diy solar

diy solar

Midnite Solar Rosie HF Inverter 27kW Surge Test

1. Is this inverter equipped with fault detection such that, if an output switch or controller fails, that it will disconnect the output with a physical relay, to prevent large DC outputs that can damage loads?

Can you provide HF inverter examples that have protective relay that prevents DC output?
 
Can you provide HF inverter examples that have protective relay that prevents DC output?
Fronius inverters for sure, had a direct communication with their engineers. Most companies don't even know what I'm talking about. Attached document details the problem but you may not be able to follow unless your an inverter design engineer.
 

Attachments

  • H2PToday2201_design_RedHillLabs.pdf
    1 MB · Views: 13
Fronius inverters for sure, had a direct communication with their engineers. Most companies don't even know what I'm talking about. Attached document details the problem but you may not be able to follow unless your an inverter design engineer.

"The closest approach possible
has been demonstrated by the attachment of a thermistor to the center, or drain lead of a MOSFET package
such as a TO3-PI as shown in Fig. 9"

Guilty as charged.
Wasn't a switch-mode MOSFET inverter application, rather a linear bipolar RF amplifier. But I put a PTC thermistor there as part of bias current temperature compensation.

That whole MOSFET "SOA" isn't safe thing occurs when MOSFET is used as a linear component, not switching.
Playing with gate voltage to balance currents I would think would still be very low Vds, but would cause more heating.

"The solution is simple. A transformerless inverter should have a mechanical relay or contactor in its output
circuit, normally open. A “dc detector” should then drive this relay such that the relay is enabled only when no
dc is detected on the inverter output. This could certainly be a defining feature for high-end transformerless
inverters. Obviously transformer-output inverters are immune to this problem"

Ahh, there it is. And here I thought you were asking me to find the protection in a schematic.
No, I'm not an inverter design engineer. But I have dug through the innards of some systems, and designed a few small things.

The SMA inverters do have relays. The first transformerless models had very loud relays, bigger than usual and with strong springs, I would say. Installation was also required to have fuses on both PV+ and PV-.
 
@SpongeboB Sinewave , how is the Rosie coming along ?

Any guesses when it may be released into the wild ??

Still needs ETL to do their thing. Last couple or 3 weeks have been devoted to getting one of the worst generators we have to work properly. Finally done I think. Not easy especially when the genny's control system wants to fight the inverter's control system at its own rate.

The new charge controllers are listed now. Just some loose ends to finish there.

For overcurrent protection, I prefer MOSFET desaturation if possible. Not as fast as some other methods but effective.

boB
 
"The closest approach possible
has been demonstrated by the attachment of a thermistor to the center, or drain lead of a MOSFET package
such as a TO3-PI as shown in Fig. 9"

Guilty as charged.
Wasn't a switch-mode MOSFET inverter application, rather a linear bipolar RF amplifier. But I put a PTC thermistor there as part of bias current temperature compensation.

That whole MOSFET "SOA" isn't safe thing occurs when MOSFET is used as a linear component, not switching.
Playing with gate voltage to balance currents I would think would still be very low Vds, but would cause more heating.

"The solution is simple. A transformerless inverter should have a mechanical relay or contactor in its output
circuit, normally open. A “dc detector” should then drive this relay such that the relay is enabled only when no
dc is detected on the inverter output. This could certainly be a defining feature for high-end transformerless
inverters. Obviously transformer-output inverters are immune to this problem"

Ahh, there it is. And here I thought you were asking me to find the protection in a schematic.
No, I'm not an inverter design engineer. But I have dug through the innards of some systems, and designed a few small things.

The SMA inverters do have relays. The first transformerless models had very loud relays, bigger than usual and with strong springs, I would say. Installation was also required to have fuses on both PV+ and PV-.
Thanks for the info on the SMA inverters (can it be substantiated/documented in any way?). I'm starting a list of known manufacturers of transformerless inverters that have that protection.

SOA, even in switching, becomes important in fault conditions, in particular shorted loads. During the interval between the short and the digital control implementing a shutdown SOA violations will take place. I had one manufacturer try to blow off this discussion by stating they had cycle by cycle current limiting. I've been using power devices for a long time and blew a lot of stuff up to learn this. One of the semiconductor products I worked on was a hot swap controller on the power input to a server that had a similar issue. And it pointed out that a really difficult scenario is where the FET (or whatever) is off, and just being turned on. You now have full supply voltage stress on the FET and essentially unlimited current. We designed some of the earliest hot swap controllers with SOA protection to make turning on into a short circuit safe.

Lastly for people who do inverter testing on YouTube and wherever, I haven't been impressed yet. You need to take your inverter, put it in a 120 deg F chamber, put a full load on it, and let it run until it fully heats up. Then repeatedly short circuit the output.

Lastly that picture in Figure 9 wasn't on any system you've described, it was on a system I worked with the guy who was doing an audio frequency power amplifier (and at Apex we built hybrid power amps where we die attached small signal bipolar transistors to the top of the MOSFET die).
 
This doesn't mention the relay, but I can hear it with my own ears.
I don't know the details of what protection they built in. They satisfied UL. Of course need to make sure it never connects 500VDC PV to the grid and saturates a transformer.


Yes, having semiconductors survive being the bonded fault doesn't seem easy. Better to dip your toes in the water first, make sure voltage and phase are what you expect.
SMA cautions not to have a transfer switch snap between AC sources; allow 5 seconds "off" so inverter disconnects, doesn't get hit with out of phase grid.
 
Fronius inverters for sure, had a direct communication with their engineers. Most companies don't even know what I'm talking about. Attached document details the problem but you may not be able to follow unless your an inverter design engineer.

The techniques are well known in power electronics. It appears off-grid battery-only HF inverters vendors don't do it for cost reasons. I am guessing that most serious users of off-grid battery inverters end up using LF inverters which are inherently immune to the problem. It may be more common for grid-tied HF inverters to have relays perhaps due to compliance reasons.
 
Last edited:
Back
Top