I still smell the burned varnish from a test we ran last year—one 6 kV strike and the dummy board turned black in half a second.
A surge protector works by grabbing extra energy and pushing it to ground, then it clamps the voltage below the level that can hurt your machines. I build these units every day in Wenzhou and test them to IEC 61643-11.
If you know how the trick is done, you can pick the right part and stop paying for specs you never use. Keep reading and I will show you the guts of the device.
Core goals: energy transfer and voltage clamping?
I once watched a 40 kA surge miss a drive by one micro-second because the MOV clicked in time—that tiny disk saved a $12,000 inverter.
The two core goals are: (1) move the surge energy to ground fast, and (2) keep the voltage that reaches the load under the safe limit written on the data sheet.
How Energy Moves Inside the Box
A surge arrives on the line. The MOV impedance drops from mega-ohms to ohms in nano-seconds. Current takes the easy path through the device, then runs down the green-yellow earth wire. The hotter the wire, the lower its impedance, so we use 6 mm² Cu and keep the lead under 50 cm. Any extra length adds 1 µH of inductance and that adds 1 kV to the let-through voltage. Customers forget this detail and blame the part when the board still dies.
Clamping Voltage vs Let-Through Voltage
People mix the two numbers. Clamping voltage is what the MOV sees. Let-through voltage is what the load sees after cable drop. I always list both on my test sheet. A part that clamps at 700 V can still let 1,200 V reach the VFD if the earth tail is 80 cm. Cut the tail, cut the pain.
Real Data from Our Lab
|
Surge Level |
MOV Size |
Earth Lead |
Let-Through |
Result |
|
20 kA 8/20 µs |
32 mm disc |
25 cm |
980 V |
PASS |
|
20 kA 8/20 µs |
32 mm disc |
80 cm |
1,450 V |
FAIL |
|
40 kA 8/20 µs |
40 mm disc |
25 cm |
1,050 V |
PASS |
The table shows that cable length beats MOV size. I tell every buyer: spend one extra dollar on short leads before you spend five on a bigger part.
Why We Add a Gas Discharge Tube in Hybrid Designs
An MOV wears out after big hits. A GDT can take more shots but is slow. We put them in parallel. The MOV starts first and clamps for the first 100 ns. Then the GDT fires and takes the bulk current. The MOV rests and lives longer. Hybrid is now our best-seller to German solar farms because the site crew wants a 20-year life, not five.
Core components and hierarchical protection mechanisms?
I open one of our Type 1+2 units and I see MOVs, GDTs, fuses, and a tiny thermal switch that clicks like a kettle when it is tired.
The core parts are: (A) varistors or GDTs that eat energy, (B) thermal disconnects that stop fires, and (C) backup fuses that clear short circuits. We stack these in three layers to match the wiring system in a plant.
Layer One: Type 1 at the Service Door
This part sees direct lightning. We use a 25 kA 10/350 µs impulse tube plus a 50 kA MOV block. The goal is to drop the strike from 1,000 kV to under 4 kV before it enters the switchboard. We mount it on a 35 mm DIN rail and bond it with 16 mm² Cu to the main earth bar. One bolt hole in the wrong place adds 2 µH and 2 kV extra. I check the drawing twice; the buyer saves a fried transformer.
Layer Two: Type 2 at Sub-Panels
This layer stops induced surges from nearby strikes or big motor switching. We pick 40 kA 8/20 µs MOVs with thermal disconnect. The part plugs in so the user can swap it without killing power. We add a green LED that goes off when the part is dead. A site manager in Milan told me he can check 50 panels in ten minutes just by walking the aisle and counting green dots.
Layer Three: Type 3 at the Load
Drives, PLCs and PCs need a local guard. We use 10 kA 8/20 µs units with let-through under 900 V. The part fits in a wall box or inside the socket strip. Cable from Type 2 to the load must stay under 10 m. If the run is longer, we add another Type 3. I once saved a $4,000 servo by adding a $9 socket SPD because the panel was 30 m away.
How the Layers Talk to Each Other
Energy is like water. If the first dam is full, the second dam must be ready. We set the voltage levels in steps: Type 1 clamps at 1.8 kV, Type 2 at 1.4 kV, Type 3 at 0.9 kV. The lower layer never starts before the upper layer, so each part shares the load. We test the full chain in our lab with three units in series and a 100 kA strike. The let-through at the end socket is 720 V, safe for any 230 V drive.
Parts List We Use Every Day
|
Part |
Role |
Spec |
Life Cycles |
|
40 mm MOV |
Clamp |
40 kA 8/20 µs |
20 big hits |
|
Thermal switch |
Fire stop |
120 °C |
One-shot |
|
6 A gG fuse |
Short clear |
50 kA breaking |
One-shot |
|
GDT tube |
Backup |
600 V spark |
100 hits |
|
LED + resistor |
Status |
2 mA drain |
10 years |
Collaboration and safety backup?
I still recall the day a thermal fuse popped and the red flag told the tech to swap the unit—no drama, no fire, just a five-minute break.
An SPD must work with breakers, earthing and cable routing. We add thermal fuses, micro-switches and remote signals so the site team knows when the part is tired and safe backup takes over.
Why an SPD Needs the Breaker as a Friend
An MOV can short-circuit when it dies. The backup fuse must clear the fault before the panel burns. We match the fuse curve to the MOV fault current. A 40 kA MOV fails at 1 kA short. We pick a 6 A gG fuse that clears in 0.1 s at 1 kA. The fuse never blows on normal surge current because that lasts micro-seconds. The math is tight, but it works. I give buyers a fuse chart so their electrician does not guess.
Remote Signaling for Big Sites
One client runs 24/7 glass furnaces. He cannot walk the plant each week. We add a micro-switch inside the SPD that flips when the thermal disc opens. The switch feeds a 24 V PLC input. A red lamp on the HMI says “SPD dead.” The operator calls us, we ship a spare cartridge, and he swaps it at the next shift change. Zero unplanned stops in two years.
Coordination with RCDs and Arc Detectors
Some engineers fear that SPD leakage will trip an RCD. We keep leakage under 0.3 mA at 230 V. A 30 mA RCD never sees it. If the site uses arc detectors, we add an EMI filter in front of the SPD so the high-frequency clamping does not fool the detector. We tested this mix at TÜV Rheinland and passed.
Key Performance Indicators?
I track three numbers on every shipment: let-through voltage, failure rate per 1,000 pcs, and swap time on site. If any drifts, I stop the line.
The top KPIs are: (1) voltage protection level (Up) measured in lab, (2) surge life count before wear-out, and (3) mean time to replace (MTTR) on live systems. I log these for every batch we sell.
Why Let-Through Is King
A 200 V drop in Up can double the life of a drive. We test every MOV disc at 100 % current and log the voltage. Discs that read high go to the solar farm line where clamping is less critical. Discs that read low go to the German PLC line. This sort adds one hour to production but cuts field faults by 40 %. I pay the hour, I save the night call.
Life Count Test We Run
We hit the same part with 20 kA every five minutes until the thermal switch pops. The record holder lasted 27 shots. We publish the curve on the data sheet. Buyers see that the part still works after ten years of normal surges. That single graph closes more deals than my best price cut.
Conclusion
Energy transfer, clamping, layers, backup and clear KPIs—that is the whole story. Pick an SPD that scores low on let-through and low on return rate, and you buy sleep.