Understanding 3 phase volt drop.......

[HappyHippyDad]

I am trying to understand 3 phase volt drop. I understand single phase VD.

Lets say we have a 70 meter run of 35mm, 5 core cable from the supply to the 3 phase consumer unit.

From the 3 phase consumer unit we have 1 x 16A triple pole MCB (so that's 16A per phase). which has a run of 12metres in 2.5mm SWA (5 core 70 degrees). Lets say it's going to a commando socket, just so it's easier to explain.
From the same consumer unit we have 1 x 20A single pole MCB which has a run of 15 meters, in 4mm SWA (3 core, 70 degrees). Let's say to a standard socket.

We'll assume max demand is 16A and 20A (no diversity)

VD from CU to standard socket = (20A x 15metres x 11 ) / 1000 = 3.3v

VD from CU to commando socket = (16A x 12m x 15) / 1000 = 2.88v

I'm not sure where to go from here. Do I use 36A as max demand and then just use the volt drop tables for 3 phase in order to find the VD between supply and CU?

If I did this it would be as follows....

VD from supply to CU = (36A x 70m x 1.1) / 1000 = 2.77v (to 2 dp)

Therefore highest VD is to the standard socket at 3.3v + 2.77V = 6.07v

The reason I am confused is because of the mixture of single phase and 3 phase equipment coming from the same CU.

 

[brianmoooore]

This is 3Ph + N, not just 3Ph. Worst case scenario is when two phases have no load, and the maximum loaded phase is actually at maximum demand, so you can calculate VD as a single phase for that particular phase.
As soon as current is drawn from the other phases, the neutral current will decrease, so VD will always be less in the neutral than in the max loaded phase, giving less VD overall..

 

[James]

Yes, you have pretty well got it.
make it simple on yourself,
use the figures for the highest loaded phase and that will give you worst case for your distribution cable.

edit,

or you could use the reply above that says what I was thinking but in a much better and more in depth way.
thanks @brianmoooore
i will go hide back under my rock!!

 

[DefyG]

VD at the 3ph DB needs to be low enough that the 3/5% limit is met at the ends of the single phase circuits so I would try to limit VD to say 1.5% at the 3ph DB so that it still gives scope to meet the overall VD limits.

400v x 1.5/100 = 6v

For 70mx35sq = 80A/ph (approx) in simplistic terms.

 

[pc1966]

It is confusing at first since the tabulated values are for different situations. So taking your 35mm cable example and looking up Table 4E4B for an example SWA situation we have:

• Single phase 1.35 mΩ/m (mV/A/m)
• Three phase 1.15

But the catch is they are on your nominal voltages of 230V and 400V respectively. So lets assume you have 70m and a 100A load in both cases (3-phase balanced here) then you get:

• Single phase VD = 1.35 * 100 * 70 = 9,450mV = 9.45V = 9.35/230 = 4.1%
• Three phase VD = 1.15 * 100 * 70 = 8,050mV = 8.05V = 8.05/400 = 2.0%

So (to rounding error, etc) you have half the voltage drop in the 3-phase case. Why? Well as @brianmoooore has already explained under balanced conditions the neutral current drops to zero and so the power lost due to N also vanishes!

Now if you have a mixed/unblanced load it is more complicated but it is always going to fall between the SP drop and the 3P drop figures depending on how much cancellation of current there is in the neutral.

The above also explains why for 3P systems you would measure the L-N and L-E PFCs and double them to get the worst case fault current. This is also known as a "three phase bolted fault" and when this happens the N has no job to do and the current is limited by the L (and to some degree the supply transformer, etc) so the 2% voltage drop above turns in to double the fault current of the 4% SP case!

If you are interested, the ratio of the tabulated values (above 1.35 / 1.15) is theoretically 2 / sqrt(3) = 1.155 to sufficient precision.

 

[pc1966]

Getting off-topic here, but if you ever have to find the max PSCC and suspect the cable has a reduced-size N, or if it is a 3+E socket so no N, then you can measure the L-L PFC and multiply by 1.155 to the the max case.

 

[HappyHippyDad]

Thanks for the replies, that's given me a bit to think about!