Filtering harmonic distortion on a 2.8KV 50Hz 230v Petrol Generator to satisfy fussy UPS

[Johnny B Good]

Hi everyone.

I've done some more research into the generator / UPS compatability
issue and can now say that I was totally on the wrong in my original
approach.

It took quite a bit of experimentation (and some serendipity) to
discover that the UPS isn't particularly concerned about the harmonic
content of the incoming supply voltage waveform.

The main parameters of concern to a UPS are the frequency (the
requirement being quite a loose tolerance of +/- 5% on either a 50 or
60Hz supply) and the voltage remaining within predefined set limits.

In my case, although the generator was running just beyond the upper
frequency limit as initially set up, this was quite easily adjusted, and
the real problem I was having turned out to a consequent of capacitive
loading on the generator's output.

Now, despite the fact that the generator design involved a conventional
2 pole rotating field energised via sliprings from an AVR module, it is
still susceptable to the self excitation effect derived from the leading
current due to the capacitor loading presented in the form of a pair of
4.7uF capacitors in the UPS's mains input circuit.

Effectively, this capacitive self excitation was hijacking the AVR
control by sending the output voltage right up to 280v, some 50 volts
above the AVR mediated setting.

What was happening was that the UPS would see the reappearance of the
genset supplied mains voltage at acceptable voltage and frequency and
attempt to switch from battery power back to mains, whereupon the
9.4uF's worth of capacitive loading would appear across the supply
causing the voltage to jump from its nice steady 230v to somewhere in
the region of 270 to 280v, way beyond the buck boost regulation range of
the UPS, forcing an immediate return to battery power allowing the
generator to once more stabilise at its AVR mediated 230 volt level
which then initiated another attempt by the UPS to switch back to mains
power and a repeat performance.

Now one way to mitigate this effect is to preload the generator output
with a suitably inductive load (an inductor wired across the output).
This will work but a 230v 50Hz rated 500mH (or lower) inductor is a far
from standard item. The best I could do was to parallel a bunch of 4H
transformer primaries (some 8 400VA transformers in all). This allowed
the unloaded UPS to switch back to generator power and remain in that
state but the loaded state would cause it to cycle between battery and
mains as before.

In this case, the problem is almost certainly due to the fact the load
actually consists of yet another UPS providing a second level of
protection to my computers and I suspect that this UPS, an Upsonic
UPS600, also has its own bunch of mains input capactors to further
agravate the capacitive loading problem. Unfortunately, I don't have the
circuit diagram for this model so can only surmise at the existance of
the extra capacitance.

I haven't, as yet, removed the redundent Upsonic UPS600 from the power
supply chain to test this theory but, when I do, I'm going to fit an
autotransformer in its place to reduce the voltage from a nominal 230
down to around the 190 to 200v mark to minimise damage to the VDR spike
protection components in the PC's PSUs.

Whilst this might provide a solution, I think the real culprits behind
this universally experienced UPS compatabilty issue are the cheapjack
penny pinching genset manufacturers who seem to have elected to use an
AVR circuit derived from automotive alternator design practice.

The automotive alternator doesn't have to contend with the possibility
of self excitation effect from capacitive loading due to the 3 phase
stator windings being directly connected to a 3 phase bridge rectifier
pack, thus allowing an AVR module that only needs to control excitation
current via a single series pass power transistor (effectively, half an
output stage) in just one direction only.

When the AVR has to contend with capactively induced self excitation,
it seems to me that it needs to be able to counter this effect and this
requires that the AVR is able to drive excitation current in _both_
directions. This can only be achieved with an output that can both
source and sink current which requires a push/pull output stage of two
series control power transistors as per a single ended audio power amp.

A 'drop in' enhanced regulator module capable of using the exisitng
excitation supply voltage and field assembly would need to be a bridged
output design (i.e. four output transitors). Although it might be argued
that this raises complexity and costs, the additional cost becomes
rather swamped out by the costs of the gross parts of the genset itself.
The complexity issue is a non- issue once a new module design has been
committed to mass production anyway.

However, the idea of using such an 'improved' AVR module might have
some fundamental flaw which I've not been able to discern (but which an
experienced genset designer might be able to point out), so I'd be quite
happy to listen to anyone who can put me straight on this matter.

So, for anyone else, whose dream of providing extended UPS runtimes
courtesy of a cheap petrol genset, have been well and truly shattered by
the genset/UPS compatabilty issue, there are two options that can be
tried.

The first being the fitting of a suitable 250mH mains voltage rated
inductor (230v 2 to 4KVA case) across the genset's output (inductive
loading is not a voltage stability issue) or else, figure out a drop-in
AVR module design capable of bucking the self excitation effect from the
capacitive loading of the UPS(es). This last _might_ be a non-starter
but the first does definitely work.

I have considered yet another means of addressing the issue. Since the
waveshape of the supply isn't particularly critical, fitting clipping
diodes across the genset's output to clip at 350v peak for a 230v supply
(and 180v peak for a 115/120v suply) might prove effective.

In practice, a stepdown transformer to allow 'amplified zenner diodes'
operating at lower voltages/higher currents would be a more workable
solution. the transformer leakage inductance helping to take the sting
out of the current spikes in the clipping diodes as well as countering
the capacitive effect.

That's three possible workarounds. Of the three, unless I _have_ missed
a fundamental flaw, the improved AVR module is the most elegant
solution.

HTH & HAND

 

[EXT]

Hi everyone.

I've done some more research into the generator / UPS compatability
issue and can now say that I was totally on the wrong in my original
approach.

It took quite a bit of experimentation (and some serendipity) to
discover that the UPS isn't particularly concerned about the harmonic
content of the incoming supply voltage waveform.

The main parameters of concern to a UPS are the frequency (the
requirement being quite a loose tolerance of +/- 5% on either a 50 or
60Hz supply) and the voltage remaining within predefined set limits.

In my case, although the generator was running just beyond the upper
frequency limit as initially set up, this was quite easily adjusted,
and the real problem I was having turned out to a consequent of
capacitive loading on the generator's output.

Now, despite the fact that the generator design involved a
conventional 2 pole rotating field energised via sliprings from an
AVR module, it is still susceptable to the self excitation effect
derived from the leading current due to the capacitor loading
presented in the form of a pair of
4.7uF capacitors in the UPS's mains input circuit.

Effectively, this capacitive self excitation was hijacking the AVR
control by sending the output voltage right up to 280v, some 50 volts
above the AVR mediated setting.

