dementeddigital
Active Member
rockinarmadillo said:
Yikes! You had a solid hit!
Automation system hit by lightning
- Thread starter kurtmccaslin
- Start date
Yikes! You had a solid hit!
You can take a West post with a grain of salt.
The Ditek clamps to ground before the voltage hits a damaging level and quickly enough (hopefully). That's the nature of it's protection. It won't solve a grounding issue, which is what needs to be addressed first.
Welded contacts are just an indication of either where the transient came in or where the cables were physically landed on the expander and how far the damage traveled on the board before the board itself fried and broke the connectivity or the transient was dissipated enough. It's well known that reed contacts have a tendency to weld if their operating voltage or amperage are exceeded. Known evil in the industry unless people move to methods like magnasphere or others.
In your case, I'd put suppression on each end of the bus and a short path to ground at each end and more distance in cable length than the ground wire. Standard best practices.
As far as your install, are all the cables that feed the west side of your structure landed on the same section of the expansion? Possibly bundled together?
The Ditek clamps to ground before the voltage hits a damaging level and quickly enough (hopefully). That's the nature of it's protection. It won't solve a grounding issue, which is what needs to be addressed first.
Welded contacts are just an indication of either where the transient came in or where the cables were physically landed on the expander and how far the damage traveled on the board before the board itself fried and broke the connectivity or the transient was dissipated enough. It's well known that reed contacts have a tendency to weld if their operating voltage or amperage are exceeded. Known evil in the industry unless people move to methods like magnasphere or others.
In your case, I'd put suppression on each end of the bus and a short path to ground at each end and more distance in cable length than the ground wire. Standard best practices.
As far as your install, are all the cables that feed the west side of your structure landed on the same section of the expansion? Possibly bundled together?
kurtmccaslin
Active Member
DELInstallations said:
Yes and no. Some were in the bundle and others were not. I lost virtually every electronic item on the west side of the garage. And a few items on the other side. The items outside of the west side of the garage were connected to that area through my network, Omni pro data bus, or water pipes.
kurtmccaslin
Active Member
A lot of great info on surge protection on this board lately! Thanks everyone.
Here is an update:
I was able to figure out that the lightning surge on Monday hit my irrigation system. Through the Hunter controller, (but it still works?). Then it fried the GFI that the controller was plugged into. It damaged a UPB switch on the same circuit, And then blew the breaker for this circuit. It seems that the surge protector at the panel took enough current to ground to blow the breaker. However, the lights on the protector show it is still functional.
Here is an update:
I was able to figure out that the lightning surge on Monday hit my irrigation system. Through the Hunter controller, (but it still works?). Then it fried the GFI that the controller was plugged into. It damaged a UPB switch on the same circuit, And then blew the breaker for this circuit. It seems that the surge protector at the panel took enough current to ground to blow the breaker. However, the lights on the protector show it is still functional.
DELInstallations said:
Clamping voltage and response times are irrelevant. Clamping voltage and response times for all effective protectors are more than sufficient. Protection is defined by an impedance to and quality of earth ground.
Better would have been a protector that connects each "low voltage wire to sprinklers" to single point earth ground. Then a direct strike to sprinklers need not connect to an earth ground electrode destructively via a sprinkler controller.
Protection is always about the path a surge current takes to earth ground. And about making that earth ground a best and most desirable path. Then current need not pass through (and damage) any other hardware.
dementeddigital
Active Member
I'm really tired. It's been a long day. This may or may not be coherent, but I wanted to respond to this tonight.westom said:
The clamping voltage and response time are very relevant. If a circuit takes a differential mode surge across two wires (ignoring other possible paths for the moment), the clamping voltage of the transient protection device across them will dictate the voltage that circuit will see. Response time is important, as the edge rate of the transient might be faster than the transient protection can start to conduct. After that, the amount of energy that the transient protection device can absorb will determine if the protection device will survive the event or not. If it doesn't the circuit which it protects may also not survive.
I've designed circuits where the input just opens (a transistor turns off) when the voltage rises above a threshold. You don't always need clamping or a path for protection. It depends on the application and the transient. I'm not sure if such a circuit would be suitable for a good lightning hit, but it could be used in conjunction with other methods of protection. (As I do for protection on the front-end of automotive circuits.)
When you look at these things, you need to look at both the current paths and the voltages across elements in the system. (Current doesn't flow without voltage...) It doesn't do you any good to divert current to ground if the voltage levels during the transient are going to rise above the breakdown voltage of the other devices in the circuit.
