The need to get 240 volts instead of using 208 volts for 240 volt
loads is not what I will be discussing here. If you want to discuss
that, please start a new thread. What I want to ask about here is
the viability of using buck/boost transformers to get that 240 volts,
either for delta connected three phase loads, or for single phase
loads.

What my calculations came up with is a wiring arrangement using the common 240/120 to 24/12 single phase buck boost transformer to get nearly the correct voltage. Actually, it would involve 3 of them. To understand what I am going to describe, you should already know how a buck/boost transformer is used in an autotransformer setup to make voltage adjustments.

Given a 208Y/120 power source, the ideal buck-boost transformer would have 120 volts input and 18.56406460551 volts out. With the output wired to boost the supply voltage, you get 138.56406460551 volts L-N and thus 240 volts L-L. Such transformers are not actually made that I can find.

My suggestion is to use a 240/120 to 24/12 transformer wired for 240 volts input and 24 volts output. The primary would be connected L-L to the source, thus powered by 208 (207.846096908265) volts. The output would then be 20.8 (20.7846096908265) volts. This output would then be wired in series with a particular L-N connections so that it's phase angle is 30 degrees. Assuming a supply labeled as:

A \ N---C / B

One B/B transformer fed from an A-B connection would have its output at the polarity that corresponds to the A input connected to line B. Showing this accurately is beyond ASCII graphics so I'll skip that to avoid introducing new confusion. If this is wired as intended the result is 120 volts at 0 degrees plus 20.7846096908265 volts at 30 degrees. This results in 138.390751136 volts at 4.306619096 degrees. Do the same on all 3 phases with 3 such transformers. The resultant L-L voltage is then 239.7 (239.699812265) volts, very close to the desired 240 volts. Normal voltage variations are greater than this. Assuming a perfectly resistive load on the 239.7Y/138.39 side of this, the supply would see a current angle displaced 4.3 degrees for a power factor of about 0.997.

I have seen many buck/boost wiring diagrams. But of all those seen, none has ever done any cross-phase connection, not even of this small amount. So I have these questions:

1. Is this an unknown method?

2. Is this prohibited by code?

3. Is there anything unsafe about this?

4. Would a power factor of 0.997 be a problem where loads often have their own power factors much worse than this?

There is another method that is a little more complex which involves phase B being boosted by have as much as above from a reduced A-B as well as from a reduced C-B. The same transformers can be used, but would have to be wired with 12 volt secondaries not connected together at all (or double the number of transformers at half the capacity). This would result in a final L-N voltage of 138.000, and a L-L voltage of 239.023, which is still very close to the target. But it would have a 0 degree voltage phase shift. If the phase shift is more of an issue then this approach is an alternative.

Or the inherint phase shift of the first setup could be used opposite to the shift of a reactive load and presumably reduce its effect.

My real interest is not so much in supplying three phase loads at 240 volts, but rather, supplying single phase loads at 240 volts. Three phase loads are more likely to be designed with the expectation of the supply voltage being 208Y/120. But single phase loads may not. Given a three phase source, getting nearly 240 volts is not hard. But a common problem is that residences in large buildings are often given 2 legs of a three phase system as "fake" single phase (where the legs are really 2 phases at 120 degrees). Getting 240 volts out of this is my real goal. One could get three phase back from this like:

A A * \ \

although that would have some substantially reactive current angles on both of the supplied phases for any load on the output C-N. The first or second examples I gave above might be applied on top of this three phase hack, but only to get one phase of 240 volts at about the A-B phase angle. Still, this is a lot of transformers and wiring at this point.

By skipping the three phase step, 1 or 2 B/B transformers could be used on a 2 phase 120/208 volt system two different ways. Probably the simplest is one transformer with 240 volts in and a pair of 16 volts secondaries out. Wire the input to the 208 volt A-B and use each of the secondaries to separately boost out A and B. This will get only 235.5589 volts total. One side effect of this is that the phase angle of A and B with respect to the neutral will increase somewhat from 120 degrees to 126 (126.008983198) degrees.

Note that the above would have 132 (132.181693135) volts L-N and would not be suitable for any L-N loads expecting 120 volts, and hence cannot be used for any 3-wire 120/240 volt loads. Such loads would probably have to use a 120 in 120/240 out autotransformer (half capacity) or a 208 in 120/240 out isolation transformer (full capacity) to get all the right voltages. And these more expensive transformers due to the higher capacities involved.