What was happening was that the UPS would see the reappearance of the
genset supplied mains voltage at acceptable voltage and frequency and
attempt to switch from battery power back to mains, whereupon the
9.4uF's worth of capacitive loading would appear across the supply
causing the voltage to jump from its nice steady 230v to somewhere in
the region of 270 to 280v, way beyond the buck boost regulation range
of the UPS, forcing an immediate return to battery power allowing the
generator to once more stabilise at its AVR mediated 230 volt level
which then initiated another attempt by the UPS to switch back to
mains power and a repeat performance.

Now one way to mitigate this effect is to preload the generator output
with a suitably inductive load (an inductor wired across the output).
This will work but a 230v 50Hz rated 500mH (or lower) inductor is a
far from standard item. The best I could do was to parallel a bunch
of 4H transformer primaries (some 8 400VA transformers in all). This
allowed the unloaded UPS to switch back to generator power and remain
in that state but the loaded state would cause it to cycle between
battery and mains as before.

In this case, the problem is almost certainly due to the fact the load
actually consists of yet another UPS providing a second level of
protection to my computers and I suspect that this UPS, an Upsonic
UPS600, also has its own bunch of mains input capactors to further
agravate the capacitive loading problem. Unfortunately, I don't have
the circuit diagram for this model so can only surmise at the
existance of the extra capacitance.

I haven't, as yet, removed the redundent Upsonic UPS600 from the power
supply chain to test this theory but, when I do, I'm going to fit an
autotransformer in its place to reduce the voltage from a nominal 230
down to around the 190 to 200v mark to minimise damage to the VDR
spike protection components in the PC's PSUs.

Whilst this might provide a solution, I think the real culprits
behind this universally experienced UPS compatabilty issue are the
cheapjack penny pinching genset manufacturers who seem to have
elected to use an AVR circuit derived from automotive alternator
design practice.

The automotive alternator doesn't have to contend with the possibility
of self excitation effect from capacitive loading due to the 3 phase
stator windings being directly connected to a 3 phase bridge rectifier
pack, thus allowing an AVR module that only needs to control
excitation current via a single series pass power transistor
(effectively, half an output stage) in just one direction only.

When the AVR has to contend with capactively induced self excitation,
it seems to me that it needs to be able to counter this effect and
this requires that the AVR is able to drive excitation current in
_both_ directions. This can only be achieved with an output that can
both source and sink current which requires a push/pull output stage
of two series control power transistors as per a single ended audio
power amp.

A 'drop in' enhanced regulator module capable of using the exisitng
excitation supply voltage and field assembly would need to be a
bridged output design (i.e. four output transitors). Although it
might be argued that this raises complexity and costs, the additional
cost becomes rather swamped out by the costs of the gross parts of
the genset itself. The complexity issue is a non- issue once a new
module design has been committed to mass production anyway.

However, the idea of using such an 'improved' AVR module might have
some fundamental flaw which I've not been able to discern (but which
an experienced genset designer might be able to point out), so I'd be
quite happy to listen to anyone who can put me straight on this
matter.

So, for anyone else, whose dream of providing extended UPS runtimes
courtesy of a cheap petrol genset, have been well and truly shattered
by the genset/UPS compatabilty issue, there are two options that can
be tried.

The first being the fitting of a suitable 250mH mains voltage rated
inductor (230v 2 to 4KVA case) across the genset's output (inductive
loading is not a voltage stability issue) or else, figure out a
drop-in AVR module design capable of bucking the self excitation
effect from the capacitive loading of the UPS(es). This last _might_
be a non-starter but the first does definitely work.

I have considered yet another means of addressing the issue. Since the
waveshape of the supply isn't particularly critical, fitting clipping
diodes across the genset's output to clip at 350v peak for a 230v
supply (and 180v peak for a 115/120v suply) might prove effective.

In practice, a stepdown transformer to allow 'amplified zenner diodes'
operating at lower voltages/higher currents would be a more workable
solution. the transformer leakage inductance helping to take the sting
out of the current spikes in the clipping diodes as well as countering
the capacitive effect.

That's three possible workarounds. Of the three, unless I _have_
missed a fundamental flaw, the improved AVR module is the most elegant
solution.

HTH & HAND

As in everything, there are the cheap and crappy, and there are the
expensive with quality -- somewhere in the middle is the reasonable and
good-enough. It is hard to determine this in a generator - but - slip rings
is a giveaway for cheap with poor quality. Generator sets consist of an
engine and a generator head. You really need to buy a generator set that
combines a quality engine such as a Honda along with a good brushless
generator head, preferably a brand name. You cannot just look at one
component and ignore the other.

 

[Johnny B Good]

The message <[email protected]>

====big snip====

As in everything, there are the cheap and crappy, and there are the
expensive with quality -- somewhere in the middle is the reasonable and
good-enough. It is hard to determine this in a generator - but - slip rings
is a giveaway for cheap with poor quality. Generator sets consist of an
engine and a generator head. You really need to buy a generator set that
combines a quality engine such as a Honda along with a good brushless
generator head, preferably a brand name. You cannot just look at one
component and ignore the other.

All that is true, but a slipring fed field design need not produce poor
results. The even cheaper designs based on an adapted asynchronous motor
which relies on the capacitive effect to excite the necessary field
currents and determine the output voltage for a given spin speed have
even worse regulation and frequency stability.

I know about the slipring eliminating designs that allow an external
AVR to regulate a synchronously generated voltage to the same precision
as that of the 'cheap slipring' designs. This type is fine when you need
very long generator head lifetimes and greater reliability. However, for
a cheap commodity genset, the life of the slipring and brush gear will
far exceed the life of the prime mover.

The charm of the slipring fed arrangement is that the AVR can be
designed to buck the capacitively induced excitation current (assuming I
haven't missed some vital fact that defeats this 'simple concept'),
whereas the brushless type would need the AVR circuitry to be
incorporated into the rotating field assembly itself.

Although I have no circuit diagrams for this PowerCraft generator, I
suppose I could try a slightly more sophisticated version of my original
12v battery powered excitation test (the one where I was trying to
determine the cause of the 1500Hz ripple component that stands out so
markedly on no load).

In this case I could make a simple Bridged output DC amplifier driven
from a potentiometer fed by a suitable bias voltage that would allow me
to feed the field with voltages over the range +12 through zero to -12v
whilst capacitively loading the generator output. This basic test should
reveal whether it _is_ possible to negate the effects of capacitive self
excitation by a modified AVR circuit.

The trouble is, I have a horrible feeling that the problem I'm trying
to address is possibly not quite so simple as it seems. The one fact
that looms large in my mind being that, aside from residual magnetism
considerations, it doesn't matter which way the current flows in the
field winding, you'll get AC output regardless.

The only control you then have being the magnitude and phase relative
to the poles on the rotor. I suspect an AVR module capable of driving
excitation current in either direction is only going to result in some
rather wild voltage amplitude excursions. I really could do with some
expert advice on this matter.

