WS-6-MINI Keeps crashing/rebooting?

[sirhc]

I understand you read my grounding posts but you should go back and read them all again, I constantly update them.

I know you said you find you have better luck with the ESD cable but this is probably because you do not have a sufficient dedicate ground on all your towers to all your radios all bonded together?

Now I am not saying people should not use ESD cable because they SHOULD however most of my towers were put in before ToughCable came out and I ran normal indoor FTP cable in plastic and metal conduits between my ground cabinet and a top side NEMA cabinet. I often ran 24 wires and used 66 Block punch downs at top and bottom so I had 24 wires pre-run and ready for expansion and as you all have seen I put a lot of radios on a tower.

I have a main ground bus at the base of the tower that connect to all the tower ground rods.
I have a ground bus in my ground cabinet that is connected to the main tower ground bus with #2
I have a #2 bond wire from the main tower ground bus and the electric service ground rods.
I have a #2 bond wire from the ground cabinet ground bus to the electric service ground rods
I have #2 that runs up the tower from the main tower ground bus to a ground bus where my antennas are located.
I have a #6 bond wire from the upper ground bus to each antenna
I have a #6 bond wire from the upper ground bus to the upper NEMA tower mounted cabinet

Anyway the reason you probably see better luck with the ESD cable is because of the copper ESD drain wire in the ToughCABLE is taking some of the current off your data lines where as my grounding system wires is doing that for me.

If you read this post you will find the exert below in that post which is another reason I hate Ethernet surge protectors and how a ground potential difference can cause Ethernet Errors.

viewtopic.php?f=30&t=1816&p=13625#p13625

Data-line surge suppressors operate by "clamping" the data line to ground whenever the voltage exceeds a certain level. After the transient surge has passed, the suppressor turns off, allowing normal data traffic to resume. Data-line surge suppressors should be connected right at the network port of the equipment being protected. The suppressor ground strap should be connected only to the equipment chassis. This ensures that there is no potential difference between the suppressor ground and the equipment ground when a transient surge is being suppressed. For this reason, data-line surge suppressors that connect to ground at the receptacle are not as effective. When installing data-line surge suppressors, they should be used in pairs to protect the equipment at each end of the cable.

Several misconceptions persist about how to use data-line surge suppressors. Ironically, some of these originate with manufacturers.

- Data-line surge suppressors do not prevent noise from interfering with or corrupting data. If the voltage on the data cable is high enough to trip the suppressor, then the data is already corrupt.

- Data-line surge suppressors do not prevent ground loops from corrupting data for the same reason.

- Data-line surge suppressors are the wrong choice when a steady-state ground potential difference exists.

If the ground voltage exceeds the suppressor trip voltage, the data line will be continuously clamped to ground. If the ground voltage is high enough, the continuous current diverted to ground will burn open the suppressor. A properly designed suppressor will interrupt the data flow when its circuits become damaged. In any event, using a surge suppressor in this application will not restore data transmission.

[6gnet]

I know you said you find you have better luck with the ESD cable but this is probably because you do not have a sufficient dedicate ground on all your towers to all your radios all bonded together?

The BIG tower does have sufficient grounding, I simply observed that ESD cable helps with errors during high winds and when TV/FM antennas on the tower are running at high power (3MW TV and 1MW radio)

Now I am not saying people should not use ESD cable because they SHOULD however most of my towers were put in before ToughCable came out and I ran normal indoor FTP cable in plastic and metal conduits between my ground cabinet and a top side NEMA cabinet. I often ran 24 wires and used 66 Block punch downs at top and bottom so I had 24 wires pre-run and ready for expansion and as you all have seen I put a lot of radios on a tower.

That answers my question about the FTP cable. We acquired roof rights on few buildings that had outside UTP cable ran to the roof and building owners are not allowing to run anything new. The grounding on these buildings is sufficient for TV/SAT systems that's shared among all appartments in the building (about 10 to 20 in each building) to run ok. In cases where the ethernet cable is FTP, we see no issues, only UTP cable runs are causing problems....
I've installed surge protector on one and so far we're ok and replaced a mini with a small toughswitch for testing on another (we had 2x UTP cables going up on a roof on this one, one of them is ethernet and second is used to power the toughswitch with a surge protectors). I'm curious to see how both will behave.....

