15-09-2015, 07:00 PM
This raises an interesting point about terminology. In that other thread, NM was referring to an "inverter", but it sounded like it also contained a battery charger as well. So perhaps it's more accurate to describe it as a UPS, given that a UPS consists of charger, batteries and inverter? (Obviously, batteries can be internal or external). I've no experience in the marine electrics world, so don't know if it's common usage to call the box an inverter. I can say that in the outside broadcast vehicles that I've seen, the charger and inverters are usually separate boxes, and referred to by their separate names.
All this reminds me of a similar thing. What do you call "the box that does your broad-band"? Chances are, you call it a router, right? Certainly, most people do in my experience.
But that box contains many things, including:
You get the idea. So why is it called a router?
Back to inverters, I remember being impressed with the APC way of implementing a cheap UPS. When you take one apart, there is only one transformer in there. When the mains is present, the transformer steps down the AC to charge the batteries. When the mains fails, a brace of relays change over, and now the transformer is acting for the inverter, stepping up an AC waveform that is generated by MOS-FETs connected to the batteries. Neat. Obviously the change-over isn't seamless, but switched-mode power supplies usually have a fairly generous "hold-up time" to cover the gap. Slightly OT, but as time goes by in the PC world, the power supply will eventually become a single 12V output, with all power conversion done on the motherboard, and that'll open up some interesting options for cheaper, more efficient UPS systems. Right now, most of the other lines are in (or nearly in) the "legacy" category - for some motherboards, it's only the PCI slots that need some of the rails. And drives, of course.
There are many ways to implement an inverter. Most of the cheaper ones use a modified square wave approach because that's simple and relatively efficient. Absolutely fine for anything that has a switched-mode supply (the vast majority of those would happily run on DC), but not so good for anything else. There are techniques to improve them, but the sine wave inverter is the preferred option. These are a lot more complex. The problem is the linear nature of the waveform, which implies heat loss, plus the magnitude of the required voltage.
To get around the efficiency problem, you can use switching techniques. A class D amplifier can give 90% or more efficiency. Then you can stick the sine wave into a 50Hz transformer and step it up to 230V as required.
But 50Hz transformers are heavy, so it's tempting to do away with them. There are many options:
Factors that drive the topology choice will include available semiconductors. E.g., high voltage MOS-FETs don't tend to switch on as fully as low voltage examples, so will have conduction losses. Obviously, that's changing as time goes by, but at a fixed point in time, you might choose to turn 24V into 170VDC with a boost converter, then apply that to the class-D stage, then have a 2:1 stepup on the ouput (in either the HF or 50Hz domains). For example...
Don't forget the need for overload protection, over-temperature shutdown, protection against reactive loads on the output, and possible "load dump" from a vehicle battery system. The more you look into these, the more you realise just how well engineered they have to be. There are relatively few firms making them (ignoring the really cheap ones). These are the ones I'm familiar with: http://www.antares.co.uk/pure-sine-wave-inverter.html - funnily enough, the more complex diagram on that page looks exactly like one that came with one of our vehicles (surely a coincidence
)
Interesting topic
All this reminds me of a similar thing. What do you call "the box that does your broad-band"? Chances are, you call it a router, right? Certainly, most people do in my experience.
But that box contains many things, including:
- An ADSL (or VDSL) modem
- A NAT router/firewall
- A network switch, usually at least 6 ports (4 of which are available to the user)
- A wireless access point
- A DHCP server
- A web server
- Perhaps a print server or a media server?
You get the idea. So why is it called a router?
Back to inverters, I remember being impressed with the APC way of implementing a cheap UPS. When you take one apart, there is only one transformer in there. When the mains is present, the transformer steps down the AC to charge the batteries. When the mains fails, a brace of relays change over, and now the transformer is acting for the inverter, stepping up an AC waveform that is generated by MOS-FETs connected to the batteries. Neat. Obviously the change-over isn't seamless, but switched-mode power supplies usually have a fairly generous "hold-up time" to cover the gap. Slightly OT, but as time goes by in the PC world, the power supply will eventually become a single 12V output, with all power conversion done on the motherboard, and that'll open up some interesting options for cheaper, more efficient UPS systems. Right now, most of the other lines are in (or nearly in) the "legacy" category - for some motherboards, it's only the PCI slots that need some of the rails. And drives, of course.
There are many ways to implement an inverter. Most of the cheaper ones use a modified square wave approach because that's simple and relatively efficient. Absolutely fine for anything that has a switched-mode supply (the vast majority of those would happily run on DC), but not so good for anything else. There are techniques to improve them, but the sine wave inverter is the preferred option. These are a lot more complex. The problem is the linear nature of the waveform, which implies heat loss, plus the magnitude of the required voltage.
To get around the efficiency problem, you can use switching techniques. A class D amplifier can give 90% or more efficiency. Then you can stick the sine wave into a 50Hz transformer and step it up to 230V as required.
But 50Hz transformers are heavy, so it's tempting to do away with them. There are many options:
- Take the DC, feed it into a boost converter to make a high DC rail. Use that to power a class D amplifier that is able to work at the high voltages. Take AC output direct from amplifier, via some filtering.
- Make a class D amplifier with a HF transformer in the output stage that does some stepping up.
- A hybrid of all options considered so far.
Factors that drive the topology choice will include available semiconductors. E.g., high voltage MOS-FETs don't tend to switch on as fully as low voltage examples, so will have conduction losses. Obviously, that's changing as time goes by, but at a fixed point in time, you might choose to turn 24V into 170VDC with a boost converter, then apply that to the class-D stage, then have a 2:1 stepup on the ouput (in either the HF or 50Hz domains). For example...
Don't forget the need for overload protection, over-temperature shutdown, protection against reactive loads on the output, and possible "load dump" from a vehicle battery system. The more you look into these, the more you realise just how well engineered they have to be. There are relatively few firms making them (ignoring the really cheap ones). These are the ones I'm familiar with: http://www.antares.co.uk/pure-sine-wave-inverter.html - funnily enough, the more complex diagram on that page looks exactly like one that came with one of our vehicles (surely a coincidence
)Interesting topic