The fact that the capacitive loading issue effect on power station
plant at Black Start capable stations which can arise with long length
unloaded transmission lines has been mentioned elsewhere in this
newsgroup rather suggests that any attempts by the AVR module to buck
the self excitation effect will be doomed to failure.

I suspect that using an inductive load to swamp out random capacitive
loadings might be the only practical solution after all. Removing the
second UPS loading from the primary UPS's output circuit might be all
that is required to achieve sufficient succes with the inductive load
I've already assembled to make it worth the effort of getting hold of a
purpose made inductor to make the whole scheme work.

Since I want to retire the old Upsonic anyway, I'll repeat my tests
next chance I get after doing so. I might have some encouraging news to
report in a few days time.

 

[Roger_Nickel]

The message <[email protected]>


====big snip====




All that is true, but a slipring fed field design need not produce poor
results. The even cheaper designs based on an adapted asynchronous motor
which relies on the capacitive effect to excite the necessary field
currents and determine the output voltage for a given spin speed have
even worse regulation and frequency stability.

I know about the slipring eliminating designs that allow an external
AVR to regulate a synchronously generated voltage to the same precision
as that of the 'cheap slipring' designs. This type is fine when you need
very long generator head lifetimes and greater reliability. However, for
a cheap commodity genset, the life of the slipring and brush gear will
far exceed the life of the prime mover.

The charm of the slipring fed arrangement is that the AVR can be
designed to buck the capacitively induced excitation current (assuming I
haven't missed some vital fact that defeats this 'simple concept'),
whereas the brushless type would need the AVR circuitry to be
incorporated into the rotating field assembly itself.

Although I have no circuit diagrams for this PowerCraft generator, I
suppose I could try a slightly more sophisticated version of my original
12v battery powered excitation test (the one where I was trying to
determine the cause of the 1500Hz ripple component that stands out so
markedly on no load).

In this case I could make a simple Bridged output DC amplifier driven
from a potentiometer fed by a suitable bias voltage that would allow me
to feed the field with voltages over the range +12 through zero to -12v
whilst capacitively loading the generator output. This basic test should
reveal whether it _is_ possible to negate the effects of capacitive self
excitation by a modified AVR circuit.

The trouble is, I have a horrible feeling that the problem I'm trying
to address is possibly not quite so simple as it seems. The one fact
that looms large in my mind being that, aside from residual magnetism
considerations, it doesn't matter which way the current flows in the
field winding, you'll get AC output regardless.

The only control you then have being the magnitude and phase relative
to the poles on the rotor. I suspect an AVR module capable of driving
excitation current in either direction is only going to result in some
rather wild voltage amplitude excursions. I really could do with some
expert advice on this matter.

The fact that the capacitive loading issue effect on power station
plant at Black Start capable stations which can arise with long length
unloaded transmission lines has been mentioned elsewhere in this
newsgroup rather suggests that any attempts by the AVR module to buck
the self excitation effect will be doomed to failure.

I suspect that using an inductive load to swamp out random capacitive
loadings might be the only practical solution after all. Removing the
second UPS loading from the primary UPS's output circuit might be all
that is required to achieve sufficient succes with the inductive load
I've already assembled to make it worth the effort of getting hold of a
purpose made inductor to make the whole scheme work.

Since I want to retire the old Upsonic anyway, I'll repeat my tests
next chance I get after doing so. I might have some encouraging news to
report in a few days time.

AVR output into the field is rectified AC?. If so then then there may be
a phasing problem . Your idea of DC exciting the field seems good.
Suggest that you do the correction in the feedback circuit rather than
hanging phase shifting components on the generator output. How much field
current do you need?. It only needs to be unipolar DC and the DC source
can be derived from the generator output. Automotive alternators get
their field current from the battery---so no issue with out of phase AC
components on the field current.

 

[Johnny B Good]

The message <[email protected]>

I rather suspect I may have answered the "vital fact that defeats this
'simple concept'" question above in the third and fourth paragraphs
below. :-(

The reason I think this can work is that it seems the self excitation
problem only occurs with leading current and not lagging (inductive)
current loadings.

AVR output into the field is rectified AC?. If so then then there may be
a phasing problem . Your idea of DC exciting the field seems good.
Suggest that you do the correction in the feedback circuit rather than
hanging phase shifting components on the generator output. How much field
current do you need?. It only needs to be unipolar DC and the DC source
can be derived from the generator output. Automotive alternators get
their field current from the battery---so no issue with out of phase AC
components on the field current.

Yes, undoubtedly, the current source used by the AVR module is a low
voltage (around 50 volts or so) rectified AC (this genset lacks electric
start, relying on recoil starting, hence no handy battery to eliminate
the reliance on residual magnetism), but this rectified ac source will
include a basic smoothing circuit (just a suitably sized electrolytic
cap across the rectifier output would be my guess), so the issue of the
phase of the ripple relative to the excitation current being fed to the
field won't arise. For the purposes of this exercise, the AVR might just
as well be energised from a battery.

When I was testing to determine whether the 1500Hz noise on no-load was
the result of a modulation of excitation current from the AVR module, I
disconnected the AVR and attached a 12v dryfit battery directly to the
slipring contacts. Fortuitously, this was just the right amount of
excitation to produce 230v on no-load (which dropped to about 100v when
loaded with my 2.5KW electric kettle test load).

Now, the important point to consider in this test is that I'd have
gotten the same result whichever way round I chose to connect the
battery. The subtle difference of phase relative to a fixed spot on the
generator shaft would have required the use of a strobe lamp to spot.

As far as the load is concerned, the polarity of the excitation current
is totally immaterial, only its magnitude being the deciding factor in
just how strong the output voltage would be for any given resistive (or
inductive) loading.

Unfortunately, capacitive loading above a trivial amount (a 4.7uF
capacitor across the 230v output was enough to raise the no-load voltage
up to some 270 odd volts which no amount of adjustment on the AVR's
trimpot would compensate). IOW, the AVR's job had been hijacked by a
capactive load that was half that which the UPS was switching in when
attempting to switch from battery power back to generator power.

I know the UPS will continue to run in pass through mode when the
Upsonic UPS (and its load) are disconnected and I have those 8 400VA
230v transformer primaries (about 500mH's worth of inductance) wired
across the generator's output so my hope is that if I remove that second
UPS from the loading equation, I might discover that I've cracked the
problem.

I'm planning on retiring that Upsonic unit anyway (it takes about 18
watts maintaining power when the battery packs are in good condition
which they no longer are so I suspect it might be taking a little bit
more than 18 watts by now).

The last endurance test I ran on this unit a couple of days back lasted
just over 11 minutes on a 200 watt load which doesn't compare very well
to a test (with new batteries 11 years ago) which ran a 394 watt load
for over 33 minutes.