Anyway the reason you probably see better luck with the ESD cable is because of the copper ESD drain wire in the ToughCABLE is taking some of the current off your data lines where as my grounding system wires is doing that for me.

We only used the cable on the TV/FM tower which has really good grounding.

If you read this post you will find the exert below in that post which is another reason I hate Ethernet surge protectors and how a ground potential difference can cause Ethernet Errors.

I'm well aware of that but there's not much choice in case where there's an existing UTP cable assuming there's no potential difference between the basement grounding cable and the roof cable (which is the case on all buildings).

-G

[sirhc]

OK first I am not trying to be a jerk BUT:

You keep saying the tower has really good grounding, I GET THAT, BUT are you getting what I am saying that the tower ground MUST BE BONDED TO THE THE SERVICE GROUND? Meaning is there a heavy ground wire from the electric service ground rods to the tower and hopefully to all your equipment?

If the Tower has great grounding but is not bonded to the electric service grounding rods then your Ethernet cable becomes the bond which is BAD.

Do you get that you should have a dedicated ground wire between your equipment in your box and the radios on the tower that is shorter and less resistive than your Ethernet cable? Having this cable in place is what allows me to use FTP instead of ESD drain wire cable but even using ESD drain wire is not GOOD grounding solution, you want a heavy dedicated ground wire.

Please read these posts again:
viewtopic.php?f=30&t=1816
viewtopic.php?f=30&t=188
viewtopic.php?f=30&t=1429

And you said you got your question answered about the FTP cable already but is there a DEDICATE #2 ground wire running from the roof to your equipment wherever it is in the building? If not then your Ethernet cable is that bond and you will fry equipment.

[6gnet]

OK first I am not trying to be a jerk BUT:

You keep saying the tower has really good grounding, I GET THAT, BUT are you getting what I am saying that the tower ground MUST BE BONDED TO THE THE SERVICE GROUND? Meaning is there a heavy ground wire from the electric service ground rods to the tower and hopefully to all your equipment?

If the Tower has great grounding but is not bonded to the tower grounding system then your Ethernet cable becomes the bond which is BAD.

Do you get that you should have a dedicated ground wire between your equipment in your box and the radios on the tower that is shorter and less resistive than your Ethernet cable? Having this cable in place is what allows me to use FTP instead of ESD drain wire cable but even using ESD drain wire is not GOOD grounding solution, you want a heavy dedicated ground wire.

Yes, maybe I wasn't clear: the BIG tower (about 1100ft) has dedicated grounding bar on each level where antennas are connected and antennas are connected to it. Each grounding bar is connected via heavy duty 1/4"x2" grounding system.
Also the entire structure is grounded and both electrical service and the structure grounding is connected at the ground level.
Antennas are at the 300ft+ level while main box with switches at around 35ft, all equipment is grounded and the tower owner even requires a shielded electrical cable to be used.
There are no issues here with failing equipment and I was simply pointing that FTP cable even with ESD drain with a length of 95m (almost 300ft) was still showing errors on ports in some conditions while the new FTP cable with ESD drain and a copper shielding from ubnt eliminated that.

Let's put it this way, a direct lightning strike at the tower does not damage our equipment on 300ft level as we've observed recently :)

https://en.wikipedia.org/wiki/Raszyn_radio_transmitter

Please read these posts again:
viewtopic.php?f=30&t=1816
viewtopic.php?f=30&t=188
viewtopic.php?f=30&t=1429

And you said you got your question answered about the FTP cable already run but is there a DEDICATE #2 ground wire running from the roof to your equipment wherever it is in the building? If not then your Ethernet cable is that bond and you will fry equipment.

If by #2 dedicated wire you mean AWG 6 type of wire, then yes, this is exactly what we have running to rooftops, in fact it's a bit thicker than AWG 6 since we measure wires with mm2 here. (one of the posts has a link to home depot #2 wire which is AWG 6, am I correct ?)

Sorry if I mixed things up but going back "rooftop setups" I wrote about earlier (as pictured on the drawing): the grounding wire on the rooftop is connected with electrical service ground. The grounding wire of the cabinet is also connected by AWG 6 wire to the same grounding bar as I drew on the picture earlier.

There are 4 buildings with identical grunding: 2 separate buildings that have FTP cables are OK, no failures there, minis are running solid stable with great uptime (only reboot was needed to upgrade the firmware to 1.4.2).

