WISP Switch - NCS - What's the use case?

[mhoppes]

What's the use case for the NCS? I understand what it is but don't understand the why.

[sirhc]

Well unless you have been living under a rock for the past year you are aware that there are massive semiconductor shortages and LONG lead times.

Drive by any car lot and see very few cars on the lot because they are missing semiconductor parts and can not build them.

Most car companies are selling cars and trucks without all the electronic features simply because they can not get all the parts.

Being in the WISP industry you should be aware that pretty much all manufacturers even the largest like UBNT are having issues getting parts and it most likely will only get worse. Now if China invades Taiwan forget it as they produce 60% of semiconductors and 80%+ of high end semiconductors such as processors, SOCs, FPGAs, and so on. And with Brandon in office now would be the time for Xi to go after it possibly when the winter olympics are over around the end of February. WHO THE HELL KNOWS IN THIS ECONOMY / UPSIDE DOWN WORLD.

Well same for us, the current sensor we use from Texas Instruments are not available anywhere. With POs we put in last year at this time we were supposed to get parts in June/July 2021. Well it is February 2022 and we do not have all the parts on the POs. Our current sensor went from supposed to get in June 2021 to September 2021, then February 2022, and now September 2023.

Our current sensors are nice feature but are not needed for the switch to "function". So we simply did not put them on the boards as we do not have them, and can not get them. JUST LIKE A LOT OF CARS AND TRUCKS WITHOUT ALL THEIR ELECTRONICS ON LOTS TODAY.

In the mean time we redesigned our switches with a newer more expensive current sensor but that requires software changes which we are working on and those current sensors are only a 12 week lead time, or so they say.

So anyway most people can use the switch without current sensors to report watts used per port, think Tough Switches, and most other passive POE switches, they do not report watts used per port.

So if someone's choice is to have the same switch without current sensors or no switch at all we are offering a choice. Best we can do in this F'd up environment.

[mhoppes]

Gotcha... seems like a fair compromise. As we're evaluating what to do in regards to equipment shortages, can you speak to the weak surge protection of the Netonix switches and how ports get zapped very easily? Has this been fixed in this generation of switches, or are the PoE transformers or whatever still weak and susceptible to stray voltage taking them out?

[sirhc]

Well when we redesigned the boards for the new current sensors we did implement changes that "should" make them more hardened against transient AC current seeking a better path to ground on the tower.

The new current sensor is what is called an isolated current sensor which means it has no electrical connection to the POE DC voltage but rather reads the voltage drop which we use to calculate watts from a magnetic coil that surrounds the electrical path. We also added what we call tie downs or filter caps on each side of each port to try and shunt AC current to ground.

But keep in mind what blows up the current sensors or other parts on the switch is transient "AC" current and since the AC current is most often traveling from the equipment box through the switch across your Ethernet wires though your radio to the tower and we happen to be the first sensitive piece of equipment in that line and get hit first if this works as we hold and the transient AC current gets to the wire then to your radio it may be the part to fail.

So short answer is yes since we were re-designing we chose a more expensive current sensor and added enhancements so we hope better handle this transient "AC" current that does the damage.

[mhoppes]

I'm talking more along the lines of these scenarios:

Plug in a 48VH powered cable while the port is hot, blows the port often.
Short out the port, blows the port.
Static electricity coming down the cable, blows the port.

Why is stray AC voltage even reaching the switch port? Shouldn't that be grounded off if the switch is plugged into a grounded outlet?

[mhoppes]

I guess one other question then too - based on what you said above, would a switch WITHOUT a current sensor be less likely to die in the ways some of them have?

[sirhc]

Plug in a 48VH powered cable while the port is hot, blows the port often.
Short out the port, blows the port.

Plugging in a live port results in fried equipment if:
End crimped wrong, shorted cable, shorted device, or incorrect POE option is ON, these are user errors.

That kind of protection is addressed with the WS3 line with our OCP. Development has slowed on this again to supply chain issues. Hard to spend money on RND being a small company when you have nothing to sell because you can not get parts.

Static electricity coming down the cable, blows the port.

ESD damage whereas it does occur actually occurs far less than you think. Again because towers usually have better grounds than electrical services or equipment boxes the path to ground is not better through the ethernet cable provided the tower has a dedicated ground run with ground bus bars attached to all antennas.

Why is stray AC voltage even reaching the switch port? Shouldn't that be grounded off if the switch is plugged into a grounded outlet?

Again if the service ground is poor and the tower ground is better and since "most" radios bond DC negative to chassis/earth ground if we clamp AC current to ground it will take the path to better ground (tower).

Look, in our lab I am able to pass 120V AC though the switch, across 100m of cable through a Rocket radio through the SMA cable to the antenna and power a 100 watt light bulb (DEAD SERIOUS HERE). The problem is when the wattage becomes to much for the components on the switch or the radio to handle that something fries.


Also when your dealing with AC ground current bleeding in rain or storm events or from bad motors or AC units or wherever often this AC current is a harmonic which Hz FAR above 50-60Hz and we find it cuts right though filter caps above 400 Hz.

Ground current from AC which remember AC is alternating current so there is no positive and negative so it just there pushing and pulling back and forth creating resistance and heat - POP
This is a very complex issue that with all due respect most people do not grasp or think it is easily overcome, and then there are many that simply do not even believe it exists and I have posted many white papers from other engineers and still here we are.

[sirhc]

I guess one other question then too - based on what you said above, would a switch WITHOUT a current sensor be less likely to die in the ways some of them have?

Actually yes, the current sensors were always the most susceptible to damage from transient AC current. The next part would be the MOSFETs.

So will you see less damage on an NCS... in theory yes, but it depends on how many watts/amps of AC current and what is its oscillation/freq. Above 400 Hz is pretty damaging. It achieves this level of Hz from harmonics.