There certainly needs to be a good and single path, and then you need to decide to what voltage levels the I/O and power need to stay below. Then you can decide what you want those lines to do in the case of a transient - open, absorb the transient, or try to divert the energy somewhere else. You need to know if the transient is differential or common mode, too. These things will dictate how you place the transient protection in the circuit. You typically see combinations of protection in a few configurations to protect against a few types of transients.
Hopefully that made some sense. If I remember, I'll come back and read this gibberish in the morning over some coffee and have a good laugh at what I just wrote.
dementeddigital said:
First, transients that overwhelm existing internal protection are longitudinal; not differential mode. Protection from differential mode already exists; is quite robust.
Second, a typically destructive surge with a fastest edge rate is the classic 8/20 microsecond transient. Rising edge is microseconds. Effective protectors respond in nanoseconds.
Yes, response time is important. View numbers. Nanosecond protector responds more than fast enough even for the fastest (microsecond) surge. Once we include numbers, protector response time is completely irrelevant.
View other relevant numbers. Microsecond transients means impedance (not resistance) is critical. Effective protector means hundreds of thousands of joules must connect to and dissipate harmlessly outside in earth. A hardwire to earth ground must be short (ie less than 10 feet). No sharp bends. Response time is why low impedance is relevant. Connection to earth ground must be low impedance (ie not inside metallic conduit). That (and not a protectors response time) is critical.
Third, destructive surges are current sources. Voltage is a dependent variable. Independent variable (what is relevant) is current. A properly earthed 'whole house' protector must conduct an entire surge (ie a direct lightning strike) and not fail. Since lightning is typically 20,000 amps, then a minimally sufficient 'whole house' protector is 50,000 amps. Another relevant number once recommendations are tempered by numbers.
Effective protectors discuss how many amps will connect to earth. Ineffective protectors are rated by how much energy it will foolishly 'block' or 'absorb'.
So what is a worst and best solution? Undersized protector fails as indicated by its light. How many joules does it claim to absorb? Hundreds? How many joules must it somehow 'block' or 'absorb'? Hundreds of thousands. Protectors that 'block' or 'absorb' a surge means a surge voltage increases. Failure due to undersizing means a 120 volt protector exceeds 900 volts. That higher voltage is due to too many amps being 'absorbed' or 'blocked' by a plug-in protector. But failure gets many to recommend that ineffective protector. "My protector sacrificed itself to save my computer." Total bull exposed by numbers.
Excessive impedance means a plug-in protector has no hardwire to earth ground. Response time means impedance of a hardwire to earth should have full attention.
Informed consumers earth one 'whole house' protector to have a 'secondary' protection layer. Then inspect their 'primary' protection layer also defined by an earth ground on a utility pole. Every protection layer is defined only by earth ground.
Concern for voltage and response time is why effective protectors must connect low impedance (ie less than 10 feet) to single point earth ground. A protector is only as effective as its earth ground. Not any ground. Single point earth ground. Then robust protection already inside all appliances is not overwhelmed.
dementeddigital
Active Member
Westom, from your comments, it seems that you're ONLY talking about whole-house transient protection. If that is the case, then much of what you say is reasonable.
While whole-house AC mains protection is important, transient protection encompasses much more than just protecting the power mains. For example, lightning can induce common-mode voltages on zone wiring. If you are a ham, HF transmissions can sneak into many places they shouldn't and cause unwanted behavior, even at modest power levels. When wiring the zones, the panel is potentially exposed to ESD (which, incidentally, has nanosecond rise times.) It doesn't really matter to me if I have an effective whole-house surge protection device on my AC mains, if I lose electronics or have them misbehave due to transients via some other path.
So transient protection is not only lightning protection. Different (but related) strategies are used to mitigate various types of EMI. Filtering and shielding are generally used for RF immunity. Transzorbs, spark gaps, and MOVs are used to absorb or divert transients. ESD can be mitigated with spark gaps, small caps, and diode clamps on the PCB. Isolation can be used to break paths, also.
Protection from differential transient already exists where? For example, on the HAI zone expansion, the only differential protection I see is a spark gap on the PCB, and some RC filtering. Since most of the transients on zone wiring are common-mode, that's probably sufficient, but it does very little for differential noise. You'd be surprised at how little protection there is in many appliances. "Robust" might be a bit of a stretch.
Not sure what you mean by, "Failure due to undersizing means a 120 volt protector exceeds 900 volts. That higher voltage is due to too many amps being 'absorbed' or 'blocked' by a plug-in protector."