Since I'd need to buy at least three 7AH 12v dryfit batteries to
refurbish just one of the two banks it can have installed and it has
started to mimic the SFX used in motion picture productions of the 40s
depicting scenes of Dr Frankenstein's Lab _and_ is a bit of overkill
anyway, I am going to decommision it in the next day or so.

Once the Upsonic is out of the picture, I'll repeat my generator
testing yet again to see if _this_ time around the inductive loading
trick will actually work. If it does, you (collectively) will be the
first to know.

 

[Johnny B Good]

The message <[email protected]>
from Roger_Nickel <[email protected]> contains these words:

====snip====

AVR output into the field is rectified AC?. If so then then there may be
a phasing problem . Your idea of DC exciting the field seems good.
Suggest that you do the correction in the feedback circuit rather than
hanging phase shifting components on the generator output. How much field
current do you need?. It only needs to be unipolar DC and the DC source
can be derived from the generator output. Automotive alternators get
their field current from the battery---so no issue with out of phase AC
components on the field current.

Oops! forgot to address this last item. In the automotive generator
case, the capacitive loading issue simply cannot arise since its output
is DC (admittedly with a low level of voltage ripple from its 3 phase
bridge rectifier pack which ripple voltage is largely suppressed by its
direct connection to the battery).

The capacitive load issue only applies when using the ac output
directly or via a step up or step down transformer.

Now that I've got a clearer picture of the problem in my mind, I'm
going to google for more information.

 

[daestrom]

Hi everyone.

I've done some more research into the generator / UPS compatability
issue and can now say that I was totally on the wrong in my original
approach.

It took quite a bit of experimentation (and some serendipity) to
discover that the UPS isn't particularly concerned about the harmonic
content of the incoming supply voltage waveform.

The main parameters of concern to a UPS are the frequency (the
requirement being quite a loose tolerance of +/- 5% on either a 50 or
60Hz supply) and the voltage remaining within predefined set limits.

In my case, although the generator was running just beyond the upper
frequency limit as initially set up, this was quite easily adjusted, and
the real problem I was having turned out to a consequent of capacitive
loading on the generator's output.

Now, despite the fact that the generator design involved a conventional
2 pole rotating field energised via sliprings from an AVR module, it is
still susceptable to the self excitation effect derived from the leading
current due to the capacitor loading presented in the form of a pair of
4.7uF capacitors in the UPS's mains input circuit.

Effectively, this capacitive self excitation was hijacking the AVR
control by sending the output voltage right up to 280v, some 50 volts
above the AVR mediated setting.

What was happening was that the UPS would see the reappearance of the
genset supplied mains voltage at acceptable voltage and frequency and
attempt to switch from battery power back to mains, whereupon the
9.4uF's worth of capacitive loading would appear across the supply
causing the voltage to jump from its nice steady 230v to somewhere in
the region of 270 to 280v, way beyond the buck boost regulation range of
the UPS, forcing an immediate return to battery power allowing the
generator to once more stabilise at its AVR mediated 230 volt level
which then initiated another attempt by the UPS to switch back to mains
power and a repeat performance.

Now one way to mitigate this effect is to preload the generator output
with a suitably inductive load (an inductor wired across the output).
This will work but a 230v 50Hz rated 500mH (or lower) inductor is a far
from standard item. The best I could do was to parallel a bunch of 4H
transformer primaries (some 8 400VA transformers in all). This allowed
the unloaded UPS to switch back to generator power and remain in that
state but the loaded state would cause it to cycle between battery and
mains as before.

In this case, the problem is almost certainly due to the fact the load
actually consists of yet another UPS providing a second level of
protection to my computers and I suspect that this UPS, an Upsonic
UPS600, also has its own bunch of mains input capactors to further
agravate the capacitive loading problem. Unfortunately, I don't have the
circuit diagram for this model so can only surmise at the existance of
the extra capacitance.

I haven't, as yet, removed the redundent Upsonic UPS600 from the power
supply chain to test this theory but, when I do, I'm going to fit an
autotransformer in its place to reduce the voltage from a nominal 230
down to around the 190 to 200v mark to minimise damage to the VDR spike
protection components in the PC's PSUs.

Whilst this might provide a solution, I think the real culprits behind
this universally experienced UPS compatabilty issue are the cheapjack
penny pinching genset manufacturers who seem to have elected to use an
AVR circuit derived from automotive alternator design practice.

If it were simply an automotive design, the regulator could reduce the
field current to nearly zero and you wouldn't be having this problem.

Most likely the regulator just can't reduce field current down low
enough. If the regulator were the only thing feeding the field current,
it should be able to almost completely shut off the field in order to
maintain voltage. But because that would require a wide range of power
from the regulator, it is common to use some current-transformers to
boost the field current as load is increased. That way the regulator
only has to supply a small portion of the field power. You may be able
to find this second source of field power and adjust it.

A simple fact of physics with synchronous generators is that capacitive
loads will have a armature current that leads the terminal voltage, and
that current will create a voltage *rise* across the stator winding's
reactance. Large utility units have circuits to prevent the generator
from operating too far in the leading power factor direction because it
can cause extra heating within the machine.

daestrom

 

[daestrom]

The message <[email protected]>


====big snip====



All that is true, but a slipring fed field design need not produce poor
results. The even cheaper designs based on an adapted asynchronous motor
which relies on the capacitive effect to excite the necessary field
currents and determine the output voltage for a given spin speed have
even worse regulation and frequency stability.

I know about the slipring eliminating designs that allow an external
AVR to regulate a synchronously generated voltage to the same precision
as that of the 'cheap slipring' designs. This type is fine when you need
very long generator head lifetimes and greater reliability. However, for
a cheap commodity genset, the life of the slipring and brush gear will
far exceed the life of the prime mover.

The charm of the slipring fed arrangement is that the AVR can be
designed to buck the capacitively induced excitation current (assuming I
haven't missed some vital fact that defeats this 'simple concept'),
whereas the brushless type would need the AVR circuitry to be
incorporated into the rotating field assembly itself.

Again with 'capacitively induced excitation current'. The rotor is
moving at synchronous speed with the magnetic field from the armature
current and so the armature's magnetic field can *not* induce any
currents into the rotor, regardless of leading/lagging.

This same slip-ring design is used on all sorts of generators, big and
small. The problem you describe is because the terminal voltage is the
vector sum of the emf created by the air gap flux and the voltage drop
across the internal reactance of the machine. The combination of
inductive reactance and a phenomenon called armature reaction combine to
appear as a large inductance in series with the 'ideal' voltage created
by the magnetic flux.

