# transformer saturation in a bucking configuration

Discussion in 'Electrical Engineering' started by [email protected], Feb 25, 2009.

1. ### Guest

As voltage increases, at what point does saturation of the core of a typical
dry-type distribution and/or buck-boost transformer begin to happen? I would
assume not for at least a 10% increase since that would be a common range of
voltage variations.

Consider a buck-boost transformer with a 240 volt winding (maybe in the form
of a pair of 120 volt windings configured in series) and a 32 volt winding
(maybe in the form of a pair of 16 volt windings configured in series) where
the 240 volt winding and 32 volt winding are in series. That would be a design
voltage of 272 volts. Now if you connect a 277 volt supply circuit across the
whole 272 effective winding (ungrounded on the 32 volt end, and grounded on the
240 volt end), you would get around 244.5 volts across the 240 volt winding.
How close is this getting to saturation? I calculate that 10% over 272 volts
works out to right around 8% of the 277 volts. If the 277 volt supply does
go up to 10%, how much saturation could be taking place?

What is the typical "percentage of voltage" before saturation begins to happen
for this class of transformer? My understanding of saturation is that it is a
gradual thing. But I've never studied it in particular, and some googling has
not yet come up with any good material (except maybe a couple papers I cannot
get more than a short abstract for at IEEE). But saturation theory is not the
whole story, anyway, since it also matters how the transformer is designed.

I have seen brief voltage swells, usually 1/2 second to no more than 3 or 4
seconds, on the utility power, that are above 140 volts. If transformers are
designed with a tight saturation tolerance, such swells could push them into
the beginning saturation range.

Is such a bucking configuration a viable way to derive around 240 volts from
around 277 volts? If not, I would have figured there would be transformers
(of the autotransformer design) marketed specifically for dropping 277 volts
to 240 volts.

2. ### Don KellyGuest

This is an autotransformer (which is simply an ordinary transformer
connected in a buck or boost arrangement). For higher voltages, there
are some insulation concerns as the Do you want to have your 277V supply
grounded at one end (that common to the 277 and 240V side as it will be
in this configuration? Not, if ,as I assume, it is taken from 2 legs of
a star. A 277/240 autotransformer would have a 240V winding and a 37V
winding and the latter would be insulated for 277V

The voltage levels should be OK - 277/272 is less than 2 % high so the
244.5 is the same % high. Exciting current will be within normal limits
but at 10% overvoltage on the 277V side, you will be pushing a 12%
overvoltage as far as the transformer is concerned. Since the
distribution at these levels is normally within 5% and 10% is allowable
for limited periods, it <should> be no problem.
However:
a) You don't specify the size of the transformer as some small
transformers (say in the 1KVA range) push the core fairly hard in
order to get lower copper loss and better voltage regulation for the
same money and the winding voltages indicate that this transformer
would typically quite small and would run a bit hot at no load because
of core losses.
b) the insulation of the 32V winding should be good for 277V as this
part of the transformer will be at the high end.
c) since it is an autotransformer, you will not have the isolation
provided by a 2 winding transformer- so watch grounding!.

Swells of 140V for a few seconds as you indicate will likely cause some
saturation but as the duration is so small, there is negligible
overheating.

3. ### Guest

| a) You don't specify the size of the transformer as some small
| transformers (say in the 1KVA range) push the core fairly hard in
| order to get lower copper loss and better voltage regulation for the
| same money and the winding voltages indicate that this transformer
| would typically quite small and would run a bit hot at no load because
| of core losses.
| b) the insulation of the 32V winding should be good for 277V as this
| part of the transformer will be at the high end.
| c) since it is an autotransformer, you will not have the isolation
| provided by a 2 winding transformer- so watch grounding!.

I guess this is more of a concern when stepping down, since if the grounded
wire comes loose from the winding in the transformer, the voltage rises to
277V relative to ground passing through.

Thinking about the effect of saturation in a bucking transformer design now
has me more concerned. If the core saturates, the bucking effect will be
reduced. So going from 277V to 277V+overvoltage is going to mean the output
goes from 244V to 277V+overvoltage. So pushing the edge with 240+32 might
be a bad idea. Probably better, then, would be 240+48, which is effectively
a 288 volt winding fed with 277, giving 230 out. The ratio would be exactly
6:5, which is the ratio between 480/277 in the USA and 400/230 in Europe, so
the voltage to utilization equipment would be essentially the same as Europe
(but at 60 Hz). It would be slightly less efficient than the 8.5:7.5 bucking
configuration in terms of transformer capacity to utlization capacity. But
it might just be the better way to go. There would be additional margin for
voltage swells given the greater risk when saturation happens. The ground
reliability issue would still be the same.

| Swells of 140V for a few seconds as you indicate will likely cause some
| saturation but as the duration is so small, there is negligible
| overheating.

So far I haven't seen a breaker trip from these events. I have had one light
bulb blow as a result of the most recent wind during a windstorm. It was a
15 watt bulb, so a normally weaker filament (I use a lot of 15 watt bulbs).

4. ### Don KellyGuest

The concern will be there stepping up or down. It is the insulation to
ground that is important. In your case, you will have 277V from the
upper terminal to neutral (assume grounded) and the 32V winding is at
the upper end- between 244 and 277 V to ground in normal operation. .
At these voltages, due to typical overinsulation at this level, it is
unlikely to be a problem- but that depends on the manufacturer who may
skimp on secondary to core insulation. If you put the 32V winding at the
"ground " end, then, in normal operation, the "low end" of theoutput
will be above ground by 32 V.

Saturation will not raise the output voltage but, rather, will reduce it
(ignoring any possible resonance at, typically, the second harmonic).

The main effect of saturation is an increase in exciting current- it
doesn't change the ratio. Note that for AC, the maximum flux is
determined by the voltage and the exciting current will be such as to
provide this flux. With saturation, this may be an increase from about
3% of the input current to 10% and in extreme cases, it will be self
limiting due to primary impedance. This increase will cause more core
losses and heating and will also affect voltage regulation to some
extent. The output at an input of 277+ x will be at or below 244+0.88x
NOT 277+x.
I use 13 watt to 23 watt cfls in preference- that way, I get a useful
amount of light. For a nightlight in the bathrooms and hallways, I use a
60mW electrofluorescent plug in- good enough to keep me from stumbling
about in the hallway or peeing on the floor.