Why single phase AC?

[HSTEd]

I didn't say that 3 phase motors, or supplies to a user have 4 cables, but the distribution network has.....if you really don't believe this just take a quick look at ANY overhead, pole supply, within towns...4 cables. 3 phases and a return.
Yes the cross of the 3 phases gives 0v but each phase requires a return to the grid, within a premises on one phase (220v) this is the neutral return, which connects to the 4th cable. Each one of the three phases uses this return. A good time can be had watching the arcing across the three live phase cables during heavy ice/snow, far less dramatic when the lower one touches the return.....

AC is used because of the greater transmission distances, DC supplies require more frequent supply inputs.

I can see a three phase distribution line out of window, three power phases and no earth.

Unbalanced earth currents are simply transferred using grounding spikes.

As to providing low voltage DC in houses you could simply provide power rigged USB sockets across the house, the currents involved as so trivial that if you used mains standard wiring there would be little problem with resistive losses.
It would mean each house could have a single 6V DC rig which would simplify things considerably.

You might be able to do the same thing with a 20V DC circuit which would be suitable for laptops.

And remember that you could build almost all plug in equipment that doesn't draw enormous currents run on 20V, switch mode power supplies are cheap these days, becoming cheaper than mains frequency transformers even thanks to having far less copper or aluminium in them.

 

[apk55]

For a domestic instalation 240V is a good compromise. It enables quite high loads such as my kettle to be supplied by reasonably sized cables. In contrast at 12 volts you would require 6mm (cooker) cable to power a 20W bulb at the far side my house 10m from the supply. There would no advantage in using DC as quite a number of appliences require AC such as my fridge which uses a capacitor start and run motor.

Computer chips require power at very low voltage typicaly between 5V and 1.8V and at very high currents and typicaly this is done with a step down buck converter on the motherboard next to the CPU as it is imposible to supply this more than a very few cm. Other electronic equipment requires different voltages so it would be imposible to pick a voltage that would be common to all, so the current arangment of local regulators for each item is the best solution.

Most domestic supply comes from a delta star transformer with 5 limbs to cope with unbanced loading on the phases. There is no need for a neutral on the supply side although one may be provided at the distribution supply transformer but used for protection purposes, by using a relay to detect ground current caused by a phse to earth fault. Where a singe phase supply is required from a three phase distribution network the transformer is simply connnected between phases.

 

[HSTEd]

For a domestic instalation 240V is a good compromise. It enables quite high loads such as my kettle to be supplied by reasonably sized cables. In contrast at 12 volts you would require 6mm (cooker) cable to power a 20W bulb at the far side my house 10m from the supply. There would no advantage in using DC as quite a number of appliences require AC such as my fridge which uses a capacitor start and run motor.

It is a good compromise voltage, but it is not really neccesary to have only one voltage these days since the primary advantage to such standardisation on a single voltage has been removed by the neccesity of having converters due to the multiple voltages required by modern electronics and by the drastic reduction in the cost of providing multiple voltages.

However, a quick survey of the equipment I have in my university hall of residence flat would reveal a large number of devices which are, or could easily be, standardised on either a USB-type power supply or on some sort of ~20V supply similar to that provided for laptops (speakers, radios, netbook, wireless headphone chargers and phone chargers for example).

Providing a provision of those two additional voltages would remove the need for a large number of duplicate converters which are almost certainly more costly and less efficient than providing two converters in each house.

Even if that proved impractical it might be an elegent solution to build such converters into socket mounts to achieve a lower profile installation than is currently achieved at present with "wall warts" everywhere.

As to the cost of providing DC motors in place of induction machines in appliances that might "require" AC in the case of an all DC grid, the cost of power electronics has dropped to the point that a primitive three phase chopper circuit could be built into the equipment relatively cheaply as induction machines are not that sensitive about power factor... or you could just use a DC motor.
Additionally many appliances with motors such as washing machines often use "universal" type motors that can run on both AC and DC.

