Voltage fluctuations

[adairw]

I have a ws6 running on the 50v 65 watt power supply.
The only thing the switch is powering is a Ignitenet metrolinq on port 2 at 24VH.

When monitoring the voltages from the gui, all three voltages are bouncing around. The 24v swing 2-3 volts.
I bring this up because 1) I've never seen it on any of our other switches.
2) the metrolinq locked up last night and we are trying to figure out why. Ignitenet doesn't think the voltage swinging a couple volts is a problem, but I'm no so sure.

This switch and metrolinq are at the base of a 1500' tall tv tower and we are inside the transmitter shack with their transmitter. So there is no doubt some stray RF floating around. Could the fluctuations be from RF and bad I2C readings? Switch is running 1.4.3rc7.

Things we will probably try,
1) replace to 50v supply
2) ground the switch (I just remembered that we grunted the crap out of the ethernet and the radio but didn't run a ground to the switch. (I can hear Chris already...)

Mostly curious if RF in the i2c bus would cause this or if it might be something else?

[lligetfa]

I would check the quality of the grounding/bonding. Sounds like an issue that jakematic was having. Not sure what his prognosis was but think he used a broad brush approach and switched to fiber ISL.

[jakematic]

Yes, I had this problem on a WS-8-150-AC viewtopic.php?f=17&t=1887&start=10

In my case there is something very weird going on electrically in the environment that grounding did not fix.
So I moved to fibre ISL after losing 2 switches, but that wouldn't work for a WS-6

My guess is ground potential difference traversing the ethernet... but I'll leave this one to the experts to diagnose.

[sirhc]

Yes this is ground potential differnce that tears up the switches.

Most often from not properly bonding service grounds to tower grounds or running dedicated ground runs that are more attractive than the Ethernet cable.

We are beefing up the current production run as best we can to prevent this issue when people do not have proper grounding.

Here is an example situation I was working on over the weekend from a customer in Canada:
Existing steel building 75' from new Trylon tower.
Existing Steel building has existing electrical service and probably old un serviced ground rods.
New tower 75' away has 3 brand new 10' ground rods.
Conduit runs between building and tower.
Tower ground rods are not bonded to building electrical service ground rods but at 75' this is pretty far but a bond would be better than nothing.
Equipment, routers, switches, and such in steel building.
Ethernet wire runs from inside building 75' under ground to base of tower then up tower to radios.
So now the tower ground rods are probably a better path to Earth ground than old service rods so ground current will flow through switch through Ethernet cables to radios to tower grounds.

What would you expect to happen?

As I said we are implementing changes to our current production run to allow the switches to handle much more ground potential current but the idea is to NOT rely on Ethernet cables to transport ground potential differences but rather equalize ground potential across your site and provide a dedicated path to ground that is better than it going through your Ethernet cables.

So we can and are beefing up our switches capacity to carry much much more ground potential current, far more than the radios can handle, so now radios will probably pop before the switches if you do not have good grounding so be prepared to climb and replace radios with dead Ethernet ports or do a better job at grounding to make sure there is no ground potential differnce. Also good grounding prevent Static and ESD Discharges from following your Ethernet cables to ground (through your switch).

We are estimating that the changes to the switches will allow up to as much as 40 Amps of ground current to flow through the switch......Can the radios handle that much current flow? Guess we'll find out.

Most of our RMAs were analyzed to be from ground potential current so we are making changes to reduce the number of RMAs but as I said these changes will now put the stress on the radios up on the tower.

Here are some good posts on grounding:
viewtopic.php?f=30&t=1816
viewtopic.php?f=30&t=188
viewtopic.php?f=30&t=1429
viewtopic.php?f=17&t=1786&start=30#p13447
https://community.ubnt.com/t5/airFiber/ ... rue#M31070
Look if you run an Ethernet cables between 2 differnt Earth Ground Potentials they will try to equalize across your Ethernet cable and this current can me a little and a ton and can vary during rain events as rain affects ground rod potentials as the dirt around the rods becomes saturated.

[Chris@edgarhighspeed.com]

We are beefing up the current production run as best we can to prevent this issue when people do not have proper grounding

I am looking for a couple of new switches and I am wondering what revision of the switches will this "Beefing UP" be completed?

