Netonix Forums

We are running a 24volt system (20amps) and using a Meanwell SD-350B 24/48 transformer to feed the WS-26-400 IDC (http://www.mantech.co.za/Datasheets/Pro ... 350-MW.pdf)

The switch is running 12 epmp 2000 radios, 3 AFX 5 radios and 1 SIAE Alfo plus.

We have had 4 times this week where the site dies, we go out and find the 24volt running, but the 48volt DC/DC has shut down. We replaced the DC/DC and a day or two later we found the DC/DC dead again.

We now installed a new WS-26-400 IDC and again after 10 hours the site died with the DC/DC tripping.

We are graphing the WS-26-400 ive gone through all ports and total wattage the switch never exceeds 200watts. (Its the only thing running off the DC/DC)

Any ideas on how I can figure out what is causing a amperage overload or voltage overload for the DC/DC to shut dwn?

The WS-26-400-ID is designed to hook to telco -48V system. It can run positive DC voltage as well 36V - 72V input range but why would you use a 350 watt power supply to power a 400 watt power supply? You can run the WS-26-400-IDC right off a 48V battery bank which is the proper way or a massive kick ass power supply system like at telco sites.

If you're going to power the switch with a power supply why not use the model designed for this which is the WS-26-DC?

However if you going to run a 24V battery system why not use the WS-26-500-DC and hook directly to the battery bank then hook your charging unit to the battery bank.

Our SMART DC switches like WS-8-250-DC, WS-8-150-DC, WS-12-250-DC, WS-26-500-DC is designed to hook DIRECTLY to a battery bank NOT to a power supply. If your going to use a power supply for some weird reason you should power the unit with a power supply with TWICE THE RATING.

So a WS-12-250-DC is you're going to power with a power supply for some silly reason should be 500 watt.

We imported several WS-26-400 on new sites we run them off a 48v rectifier system. The 26-400 suits the ground requirements etc.

However this site cant currently fit the rectifier so we ran it off the 24v system and stepped up off that DC. We have some other sites running in the same fashion.

Have look at the graph below, the breaks are where the PSU tripped from overload or over voltage its running nowhere near 350watts. Also these are new PSU's and a new netonix after the first event. Is it possible some damaged radio could be doing something strange causing an overvoltage? What can I check to try figure out the cause.

So if I follow you the Meanwell SD-350B 24/48 transformer is turning or OFF?


Most power supplies have over current protection circuits that shut them down.


Sometimes the over current protection allows for "spikes" to handle inrush and such ignoring current draw spikes for a specific persion of time.


Example:
A 200 watt power supply may allow up to 150% (300 watts) for X microseconds before it considers it overcurrent or a possible dead short and shuts down.

Not all power supply over current protection allows for surges or spikes to handle inrush demands.

It's possible your Meanwell unit is bad, test by swapping it out.
It's also possible the WS-26-400-IDC is bad and draws to much current, I doubt this but also swap it out and test.

However this site can't currently fit the rectifier so we ran it off the 24v system and stepped up off that DC. We have some other sites running in the same fashion.

But what I am saying is if you're going to use the Meanwell SD-350B 24/48 which is an ISOLATED power supply you might as well use the WS-26-DC and save some money. All your doing is using more wattage than necessary running the power through 2 power supplies which is losing efficiency with 2 power supplies 1 feeding another and doubling the efficiency loss and number of parts that could fail.

It seems like we found an AFX5 radio that has some surge damage on it that is causing this issue. The radio sometimes draws up to 40watts and we saw it draw 100watts on the port once (on the first WS-26-400-IDC) the ports where this AFX was connected to are blown on 2 x WS-26's now. So it seems somehow the AFX was causing a short the switch was staying on but causing the meanwell to trip.

Tell me, how does current limitation work on the netonix will it shutdown the port on overload, does it auto recover on fault restore or what is the design ?