With a resistive load/current, the voltage drop across the machine
reactance is at 90 degrees to the generated EMF and the resulting sum is
the terminal voltage. With a lagging load, the voltage drop across the
internal reactance is more than 90 degrees out of phase with the
generated EMF and the terminal voltage drops very quickly with
increasing load. With a leading load, the voltage drop across the
internal reactance is less than 90 degrees out of phase and it adds
considerably to the EMF, resulting in a rise in terminal voltage.

The key is how much control the AVR has over the field current. If
there is a current-transformer-supplied boost to the field (to limit the
size requirements of the regulator), then it alone will provide more
field current under leading power factor conditions than is needed for
rated voltage output.

If you figure out what field current is required to get rated voltage
with no load, you'll find that you will have to lower the field current
*below* that point to maintain rated voltage with a largely capacitive
load.

Although I have no circuit diagrams for this PowerCraft generator, I
suppose I could try a slightly more sophisticated version of my original
12v battery powered excitation test (the one where I was trying to
determine the cause of the 1500Hz ripple component that stands out so
markedly on no load).

In this case I could make a simple Bridged output DC amplifier driven
from a potentiometer fed by a suitable bias voltage that would allow me
to feed the field with voltages over the range +12 through zero to -12v
whilst capacitively loading the generator output. This basic test should
reveal whether it _is_ possible to negate the effects of capacitive self
excitation by a modified AVR circuit.

Fire it up, connect the caps, and open-circuit the field (put a switch
in the lead to one of the brushes). Voltage decay to near zero
(depending on the amount of residual magnetism). The capacitive load by
itself cannot supply excitation unless there is some form of boost
circuit that takes a portion of the generator output current and feeds
it to the field.

The trouble is, I have a horrible feeling that the problem I'm trying
to address is possibly not quite so simple as it seems. The one fact
that looms large in my mind being that, aside from residual magnetism
considerations, it doesn't matter which way the current flows in the
field winding, you'll get AC output regardless.

That is certainly true.

The only control you then have being the magnitude and phase relative
to the poles on the rotor. I suspect an AVR module capable of driving
excitation current in either direction is only going to result in some
rather wild voltage amplitude excursions. I really could do with some
expert advice on this matter.

No need to try and reverse the current. Just need a way to be able to
lower it down to zero.

The fact that the capacitive loading issue effect on power station
plant at Black Start capable stations which can arise with long length
unloaded transmission lines has been mentioned elsewhere in this
newsgroup rather suggests that any attempts by the AVR module to buck
the self excitation effect will be doomed to failure.

The problem we have with unloaded transmission lines is that when you
energize them at one end at nominal voltage, the other end can be higher
than nominal. Seen from the substation, the unloaded line is quite high
in voltage and as it is loaded the voltage drops quite a bit.

But the *generator* end of the line, with a regulator on the generator
that is capable of varying the field current below the 'no-load' value,
has no problem.

daestrom

 

[daestrom]

The message <[email protected]>
from Roger_Nickel <[email protected]> contains these words:

Yes, undoubtedly, the current source used by the AVR module is a low
voltage (around 50 volts or so) rectified AC (this genset lacks electric
start, relying on recoil starting, hence no handy battery to eliminate
the reliance on residual magnetism), but this rectified ac source will
include a basic smoothing circuit (just a suitably sized electrolytic
cap across the rectifier output would be my guess), so the issue of the
phase of the ripple relative to the excitation current being fed to the
field won't arise.

Quite right. Besides, the field winding itself is a large inductance
that will filter out the ripple. In utility-sized units we use the AC
output and feed it through phase-controlled SCR's to control the field.
This sort of chopped sine wave supply is terrible for a DC power
source, yet we feed it directly to the field with a 'free-wheeling'
rectifier to allow the field current to bypass the SCR's during the part
of the cycle when they are off. Works like a charm.

When I was testing to determine whether the 1500Hz noise on no-load was
the result of a modulation of excitation current from the AVR module, I
disconnected the AVR and attached a 12v dryfit battery directly to the
slipring contacts. Fortuitously, this was just the right amount of
excitation to produce 230v on no-load (which dropped to about 100v when
loaded with my 2.5KW electric kettle test load).

Now, the important point to consider in this test is that I'd have
gotten the same result whichever way round I chose to connect the
battery. The subtle difference of phase relative to a fixed spot on the
generator shaft would have required the use of a strobe lamp to spot.

As far as the load is concerned, the polarity of the excitation current
is totally immaterial, only its magnitude being the deciding factor in
just how strong the output voltage would be for any given resistive (or
inductive) loading.

Another proof of this is that on mid-sized machines, we reverse the two
leads going to the slip rings so that the + and - leads are connected to
opposite rings for about six months and then reverse them back. This
helps prolong the slip-rings as always having one polarity causes a form
of electro-chemical wear on one ring faster than the other. So
periodically reversing them evens out the wear.

daestrom

 

[Johnny B Good]

The message <[email protected]>

====big snip====

If it were simply an automotive design, the regulator could reduce the
field current to nearly zero and you wouldn't be having this problem.

I'm sure you're right. My problem is that this subject is way more
complex than I'd originally imagined. I have googled for information
regarding alternator performance issues with reactive (notably,
capacitive) loadings and my head hurts. :-(

Most likely the regulator just can't reduce field current down low
enough. If the regulator were the only thing feeding the field current,
it should be able to almost completely shut off the field in order to
maintain voltage. But because that would require a wide range of power
from the regulator, it is common to use some current-transformers to
boost the field current as load is increased. That way the regulator
only has to supply a small portion of the field power. You may be able
to find this second source of field power and adjust it.

I guess you're referring to the equivilent of a compound DC machine
where a combination of parallel and series field windings are used to
either stabilise voltage output in the generator case, or rpms in the
motor case.

Unfortunately, I don't have a circuit diagram to peruse. However, if
the excitation voltage supply to the AVR module is, in part, derived
from a current transformer to boost avialable field voltage (leaving the
AVR to act as a 'trimming circuit'), I could test this by powering the
field solely from a seperate battery source and test whether capacitive
loading still produces the same symptom.

A simple fact of physics with synchronous generators is that capacitive
loads will have a armature current that leads the terminal voltage, and
that current will create a voltage *rise* across the stator winding's
reactance. Large utility units have circuits to prevent the generator
from operating too far in the leading power factor direction because it
can cause extra heating within the machine.

Yes, this was one of the many issues I saw described in my google quest.

BTW, I did get around to removing that 'surplus' UPS from the main
UPS's load and ran some more tests yesterday. The initial behaviour from
the master UPS was more encouraging in that it would pass the generator
power through for 15 to 30 seconds at a time between 'flipping out' and
briefly resorting to battery power. On balance, it was spending more
time on generator power than on battery.

