on, and they made a big deal about using this chip. Courtesy of the Wayback Machine, here is how they used to show the unit:

What I noticed on the manufacture's site is that they had quietly changed the chip. The newer units use a different chip, the Silan SD4950. Waveshare still advertises it as having a fully isolated SMPS, but they don't tout the chip, and Silan offers scant detail on the chip (at least in english). I contacted Waveshare to ask if they still had any inventory of the older units, they indicated that the supply had been exhausted.

I also tried several resellers to see if I could find some of the original units still in the pipeline, but did not find any that could be confirmed. The closest I got was placing an order with a seller on eBay who had initially confirmed his inventory was the original version. The order cancelled several hours later after the guy went to the warehouse with a screwdriver and magnifying glass and opened one up to double check. I really appreciate that guy.

Anyway, I've given up on the idea of testing this unit.

snell · February 4, 2024, 5:40pm

Just to add a bit of my own experience (I do not have extensive testing to show graphs though)... I was using my C-8 for some time connected to a DSLRKIT Gigabit USB Type C Active PoE Splitter and was having Z-Wave issues for some devices despite quite an extensive set of repeaters and such... and I could never figure it out until I saw an offhand reference at some point on the forum about PoE and Z-Wave with the C-8.

I ALSO had a power failure (~3 hours long) that prompted me to look at additional UPS capacity. I had a C-7 that got it's power from a power bank that was charged by that same PoE splitter. I bought additional power banks and placed them on every single Hubitat (including my C-8 and my new C-8 Pro). Ever since then the C-8's Z-Wave has been absolutely rock solid. I did a repair and EVERYTHING came back and shows happy for the first time in... well... ever.

So my anecdotal data is that my C-8's Z-Wave performance was also impacted by the PoE supplying it power directly, but the extra stage (the power bank) has resolved that.

Please note (if anyone else looks to add a power bank in between their power source and hub) that many power banks DO NOT provide power to devices while the power pack is charging. You need to find one that does or else you just wasted money for this purpose.

EDIT:
After further investigation the power bank I previously linked had a "fatal flaw" so I can no longer recommend it (at least whatever model of it is being sold now). When it loses power it stops providing power for a matter of seconds... then starts again. But this is long enough to cause the Hubitat to power cycle. It does a similar thing (shorter, but still enough) when it gets power again. So... it WILL power the Hubitat while charging and it WILL power the Hubitat during an extended power failure, but the Hubitat is going to power cycle at the start and end of the power failure.

lcw731 · February 4, 2024, 5:46pm

I must be in that group. All my hubs (1 C8 and 2 C7's) are powered either with the Uctronics or POE Texas splitters listed on a Monkerlink Managed switch (They have been for some time), and I have never NOTICED any issues. Would you recommend swapping them out for one of the Procet Splitters?

dennypage · February 4, 2024, 5:47pm

I believe that's the same as a UCTRONICS UC-3AF-USBC.

Finding always-on battery packs is kinda tough, and that one is reasonably priced. Had I seen that one before, and knew it worked, I might skipped all of this.

dennypage · February 4, 2024, 6:03pm

FWIW, all of my C7s were powered by PoE Texas splitters, which are relabeled UCTRONICS units. I never had any issue with them at all. I didn't run into problems until the C8, and even then they were sporadic.

I hesitate to recommend things... but if your current setup is adequate to your needs, why change it?

lcw731 · February 4, 2024, 6:15pm

Good point. In my day job I do a lot of looking for potential problems and trying to head them off before they happen. Sometime that spills over.

dennypage · February 4, 2024, 6:27pm

I completely understand. Me too.

ritchierich · February 4, 2024, 10:45pm

@dennypage Thank you very much for this post. I appreciate the level of detail and the time you spent diagnosing this challenging issue. I know the community will benefit from this.

danabw · February 4, 2024, 10:48pm

Lots of time/work in this, thanks for doing the (intellectually) heavy lifting for the rest of us!

jtp10181 · February 5, 2024, 2:45am

I don't even use PoE but this is great. I love that you did not just accept it doesn't work and figured out the why, with lots of testing. I am the same way with things, I don't accept something doesn't work without details on why.

danabw · February 5, 2024, 4:12am

Wait until you have children. You will give up, trust me.

snell · February 5, 2024, 4:45am

Not only do I agree with @danabw with the kids aspect... but also plenty of the work/projects I have done over the years have ended with "forget it, I am not bothering/wasting my time anymore, I am just replacing the stupid thing and moving on..." especially if it was "tech support" for family/friends.

kevin · February 5, 2024, 11:14am

Did you ascertain at all if RF interference / proximity from the POE splitters might have had anything to do with degrading the Z-Wave performance?

bill.d · February 5, 2024, 11:42am

Great thread and I don't even use a POE splitter. If it were my company I'd make 'resilience' my next #1 priority on both hardware and software sides. The price is getting high enough where more resilience should be expected as a customer.

ronv42 · February 5, 2024, 11:57am

Great review. I find that the components are constantly changing on these adapters. What I purchased years ago has a completely different board and chipset if I purchase it today. The manufactures blame it on the supply chain issue during covid. I also think it is to save a few pennies at the risk the device will go out of specification.

I remember having to use "Avaya" injector banks with their phones because Avaya wouldn't support the power delivered by the switch directly for their phones. We experimented with many Cisco switches at the time to eliminate the expensive Avaya hardware and there were many times a phone wouldn't complete bootup in that configuration. Funny that Cisco VoIP phones never had an issue with power from Cisco switches.

dennypage · February 5, 2024, 6:19pm

I didn't see any direct evidence suggesting RF interference as an issue. No packet errors in the sniffer. The problems seen were lower transmit power level and/or delay of transmissions by the hub.

When testing with the multimeter in-line, the splitters were more than 12 inches from the hub. That's quite a distance. Regardless of distance, if there were sufficient RFI to disrupt the hub and cause packet transmission delays, I would expect other aspects of the hub to be affected as well which I don't see.

Of course there could be noise transmitted via the usb power connection that could interfere with the Z-Wave chip, but I don't have equipment capable of measuring that.

eaton.blumenstein · February 5, 2024, 6:19pm

Ok, so this all makes a lot of sence, and I think some of this might be my issue now that I moved to a Pro.

with that being said, do you think something like this would help. my PoE to USB adapter has a USB-A end on it, from there I plug in an A to C cable to the hub itself.

if I put this in between , it could resolve some of the issues, as it makes sure that it stores a little bit of power to keep a smooth output. This adapter could eliminate issue.
Let me know what you think, I do not have the ability to do the testing you did, but this makes me think this could resolve some of the issues with low power fluxuations. if the device asks for more power, it comes from the reserve to fill it.

EDIt: FYI this might not be USB-C, might need to get another adapter from the Micro to C. but I assume it will still work. .