Easiest way to protect appliances against DC voltage on AC output on non-isolated off-grid HF inverters?

[fmeili1]

I'm thinking about how to protect my appliances in case one or more of the MOSFETs/IGBTs in the H-bridge break and there is a continuous high DC voltage at the AC output of such an inverter.

I've already installed AC surge protectors on the AC-out of the inverter to just protect AC surges. But I'm sure they will not protect against (high) DC voltage on the output side (the high DC voltage from the DC-bus will usually be the same as the AC peak voltage of the AC sine wave, so I'm pretty sure, the AC surge protector will not do anything in such a case).

I don't know if usual non-isolated off-grid HF-inverters do have a DC protection integrated on their AC-out side.

Does a kind of special breaker/protector device exists to install it on the output of my AC-out's to protect against DC voltage (I think even a lower than DC bus-voltage would hurt some appliances so it will not be enough to just protect about high DC voltage, it should detect and trip in a wide range of DC voltages maybe >1VDC or >5VDC or >12VDC or something else... I don't know)?

From the electrical basics, a simple inductor between L and N on the output side of the inverter could be used to produce a short in case of DC voltage, which may trip the overcurrent protection inside the inverter (which hopefully exists?) - but I don't know if this is the best idea.

Maybe the scenario described doesn't really exist and/or doesn't cause any problems at all and I'm just thinking that this could be a problem...

Can someone shed some light to this?

 

[wpns]

Inductor is going to be huge, series capacitor will also be huge. How about an averaging RC network that triggers a crowbar circuit?

DC bus is typically much higher than AC, maybe the appropriate surge protector would conduct.

How often have you heard of this happening?

 

[timselectric]

While anything is possible in a total catastrophic failure.
I don't think this is something you should be concerned about.

How often have you heard of this happening?

I have never.

 

[Luk88]

Inductor is going to be huge, series capacitor will also be huge. How about an averaging RC network that triggers a crowbar circuit?

DC bus is typically much higher than AC, maybe the appropriate surge protector would conduct.

I think this is correct. Few times I heard about exact voltages used it was in the region of 450-500V. This is in inverters that use a common bus.

While anything is possible in a total catastrophic failure.
I don't think this is something you should be concerned about.

I have never.

Me too.

If such 450V-500V was connected to the AC output I think surge protection would short and that would trigger the overcurrent protection. But how quickly? Probably not quickly enough to prevent damage? I heard about a couple of "electrician mistakes" where interphase voltage (415AC) was mistakenly sent onto 240V circuits (by connectinglp00single time the people it happened to were replacing everything. Tvs, fridges,0 computers. Everything.

This was AC. I believe high voltage is much more of a problem than just DC. Most devices such as tvs, computers, even inverter based fridges and washing mashines will have a switching PSU that rectified the AC to DC (usually this ends up being 1.4 times AC voltage - so 336V here). Such devices litld be fine if they get DC up to their

 

[mciholas]

Simplest solution for the people worried about this is to pass the AC through an isolation transformer. If DC appears on the primary, it won't pass to the secondary.

There will be an efficiency hit and a cost.

I would expect an inverter to have a relay that drops out when the output is out of spec. This belief comes from seeing about 10 relays inside my inverter and they must be doing something like that.

Mike C.

 

[fmeili1]

Thank you for your answers. I'm probably thinking about a problem that isn't relevant in practice.

I've never heard of such a case myself, but if you look at the H-Bridge circuit you can imagine that this problem could exist. I can't estimate how often it happens that a defective MOSFET's/IGBT's becomes permanently conductive between D/S or C/E instead of permanently blocked.

Instead of passing the entire power through such a protective circuit (very large components, etc.), you could simply install a measuring circuit that detects the DC voltage component in the AC voltage and then triggers the emergency shutdown on the inverter... but even if this existed, it is wondering if that would react quickly enough...

But since you haven't heard of such cases either, I'll just ignore this case and hope it will never happen.

 

[Warpspeed]

It happened to me with a very early Warpverter prototype.
It killed numerous household appliances, and fixing the resulting carnage was neither fun or cheap.

Many of the smaller low cost crappy Chinese high frequency inverters do not have sufficient power from the high voltage rail to do much damage, so its no real surprise that these just usually go pop and stop working.
Something a bit more serious that can feed some significant high voltage dc power direct onto the ac output can be a lot more destructive.

My own thoughts are, that any reasonably high powered inverter should have an output transformer.
In fact for grid tie here in oZ its a legal requirement.

 

[fmeili1]

It happened to me with a very early Warpverter prototype.
It killed numerous household appliances, and fixing the resulting carnage was neither fun or cheap.