DC would have important other advantages, as an AC line's average voltage is only 77% its rated voltage (the voltage the insulation must protect against) DC lines are inherently more efficient at carrying energy and could make do with smaller cables per unit power transferred.
Additionally with the low cost of DC converters it would now be possible to replace transformers in the grid with choppers tapping kHz range transformers that would have a similar total cost.
Additionally there is far less worry about "power factors" in a DC system as the answer to line voltage is to attach more capacitance, rather than the calculations neccesary in a highly complex AC system.

Really if we were to build an electricity grid from scratch with no accounting for the availability of equipment off-the-shelf, there is little reason to go for AC.

 

[JGR]

It is a good compromise voltage, but it is not really neccesary to have only one voltage these days since the primary advantage to such standardisation on a single voltage has been removed by the neccesity of having converters due to the multiple voltages required by modern electronics and by the drastic reduction in the cost of providing multiple voltages.

However, a quick survey of the equipment I have in my university hall of residence flat would reveal a large number of devices which are, or could easily be, standardised on either a USB-type power supply or on some sort of ~20V supply similar to that provided for laptops (speakers, radios, netbook, wireless headphone chargers and phone chargers for example).

Providing a provision of those two additional voltages would remove the need for a large number of duplicate converters which are almost certainly more costly and less efficient than providing two converters in each house.

Even if that proved impractical it might be an elegent solution to build such converters into socket mounts to achieve a lower profile installation than is currently achieved at present with "wall warts" everywhere.

As to the cost of providing DC motors in place of induction machines in appliances that might "require" AC in the case of an all DC grid, the cost of power electronics has dropped to the point that a primitive three phase chopper circuit could be built into the equipment relatively cheaply as induction machines are not that sensitive about power factor... or you could just use a DC motor.
Additionally many appliances with motors such as washing machines often use "universal" type motors that can run on both AC and DC.

DC would have important other advantages, as an AC line's average voltage is only 77% its rated voltage (the voltage the insulation must protect against) DC lines are inherently more efficient at carrying energy and could make do with smaller cables per unit power transferred.
Additionally with the low cost of DC converters it would now be possible to replace transformers in the grid with choppers tapping kHz range transformers that would have a similar total cost.
Additionally there is far less worry about "power factors" in a DC system as the answer to line voltage is to attach more capacitance, rather than the calculations neccesary in a highly complex AC system.

Really if we were to build an electricity grid from scratch with no accounting for the availability of equipment off-the-shelf, there is little reason to go for AC.

It isn't *necessary* to provide only one voltage, but it is not necessary to provide more than one either. Wiring a house/etc is a non-trivial operation, and having N different sets of wiring just adds to costs.

5V or 20V is OK if you device wants exactly 5V or 20V, if it wants 9V or 3.3V, then you're back to buck converters which can be happily fed from mains AC. Also the cumulative current of all these LVDC devices is not something that you can just hand-wave away. When even the humblest of phone chargers wants an amp these days, you're going to find yourself supplying large currents at very low voltage over significant distances, which is not efficient, and probably prone to voltage fluctuation with load.

Building converters into sockets is unnecessarily expensive, whereas it's trivial to just stick a converter into the socket itself if one is needed.

DC motors are more expensive, require more maintenance and don't perform as well as 3-phase electric motors.

The DC/AC insulation difference and cross-section issues are really not a big problem in most cases.

The power factor correction calculations in an AC system are not difficult, you could probably get away with not doing any calculations at all and still correct most of it.
Attaching a capacitor to a DC system only deals with pulsed loads. It won't deal with voltage sag from constant loads.

Large scale HV inverters are not quite as cheap as I think that you think they are. The break-even point for HVDC over HVAC for land links is rather long due to the cost of the converters at each end. Also, connecting intermediary taps is much more difficult.