2 other buildings with UTP cables do have problems, minis seem to go bad on them really fast and what bothers me is the fact that mini goes bad and not the other switch that's connected at the other end of the ethernet cable. I understand that all equipment must be and is grounded on both sides of the ethernet cable: the rooftop and the basement and I can assure you there's no potential difference between devices on both ends of the ethernet cable and there's no excess current running through data wires.

Now assuming that grounding is OK my whole point is that minis seem to be very sensitive, more sensitive than other netonix switches:
- even if there was potential difference between mini and the 12 port netonix switch, wouldn't it damage both the mini and the port on the 12 port switch ?
- in case there's any type of ESD that hits the UTP cable, shouldn't it damage both the mini and the 12 port switch as well ?
- without seeing the schematic for the mini and since there are voltage fluctuations observed and no errors on the port itself after the event that damages the mini, do I understand correctly that the power part of the input ethernet port gets damaged and not the data part?

I also noticed that out of 3 damaged minis:
- one has fluctuating voltage but sill runs solid with all PoE ports on and set to 24V
- one is stable with less than 2x 24V poe ports active and freezes if 3 ports are powered on.
- one is stable with less than 3x 24V poe ports active and freezes if 4 ports are powered on.

Perhaps some conclusions can be drawn from this to improve minis or at least to suggest how we can protect them from getting damaged in case it's not possible to replace UTP cable with FTP type.

Wow this is long post ;) one of the longest I've written so far lol

-G

[sirhc]

OK thanks for clearing that up, seems your grounding is in order at first glance.

Now the question of why does a MINI appear to get damaged when not using ESD drain wire cable...

First question and I am sure you answered it already, is this MINI in the ground box or in the tower?

If the MINI is in the ground box be sure to connect the Earth Ground lug because that is the only path to Earth Ground it has if powering with the barrel adapter. If you are powering via a POE brick with ESD cable it can get access through the POE brick but still good to connect lug to ground.

If the MINI is on the tower and you use just standard FTP cable with no ESD drain wire then the switch does not have access to Earth ground and I can see this as causing a problem for the MINI as ESD discharges, Static Charges, and even stray voltage/current would be forced to ride the data pairs down and into the switch and could cause issues.

If you mount a WS-6-MINI say on a roof powered from below via POE in on port 1 but the cable is just standard cable with no ESD frain wire then the switch is floating. Now you use ESD cable to the radios so the switch has access to the tower Earth ground but stray voltages or ground potential would ride the data pairs down and could be bad. Think about if there is stray voltage at the switch then it would either have go all the way to the radio to get to the tower ground then run that length of ground cable down to Earth Ground.

It all comes down to having everything bonded together with heavier less resistant cable even if under load from a transient voltage from a near strike to attempt to keep the ground potential as close together at any point as possible. No matter where there is transient voltage there must always be a shorter path to ground on an intended ground cable so that current does not take the Ethernet cable to ground.
The fact that you see less damage when using ESD cable might be indicating that the grounding system path you have is not the shorter and less resistant than the Ethernet cable path to ground.

Do you always make sure from any piece of equipment such as a radio that there is a dedicated ground path that is "shorter" than the Ethernet cable??

I always make sure my Ethernet cable run is at least a 10% longer path than the intended ground wire run by adding a larger service loop.

Also some people think that steel is OK in the middle between and Earth Ground cable and the piece of equipment and this is so wrong as steel is MUCH higher resistant than copper plus the insertion loss of the bonding point which may be corroded.

[6gnet]

First question and I am sure you answered it already, is this MINI in the ground box or in the tower?

Minis are up on a chimney and are grounded with a ground lug connected to a copper bar that's grounded with a AWG 6 grounding cable.
All FTP cables that we use have ESD drain.
AWG 6 ground cable is shorter than any ethernet cable since we have a service loop by the switch.

If the MINI is in the ground box be sure to connect the Earth Ground lug because that is the only path to Earth Ground it has if powering with the barrel adapter. If you are powering via a POE brick with ESD cable it can get access through the POE brick but still good to connect lug to ground.