So it can really happen that a MOSFET/IGBT breaks down in such a way that it switches permanently on and does not break permanently between D/S or C/E - even if it's a rare situation (no one else of the previous posters has heard of such a problem so far).
I can imagine it was no fun to repair/replace your damaged appliances and other damaged devices!

Many of the smaller low cost crappy Chinese high frequency inverters do not have sufficient power from the high voltage rail to do much damage, so its no real surprise that these just usually go pop and stop working.
Something a bit more serious that can feed some significant high voltage dc power direct onto the ac output can be a lot more destructive.

I use such low cost HF inverters (EG4-6500EX, based on Axpert Max MKS2-65000) with max. 6.5kW (six in parallel split phase to get 39kW). I don't know, if these inverters would just break/pop in case of such an event - but I hope because it would be much cheaper to replace an inverter instead of many broken appliances, computers, TV's, routers, cameras, 3D printer, pool pump, central heat pump, mini splits, ... everything!
Such an event would be the worst case super disaster.

My own thoughts are, that any reasonably high powered inverter should have an output transformer.
In fact for grid tie here in oZ its a legal requirement.

None of these low cost inverters have output transformers. I looks like I have to start again thinking about whether it is possible to develop a protective circuit against high DC voltage on the AC out - without adding 6 times 6.5kW transformers on the AC-out.

 

[sunsurfer]

Can someone shed some light to this?

Easiest solution is to buy an inverter with an output transformer.

I wouldn't worry about it but next time get something with some copper under the hood if it bothers you.

 

[sunsurfer]

None of these low cost inverters have output transformers. I looks like I have to start again thinking about whether it is possible to develop a protective circuit against high DC voltage on the AC out - without adding 6 times 6.5kW transformers on the AC-out.

I agree, its all about making them lower cost. The big issue is likely the life of the inverter and not a shorted FET. I want my inverters to be reliable and HF inverters have shorter lives then LF. And yea, its kinda weird they even make HF inverters if you think about it.

 

[cs1234]

Do Victron inverters potentially have this issue? Or does their toroidal transformer design block it?

How about other popular inverters on the forum, the EG418KPV or Sol-Ark 15k?

 

[sunsurfer]

Do Victron inverters potentially have this issue? Or does their toroidal transformer design block it?

How about other popular inverters on the forum, the EG418KPV or Sol-Ark 15k?

There transformers are on the output like typical LF inverters so it provides isolation from the driver circuits.

You can still get interaction if there is a problem but at least its through the transformer which is not directly wired.

 

[sunsurfer]

It happened to me with a very early Warpverter prototype.
It killed numerous household appliances, and fixing the resulting carnage was neither fun or cheap.

Many of the smaller low cost crappy Chinese high frequency inverters do not have sufficient power from the high voltage rail to do much damage, so its no real surprise that these just usually go pop and stop working.
Something a bit more serious that can feed some significant high voltage dc power direct onto the ac output can be a lot more destructive.

My own thoughts are, that any reasonably high powered inverter should have an output transformer.
In fact for grid tie here in oZ its a legal requirement.

Seems like burning to "OPEN" is going to be the more common failure. I am also surprised they let HF grid tie in the states. I guess they figure its rare enough of an issue, no need to worry.

Appliance damage from the grid could be more likely then from an inverter. Something to think about....

 

[Warpspeed]

Appliance damage from the grid is far from unknown.

The usual sequence of events is that a large heavy vehicle knocks over a power pole.
High voltage lines (6.6Kv or whatever) at the top of the pole, come in contact with the 230v lines located further down.
That will fry just about every appliance in your home, switched on or not.

 

[mciholas]

The usual sequence of events is that a large heavy vehicle knocks over a power pole.
High voltage lines (6.6Kv or whatever) at the top of the pole, come in contact with the 230v lines located further down.
That will fry just about every appliance in your home, switched on or not.

A few years after my brother moved into their new house, the utility replaced their pad transformer for no apparent reason.

They set the secondary tap wrong. Fried just about everything in the house. Was a big mess and while the utility did end up covering the costs, it was a great hassle and very disruptive.

Basic common sense is to pull the meter, replace the transformer, CHECK THE OUTPUT, then put the meter back into service. Nope, just got the wires wrong and didn't care.

This happens more than you would think.

Some utilities don't cover their mistakes, too:

Duke Energy equipment failure blew appliances in an Orlando home. Why the company says it’s not liable

An Orlando couple says an equipment failure in their neighborhood sent a surge of electricity into their home and fried many of their appliances. The company that handles claims for Duke Energy admitted Duke had an equipment failure, but the company says Duke is not liable for the losses.

www.clickorlando.com

Mike C.

