3 hours ago, Gnome said:

Hahaha.  I know it seems naive to assume that, but coming from electronic engineering side of this, let me tell you that a lot of these ratings inform customer buying.  So they always take the most enthusiastic value.

They'll typically give the efficiency rating the data sheet at the most optimistic ambient temperature with the most ideal input battery voltage with the most ideal load percentage.  Those 3 components together can cause an enormous shift in efficiency.  You can actually see this, for example, from the Victron inverters where they have a crazy efficiency rating but when you look at the actual data closely you'll notice they use a very high input voltage and very rapidly derate based on temperature.

If you want a really good example of this that is REALLY moving fast, take a look at computer power supplies.  They are years ahead of inverters, they are using GAN FETs and have really incredible efficiency ratings and even more impressive claims.  When the devices are actually tested you realise that the maximum rating rarely applies and when it does apply it is for an incredible narrow band.

Unfortunately I don't see a good amount of independent reviews where the efficiency claims are tested and an efficiency graph is generated.

Before this kind of testing became common place with computers the manufacturer outright lied, all of them without exception.  And even now that it is very tightly controlled by specifications they still only claim the best they can achieve which is very far off from worst case.

Inverters are not special, they operate on a lot of the same principles and have all the same constraints and have most of the same common components.  And since people buy using that single value, it is important to give the most optimistic value there or you are hurting your company when the competitors are all doing it.  This is true for all electronics, they all work like this.

So much misinformation in one post...

Then you mention Victron inverters. Those are boat inverters. The same ones they used on the Ark. They have the worst losses because they are LF inverters, where primarily the copper loss ceiling hits you. (not to mention that victron undersize the transformers in them, so it hits hard.) An inverter and a computer PSU, though it seems to be a good comparison at first, is not in fact a good comparison for losses. Do a design and calculate the losses for both and you'll quickly see why. And very few computer PSUs are using GANs - they are difficult to work with and expensive. (Though they are starting to be used in bricks)

The efficiency of grid connected inverters are regulated - at least in some countries. So they have to be accurate in order to be certified.

 

2 hours ago, P1000 said:

Then you mention Victron inverters. Those are boat inverters. The same ones they used on the Ark. They have the worst losses because they are LF inverters, where primarily the copper loss ceiling hits you.

https://www.victronenergy.com/inverters/inverter-rs-48-6000-230v-smart

But I guess I know nothing 🤷‍♂️ Of all the inverter companies they are the most honest with their specifications.  But looking at their "maximum efficiency" then looking at their stated max outputs it becomes quickly obvious that they applied those at the most ideal circumstances.

2 hours ago, P1000 said:

An inverter and a computer PSU, though it seems to be a good comparison at first, is not in fact a good comparison for losses. Do a design and calculate the losses for both and you'll quickly see why. And very few computer PSUs are using GANs - they are difficult to work with and expensive. (Though they are starting to be used in bricks)

The efficiency of grid connected inverters are regulated - at least in some countries. So they have to be accurate in order to be certified.

I'm not going to argue with you, it is going to go back and forth.  The reality is they aren't that efficient.  I have a ton of electronics experience and I actually have the equipment to test it, specifically a professional level load tester which I've used to test few of inverters.  SunSynk and Axpert types so far.  They aren't nearly as efficient as they claim while on battery.  Not nearly.

As for the efficiency and design of computer power supplies, again an area I have a lot of knowledge in, specifically calculating loads for data centers and being intimately involved in technology specifications, selections and de-ratings (hint hint, for the biggest data center owner in the world).  Including how the designs look and how they compare to inverters to a lot of respects.  But what is the point in arguing with someone that knows best.

This is also why I know that de-ratings need to be applied to the equipment without exception and using the values you have in the public data sheet, specifically the maximum stated efficiency, is pretty far off the mark.

Unless ofc you are a bakkie brigade installer as we call them, then YOLO.  The maximum stated efficiency and most enthusiastic values why not.

Edited October 5, 20223 yr by Gnome