Minis are powered via PoE from another Netonix switch, mini gets damaged, the other switch does not ;)

If you mount a WS-6-MINI say on a roof powered from below via POE in on port 1 but the cable is just standard cable with no ESD frain wire then the switch is floating. Now you use ESD cable to the radios so the switch has access to the tower Earth ground but stray voltages or ground potential would ride the data pairs down and could be bad. Think about if there is stray voltage at the switch then it would either have go all the way to the radio to get to the tower ground then run that length of ground cable down to Earth Ground.

This setup seems to fry minis:

Atenna -FTP w/ ESD drain-> Mini (grounded) -> UTP cable -> 12port netonix (grounded)

This setup works ok:

Antenna -FTP w/ ESD drain-> Mini (grounded) -> FTP w/ ESD drain -> 12port netonix (grounded)

Which points to ESD hitting the UTP cable....stray voltage on data wires and voila. We need to be able to protect mini somehow without replacing the cable hence my questions about surge protectors in previous posts.

Since you don't have much experience with surge protectors, perhaps we can figure out how much voltage on data wires is too much (50, 60, 80, 90V) so I can come up with a surge protector that will help.

Also if it's a stray voltage on data wires from ESD, I'm still bugged by mini frying and not the other end of the cable ;)

Before we've used Netonix switches, we had either Cisco (with power bricks) or Ubnt (PoE) on one side of the ethernet cable and antennas on the other side and even with 50 meter UTP cables never had a single port that was fried (luck?).

After we started to use Netonix, haven't had a single 10/12 port netonix switch that had problems (regardless of the location, up on a roof or tower or in the basement) but these minis which I love so much don't seem to like us ;)

Do you always make sure from any piece of equipment such as a radio that there is a dedicated ground path that is "shorter" than the Ethernet cable??

Yes

Also some people think that steel is OK in the middle between and Earth Ground cable and the piece of equipment and this is so wrong as steel is MUCH higher resistant than copper plus the insertion loss of the bonding point which may be corroded.

No steel or aluminum cable or pipe, only copper (I know who the hell uses aluminum, but you would be surprised what kind of left overs you can find here).

The fact that you see less damage when using ESD cable might be indicating that the grounding system path you have is not the shorter and less resistant than the Ethernet cable path to ground.

I don't see any damage with FTP (w/ ESD drain) cable. I do see damage with UTP cable..... however the mini gets damaged not the switch on the other end of the UTP cable.

Would it help to send 2 damaged boards back ?

-G

[sirhc]

Well you can get a any damaged WS-6-MINI repaired for $75 which covers us putting in a new switch board.

We repair all switches at cost (actually we lose a couple dollars sometimes)

Repair costs at a glance (new switch board):
WS-6-MINI $75
WS-10-250-AC $125.00
WS-12-ANY $150.00

250W power supply $75 (we make a couple dollars here)
400W power supply $100 (we make a few dollars here)

We are improving ESD protection with every board Rev since Rev D based on information we find by diagnosing fried boards in the RMA channel

Current Rev in Production is F

We will not know if the enhancements improve things for at least 1 year after we compare to previous year to see if an improvement was realized.

So RMA those switches

In July we will offer replacement boards for sale on our web site to avoid the return shipping fees for people.

[david.sovereen@mercury.net]

We just had a port on a WS-12-DC fry. The port won't power up an AP. Not sure if it will provide Ethernet link. Is this repairable, and if so, is it the $150 listed above? What is the process for sending in units for repair?
Thanks,
Dave

[sirhc]

We just had a port on a WS-12-DC fry. The port won't power up an AP. Not sure if it will provide Ethernet link. Is this repairable, and if so, is it the $150 listed above? What is the process for sending in units for repair?
Thanks,
Dave

Yes a new board is $150

Follow the RMA instructions in this thread: viewtopic.php?f=6&t=1259

[Travis_WVN]

I'm seeing this or a similar error on a WS-6-MINI. I'm sitting in a hotel room putting a bunch of radios together for a workshop later today, and the MINI is only acting as a network switch NOT using it for POE. Powered from the optional barrel-jack power supply.

I can hook 4 devices to it - as soon as I connect #5, the switch goes into a partial reboot cycle over and over until I cut power and remove one of the device cables.

Firmware is the latest version, 1.4.2. Voltage in the status tab does swing pretty significantly with no real loads on it.

Any ideas? In a couple hours I'm supposed to start getting set up to show people how to build private broadband networks, and expecting to promote the Netonix line as the gold standard. Got a bunch of them working fine at my WISP, so this is a surprise