Most AC protection is MOV-based (or a combination MOV and spark gap). The MOV will begin conducting at a certain voltage, and thus clamp the voltage to a certain level. A MOV line-to-ground will allow the voltage on the line to rise to a certain point, and then begin conducting current to ground. Undersizing a device isn't going to make the voltage increase during a transient. Undersizing it will cause it to exceed its power dissipation capability during the transient and then degrade or completely fail shorted or open. This is why "amps" isn't a useful number in transient protection - joules is more useful. Anything can conduct 20,000 amps for a femtosecond.
In any case, good discussion on transients and transient protection. Useful stuff.
While whole-house AC mains protection is important, transient protection encompasses much more than just protecting the power mains. For example, lightning can induce common-mode voltages on zone wiring. If you are a ham, HF transmissions can sneak into many places they shouldn't and cause unwanted behavior, even at modest power levels. When wiring the zones, the panel is potentially exposed to ESD (which, incidentally, has nanosecond rise times.) It doesn't really matter to me if I have an effective whole-house surge protection device on my AC mains, if I lose electronics or have them misbehave due to transients via some other path.
So transient protection is not only lightning protection. Different (but related) strategies are used to mitigate various types of EMI. Filtering and shielding are generally used for RF immunity. Transzorbs, spark gaps, and MOVs are used to absorb or divert transients. ESD can be mitigated with spark gaps, small caps, and diode clamps on the PCB. Isolation can be used to break paths, also.
Protection from differential transient already exists where? For example, on the HAI zone expansion, the only differential protection I see is a spark gap on the PCB, and some RC filtering. Since most of the transients on zone wiring are common-mode, that's probably sufficient, but it does very little for differential noise. You'd be surprised at how little protection there is in many appliances. "Robust" might be a bit of a stretch.
Not sure what you mean by, "Failure due to undersizing means a 120 volt protector exceeds 900 volts. That higher voltage is due to too many amps being 'absorbed' or 'blocked' by a plug-in protector."
Most AC protection is MOV-based (or a combination MOV and spark gap). The MOV will begin conducting at a certain voltage, and thus clamp the voltage to a certain level. A MOV line-to-ground will allow the voltage on the line to rise to a certain point, and then begin conducting current to ground. Undersizing a device isn't going to make the voltage increase during a transient. Undersizing it will cause it to exceed its power dissipation capability during the transient and then degrade or completely fail shorted or open. This is why "amps" isn't a useful number in transient protection - joules is more useful. Anything can conduct 20,000 amps for a femtosecond.
In any case, good discussion on transients and transient protection. Useful stuff.
I can only speak for myself but this is all no help to me and may as well all be in a foreign language. I'm sure that a surge device's effectiveness can be expressed in simple, legible English, no?
" Undersized protector fails as indicated by its light." - what does that even mean?
On my outdoor cable I am happy to use a clamping voltage that is a percentage above the expected circuit voltage and has been used successfully in the past. Ditek has devices that they recommend for rs-485 and IP cables and that is good enough for me.
Mike.
" Undersized protector fails as indicated by its light." - what does that even mean?
On my outdoor cable I am happy to use a clamping voltage that is a percentage above the expected circuit voltage and has been used successfully in the past. Ditek has devices that they recommend for rs-485 and IP cables and that is good enough for me.
Mike.
dementeddigital
Active Member
It's actually not as bad as you'd think. The math isn't complicated.mikefamig said:
A joule is a watt-second. If a device is rated for 1 joule, then it can dissipate 1 watt of power for 1 second before exceeding its capability.
Watts are power, and can be calculated one way as volts x current. If the 1 joule device above clamps the voltage to 5V, then it can conduct 200 mA for one second because 1 watt / 5 v = 0.2 A. If the transient you need to protect against lasts 0.1 seconds, the same 1 joule device can dissipate 10 watts of power for that time, and if it clamps at 5V it will conduct 2A for that 0.1 seconds before it exceeds its capability.
Ditek is a great company, and their recommendations are likely good enough. If you want to dig deeper, the math isn't bad. I'd recommend putting the calculations into Excel, and that will allow you to play with voltage levels, joule ratings, and time to see if the protection is sufficient.
NeverDie
Senior Member
A lot of people advocate for running home automation software on a vmware platform, and I thought the arguments were pretty good. However, it was the lightning risk that tilted my decision to put it on a cheap dedicated platform instead. The reason is simple: when lightning is on the way, I unplug and disconnect the high end equipment. The storms usually don't last all that long anyway. I still want access to lighting controls, though, so I leave the dedicated automation platform connecedt and powered on, albeit with a surge protector and a UPS. If the low-end platform gets fried, at least the replacement cost is low. I'd feel much worse if a vmware megabox got hit.
