Step-down Transformer question, connecting 3PH String inverters to a 208/120V service

Status
Not open for further replies.
Quick summary: 208/120V Gnded-Wye service, using Solectria PVI-28kW 3P 480/277V inverters, so I need to install a step-down transformer. The inverters require a 480/277V Gnded-wye on the inverter-side of this transformer.
Questions: Does it matter whether the utility-side of this transformer is 208/120V Gnded-wye or ungnded-wye? I believe the inverters will function fine either way, but is there an advantage one way or the other? If I were to use a gnded-wye to gnded-wye transformer, would it be better to still keep the X0 and H0 terminals separate, or could/should they be bonded together? Is there a need to establish a GEC on the inverter-side of this transformer?

Thank you.
 
Questions: Does it matter whether the utility-side of this transformer is 208/120V Gnded-wye or ungnded-wye?
I believe the inverters will function fine either way, but is there an advantage one way or the other?
If I were to use a gnded-wye to gnded-wye transformer, would it be better to still keep the X0 and H0 terminals separate,
Is there a need to establish a GEC on the inverter-side of this transformer?

I'm pretty sure it does matter, but I can't explain why.
If you look at page 28 here, it's calling for gnded Y to delta, the delta being the grid side.
http://solectria.com//site/assets/f...llation_and_operation_manual_generation_i.pdf

Doesn't the service count as 208/120 gnded Y, so you'd just want to go with that?

Not sure if Solectria work similar to SMA, but with SMA the XO is grounded on the inverter side and kept separate.

Yes, page 28 says GEC required.
----
The problem is, 480/277 to 208/120 transformers aren't that common it seems.

If you had three of those 28kW inverters, you could possibly use 3 of these xfmrs wired Y on the inverter side (neutral from all 3 inverters >> neutral on all 3 xfmrs and L1 from each inverter to xfmr 1, each L2 to xfmr 2 and each L3 to xfmr 3.
http://www.temcoindustrialpower.com/products/Transformers_-_General/HT0026.html

I'm planning on using 3 SMA Tripower 20kWs- if I was going to step them down to 208/120V, I'd use 3 of these, wired 277V L-N on the inverter side and 120V L-N on the grid side, because the grid connection will be 3 legs of 120V L-N, I believe.
http://www.temcoindustrialpower.com/products/Transformers_-_General/HT0073.html
 
I'm pretty sure it does matter, but I can't explain why.
If you look at page 28 here, it's calling for gnded Y to delta, the delta being the grid side.
http://solectria.com//site/assets/f...llation_and_operation_manual_generation_i.pdf

Doesn't the service count as 208/120 gnded Y, so you'd just want to go with that?

Not sure if Solectria work similar to SMA, but with SMA the XO is grounded on the inverter side and kept separate.

Yes, page 28 says GEC required.
----
The problem is, 480/277 to 208/120 transformers aren't that common it seems.

If you had three of those 28kW inverters, you could possibly use 3 of these xfmrs wired Y on the inverter side (neutral from all 3 inverters >> neutral on all 3 xfmrs and L1 from each inverter to xfmr 1, each L2 to xfmr 2 and each L3 to xfmr 3.
http://www.temcoindustrialpower.com/products/Transformers_-_General/HT0026.html

I'm planning on using 3 SMA Tripower 20kWs- if I was going to step them down to 208/120V, I'd use 3 of these, wired 277V L-N on the inverter side and 120V L-N on the grid side, because the grid connection will be 3 legs of 120V L-N, I believe.
http://www.temcoindustrialpower.com/products/Transformers_-_General/HT0073.html

Thank you for the input. I've already spoken to Solectria, and despite what is shown on page 28 of their installation manual, all of the connections indicated are not actually necessary. Although they show a delta configuration on the utility-side of the step-down, the inverter itself doesn't care - I think they've included this just because it is a common xfmr configuration. Given that they do show a delta configuration on the utility-side as shown, then I believe it would fall under the NEC requirements for a "seperately-derived system", which would necessitate a typical GEC & grounding electrode on the inverter-side. Regarding your suggestion of three 1P xfmrs, I'm a little unclear how this will be more effecient than one 3P xfmr? Physical space is a concern, and it seems like three 1P xfmrs will take up quite a bit more room.

