Sizing an HVAC only panel

[emahler]

we have a situation where a brand new building is a 208/120V service, but someone goof'd and ordered 460V RTU's...the RTU manufacturer is asking us to look at solutions that don't involve replacing all the RTU's (which are already on the roof and installed)....

One possible solution is to install a transformer and step up to 480V for a panel to feed the unit. There are 6 units total. 1-15T, 4-8.5T and 1-6.5T. If my memory is correct, I need to size the panel as follows:

300% of the largest RLA plus the RLA of the other loads. In this case the RLA of the 15T is 42.6A, the 8.5T are 28.2A and the 6.5T is 19.6A (all at 480V)

(42.6x3) + (28.2x4) + 19.6 = 260A @ 480V...

is it this simple?

 

[kwired]

we have a situation where a brand new building is a 208/120V service, but someone goof'd and ordered 460V RTU's...the RTU manufacturer is asking us to look at solutions that don't involve replacing all the RTU's (which are already on the roof and installed)....

One possible solution is to install a transformer and step up to 480V for a panel to feed the unit. There are 6 units total. 1-15T, 4-8.5T and 1-6.5T. If my memory is correct, I need to size the panel as follows:

300% of the largest RLA plus the RLA of the other loads. In this case the RLA of the 15T is 42.6A, the 8.5T are 28.2A and the 6.5T is 19.6A (all at 480V)

(42.6x3) + (28.2x4) + 19.6 = 260A @ 480V...

is it this simple?

Minimum feeder ampacity is determined by 430.24:

430.24 Several Motors or a Motor(s) and Other Load(s).


Conductors supplying several motors, or a motor(s) and other load(s), shall have an ampacity not less than the sum of each of the following:
(1) 125 percent of the full-load current rating of the highest rated motor, as determined by 430.6(A)
(2) Sum of the full-load current ratings of all the other motors in the group, as determined by 430.6(A)
(3) 100 percent of the noncontinuous non-motor load
(4) 125 percent of the continuous non-motor load.

Note there is a difference between RLA of a compressor motor and MCA of the unit assembly, 125% of largest full load current rating of highest rated motor would be according to RLA of that particular compressor, not the MCA of that particular unit. That MCA is great for sizing individual branch circuits, but to size a feeder for multiple units if you go by MCA your calculation will result in larger feeder then you could run to be in compliance with NEC. Not sure if your figures are MCA or RLA, or where you came up with the 300%, you need FLA of blower motors factored into this equation, as well as any other loads (like electric heat strips if there are any).

It is possible with such a feeder to have overcurrent protection device that is higher then the minimum conductor ampacity required.

The transformer should only need to be sized to the actual full load it supplies and should be able to be loaded to 100% of it's VA rating.

 

[Sahib]

. Not sure where you came up with the 300%,

The OP is sizing panel disconnect for which up to 400% of FLA of largest motor is allowed by code.

 

[kwired]

The OP is sizing panel disconnect for which up to 400% of FLA of largest motor is allowed by code.

The panel needs protected at or below it's rating, the feeder supplying all those units can have short circuit/ground fault protection that is the highest branch circuit device rating plus all other loads - which may well be more then the ampacity of the feeder conductor. If there is trouble starting motors, it may be able to be increased.

Where do you get 400% from?

Just throwing figures out there that are not based on OP's loads, if you have a minimum ampacity necessary of 175 amps but need SCGF protection of 400 amps, you are going to need a 400 amp panel. Most of us would still feed it with 400 amp conductors just for some future load possibilities, some maybe because they don't know it could be supplied with 175 amp conductors even though it is on a 400 amp OCPD.

 

[Sahib]

Where do you get 400% from?

430.52C

 

[Smart $]

430.52C

That's fine for motors, but not for refrigerant motor compressors.