 

[HSTEd]

5V or 20V is OK if you device wants exactly 5V or 20V, if it wants 9V or 3.3V, then you're back to buck converters which can be happily fed from mains AC. Also the cumulative current of all these LVDC devices is not something that you can just hand-wave away. When even the humblest of phone chargers wants an amp these days, you're going to find yourself supplying large currents at very low voltage over significant distances, which is not efficient, and probably prone to voltage fluctuation with load.

Since chargers are already standardised at voltages not directly used by the electronics anyway, you don't really loose very much by simply setting the "intermediate" voltage to some standard for each power demand range.

For instance laptops can't directly use 19V, so you really need two converters in the present setup, we are only touching the "upstream" one.

Building converters into sockets is unnecessarily expensive, whereas it's trivial to just stick a converter into the socket itself if one is needed.

"Wall warts" have all manner of problems, such as the risk of damaging the converter since it just hanging off the wall from the plug, risk of non standard equipment causing all sorts of power factor issues and the like.
Building it into the socket allows the form factor of the system overall to be drastically reduced, which is important in the wire congestion that prevails in many houses these days.

DC motors are more expensive, require more maintenance and don't perform as well as 3-phase electric motors.

And yet DC capable motors are used extensively in all manner of domestic appliances, such as the aforementioned washing machines and hairdryers and similar equipment. (universal motors that can run on either AC or DC)

No domestic appliances use true three phase motors and if we are talking large industrial motors, many of them use DC anyway (see rolling mills) or use variable speed motor drives anyway, making what it is powering the converters irrelevent.

The DC/AC insulation difference and cross-section issues are really not a big problem in most cases.

DC cables would only contain 70% of the material required in AC cables of the same rated voltage and power draw.
That sounds significant to me.

The power factor correction calculations in an AC system are not difficult, you could probably get away with not doing any calculations at all and still correct most of it.
Attaching a capacitor to a DC system only deals with pulsed loads. It won't deal with voltage sag from constant loads.

Surely power factor calculations can't deal with voltage sag as that is a result of the resistance of the power line and the current being drawn?

Large scale HV inverters are not quite as cheap as I think that you think they are. The break-even point for HVDC over HVAC for land links is rather long due to the cost of the converters at each end. Also, connecting intermediary taps is much more difficult.

The breakeven point was apparently down to 70km in 1999, and the price of converter stations has dropped significantly since then, thanks to the deployment of systems such as HVDC Light.

Also, if you were going for a DC grid today you could use single level converters (with 6.5kV rated transistors) to provide 3kV to every single house with the converter being integrated into the electricity meter.

This would allow you some rather impressive savings as you might imagine.

 

[apk55]

Having cheap standard sockets supping a standard voltage which any appliance can be plugged in is the advantage of the current system. I can plug in any appliance in any socket anywhere in the country without problems. Houses and offices can be fitted with a surplus of sockets so there is a lot of flexibility if things are re arranged. Sockets with a built in converter are going to expensive and bulky. An external converter can be matched to the appliance and switched off when not in use.

While DC to AC inverters are established technology, building units on a domestic scale will not be cheap. In order to get a reasonable sine wave on the output and efficient operation several square wave inverters in series and running at slightly different phase angles are required, or a less efficient pulse width modulation system would be required. This is not a problem on a large scale (several megawatt) installation but would be costly in a domestic 10 to 20KW installation. And with a several KV input voltage this box would be bulky. No the existing system of having a transformer on the street corner supping a number of houses I still think is the best and it has proved robust and reliable. The use of a DC link on our national grid is understandable and can improve the stability of network particularly with large power flows over significant distances.

 

[sbt]

Wiring a house/etc is a non-trivial operation, and having N different sets of wiring just adds to costs.

Tell me about it - I live on a boat and have 240V AC (Shore or Generator supplied) and 12V DC (Batteries charged by Engine or Shore via a charger). When cruising some equipment requires 12V DC to 240v AC (via an Inverter) to some random voltage DC via the equipments power supply.