Paging ARW - generators bla bla

[Dave Liquorice]

Surprised you don't have solar thermal for hot water with electric back up. Plenty of sun in Spain, unlike the UK...

[Paul]

Or better still connect earth and neutral at the isolator. That means that I can just hard wire a cable with normal Spanish plug to plug into the genny as it won't matter which phase connects to earth/neutral as they are the same. Whichever one happens to be connected to earth via the cable will become my neutral. It seems so simple now ;-)

Listen out for the bang...

[Johny B Good]

The higher than normal distortion in the generator's output is the least of your worries (it's much better than the quasi sinewave or modified squarewave of the cheaper inverters commonly used in sub

500VA UPSes and 12v inverter supplies).

The real problem with such gensets is their susceptability to overvolting when driving a load with a relatively small amount of leading current (capacitive loading). Such excess capacitive loading can cause a 230v generator head to output in excess of 270v which the AVR module can no longer control. Resistive and inductive loads are free of this effect (incandescent lamps, electric toasters or motors of almost any type).

Unfortunately, most electronic kit tends to demand some amount of leading current with older UPS designs being the worst offenders by placing a considerable capacitive load across their incoming mains connections.

The more expensive inverter gensets are free of this overvolting effect on both leading and lagging current loads so are the only viable emergency backup to UPS protected IT kit in the 2 to 5KVA genset range.

The larger 10KVA and up gensets aren't free of this unfortunate side effect, they only perform acceptably when the IT kit loading is only a tiny fraction of the genset's VA rating. This is the basis of the myth that only oversized gensets are of good enough quality to power IT kit. In all honesty, the purity of their waveforms is little better than the cheap 2 to 3 KVA home emergency gensets commonly sold in hadware retail outlets Like Home Depot in the US of A and B&Q (and Aldi) here in the UK.

[John Rumm]

Yup that looks like the kind of thing you need. ;-)

[Johny B Good]

====snipped====

Thank you very much for that praise. :-)

I think there is some confusion here. When Paul mentioned he could trip the rcd by earthing either pin in his house sockets[1] I concluded that he may have been on a bi-phase supply.

My reference to the generator head having a pair of 110v stator windings wasn't intended to imply a centre tapped bi-phase option, just to indicate why a neon lamp test would indicate such a bi-phase supply due to the fairly balanced nature of the capacitive leakage to the frame and surrounding ground he was stood upon. Once you've strapped either end of the series connected windings to the frame and earth, any resulting leakage current would merely be microampres in magnitude and of no consequence.

In this case, it seems that the two windings have been linked together in the generator head with only the 220v connections brought out to the panel. IME, it seems to be standard practice to leave the generator windings floating so that the end user remains free to deal with this problem as they deem appropriate.

Whilst it's true that the waveshape can noticablly depart from a pure sinewave when under load, that's not really the issue. The real issue is such generators' reaction to, well, reactive loads (leading current, capacitive loads to be more specific) whereby they go into an uncontrollable overvolt condition which the AVR cannot compensate for.

Lagging current (inductive) loads aren't a problem in this regard other than the more prosaic issue of low PF loadings reducing the usable real power output (wattage) in the face of a large reactive current demand on the total VA available from the genset.

[1] I may have misinterpreted exactly what Paul was trying to describe so the question of a bi-phase supply on his property is most probably a red herring.

[John Rumm]

Although reading what I wrote there, does not actually make sense to me since I think I missed a sentence out somewhere ;-)

For clarification, the connection between earth and one end of the winding obviously has nothing to do with the transfer switch. That is required for the final connection of the genny to the house electrics.

[John Rumm]

So long as there is no "normal" mains input to the UPS, that ought to prevent it from trying to "charge" your genny like its a battery.

Some UPSs can be difficult to power up without mains present in the first place though.

Power throughput should match the VA rating of the UPS so long as the genny can supply adequate amps.

[John Rumm]

Yup that makes sense when the socket on the genny has no preferred live and neutral pins. Typically we are dealing with setups using UK style connectors which have very definite polarisations.

(might be worth labelling the input carefully though so that someone in future knows not to connect a genny that already has a link made potentially the other way around)

;-)

[Johny B Good]

5 ohms seems rather a high resistance on a 220v 2500VA genset (645 watts at maximum load dissipated in the stator windings (over 20% of its 2500VA output wasted as I squared R losses).

Ignoring the 50 or 60 watts maximum dissipation in the rotor field winding, I'd expect no more than 250W to be safely dissipated in the stator windings at full load (assuming a 90% efficiency for the genset head - a not unreasonable figure for a small commodity generator).

I'd have anticipated a much lower reading (2 ohms or so, maybe as high as 3 for a really inefficient stressed to the limit design). I'd expect it's the meter that's giving a high reading rather than the genset being only (electrically) 78% efficient at full load.

[Paul]

That was with the 3 quid multimeter recommended hereabouts ;-)

I'll try it later with the Avometer.

[Fredxxx]

The meter may be 1 ohm or so out, but I would be surprised if the measurements aren't incorrect.

If you think about it, efficiency is wholly dependent on size, and so it's hardly surprising on a cut price generator. It only has a cursory specification to fuel usage, where 80% electro/mechanical efficiency is actually quite good.