When I went back into the office to check what was happening, the
remaining UPS, the little BackUPS 500 which protects the fileserver (a
57 watt load), was regularly switching to battery and back to 'mains'
power. The rate was faster than I was observing on the master UPS so no
obvious synchronicity, but I decided to transfer this UPS to the normal
mains feed to take it out of the load equation (and allow its single 7AH
12v dryfit battery to recover from the abuse it had recieved).

At that point my main desktop PC was still off. Switching it on didn't
make any obvious difference to the state of the protected supply, but a
return to the basement showed that the master UPS was now mainly running
off its battery so I shut the office PC down to allow the generator test
to run without further depleting the UPS battery.

I then decided to check how the master UPS would behave with a couple
of 150 Watt Tungsten filiament GLS lamps as a resistive load. The UPS
carried on as before but I did observe a strong 25Hz flicker on the
lamps which more or less disappeared when I added a third lamp to
increase this loading to 450 watts. There is obviously an issue of
stability with certain loads and the AVR module.

When I booted the office PC back up again, the damn master UPS was back
into its rapid cycling act again. With the lamp load to give a visual
clue (digital mulimeters just don't cut it), I could see a flick of
intensity as the generator output was being briefly switched through (at
which point I started cursing the inadequacies of small gensets and ATX
PSUs).

I then had no choice but to transfer the master UPS back to mains power
and shut the generator down and park it back in the garage (after
letting it cool down). TBH, I was rather horrified to discover that the
office PC was also proving to be a 'problematic load' which left me
rather downhearted over the whole exercise.

It was only an hour or so later that I discovered that my test load had
also inadvertently included the 'house PC' which I'd forgotten that I
had left running (the monitor, a 17 inch CRT, would have been in standby
at that point so no more than 100 watts load from the PC itself). This
mitigated my disappointment a little and leaves me wondering about the
fact that the office PC's PSU has had a PFC choke-ectomy[1].

Although some consider that narrow rectifier conduction angles (even
with a PFC choke) represents a less than unity PF loading, it can't
produce leading out of phase current. The current waveform may be spikey
but it is at least always in the same sense (or phase) as the voltage.

Whatever reactive loadings the typical ATX psu filter presents, it is
very rarely more than half a microfarad's worth, so it would seem that
the narrow conduction angles of the rectifier pack can create havoc with
the AVR module on a cheap and cheerful synchronous generator.

This could be tied in with what you were saying about the AVR module
being supplemented by a current transformer source (or equivilent) and
an overly long LR time constant in the field inductor circuit.

Having experienced the difficulties of using a simple synchronous
alternator to power electronic equipment, with their narrow conduction
angle loads (aside from random capacitive loadings introduced by UPSes),
I'm beginning to see the appeal of using gensets based on the PM
generator driving an electronic inverter to produce clean and stable 50
or 60 Hz pure sine wave power. Unfortunately, this type of genset
doesn't come cheap.

[1] I'm pretty certain I had to remove the PFC choke when I transplanted
the innards of the PSU into another PSU case that had its intake vent
slots where I needed them to be. I'll have to check to confirm this. If
I'm not mistaken, I can try adding the choke back into the circuit (even
if it means dangling it outside of the PSU case) and run another test.

Actually, it'll be simpler to just use another PC that _hasn't_ had its
PFC removed as a test load and _then_ decide whether it's worth
reinstating said choke.

 

[daestrom]

The message <[email protected]>


====big snip====



I'm sure you're right. My problem is that this subject is way more
complex than I'd originally imagined. I have googled for information
regarding alternator performance issues with reactive (notably,
capacitive) loadings and my head hurts. :-(


I guess you're referring to the equivilent of a compound DC machine
where a combination of parallel and series field windings are used to
either stabilise voltage output in the generator case, or rpms in the
motor case.

No, that's a different critter. Those have two windings, one shunt
winding (many turns of small wire) that is controlled and the other is
few turns of heavy wire that carries load current.

In many large AC regulators, they sometimes take a current transformer
from the output, rectify the secondary and put it in series with the
regulator so that it boosts the field current. Since the required field
current rises as you put load on the machine, and the current
transformer's secondary current also rises as you load the machine, this
arrangement supplies much of the additional power needed. By doing
this, the regulator only has to boost/buck the CT current a small amount
to 'fine-tune' the voltage output.

But the more I think on it, the more I think your unit is probably too
small for them to have used this sort of feature. But if you had a
drawing it would be easy to recognize it.

Unfortunately, I don't have a circuit diagram to peruse. However, if
the excitation voltage supply to the AVR module is, in part, derived
from a current transformer to boost avialable field voltage (leaving the
AVR to act as a 'trimming circuit'), I could test this by powering the
field solely from a seperate battery source and test whether capacitive
loading still produces the same symptom.

Exactly so.

A simple fact of physics with synchronous generators is that capacitive
loads will have a armature current that leads the terminal voltage, and
that current will create a voltage *rise* across the stator winding's
reactance. Large utility units have circuits to prevent the generator
from operating too far in the leading power factor direction because it
can cause extra heating within the machine.

Yes, this was one of the many issues I saw described in my google quest.

BTW, I did get around to removing that 'surplus' UPS from the main
UPS's load and ran some more tests yesterday. The initial behaviour from
the master UPS was more encouraging in that it would pass the generator
power through for 15 to 30 seconds at a time between 'flipping out' and
briefly resorting to battery power. On balance, it was spending more
time on generator power than on battery.

When I went back into the office to check what was happening, the
remaining UPS, the little BackUPS 500 which protects the fileserver (a
57 watt load), was regularly switching to battery and back to 'mains'
power. The rate was faster than I was observing on the master UPS so no
obvious synchronicity, but I decided to transfer this UPS to the normal
mains feed to take it out of the load equation (and allow its single 7AH
12v dryfit battery to recover from the abuse it had recieved).

At that point my main desktop PC was still off. Switching it on didn't
make any obvious difference to the state of the protected supply, but a
return to the basement showed that the master UPS was now mainly running
off its battery so I shut the office PC down to allow the generator test
to run without further depleting the UPS battery.

I then decided to check how the master UPS would behave with a couple
of 150 Watt Tungsten filiament GLS lamps as a resistive load. The UPS
carried on as before but I did observe a strong 25Hz flicker on the
lamps which more or less disappeared when I added a third lamp to
increase this loading to 450 watts. There is obviously an issue of
stability with certain loads and the AVR module.

When I booted the office PC back up again, the damn master UPS was back
into its rapid cycling act again. With the lamp load to give a visual
clue (digital mulimeters just don't cut it), I could see a flick of
intensity as the generator output was being briefly switched through (at
which point I started cursing the inadequacies of small gensets and ATX
PSUs).

