Quick inverter comparison (cheap vs less-cheap)

I was asked to also plonk this down here, true to my chosen nom de plume.

I got roped into helping someone with an inverter (turned out to be just a dead lead-acid battery), but in the process I had to blow out all the dust and give it a test. So when opportunity knocks, we make a (poor!) video!


It is crazy how many of these cheaper transformer based ups units are available at the metal recyclers, with the only issue being one or more dead lead acid batteries.

I have picked up a handfull that I have refurbished and handed out to family and friends. Still have two of the same 800w effective models left, that just needs new cells.

Thanks @plonkster !! Really insightful! Would be interesting to see for more and some of bigger capacity products out there…

If you have enough of these for use as doorstops then I think the next best place for them is at the recyclers… :sunny:

All true, but with so many people working from home these days, one would think that any temporary ups/backup solution would be quite popular.

Maybe they are getting replaced with the larger inverter trolleys, giving the non d.i.y. users longer “off grid” capabilities.

A lot of my friends & colleagues were using them solely to run their internet connections. They’ve now all switched to the small “Wifi UPS” models after the LA died.

1 Like

People buy the trollie UPS’es, great excitement, costs wot ±R10k
Then they add more and more onto it, 1-3 years later, the batteries die.
The cost of replacing the 2 x 12v 100ah batteries, or some such, the plan stops that day.

Some make the sums, they “made a profit” as their business continued, helped them “earn”, so they replace the batteries, or upgrade.

For most using it for convenience, ±R10k down the tubes.

I’m now sitting with people about UPS’es. We don’t have R10k … what are the options for Wifi, laptop, screen and maybe even TV?

There is no cheap long-lasting plug and play never look back option.

So often once people and family hear that you have solar installed, then you become the “expert” and they want you to solve their load shedding problems, as cheaply as possible.

A lot of the times I do not mind, but some expect you to have THE ANSWER (its 42) that will magically restore power in their lives. They do not like to hear that it will cost money to get to the perceived level of convenience they want.

For the more techy friends, I now just point them in a direction or website and leave them to come back with “better” questions and then they start their informed journey.

1 Like

I usually tell them… if you go to the hardware store now and buy the cheapest “trolley” they have on the floor there, that is going to set you back 8.5k… and then you still have to buy batteries. So you’re looking at spending between 10k and 20k. What kind of budget did you have in mind?

“Naaah, I had more like 5k in mind”…

Then I tell them to look at the Flexopower… :slight_smile:

1 Like

If someone is only willing to spend 5k, they can’t really be that inconvenienced by load shedding, so why even bother.

1 Like

Interesting review. I don’t understand why these type of devices that are modified sine still exist (actually I do - it’s because of the marketing department, but it does not make sense from an engineering viewpoint). What does not make a lot of sense is why the losses is so high. I know why it is low for the victron - but the mecer’s topology has no reason not to be as efficient. Ignoring the inefficiency of some loads on modified sine, the mecer has a DC-DC converter followed by an H-bridge to convert it into modified sine. A DC-DC converter is easy to make efficiently - especially under light loads (easier than with the victron’s design) and the H-bridge has very little no-load losses.


Good spot. Indeed, I had that case open earlier (to blow out all the dust bunnies), and it indeed has an HF design. There is no reason for it to be that inefficient.

In the olden days, the square wave setup was a lot more efficient (because back then we didn’t have PWM and MOSFETs were too expensive), so banging a large silicon resistor on and off was how you did it. Nowadays I think it just makes shaves off cents from the manufacturing process, you very likely need a few less components in the output stage (eg no reference signal, comparator, etc) if you are not making a real sinewave.

I suspect that , most of the time , this is a result of people not understanding batteries. They assume that if the inverter can supply 2 kw they can draw 1 kw and then the batts get hammered to 5% soc in no time.
I know of at least two people that has solar system that also had to replace the batteries after two years…
Buddy of mine was looking at options and ended up buying the “trolley” kit to run the tv and two lights.
Max load is about 300w and the system is now about 5-6 years old and is till running.

One other thought that comes to mind as I am typing this. I recon most of those kits was bought in bulk and is only selling when there is a load shed run. So there is a good chance that some of those kits is 3 or 4 years old without any maintenance on the batteries.

Indeed, but then he needs to remember that he paid R16k (don’t you need to add batteries yourself?) to service an average load of 150W say.

Now this might be exactly what he needed, but a trolley is a trolley. It has to be parked somewhere in the house, take up a lot of space, only service loads close to it, and generally looks a bit ugly.

It also can never pay for itself. You can probably get a built-in inverter system, with PV and lithiums, that backs up all lights, electronics, fridges and other kitchen stuff for say 10 times that. Yes it is a large initial outgo, but it will actually turn into an investment without all the downsides of a trolley based UPS.

The Axpert and it’s clones all have a relatively low efficiency compared to low-frequency transformer based inverters such as the Victron inverters because of their topology. In the Axpert inverters, the DC-DC converter is made up of two full-bridge (also known as an H-bridge) inverters with a high-frequency transformer between them. This allows it to push current in both directions (i.e. From the battery to the DC bus and from the DC bus to the battery). It then has a 3rd full-bridge to generate the AC output from the DC bus.

In low-frequency transformer based designs there is only 1 full-bridge inverter on the low-voltage side of the transformer, so you have 1/3rd of the power electronics losses. (of course there are other losses in the inverter as well, but the losses in the switching components are the largest)

Yes, I am fully aware of the topologies involved. There is no reason, despite the need for 3 H bridges, for it to be less efficient.

The fact that it has 3 H bridges is the reason that it is less efficient. I don’t see how it could not be.

1 Like

It does not have to be, it is easier(read cheaper) to make the 3 H bridges more efficient than good efficiency in a low frequency design. The reason is that you need a toroid in a low frequency design to get good efficiency as well as a fairly big inductor. The victron does not have a big inductor, but they have a partially wound window on the toroid, which increases leakage inductance so you don’t need an additional expensive inductor. The downside is that you have to use thinner gauge copper, so your copper losses go up, but that is only relevant at high loads.

Either way, I had a look at the datasheet of the Mecer, and I do not think that they employ a bidirectional DC-DC converter in this case; due to the low charge current, I think they have two DC-DC converters, one for charge and one for discharge. It is a lot easier to design, so it makes sense that they have gone that route.

Most that I have seen sell without batteries…

They are around 9k-10k without batteries, so they end up around 15k (ish) after adding batteries.

We should not confuse the inverter I tested above, which I don’t think is a Voltronic (it is something else rebadged as a Mecer), with the common-as-weeds 5kVA Voltronic aka Axpert. The Axpert actually has a fairly efficient buck/boost stage that competes pretty well with the LF designs of other makers (such as Victron, MLT, etc). At least as far as I remember.

Agreed on this topic. The inverter tested has a fairly modern HF boost stage, so that really was disappointing.

It looks like the charger side is a common transformer of maybe 4A or so. I don’t have the inverter here anymore, it has gone back to its owner, but when I had it open, I did notice the iron transformer in there too, and it looks larger than it needs to be for a simple “power supply” (as some inverters have). I suspect this is another slow-recharge UPS… which is also a bad thing once you hit stage 4 load shedding.

1 Like

Just for fun…

R11k for the kit.But I am sure the batts will be suspect

1 Like