1
\$\begingroup\$

TL;DR If you wish to use DC surge protection products such as these (datasheet), do you need to bond the DC supply to ground? When they trigger, where does the excess current in a surge actually go?

Example of a 2-pole surge protector

I have a remote property I am planning to set up with a 48 V battery system being charged via solar, powering some cameras and an Internet connection. There is no mains power for many miles/kilometres and I won't need to run an inverter, so this is a purely DC system. The site is on the top of the tallest hill in the area, and it has an existing antenna mast that is the highest point in the area. So I imagine it's going to get hit by lightning from time to time.

I am therefore looking at different types of lightning and surge protection devices, and from the surge protection perspective these SPDs shown in the pictures above and below seem to be worth having. They come in many variants including for 48 VDC.

However as I understand it, they are essentially MOVs connected between the DC conductors and PE, which would mean any surge would have to have a path from the DC conductors to ground - otherwise if the DC supply was isolated, there would never be a voltage difference between any given conductor and ground, so the surge current would never flow into PE.

Is my understanding correct, or have I missed something? It would seem that I would need to bond one of the DC conductors to PE in order to guarantee the surge protector will pass current when the voltage goes too high. Although in my case I would be bonding it to the battery bank's negative terminal, and I don't necessarily want high voltage spikes dumped into the batteries (although perhaps they would desulphate the batteries for me!)

So I'm a bit confused about how various surges in the system get diverted and where the excess current goes. Like a direct lightning strike going to ground makes sense, but if that strike is nearby and induces a higher voltage in the DC conductors, where does that go? I guess your only option is to short it out and burn it off as heat until the voltage returns to normal?

There are also 3-pole versions of these devices (datasheet), however I am not sure what the difference is between the two and three pole variants. The ratings are all the same, but the three-pole version seems to put an extra MOV on the PE (so it's effectively in series with the other MOVs) so I'm not sure what the purpose is of that?

Example of a 3-pole surge protector

\$\endgroup\$
5
  • 1
    \$\begingroup\$ Please edit the question to include a link to the manufacturers datasheet for the DC surge protection product. \$\endgroup\$ Jan 6 at 9:16
  • \$\begingroup\$ Consider your wires being the secondary of a transformer, the primary being the nearby lightning strike, the secondary voltage will rise until the insulation breaks over. The MOV limits this voltage and burns the energy off as heat. That’s why the protectors have an energy rating. Ideally, you want protection on both ends of the cable and a heavy, common PE between the two ends. The energy then gets routed into the loop formed by the PE and the conductor completed by the MOVs. \$\endgroup\$
    – Kartman
    Jan 6 at 10:01
  • \$\begingroup\$ @Kartman: So let's say we have a 48 VDC battery powering a lamp, and we inject 200 VDC onto the positive conductor, we can simulate a surge. If we add SPDs/MOVs at the lamp end and the battery end, they will start to conduct from the positive rail to PE, quickly raising PE to 200V at both ends of the cable. If the PE is connected to ground but not bonded to the supply, that 200 volts won't go anywhere (no current will flow so the voltage will not drop) causing the voltage spike to remain? \$\endgroup\$
    – Malvineous
    Jan 7 at 7:13
  • \$\begingroup\$ @ChesterGillon: I have added links to the datasheets as requested - my apologies, I thought these were well-known devices and I was only asking in the general sense. \$\endgroup\$
    – Malvineous
    Jan 7 at 7:14
  • \$\begingroup\$ @Malvineous it pays to draw a diagram to be clear what is being injected where and how. Along with the schematic of the SPDs, it is then simple to propose a transient situation and then analyse the voltages, current flows etc. whilst it is difficult to fully quantify the effect of a nearby lightning strike(direct strikes usually cause significant physical damage) being able to analyse possible scenarios gives a bit more confidence in a given solution. \$\endgroup\$
    – Kartman
    Jan 7 at 13:37

1 Answer 1

1
\$\begingroup\$

Does a 48 VDC system need to be grounded for lightning protection?

There may be specific requirements in your jurisdiction.

Other than that, the temporary grounding path is provided by the lightning protector device. Otherwise, you could put the an overvoltage protector between (+) and (-), as the DC source would be already ground-referenced.

The lightning protector maintains isolation of the DC circuit from the ground, and only establishes the connection when an overvoltage to ground exists.

The lightning protector may also provide a lower voltage rated protection between the (+) and (-), but that's mainly to help the circuit survive, and not needed for electrical safety per se. Of course the circuit not blowing up may improve some aspect of safety.

When they trigger, where does the excess current in a surge actually go?

To ground. And it's not really "excess" current, it's just current, because until the protector triggers, there is no current yet. Just a voltage going up, potentially very quickly.

The whole point of those protectors is to keep the circuit at a low enough voltage from the ground that insulation systems elsewhere in the circuit won't break down to ground. Especially if a person's body is on the path between the circuit and the ground!

It is imperative to provide a low impedance path from the PE terminal of the protector to the building/structure earthing system, or to a dedicated grounding rod in absence of the former.

For example, suppose this protection is used at the entrance of a 48V supply line to a wooden or masonry shed, with no other electrical system in the shed. The protector's PE terminal would need to be connected to an earthing rod. Otherwise, the PE terminal could be connected to the earth/grounding bar of the electrical panel in the shed. If the shed has a metal frame/structure, then the metal should already be properly grounded using grounding rod(s) per local requirements and good engineering practice. The overvoltage protector could be connected to the nearest grounding point where the structure is grounded. It's not unusual for structures to have redundant grounding, so there may be two or more grounding points that could be used.

This is all rather general. Your jurisdiction will have minimum requirements mandated by building codes, and good engineering practice may dictate going above these minimal requirements.

Personal note: Un-earthed DIY-made (or clueless-contractor-made) steel or aluminum structures make the hair stand on the back of my neck. When a thunderstorm is approaching, this may be quite a literal and involuntary reaction, too.


If the DC side is electrically isolated then it seems an "overvoltage to ground" state could never exist, except perhaps in the case of a direct lightning strike?

Paraphrasing a bit, that's like the whole point of this :) Floating DC circuits can get the equivalent of "lightning" from electrostatically charged people and things, from energized conductors coming in contact with the circuit inadvertently, from failures of devices attached to the circuit, from nearby lightning strikes (rather than direct), from electrostatic induction in high potential gradient fields that form near ground with charged clouds overhead, etc.

presumably this is only of any use if the DC system is already referenced to ground

If the DC system is already bonded to ground, there's no need for additional lightning protection to ground, only between points within the circuit itself.

"Ground referenced" and "reliably connected to protective earth" are two different things. A 48V lighting circuit connected through a puny little wire to ground may be ground referenced, but the "reference" connection will act as a fuse and unprotect the circuit as soon as anything interesting happens where the circuit would actually need real protection.

If the DC side is electrically isolated then it seems an "overvoltage to ground" state could never exist

Have you ever stepped out of a car, touched the chassis, and got zapped? That's how a capacitor works. There is insulator between the plates, and the voltage between the plates can be pretty damn high until something breaks down. You could even say that isolation in that case is the whole reason for there being overvoltage to ground in the first place. Otherwise, the circuit would be near ground potential already!

In our case, we'd like that some protector breaks down before we do :)

\$\endgroup\$
0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.