[sirhc]

We are beefing up the current production run as best we can to prevent this issue when people do not have proper grounding

I am looking for a couple of new switches and I am wondering what revision of the switches will this "Beefing UP" be completed?

All switches shipping as of September 9th have this new configuration.

There are some units in the channel that do not have the modification.

But I want to stress this is an attempt to help people who do not practice good grounding.

Like when they do not bond their service ground rods to their tower ground rods, or if they are pulling electricity from a distant source where bonding the ground rods is not practical they should not bring the Earth ground but simply use the tower Earth Ground that way the Ethernet cable is not the bond between service ground rods and tower ground rods as show in the picture below:


However now that we are beefing up the switches ability to pass this ground current the next device this ground current will have to pass through will be the radio on the tower.

The switch can now pass between 2 amps to 48 amps depending on the model and how many Ethernet cables are being used between the switch and the tower.

So now maybe people with poor grounding will lose less switches but they may find themselves climbing the tower to replace blown radios.

The best/smartest thing people can do is truely read my grounding posts and spend a little time and money making sure they follow good grounding practices rather than hoping all their equipment can handle the ground current traversing the cables bonding the two ground potentials.

We are also modifying all RMA units that are sent back to us as of 9/9/2016 so any RMA units we had on the bench on that date we modified.

Here are some good posts on grounding:
viewtopic.php?f=30&t=1816
viewtopic.php?f=30&t=188
viewtopic.php?f=30&t=1429
viewtopic.php?f=17&t=1786&start=30#p13447
https://community.ubnt.com/t5/airFiber/ ... rue#M31070

Another IMPORTANT defence against ESD and Static Discharge damage is to have a dedicated #2 Green ground wire that runs up the tower to an isolated ground bus then #6 Green to each antenna mount and make sure the Ethernet service loop insures the path to ground is shorter through the intended ground wire otherwise static charge that builds up on the antennas will follow the Ethernet cable to ground. SO many people think the steel tower is a good path to ground which it is NOT. Steel has a much higher resistance than the Ethernet copper wire and most towers are bolted together painted or galvanized sections and those connections are old and corroded and the law states current will take the path of least resistance.

Another thing people need to learn is most damage is not from surges from the AC power grid as a simple surge protector will protect you there.

Most damage is from ground potential current flowing over your Ethernet cable or ESD or Static Discharges (more so static than ESD as near strikes are not so common as wind during a quick atmospheric humidity swing.) Static charges building up on towers and antennas happens all the time. It is not a question of if it will happen but how often and how will the current go to ground?

To many WISPs make a deal with a property owner to put up a tower and of course they drive in NEW ground rods for their tower then get electricity from the existing building with 25 year old electric service and ground roads that are no longer sufficient and suddenly all the excess current from that old service sees those nice new ground rods as a better path to ground so then all that current wants to run across your cables - SOMETHING WILL POP

[Chris@edgarhighspeed.com]

Excellent. How will I be able to tell if I receive a switch that was updated vs one that wasn't?

[sirhc]

Excellent. How will I be able to tell if I receive a switch that was updated vs one that wasn't?

Well any switches shipping from here since 9/9/2016 and any shipping from Streakwave as of yesterday. We actually went back and opened and modified every switch we had in the warehouse last week.

Or if the switch has a console port you can use a multimeter and test for near 0 ohms between Pin 5 of the console port and the RJ45 ESD shielding or chassis ground.

[jakematic]

Most damage is from ground potential current flowing over your Ethernet cable or ESD or Static Discharges (more so static than ESD as near strikes are not so common as wind during a quick atmospheric humidity swing.)

My yesterday:

• meh... I can get away with regular ethernet on that outdoor device... it's 'protected'... 'enough'... hate those ToughCable ends... need to buy Netonix ends...
• wind whips flag as storm rolls in
• remembered your post on static and discussion we had
• told self "Are you insane?!"

[Chris@edgarhighspeed.com]

Can someone answer me this? If I have a switch with bouncing voltages and I remove the switch, and plug it in on the bench. Should the voltages still bounce?

Are the fluctuating voltages only apparent when a grounding issue is occurring or once the switch is damaged from a grounding issue will the voltages always fluctuate until the switch is fixed.

I am trying to figure out if I still have a grounding issue or if I wreaked both switches prior to upgrading the grounding.