Damage of RS-485 Hub connected to few RS-485 units

[techie]

Carnage

Can anyone explain what happened to this circuit. It is an RS-485 Hub connected to a few RS-485 units on its 5 ports. Most of the connected circuits were also damaged. All circuits were opertaing on 12V and none of the 12V transformers were damaged.

Thanks for any answers

 

[solvarg]

My guess is, either lightning or floating ground with associated different potentials.
What is connected on the 485 network?
How is the power supplied to the various nodes?
How is signal ground connected to PE?

Regards

 

[techie]

Thanks for your input, solvarg.
This hub connects on one port, an isolated 232to485 convertor and then to a PC. One two of its ports (last two were unused), about 10 each data aquisition cards are connected. They each have their own 12v power supplies (transformer based). The data acquisision cards have some sensors connected to them. The connection to these cards is made using 3 wires, 2 for RS-485 and 1 for ground (or common).
The signal ground is connected to Earth at each data acquisition card.

The installation was working fine for almost a year when suddenly this happened. Before repairing all the damage, I do want to know the cause so that this does not happen again.

 

[solvarg]

What is the geographical distance between the nodes?

 

[techie]

the distance is hardly a few meters between nodes. The total distance from end to end is probably 30 to 50 meters. All nodes are in one big factory Hall.

Solgvarg, Your pointing to PE connection rings some bells. Can you shed some light on it. Is it possible that maybe the grounds of two nodes are at a largely different potential causing a huge current to flow thorough the data lines.

 

[penrico]

I think it was a lighting coming from the power supply. In the picture you destroy all the semiconductors, so only if a very high current comes from the power supply will destroy in this manner.

Remember, allways protect your circuits from lightings, much if you use long wires.

 

[techie]

I doubt any such occurance of Lightning. No thunderstorm or any such phenomenon was there on the day it happened.

 

[WA]

Initially,

I think to look at electrolitic capactior.

if this shorts out (electrolite drying out, leaking etc...)

I don't have schematic so I can only guess.

if capacitor shorts, then the 3 terminal regulator now has big problem.

I think regulator now shorted out and took out several flat packs.

I think that shorting is problem as the burned out resistors (severly) is tell-tale sign.
It has been my experience reguarding troubleshooting and repairing circuits.

I'm sorry if I can't help more, but glad to offer my thoughts.
WA

 

[penrico]

If it possible of a fast peak voltage shot?????

Is normal in power lines to have peaks of 3000v with 10 uSecs to 100uSecs of duration, if you don't protect agains its, you will be in troubles.

Transformers, will not damage at all, but semicondutors, died...

 

[techie]

The power input is from a 12V adapter. Here the circuit is fine. The input has a 1N4007 diode which is not damaged. Then there is a 1000uF/25 cap which is still OK. Then is the 7805 which is damged. All the ICs are on 5V and aslo burst.

 

[penrico]

I suggest, allways protect the input in power line with PTCs, varistors, and gas or ceramic dischargers to earth.

If you use it, only very big electrostatics discharges will affect.

 

[jzaghal]

Electrical Surge !!

Hi,

It is definitely a high voltage of some sort.
I believe it came through the transformer as a very
high surge.

Bye the way, any other problems to electronic
equipement arround ???

Bye.

 

[techie]

No damage to other electronic equipment around was noticed. Even a very high surge on power lines cannot blast the IC's. It had to come somewhere from the data lines.

 

[WA]

techie

The power input is from a 12V adapter. Here the circuit is fine. The input has a 1N4007 diode which is not damaged. Then there is a 1000uF/25 cap which is still OK. Then is the 7805 which is damged. All the ICs are on 5V and aslo burst.

Now suspicions comfirmed.

inputs not problem. you say 7805?

7805 blow up because of something else, I notice two smaller capacitors(ceramic), are they on 5V line?

I say this because (my rule of thumb) flat packs don't "usually" go. associated circutry with flat pack take them out.

Again I don't think external source is cause, I think something internally just quit. Unfortunatelly, nothing work forever and if this is the case, nothing could have prevented this from happening.

Again this is just my experience reguarding failures like this.
WA

 

[jzaghal]

Surge again !!

Hi,

You see even if all the dc at the input of the regualator was present
at it's output will not make this mess !!

Also cannot be from the data cables connected to the unit.

I have seen bad situations like this due to surges, and transformers
were ok.

Bye.

 

[techie]

the 2 small ceramics are 33pf that are with the Xtal of oscillator. The ceramics on 5v.... wait I will check them again tomorrow.... strangely they look undamaged. They are 50v rating so maybe the surge was less than that.

 

[jzaghal]

Re Ceramic capacitors !!

Hi,

You see these cannot blow, because they are not connected to the power supply.

However any semiconductor component connected to the power supply is gone !!

Bye.

 

[techie]

the ceramics are connected directly accross 5v and GND. Why cant they blow

 

[sparkytron]

Is there any way that the power supply voltage as seen by the semiconductors reversed polarity?

 

[jzaghal]

Re Ceramic capacitors !!

You see I said semiconductors sharing the supply have blown.
Semiconductors are very sensitive devices, remember that
you can blow up a cmos ic and not a capacitor or resistor.
Don't forget also that small capacitors in the picofarad or micro-
farad range have a high impedance at a surge tthat is very fast,
hence they pass very little current throu them, and they don't
blow, whereas the semiconductors retain their low resistance and
cannot change it like a capacitor or inductor.

Hope you got it !!