Hedges
I See Electromagnetic Fields!
- Joined
- Mar 28, 2020
- Messages
- 20,906
Others have observed same, even got a shock from it.
The architecture may have a high voltage rail that PV feeds and inverter operates from. Slight leakage resistance could present high voltage at the terminals but hopefully low current.
If you connected an incandescent light bulb from the terminal to ground, you might find it pulls voltage to zero (because very little current is sourced.)
When working on it, need to shut off inverter, ideally disconnect all power sources (battery, PV, grid or generator) and measure voltage, wait until it has discharged before touching anything.
I wondered if SPD could see PV negative biased up to that 189 VDC and PV positive biased to 189 VDC + PV voltage, which could be excessive relative to ground. But I don't think circuit adding those voltages together makes sense. Still, you never know, which is why I suggested connecting (replacement) parts with SPD not included and check voltages.
But ... the failure happened at night, so no PV voltage present! Only source of energy to blow up SPD and start a fire is battery and inverter? Or is grid present?
An SPD can take spikes, meant to absorb and dissipate brief transients. Limited number of Joules (watt seconds). If over-stressed, or maybe just after many hits, they can be damaged and short out. It is possible that inductive switching transients send spikes down the line. Your MPPT would not have been operating at night (I don't think) but inverter could have some coupling to it. Have access to an oscilloscope to look at those terminals? MPPT has inductor, inverter has inductor, and switching circuit will cause voltage transients. If they spike high they could keep hitting the SPD.
I finally stopped procrastinating and ordered a quantity of the following.
My grid AC input has a large SPD system. PV input, some model inverter had them but other similar models did not. I will connect three of these MOV in a delta configuration and put wires into PV+, PV-, Ground terminals for each inverter.
This MOV includes a thermal disconnect (maybe PTC fuse?) and they caution against overheating during soldering. They have a lead connected between MOV and PTC, which could be connected to LED indicators, providing "Protected" and "Failed" indicators like the Midnight SPD have. That is better because you can observe LED color at a glance. I'll just leave them accessible to connect DMM test leads. I don't expect a near lightning strike, not so common in my area. But maybe the next Carrington event creating a voltage across the loop formed by my PV string, it will protect against.
Data sheet says, "Notes: Average power dissipation of transients should not exceed 1.5 watts."
In other words, it can dissipate 1.5W as heat. Deposited energy (Joules) has to be infrequent enough to allow cooling off. If a switching supply operating at a few hundred Hz or kHz keep hitting it with spikes, it could gradually heat up then fail.
Also, can take 20,000A but graphs show a single 20,000A transient lasting 20 microseconds, or fifteen 10,000A transients lasting 20 microseconds is all it can take before wear-out.
The architecture may have a high voltage rail that PV feeds and inverter operates from. Slight leakage resistance could present high voltage at the terminals but hopefully low current.
If you connected an incandescent light bulb from the terminal to ground, you might find it pulls voltage to zero (because very little current is sourced.)
When working on it, need to shut off inverter, ideally disconnect all power sources (battery, PV, grid or generator) and measure voltage, wait until it has discharged before touching anything.
I wondered if SPD could see PV negative biased up to that 189 VDC and PV positive biased to 189 VDC + PV voltage, which could be excessive relative to ground. But I don't think circuit adding those voltages together makes sense. Still, you never know, which is why I suggested connecting (replacement) parts with SPD not included and check voltages.
But ... the failure happened at night, so no PV voltage present! Only source of energy to blow up SPD and start a fire is battery and inverter? Or is grid present?
An SPD can take spikes, meant to absorb and dissipate brief transients. Limited number of Joules (watt seconds). If over-stressed, or maybe just after many hits, they can be damaged and short out. It is possible that inductive switching transients send spikes down the line. Your MPPT would not have been operating at night (I don't think) but inverter could have some coupling to it. Have access to an oscilloscope to look at those terminals? MPPT has inductor, inverter has inductor, and switching circuit will cause voltage transients. If they spike high they could keep hitting the SPD.
I finally stopped procrastinating and ordered a quantity of the following.
My grid AC input has a large SPD system. PV input, some model inverter had them but other similar models did not. I will connect three of these MOV in a delta configuration and put wires into PV+, PV-, Ground terminals for each inverter.
This MOV includes a thermal disconnect (maybe PTC fuse?) and they caution against overheating during soldering. They have a lead connected between MOV and PTC, which could be connected to LED indicators, providing "Protected" and "Failed" indicators like the Midnight SPD have. That is better because you can observe LED color at a glance. I'll just leave them accessible to connect DMM test leads. I don't expect a near lightning strike, not so common in my area. But maybe the next Carrington event creating a voltage across the loop formed by my PV string, it will protect against.
Data sheet says, "Notes: Average power dissipation of transients should not exceed 1.5 watts."
In other words, it can dissipate 1.5W as heat. Deposited energy (Joules) has to be infrequent enough to allow cooling off. If a switching supply operating at a few hundred Hz or kHz keep hitting it with spikes, it could gradually heat up then fail.
Also, can take 20,000A but graphs show a single 20,000A transient lasting 20 microseconds, or fifteen 10,000A transients lasting 20 microseconds is all it can take before wear-out.