I then had no choice but to transfer the master UPS back to mains power
and shut the generator down and park it back in the garage (after
letting it cool down). TBH, I was rather horrified to discover that the
office PC was also proving to be a 'problematic load' which left me
rather downhearted over the whole exercise.

It was only an hour or so later that I discovered that my test load had
also inadvertently included the 'house PC' which I'd forgotten that I
had left running (the monitor, a 17 inch CRT, would have been in standby
at that point so no more than 100 watts load from the PC itself). This
mitigated my disappointment a little and leaves me wondering about the
fact that the office PC's PSU has had a PFC choke-ectomy[1].

Although some consider that narrow rectifier conduction angles (even
with a PFC choke) represents a less than unity PF loading, it can't
produce leading out of phase current. The current waveform may be spikey
but it is at least always in the same sense (or phase) as the voltage.

There are two sub-parts to a non-unity power factor. You can use simple
passive reactances to shift the phase-angle, or you can have non-linear
devices (diodes, thyristors) that cause a significant harmonic content.
Since 'power factor' is defined as real-power / apparent power, either
of these forms of distortion qualify as a non-unity power factor.

Of course capacitive or inductive reactances can be cancelled out using
a suitable amount of the opposite reactance. But harmonic content has
to be filtered out using different techniques.

Most simple UPS's present a high harmonic load to the supply line
because of the simple, full-wave bridge design used to feed the internal
DC bus. There are higher-end models that have power-factor correction
components that basically filter out the harmonics.

AFAIK, if the UPS is of the continuous inverter type, it always supplies
the load from the inverter and the load's power factor isn't reflected
in the line supply current. But some UPS (and yours sounds like this
type), supply the load directly from the line and switch to the inverter
through a relay. Obviously the load on this type is 'seen' by the line
when its operating 'on the mains'.

Whatever reactive loadings the typical ATX psu filter presents, it is
very rarely more than half a microfarad's worth, so it would seem that
the narrow conduction angles of the rectifier pack can create havoc with
the AVR module on a cheap and cheerful synchronous generator.

This could be tied in with what you were saying about the AVR module
being supplemented by a current transformer source (or equivilent) and
an overly long LR time constant in the field inductor circuit.

Having experienced the difficulties of using a simple synchronous
alternator to power electronic equipment, with their narrow conduction
angle loads (aside from random capacitive loadings introduced by UPSes),
I'm beginning to see the appeal of using gensets based on the PM
generator driving an electronic inverter to produce clean and stable 50
or 60 Hz pure sine wave power. Unfortunately, this type of genset
doesn't come cheap.

[1] I'm pretty certain I had to remove the PFC choke when I transplanted
the innards of the PSU into another PSU case that had its intake vent
slots where I needed them to be. I'll have to check to confirm this. If
I'm not mistaken, I can try adding the choke back into the circuit (even
if it means dangling it outside of the PSU case) and run another test.

Actually, it'll be simpler to just use another PC that _hasn't_ had its
PFC removed as a test load and _then_ decide whether it's worth
reinstating said choke.

Have you considered putting a resistive load such as your tungsten lamps
on the generator *upstream* of the UPS's?? An additional, real power
load on the generator might help it stabilize.

daestrom

 

[Johnny B Good]

The message <[email protected]>

No, that's a different critter. Those have two windings, one shunt
winding (many turns of small wire) that is controlled and the other is
few turns of heavy wire that carries load current.

That's exactly what I meant, only instead of a series winding to carry
the load current and compensate for generator droop with increasing load
when driven at a fixed speed, the current transformer supplies
additional excitation current to help compensate for increased loading,
leaving the AVR the less arduous task of trimming the output voltage to
a more precise level than the CT circuit could do alone.

In many large AC regulators, they sometimes take a current transformer
from the output, rectify the secondary and put it in series with the
regulator so that it boosts the field current. Since the required field
current rises as you put load on the machine, and the current
transformer's secondary current also rises as you load the machine, this
arrangement supplies much of the additional power needed. By doing
this, the regulator only has to boost/buck the CT current a small amount
to 'fine-tune' the voltage output.

But the more I think on it, the more I think your unit is probably too
small for them to have used this sort of feature. But if you had a
drawing it would be easy to recognize it.

Exactly so.

A simple fact of physics with synchronous generators is that capacitive
loads will have a armature current that leads the terminal voltage, and
that current will create a voltage *rise* across the stator winding's
reactance. Large utility units have circuits to prevent the generator
from operating too far in the leading power factor direction because it
can cause extra heating within the machine.

Yes, this was one of the many issues I saw described in my google quest.

BTW, I did get around to removing that 'surplus' UPS from the main
UPS's load and ran some more tests yesterday. The initial behaviour from
the master UPS was more encouraging in that it would pass the generator
power through for 15 to 30 seconds at a time between 'flipping out' and
briefly resorting to battery power. On balance, it was spending more
time on generator power than on battery.

When I went back into the office to check what was happening, the
remaining UPS, the little BackUPS 500 which protects the fileserver (a
57 watt load), was regularly switching to battery and back to 'mains'
power. The rate was faster than I was observing on the master UPS so no
obvious synchronicity, but I decided to transfer this UPS to the normal
mains feed to take it out of the load equation (and allow its single 7AH
12v dryfit battery to recover from the abuse it had recieved).

At that point my main desktop PC was still off. Switching it on didn't
make any obvious difference to the state of the protected supply, but a
return to the basement showed that the master UPS was now mainly running
off its battery so I shut the office PC down to allow the generator test
to run without further depleting the UPS battery.

I then decided to check how the master UPS would behave with a couple
of 150 Watt Tungsten filiament GLS lamps as a resistive load. The UPS
carried on as before but I did observe a strong 25Hz flicker on the
lamps which more or less disappeared when I added a third lamp to
increase this loading to 450 watts. There is obviously an issue of
stability with certain loads and the AVR module.

When I booted the office PC back up again, the damn master UPS was back
into its rapid cycling act again. With the lamp load to give a visual
clue (digital mulimeters just don't cut it), I could see a flick of
intensity as the generator output was being briefly switched through (at
which point I started cursing the inadequacies of small gensets and ATX
PSUs).

I then had no choice but to transfer the master UPS back to mains power
and shut the generator down and park it back in the garage (after
letting it cool down). TBH, I was rather horrified to discover that the
office PC was also proving to be a 'problematic load' which left me
rather downhearted over the whole exercise.

It was only an hour or so later that I discovered that my test load had
also inadvertently included the 'house PC' which I'd forgotten that I
had left running (the monitor, a 17 inch CRT, would have been in standby
at that point so no more than 100 watts load from the PC itself). This
mitigated my disappointment a little and leaves me wondering about the
fact that the office PC's PSU has had a PFC choke-ectomy[1].

