Z-Wave RF optimization

rafale77

The expertUI of Z-way has a very interesting menu called analytics which I have not seen on any other controller.
My engineering curiosity and OCD tendency to want to optimize everything lead me to dig into it to see if I could get things working even better... even though nothing seem to be broken since my move from the UI firmware to z-way. So the background noise menu was of particular interest.

So here it is: This was my first shot of the starting block on the left side of this graph and you can see that at around 21:00 I did something to reduce the noise. My zwave.me uzb is connected to the vera through a usb extension cable and a usb hub but initially the vera and usb hub were sitting only a 6-7 inches (20cm) away from the uzb. Keep in mind the scale on both channels. What I did was to move the vera away from the uzb stick, to about 80cm but the big step was for me to use some metal leftover amazon gift card boxes I had from christmas and covered the usb hub with them. The step back up was me moving these macgyver shields around.

Next step is still goofing around with the shields to see if I could optimize further and while I initially made it worse, you can see that it got a little better later on. That USB hub is quite noisy.

After sometime, I just thought that the whole thing is just too sensitive any movement around that area could potentially cause problems. So I decided to remove the hub connected to the vera which really was just allowing to extroot the vera. Since I am no longer using the vera as a vera but only as a remote serial bridge, which doesn't write anything on its drive, I took it out. Again look closely at the scales. It did take the noise down another notch.

Now the stick is still sitting 14inch away from a UPS with a bunch of wall warts connected to it, powering a variety of devices like a hue 2 hub. I had the idea of replacing just the wall wart (AC-DC power supply) for the vera with a POE splitter. Since channel 2 is already hovering around the detection limit, I am only showing channel 1 here

This last step alone enable me to flip this other statistics menu:

It not only significantly decreased the CRC error rate but also made the foreign network impact greater that the CRC error bar negating the success on reception for the first time. When I took a zniffer to look into what these CRC error frames were, I discovered that they were all from my neighbor's network which is distant and have obviously poor signal strength. (Yes I can see his HomeId from the zniffer). The fact that I see less error means that my controller's range is greater. It is now able to recognize frames which are from my neighbor's instead of saying they are corrupted... Also The back off frames number decreased indicating less wasted/repeated frames on the network.

CatmanV2

Dude you are so over it!

One of the things that I like is the multi way USB chargers for driving things like Raspberry Pis. I have nothing like the analytical skills you have, but removal of wall warts seems a good thing!

C

rafale77

I will be posting some more test results, not just for zwave but more generically about these wall warts, their quality and how to replace them. I am waiting on more samples.
Focus will be on:
Hi frequency RF emissions of the DC output cable. (which likely corresponds to noise in the power output)
Low frequency RF emissions of the DC output cable. (which would correspond to ripple in the power output)
I have also some disturbing results from my microwave oven... It is so poorly shielded that it shutdown wifi, zigbee, BT and another of my security system running 2.4GHz when it runs by completely saturating the 2.4GHz band. I bought a shielding fabric to cover it top see if makes a difference.