Makes me wonder… how many people with chronic mesh problems may have bad/spattering power supplies? I was just remarking to @HSD99 that I’m always a bit mystified that I have a fairly stable system, when others have such bad luck with significantly fewer devices. I’ve run my Plus on PoE for years now. Maybe…?
1 Like
It could be a reason and is related to what support used to call the “baby monitor” check. 
Not in my case I have been extremely anal about locating the zwave antenna exactly at the same location even down to using exactly the same zwave dongle on the vera and z-way and exactly same environment. The difference was night and day, the vera was doing its “you know what”… and you reported it in your logs too.
1 Like
I run my Veras from high-quality 4A power supplies from a known and trusted power supply vendor. These supplies have much lower noise and ripple, Class B EMI, superior regulation, etc. The 4A capacity will allow for a ~ 100 msec ride-through on loss of AC mains. The supplies are connected to a commercial AC relay system that pings each Vera and will recycle the power on lost of ping. The relays are controlled via the Web. The AC source is a high-quality commercial UPS with extensive input line filtering and protection. All of this guarantees a stable, high quality 12 VDC input to the Veras.
I don’t use PoE splitters as they are not defined by the IEEE standard. Many of them violate the isolation requirements of IEEE Std 802.3 and can damage equipment.
1 Like
Hmm that’s good insight I did not know of. It is pretty convenient though… Maybe I need to look at alternatives… If I was to build a house today, I would definitely wire it entirely with all DC power converting at the entry point which would also enable direct powering from solar panels.
1 Like
Here is a calculator
1 Like
Purist. Kill-joy. Meh. 
Our smartest customers in the data center ordered DC power. It was a bit of a challenge finding 48VDC power supplies for the servers, but I don’t recall any getting replaced, ever, either. But a good telco-grade DC power system will beat anybody’s UPS any and every day of the week, and cost a lot less.
My old mentor was exploring/promoting the distribution of 330VDC for data center power, because, as it turns out, you can pretty much feed it directly to many standard switching power supplies and it (a) wouldn’t know the difference (basically just bypassing their own rectifiers), and (b) eliminated power factor problems. I think the force of that effort died with him, though. It’s been 8 years and I’ve never heard it come up again.
2 Likes
Maybe you can takeover the baton and make his vision a reality!
No way! That probably contributed to his early demise! 
Here’s the scoop: “While some advocate using DC over AC power citing power efficiency benefits, the availability of DC power delivery systems outside of telecom equipment racks is not widely supported and the proposed efficiency gains are similar to the gains achieved by operating at 240V AC over 120/208V AC. Thus, standardizing on 240V AC power can achieve similar benefits at a much lower cost, with less complexity, and provide greater overall compatibility.”
HV DC has a lot of safety issues associated with it. Not for nothing did Westinghouse win out over Edison!
48 VDC telecom supplies with 240 VAC front ends are the way to go. @rigpapa is correct. DC was the next big thing in ~2014, but reality set in.
I chaired the IEEE 802.3af DTE Power over the MDI (aka PoE) standard project back in the day, so call it a protective interest! 
Oh, you would have hated his original 540VDC proposal, then (p. 68-69).
So the 100s of network companies that put out POE devices like Access Points, Cameras, switches, routers and monitoring equipment are not the future? Wiring up ac outlets from someone 50 year old breaker box is clean regulated power? 
That’s not the question. The question is using dangerous high-voltage DC (e.g.380 V) as data center power. to supply hundreds of kW. 802.3af was designed to eliminate the need for normal mains power for certain network devices, such as Wi-Fi access points, that are frequently installed in areas (ceilings) where it is costly to provide mains voltage.
Since the existing structured wiring plant and category cable was now used for power delivery in addition to data communications, the project paid a great deal of attention to the issues of DC power distribution and cable heating, hence the voltage (42-57 VDC) and current limit (.35 A) chosen. It has done a wonderful job in that area, and follow-on projects (802.3at and 802.3bt) have increased power to almost 90W. At that power level there are restrictions on the cable types that may be used due to wire heating.
I plug everything directly into the wall using cheap-ass wall warts, and then wait for the grid to go down on a regular basis (which in my hometown is rather often). No UPS. No filters. No surge protectors. Lightning, spikes, brown-outs… BRING 'EM! Exploding transformers are our fireworks here!!
But I will admit this talk of PoE has me highly curious. Thing is, it costs money and means unplugging stuff, so imma prolly pass. 
Funny, I have my Vera Plus, Linux Edge, Hue hub and Google WiFi router all sitting within 12 inches of one another (right next to my Synology NAS and camera NVR), and none of them seem to mind the proximity.
On the plus side – getting serious now – I’ve never met another person who uses Z-Wave, has Hue lights, nor heard of Vera (or MiOS or MCV or ezlo). None. Zero. Nada. Thus I’m confident none of my immediate neighbors have any such technology installed, much less competing for airtime on any of the frequencies my stuff uses (except for WiFi, which I’m exceedingly careful to keep separated channel-wise).
Guess I’ve been lucky all these years!!
- Libra
Hmm I think you may have taken my statements a bit too far.
Again, I don’t have any zwave problem at the moment. As you can even see from my graphs above, that single POE splitter insertion decreased the noise level only by very little. It is my engineering OCD mind which got me to try out some theories and I just posted results which exceeded what I was expecting.
To give a easier to understand perspective, you can compare these to sound.
Basically the unshielded cable moving DC current with lots of ripples from bad wall warts are emitting low level noise. If the other guys nearby speak loud enough, the vera will hear it. If the other guy is too far away, its voice will be low and the ability of the vera to hear it will have a strong dependence of how loud that little wart nearby is even though it is relatively quiet.
There are only 3 dimensions in which you can move to change things:
-Space: move the noise away so the zwave antenna can’t hear it or move the other guy closer or get someone in between to repeat.
-Time: ask the initial guy to repeat hoping the next time you can hear it better and the noise is not there any more.
-Frequency: Move one or the other to another frequency like get the guy to speak in ultrasound for example so the antenna can’t hear it anymore. In this case I have no idea what the emission spectrum of these warts look like and I can’t control it, the zwave one is also fixed so it’s not an option.
Ultimately removing that little noise maker nearby helps the zwave controller hear that guy from far away. The idea of a mesh is to have some other guy in between who heard the far away guy clearly to repeat what he heard so that the vera can hear it louder (closer) but that means that air time is consumed because the same message was repeated as many times as you need (hops), so during that time no one else can be heard if someone wanted to speak, albeit in different space but as you get closer to the controller the available space is also smaller.
What this does is optimizing and making the network more efficient(a pet peeve of mine), allowing the network to get bigger (Scale). In most cases you probably won’t see much difference except for maybe some response lag in extreme situations. What happens in a model where every few rooms in the house would each have its controller though is a different story. My neighbor is a good 20m away and through a number of walls. Imagine what would happen if you have a number of networks in the house with zwave’ s relatively long range all interfering with one another competing for air time and making all these devices repeat because they can’t hear one another… Cacophony! Andrea Bocelli and the Castafiore singing out loud in every room and repeating the songs because they got interrupted by the singer next door.
3 Likes
This wins the forum for the day!
4 Likes
I bough a 48v to 5v micro-usb from this vendor: https://dslrkit.aliexpress.com/store/100107?spm=a2g0o.cart.0.0.1f0c3c006yB8Wt
Work like a charm.
Will buy 48v to 12v for my vera edge and vera plus. They should work as expected too. One model will be Fast Ethernet (100MB) for the Vera edge and one will be Gigabit for the Vera Plus.
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.