Although some consider that narrow rectifier conduction angles (even
with a PFC choke) represents a less than unity PF loading, it can't
produce leading out of phase current. The current waveform may be spikey
but it is at least always in the same sense (or phase) as the voltage.

There are two sub-parts to a non-unity power factor. You can use simple
passive reactances to shift the phase-angle, or you can have non-linear
devices (diodes, thyristors) that cause a significant harmonic content.
Since 'power factor' is defined as real-power / apparent power, either
of these forms of distortion qualify as a non-unity power factor.

Either forms of 'non-unity PF' are bad news to the Public Supply
Utilities (PSU) plant. Even 15 years ago, before the whole world and
their dog had a PC or three, I noticed that the PSU mains voltage
waveform was visibly distorted as displayed on an oscilloscope (the 50Hz
test signal from the 'scope's own transformer).

At first I thought this was distortion in the 'scope's transformer (it
was an ancient all valve(vacuum tube) design) but running it off the UPS
showed a dramatic improvement and a trace that was classic sinewave. I
was rather surprised that the mains supply departure from sinewave was
so bad it was readily visible as a softly clipped sinewave. However, at
least the harmonic content was essentially just odd order products with
almost no even harmonic content.

Of course capacitive or inductive reactances can be cancelled out using
a suitable amount of the opposite reactance. But harmonic content has
to be filtered out using different techniques.

The problem is best addressed in the smpsu ht rectifier pack by
incorporating the so called PFC choke between the rectifier output and
the HT smoothing capacitor(s), essentially a series filter element in a
2 pole LPF. This basically takes the sting out of the narrow conduction
angle that otherwise results when the rectifier feeds directly into the
smoothing capacitor(s), extending the conduction angle by a factor of
two or three times, depending on the PFC inductance value chosen against
the existing smoothing caps.

Most simple UPS's present a high harmonic load to the supply line
because of the simple, full-wave bridge design used to feed the internal
DC bus. There are higher-end models that have power-factor correction
components that basically filter out the harmonics.

AFAIK, if the UPS is of the continuous inverter type, it always supplies
the load from the inverter and the load's power factor isn't reflected
in the line supply current. But some UPS (and yours sounds like this
type), supply the load directly from the line and switch to the inverter
through a relay. Obviously the load on this type is 'seen' by the line
when its operating 'on the mains'.

Well, all my UPSes are of the line interactive type (I think that's the
'technical term') where the mains is simply switched through to the load
(with or without buck boost control) until the mains quality mandates
the UPS to switch its load to the inverter. The other type where the
load is _always_ driven by the inverter is far too expensive for me,
both in terms of acquisition and running costs. Indeed, you're unlikely
to see that type in any rating below 3KVA.

Whatever reactive loadings the typical ATX psu filter presents, it is
very rarely more than half a microfarad's worth, so it would seem that
the narrow conduction angles of the rectifier pack can create havoc with
the AVR module on a cheap and cheerful synchronous generator.

This could be tied in with what you were saying about the AVR module
being supplemented by a current transformer source (or equivilent) and
an overly long LR time constant in the field inductor circuit.

Having experienced the difficulties of using a simple synchronous
alternator to power electronic equipment, with their narrow conduction
angle loads (aside from random capacitive loadings introduced by UPSes),
I'm beginning to see the appeal of using gensets based on the PM
generator driving an electronic inverter to produce clean and stable 50
or 60 Hz pure sine wave power. Unfortunately, this type of genset
doesn't come cheap.

[1] I'm pretty certain I had to remove the PFC choke when I transplanted
the innards of the PSU into another PSU case that had its intake vent
slots where I needed them to be. I'll have to check to confirm this. If
I'm not mistaken, I can try adding the choke back into the circuit (even
if it means dangling it outside of the PSU case) and run another test.

Actually, it'll be simpler to just use another PC that _hasn't_ had its
PFC removed as a test load and _then_ decide whether it's worth
reinstating said choke.

Have you considered putting a resistive load such as your tungsten lamps
on the generator *upstream* of the UPS's?? An additional, real power
load on the generator might help it stabilize.

That was one of the first things I tried when I first discovered the
UPS/generator compatability issue. I seem to have had better results
with my makeshift 470mH inductor load (those 8 400VA transformer
primaries I've got stacked up in the basement). Even this suffers
surprisingly high losses, around 17 watts per transformer.

The trouble with using a resistive 'dummy load' is that it usually
needs to be 50 to 70 % of the generator's maximum continuous rating to
do any good, and even this 'quick fix' doesn't always solve the problem.
Also, such advice is usually coupled with the exhortation to use an
oversize generator (by a factor of 3 or more) for such loads.

Now me, I'd rather sort out a solution that more intelligently solves
the problem rather than one which simply uses sheer brute force to
hammer it into submission. The inductive loading looks the most
promising at the moment (but I've yet to run the battery excitation
test, so might discover a better solution in a genset mod).

I'd like to double up on those transformer primaries but I've only got
another 5 of those transformers to hand. Still, that might be enough to
indicate whether it would be worth making an air cored 220mH inductor.

I only need to get hold of some 600 yards of enamelled 10 AWG copper
wire according to the calculator here
<http://www.pronine.ca/multind.htm> if I want to keep the losses below
the 20 watt mark or I can use 1446 feet of 15 AWG wire if I decide to
allow the copper loss to rise to 51 watts (and more with the inevitable
increase in temperature in the coil). Actually, I might need to go for
a lower loss coil just to avoid the overheating issue.

Air cored inductors might require an expensive amount of wire as well
as producing some impressively large coils, but at least the design
parameters are pretty cut and dried compared to those required in air
gapped iron core designs.

A 220mH inductive load draws 3.3333 amps from a 230v sinewave supply
which is just less than a third of the generator's maximum continuous
rating. I think this is probably the best compromise to counter the
capacitive loadings without adding excessive inductive VA loading on
the generator's output.

At the end of the day, If I can find a simple (and, more importantly,
generic) solution that solves this issue of powering computer loads
(preferably including a line interactive UPS) from such cheap gensets,
this will not only benefit me but others who have also experienced
similar problems (as well as those who are considering adding a cheap
standby generator to their existing battery backup power system).

Incidently, winding an air cored inductor for use on a 120v 60Hz
generator is even easier since you'll only need to use half the number
of turns. Actually, for the same power rated genset (a 2.5KVA continuous
rated one in my case) running at 115v instead of 230 volt requires just
one quarter of the inductance (55mH instead of the 220mH I've chosen).
For a 120v 60Hz 2.5KVA rated genset, 50mH would be a suitable choice,
imo. 682 feet of 12 AWG wire should suffice to make up such an air cored
inductor (cold temp losses equating to 44 watts in this case).