What my calculations came up with is a wiring arrangement using the common 240/120 to 24/12 single phase buck boost transformer to get nearly the correct voltage. Actually, it would involve 3 of them. To understand what I am going to describe, you should already know how a buck/boost transformer is used in an autotransformer setup to make voltage adjustments.

Given a 208Y/120 power source, the ideal buck-boost transformer would have 120 volts input and 18.56406460551 volts out. With the output wired to boost the supply voltage, you get 138.56406460551 volts L-N and thus 240 volts L-L. Such transformers are not actually made that I can find.

My suggestion is to use a 240/120 to 24/12 transformer wired for 240 volts input and 24 volts output. The primary would be connected L-L to the source, thus powered by 208 (207.846096908265) volts. The output would then be 20.8 (20.7846096908265) volts. This output would then be wired in series with a particular L-N connections so that it's phase angle is 30 degrees. Assuming a supply labeled as:

A \ N---C / B

One B/B transformer fed from an A-B connection would have its output at the polarity that corresponds to the A input connected to line B. Showing this accurately is beyond ASCII graphics so I'll skip that to avoid introducing new confusion. If this is wired as intended the result is 120 volts at 0 degrees plus 20.7846096908265 volts at 30 degrees. This results in 138.390751136 volts at 4.306619096 degrees. Do the same on all 3 phases with 3 such transformers. The resultant L-L voltage is then 239.7 (239.699812265) volts, very close to the desired 240 volts. Normal voltage variations are greater than this. Assuming a perfectly resistive load on the 239.7Y/138.39 side of this, the supply would see a current angle displaced 4.3 degrees for a power factor of about 0.997.

I have seen many buck/boost wiring diagrams. But of all those seen, none has ever done any cross-phase connection, not even of this small amount. So I have these questions:

1. Is this an unknown method?

2. Is this prohibited by code?

3. Is there anything unsafe about this?

4. Would a power factor of 0.997 be a problem where loads often have their own power factors much worse than this?

There is another method that is a little more complex which involves phase B being boosted by have as much as above from a reduced A-B as well as from a reduced C-B. The same transformers can be used, but would have to be wired with 12 volt secondaries not connected together at all (or double the number of transformers at half the capacity). This would result in a final L-N voltage of 138.000, and a L-L voltage of 239.023, which is still very close to the target. But it would have a 0 degree voltage phase shift. If the phase shift is more of an issue then this approach is an alternative.

Or the inherint phase shift of the first setup could be used opposite to the shift of a reactive load and presumably reduce its effect.

My real interest is not so much in supplying three phase loads at 240 volts, but rather, supplying single phase loads at 240 volts. Three phase loads are more likely to be designed with the expectation of the supply voltage being 208Y/120. But single phase loads may not. Given a three phase source, getting nearly 240 volts is not hard. But a common problem is that residences in large buildings are often given 2 legs of a three phase system as "fake" single phase (where the legs are really 2 phases at 120 degrees). Getting 240 volts out of this is my real goal. One could get three phase back from this like:

A A * \ \

*/ \ N ... N C /*/ B Balthough that would have some substantially reactive current angles on both of the supplied phases for any load on the output C-N. The first or second examples I gave above might be applied on top of this three phase hack, but only to get one phase of 240 volts at about the A-B phase angle. Still, this is a lot of transformers and wiring at this point.

By skipping the three phase step, 1 or 2 B/B transformers could be used on a 2 phase 120/208 volt system two different ways. Probably the simplest is one transformer with 240 volts in and a pair of 16 volts secondaries out. Wire the input to the 208 volt A-B and use each of the secondaries to separately boost out A and B. This will get only 235.5589 volts total. One side effect of this is that the phase angle of A and B with respect to the neutral will increase somewhat from 120 degrees to 126 (126.008983198) degrees.

Note that the above would have 132 (132.181693135) volts L-N and would not be suitable for any L-N loads expecting 120 volts, and hence cannot be used for any 3-wire 120/240 volt loads. Such loads would probably have to use a 120 in 120/240 out autotransformer (half capacity) or a 208 in 120/240 out isolation transformer (full capacity) to get all the right voltages. And these more expensive transformers due to the higher capacities involved.

--

|---------------------------------------/----------------------------------|

| Phil Howard KA9WGN (ka9wgn.ham.org) / Do not send to the address below |

|---------------------------------------/----------------------------------|

| Phil Howard KA9WGN (ka9wgn.ham.org) / Do not send to the address below |

Click to see the full signature.