But what I am saying is if you're going to use the Meanwell SD-350B 24/48 which is an ISOLATED power supply you might as well use the WS-26-DC and save some money. All your doing is using more wattage than necessary running the power through 2 power supplies which is losing efficiency with 2 power supplies 1 feeding another and doubling the efficiency loss and number of parts that could fail.

Agree with you fully, its not a normal setup we are pairing these with a Delta rectifier system just this site doesnt have one yet. We have limited stock of the netonix range in our country so we bought many of the IDC as we can use it in different scenarios.

Currently our switches use Polyfuses for protection against shorts.

You can read about PolyFuses and how they work by using Google. We have to size the polyfuse to deliver .75A with the switch operating in a 55C environment which makes it oversized as they work on thermal characteristics plus they are SLOW to open the circuit and thus the damage is done to the switch in that time. The polyfuse is simply there for safety measures not to protect the switch.

Poly fuses are not fast at detecting over current or a short and until they detect the short (by thermal means) the switch will take damage, usually blowing the Port's Ethernet Transformer which is only rated to 1 watt per pair. This damage can "usually" be repaired via RMA for around $50 for 1st port and $25 for each additional port. But sometimes the transformer fuses in a way that allows current back into the SOC and damages the port PHY which makes the board unrepairable without replacing the SOC which we currently do not that the equipment here to replace that part and have to replace the entire board. We are in the process of buying this piece of equipment for RMA department but this is not a cheap piece of equipment.

You can also learn about Ethernet Transformers and how they work with a Google search but to simplify things they use what is called magnetic coupling and there is no electrical connection between the cable and the PHY but if the coils burn up in a manner that shorts the current is free to pass into the PHY and POOF

However with our switch a dead short cable will most often only fry the single port transformer but "our" lab testing showed the ToughSWITCH Pro often suffered a complete board failure with a dead short cable?

We are working on what we call "active" over current protection for our next line of switches which is designed to protect the switch itself from shorted cables but this is harder than it seems but we are working on it. The problem is you have to deal with inrush demands which is often 200% above the rating you want for microseconds and the big problems is the current sensors often get fried due to a voltage differential on either side of the sensor. As I said we are working on it.

THe #1 Cause of WISP equipment failure is Ground Current, read the below posts TWICE:
http://forum.netonix.com/viewtopic.php? ... 279#p19279
http://forum.netonix.com/viewtopic.php?f=30&t=1816
http://forum.netonix.com/viewtopic.php?f=30&t=188
http://forum.netonix.com/viewtopic.php? ... =30#p13447
http://forum.netonix.com/viewtopic.php?f=30&t=1429

Also NEVER use Ethernet surge protectors with "passive" POE devices and switches as how they work is to clamp all 4 pair together and to ground which is a DEAD SHORT. Active 802.3af/at devices and switches can deal with this but passive POE switches using PolyFuses like our WISP Switch line and the UBNT ToughSWITCH line can not and switch damage will occur. Better to spend your money on BETTER GROUNDING.

So after taking what we thought was the problem (AFX5) we put that radio on AC mains and the site ran fine for 6 days and tripped again.

We now find a new radio doing the same thing as the AFX5 in question :



The epmp2000 was on port 7, we move it to port 8 it totally blew the port.

Now its on port 9 and look at the wattage. You unplug the ethernet at the switch plug it back and the radio runs fine draws normal wattage. Then the DC/DC then trips when it comes online after power cycling it the wattage then shows 82watts.

1. Their is FM on the tower two di-poles we have moved our equipment as far away as possible, we used good cat6 and dont have link issues anymore.
2. Their has been alot of rain recently.
3. The original AFX that we thought faulty : the ethernet cable made a small arc on the brush panel when we were working in the cabinet when it was not plugged into the switch or POE.
4. We arent sharing our brackets with other operators equipment, there are SIAE's etc above and below on different platforms. All earthed to the tower itself.

Its almost like there is some kind of charge or static that is building up on the cables and causing this ?

Any ideas of what else we can try here?