430.24 is good for establishing minimum conductor ampacity. However, if the conductors are sized no larger than this minimum, the maximum OCPD for a multi-RTU feeder (and thus, the minimum panel rating) is determined by 430.62(A), which in turn requires using the value determined under 440.22(A)... and it requires the greater of nameplate RLC or BCSC to be used. Maximum unit OCPD rating is 175% of nameplate but permitted up to 225% if 175% results in start tripping.

 

[Smart $]

...
430.24 is good for establishing minimum conductor ampacity. However, if the conductors are sized no larger than this minimum, the maximum OCPD for a multi-RTU feeder (and thus, the minimum panel rating) is determined by 430.62(A), ...

Forgot to add that if your feeder conductor ampacity is greater than the minimum you can base the feeder ocpd on that ampacity under 430.62(B). The minimum panel rating would be adjusted accordingly.

 

[emahler]

bear in mind, we don't have the individual compressor name plates/ratings....we have the manufacturer cut sheets which simply give us the MCA and MOCP. Spoke with an engineer today who leaned towards sizing to panel according to the Total MOCP of the 6 RTU's. His reasoning was that the MOCP from the manufacturer takes into account all the % and allowances. But he didn't run actual calculations.

Based on the MOCP's for the units, he felt a 250A 480V panel would be fine. He proposed to simply add the MOCP (there is the possibility they will all be running at once, and a slight possibility they could all start at once) and multiply x 125% for a fudge factor.

does this sound plausible?

 

[steve66]

we have a situation where a brand new building is a 208/120V service, but someone goof'd and ordered 460V RTU's...the RTU manufacturer is asking us to look at solutions that don't involve replacing all the RTU's (which are already on the roof and installed)....

One possible solution is to install a transformer and step up to 480V for a panel to feed the unit. There are 6 units total. 1-15T, 4-8.5T and 1-6.5T. If my memory is correct, I need to size the panel as follows:

300% of the largest RLA plus the RLA of the other loads. In this case the RLA of the 15T is 42.6A, the 8.5T are 28.2A and the 6.5T is 19.6A (all at 480V)

(42.6x3) + (28.2x4) + 19.6 = 260A @ 480V...

is it this simple?

I would just use the minimum circuit ampacity (MCA) for the largest unit, and the rated load amps for every other unit.

I get something like 186 amps @ 480V. If they are all 3 phase, that's just a little more than 150KVA.

The owner is going to be paying for the transformer losses forever.

 

[GoldDigger]

bear in mind, we don't have the individual compressor name plates/ratings....we have the manufacturer cut sheets which simply give us the MCA and MOCP. Spoke with an engineer today who leaned towards sizing to panel according to the Total MOCP of the 6 RTU's. His reasoning was that the MOCP from the manufacturer takes into account all the % and allowances. But he didn't run actual calculations.

Based on the MOCP's for the units, he felt a 250A 480V panel would be fine. He proposed to simply add the MOCP (there is the possibility they will all be running at once, and a slight possibility they could all start at once) and multiply x 125% for a fudge factor.

does this sound plausible?

I would say that is conservative to the point of overkill.

If all of the units might run at the same time, I would agree to sizing the feeder and panel according to the sum of the MCAs, which would itself still be conservative by NEC standards.

Only if you see that it is likely that all units would start at exactly the same second would I even consider going beyond that.
If you have a serious concern that all the units might start at once (such as after a shutdown period or on power restoration), then coordination to the sum of the MOCPs might be reasonable to prevent any chance of nuisance tripping at the feeder level.

There is no justification whatsoever to applying a 125% factor to the sum of the MCAs, much less the MOCPs.

JMO

 

[kwired]

If there is good chance of starting all at same time, put on delay timers in the motor control circuit, stagger them so they don't all start at same time. They only need to be a few seconds apart, and the users would seldom if ever notice this.

 

[Smart $]

....does this sound plausible?

Minimum per Code is MOCP of unit with largest motor + MCA* of others.

*Code states it as RLC or branch-circuit selection current, whichever is larger.