-
3
- #1
For about 40 years, I have been working with energy management and other control systems where control wires and cables must be run underground for short and long distances. Lightning and transient suppression on such lines has always been a problem, and no low cost solution will protect such wiring from a direct lightning strike. However, there are a number of economical options for protecting wiring and signalling from all but very nearly direct strikes. Here are some techniques to try.
1) Unprotected CMOS 5V logic inputs are a problem and are REALLY NOT a good idea to <ever> use. However, if you MUST use these inputs, from the buried cable use a 100 ohm 2W metal film resistor to a PARALLEL PAIR of Vishay P6KE7.5CA transient protection diodes (one side grounded) to a 5K, 1/4W resistor then to the logic input. This has never (yet) failed on me from a lighting strike. A 5K or so pullup from the ungrounded side of the PAIR of Vishay P6KE7.5CA diodes to a 5v PSU is needed on this input as a pullup. I have had to use this technique a few times, but I do not recommend it. And.. Such underground positive voltage logic wiring can fail from corrosion if water gets into the buried copper wire from any sort of puncture. Note also that you can/will get unwanted logic transitions from outside transients and your system must make allowances for such spurious signals.
2) Unprotected 24VAC control wiring such as remote relay actuation. I have had success with using a 1N6289 (series 1N6267 thru 1N6303A) suppressors on both ends of a control wire to ground. These small devices will absorb about 1.5KW peak current transients. This is good enough for all but a "direct strike" on wiring such as 18ga to 22ga. I have had but a few failures of this device.
3) Ethernet, RS422, and RS485 cables. I have used the APC model PNET1 lightning arrestor successfully. However, I HAVE had a few failures of these devices, but so far, none of the protected equipment has been damaged. A protector is required on BOTH ends of any such circuit which runs underground even for a short distance. My longest underground ethernet run so far (10mbps) has been about 300ft. On longer runs, be sure to use individually shielded pairs and ground the shield on both ends. This helps a lot to keep down large voltage transients.
4) In my Building Automation Equipment work, I frequently have a few low voltage control wires going to outside switches used for such things as turning lights on and off. The signal inputs of the building automation equipment I use is
a) opto-isolated and rated at 1KV breakdown. I have had many of these inputs to fail from transient voltage surges when tied directly to buried cable conductors. The underground cables are usually from 7 to 25 conductor 18ga and rated for direct burial. However, I have found that if these 28vdc (idle), 0vdc (activated) signals are run underground for many hundreds or thousands of feet from the controller to a remote pole switch, a wire WILL occasionally fail due to corrosion where water somehow gets into the cable over 5 to 20 years. I developed a circuit which has (so far) taken care of this problem. Note: I never use 24ga or 26ga wires underground in industrial environments. You are just asking for corrosion problems if you do. 16ga, 18ga, or 20ga have much longer life should corrosion problems happen.
See attached schematic.
This circuit includes a) Four logic (switch) inputs which inputs are normally biased NEGATIVE so as to inhibit corrosion on any external wiring which may be in contact with water. (Note: Corrosion to buried copper conductors does not occur when the wire is negative with respect to earth ground.) When one of the four inputs on P2 is grounded at the distant end, the corresponding output on P4 is pulled low by the 2N7000 transistor. The 22K pull down resistor to the negative voltage keeps the wire at a voltage in the range of -30vdc when idle. Your design must still contend with spurious logic transitions from external transients. This problem cannot be overcome in a simple circuit such as this.
b) Included on the board is a two wire surge protector for a RS422/485 circuit. This filter is suitable for RS422 and RS485 circuits and inserts a 27 ohm, 2W resistor in series with the input from the buried cable. So far, I have not had a failure of this circuit or the associated RS485/422 devices in about 15 years use of this circuit.
So.. There really ARE practical and low cost means for protecting signal circuits from voltage surges from lightning and other transient voltages. None is 100% good in a “worst case” scenerio. But I have found the above methodology to be both practical and almost troublefree over a long period of time in a lot of different applications
thread236-328817
1) Unprotected CMOS 5V logic inputs are a problem and are REALLY NOT a good idea to <ever> use. However, if you MUST use these inputs, from the buried cable use a 100 ohm 2W metal film resistor to a PARALLEL PAIR of Vishay P6KE7.5CA transient protection diodes (one side grounded) to a 5K, 1/4W resistor then to the logic input. This has never (yet) failed on me from a lighting strike. A 5K or so pullup from the ungrounded side of the PAIR of Vishay P6KE7.5CA diodes to a 5v PSU is needed on this input as a pullup. I have had to use this technique a few times, but I do not recommend it. And.. Such underground positive voltage logic wiring can fail from corrosion if water gets into the buried copper wire from any sort of puncture. Note also that you can/will get unwanted logic transitions from outside transients and your system must make allowances for such spurious signals.
2) Unprotected 24VAC control wiring such as remote relay actuation. I have had success with using a 1N6289 (series 1N6267 thru 1N6303A) suppressors on both ends of a control wire to ground. These small devices will absorb about 1.5KW peak current transients. This is good enough for all but a "direct strike" on wiring such as 18ga to 22ga. I have had but a few failures of this device.
3) Ethernet, RS422, and RS485 cables. I have used the APC model PNET1 lightning arrestor successfully. However, I HAVE had a few failures of these devices, but so far, none of the protected equipment has been damaged. A protector is required on BOTH ends of any such circuit which runs underground even for a short distance. My longest underground ethernet run so far (10mbps) has been about 300ft. On longer runs, be sure to use individually shielded pairs and ground the shield on both ends. This helps a lot to keep down large voltage transients.
4) In my Building Automation Equipment work, I frequently have a few low voltage control wires going to outside switches used for such things as turning lights on and off. The signal inputs of the building automation equipment I use is
a) opto-isolated and rated at 1KV breakdown. I have had many of these inputs to fail from transient voltage surges when tied directly to buried cable conductors. The underground cables are usually from 7 to 25 conductor 18ga and rated for direct burial. However, I have found that if these 28vdc (idle), 0vdc (activated) signals are run underground for many hundreds or thousands of feet from the controller to a remote pole switch, a wire WILL occasionally fail due to corrosion where water somehow gets into the cable over 5 to 20 years. I developed a circuit which has (so far) taken care of this problem. Note: I never use 24ga or 26ga wires underground in industrial environments. You are just asking for corrosion problems if you do. 16ga, 18ga, or 20ga have much longer life should corrosion problems happen.
See attached schematic.
This circuit includes a) Four logic (switch) inputs which inputs are normally biased NEGATIVE so as to inhibit corrosion on any external wiring which may be in contact with water. (Note: Corrosion to buried copper conductors does not occur when the wire is negative with respect to earth ground.) When one of the four inputs on P2 is grounded at the distant end, the corresponding output on P4 is pulled low by the 2N7000 transistor. The 22K pull down resistor to the negative voltage keeps the wire at a voltage in the range of -30vdc when idle. Your design must still contend with spurious logic transitions from external transients. This problem cannot be overcome in a simple circuit such as this.
b) Included on the board is a two wire surge protector for a RS422/485 circuit. This filter is suitable for RS422 and RS485 circuits and inserts a 27 ohm, 2W resistor in series with the input from the buried cable. So far, I have not had a failure of this circuit or the associated RS485/422 devices in about 15 years use of this circuit.
So.. There really ARE practical and low cost means for protecting signal circuits from voltage surges from lightning and other transient voltages. None is 100% good in a “worst case” scenerio. But I have found the above methodology to be both practical and almost troublefree over a long period of time in a lot of different applications
thread236-328817