 

[fmeili1]

I agree, its all about making them lower cost. The big issue is likely the life of the inverter and not a shorted FET. I want my inverters to be reliable and HF inverters have shorter lives then LF. And yea, its kinda weird they even make HF inverters if you think about it.

In my opinion, because of the necessary solar expansion in the world, the number of required inverters will explode. If all of these needs a "many kW sized" transformer it would cost a lot of additional resources like copper and core material which results in increased costs. I know what you mean about your "weird" feeling with an transformerless inverter design, but I think, in the end the costs will make the decision and as more reliable the transformerless units will be, this design will probably serve the mass market at the end. I'm sure the design could still be improved to make them "nearly" as reliable as transformer based inverters (technology evolves and things like e.g. Silicon Carbide MOSFET's, super capacitors, will help to increase the surge capabilities, the efficiency, etc.).

Update:
Beside enhancements in electronic components like SiC MOSFETS and super capacitors, I think that new H-Bridge design enhancements will make the HF inverters in future more reliable, also. E.g I've read something about the "Aalborg inverter" and the newer "Modified Aalborg inverter" which are promising...

 

[cs1234]

We wouldn't necessarily need such large inverters if the rules for backfeeding and installing a small system with batteries were loosened up a bit. Even a small GTIL and 1000watts of panels would help tremendously if nearly every house had one. That would cover most of their load during non peak times.. and a good portion of it during busier times could be shaved off.

 

[sunsurfer]

In my opinion, because of the necessary solar expansion in the world, the number of required inverters will explode. If all of these needs a "many kW sized" transformer it would cost a lot of additional resources like copper and core material which results in increased costs. I know what you mean about your "weird" feeling with an transformerless inverter design, but I think, in the end the costs will make the decision and as more reliable the transformerless units will be, this design will probably serve the mass market at the end. I'm sure the design could still be improved to make them "nearly" as reliable as transformer based inverters (technology evolves and things like e.g. Silicon Carbide MOSFET's, super capacitors, will help to increase the surge capabilities, the efficiency, etc.).

Update:
Beside enhancements in electronic components like SiC MOSFETS and super capacitors, I think that new H-Bridge design enhancements will make the HF inverters in future more reliable, also. E.g I've read something about the "Aalborg inverter" and the newer "Modified Aalborg inverter" which are promising...

Copper supply is the least worry, silver on the other hand...there is not enough for everyone to have solar. Thats the bigger issue or elephant in the room as the saying goes.

I will have to read those links. There is definitely some cool stuff that can be done with fast switching HF inverters and solid programing. Personally, I've learned over the years to keep things simple and leave the complicated to the "keeping up with the Jones" crowd. I find simple and late adoption is the less painful route.

The one thing an HF can never do is provide physical isolation.

 

[fmeili1]

Copper supply is the least worry, silver on the other hand...there is not enough for everyone to have solar. Thats the bigger issue or elephant in the room as the saying goes.

Full Ack - but there will be a cooper shortage too in future!

Btw. Silver is, in my opinion, the most undervalued and underestimated material in our modern world and substitutes are often not possible. From a rarity standpoint, the price should be about 1/10 of the price of gold... that would currently be about $230 per ounce - but it's only $29 now... Gold has a known resource life of just under 36 years and silver only 24 years... but that's off-topic...

I will have to read those links. There is definitely some cool stuff that can be done with fast switching HF inverters and solid programing. Personally, I've learned over the years to keep things simple and leave the complicated to the "keeping up with the Jones" crowd. I find simple and late adoption is the less painful route.

That's in general true! KISS - Keep it simple, stupid

The one thing an HF can never do is provide physical isolation.

Right, even the best HF design will not be able to provide this without a transformer... as far as I know.

 

[Luk88]

Full Ack - but there will be a cooper shortage too in future!

For sure. Unless we get really good with recycling, not pretend to do so.

At least recycling things like transformers is pretty easy.

Btw. Silver is, in my opinion, the most undervalued and underestimated material in our modern world and substitutes are often not possible. From a rarity standpoint, the price should be about 1/10 of the price of gold... that would currently be about $230 per ounce - but it's only $29 now...

Sadly(or favourably to the industry) , we have VAT on silver here and no VAT on gold(other taxes elsewhere) . If there was no VAT on silver I bet the price would go way way up due to people using it as store of wealth.

Right, even the best HF design will not be able to provide this without a transformer... as far as I know.

Other than things already mentioned there is one that low frequency transformers have a certain efficiency and (allegedly) it is not possible to improve. This is really evident at low power levels.

I read at rated load LF transformers can be 99% efficient, but same transformer drops under to 50% at 1% load.