To share a little more detail, my main concerns are the following:
- Wye/Wye xfmrs aren't very common, and I wonder if it would be more cost-effective to design using a delta/wye xfmr?
- Between the step-down transformer and the POI, the utility requires both a disconnect and an inline utiliity meter. The meter itself requires a neutral connection, so I have to bring the 3P/4W from the POI to at least the meter. It just seems strange to pull a neutral all the way to the meter, then not continue it the final 5ft to the step-down xfmr, only to re-establish a neutral again on the inverter-side of the xfmr.
- It seems cleaner to simply run 3P/4W on both sides of the xfmr, and configure it Wye/Wye. If I do that, though, the question I have is whether to bond the two neutrals or not. If I do, then I don't believe it qualifies as a "separately-derived system", and question whether I should establish a new GEC on the inverter-side of the xfmr. If I don't bond them, then I believe a new GEC would be required.

My inclination is to do the following: 1x28kW + 1x23kW inverter combined in a panelboard, 480/277V Wye xfmr primary --> 208V Delta xfmr secondary, 3P/3W to the meter housing, meter housing, 3P/4W to the POI. Bond and ground the 480/277V-side of the xfmr per the rules for separately-derived systems. Does this make sense?

Sorry for the wall of text. Thank you for your thoughts,
Jon
 
Quick summary: 208/120V Gnded-Wye service, using Solectria PVI-28kW 3P 480/277V inverters, so I need to install a step-down transformer. The inverters require a 480/277V Gnded-wye on the inverter-side of this transformer.
Questions: Does it matter whether the utility-side of this transformer is 208/120V Gnded-wye or ungnded-wye? I believe the inverters will function fine either way, but is there an advantage one way or the other? If I were to use a gnded-wye to gnded-wye transformer, would it be better to still keep the X0 and H0 terminals separate, or could/should they be bonded together? Is there a need to establish a GEC on the inverter-side of this transformer?

Thank you.

My first thought would be to use a 208 Delta on the utility side, and a 480 wye on the inverter side. This is just based on the fact that 99% of non utility transformers are wired with delta primary and wye secondary. In fact it is usually said to be bad to connect the neutral to the wye primary of a transformer , but I forget why (nor why utilities usually DO do it). My second choice would be a wye/wye and leave the 208 XO floating and unconnected.

If you are having trouble finding either of these two configurations or have to pay a lot for it because it is uncommon, you may want to look at a bank of single phase transformers as others have suggested as you will have much more flexibility in getting the connections you need.

Regarding connecting the XO and HO, I cannot think of a reason to do that. As far as I know, that is pretty much never done in low volt transformer installations (utilities using an MGN system do it all the time). Note that IF you had a wye/wye, and IF you connected the neutral to the utility side wye point, both things pretty rare to have/do in low volt, the XO and HO will be indirectly connected through the grounding electrode conductor/GES.
 

GoldDigger

Moderator
Staff member
Location
Placerville, CA, USA
Occupation
Retired PV System Designer
The problem for non utilities is that, unlike POCO, you may have little control over the exact voltage balance of the incoming wye. If the secondary is also a wye the matching imbalance on the secondary is not usually a problem.
But if the secondary is delta, the resulting imbalance can cause major circulating currents both in the secondary and in the primary neutral. Leaving the primary wye point floating allows the phase to wye point voltages to adjust to prevent the circulating currents.

There are more detailed discussions of the vector equations in older threads.
 
The problem for non utilities is that, unlike POCO, you may have little control over the exact voltage balance of the incoming wye. If the secondary is also a wye the matching imbalance on the secondary is not usually a problem.
But if the secondary is delta, the resulting imbalance can cause major circulating currents both in the secondary and in the primary neutral. Leaving the primary wye point floating allows the phase to wye point voltages to adjust to prevent the circulating currents.

There are more detailed discussions of the vector equations in older threads.

So you want to float/not connect the primary neutral with a wye/delta, but ok to do with wye/wye?
 

Smart $

Esteemed Member
Location
Ohio
Have you considered using the 3Ph 208V Solectria 14 kW inverter and eliminating the need for a transformer?
Likely the route I would have chosen from the get go. I cannot understand why PV designers plan a system for an AC voltage other than that of the service or an existing SDS.
 

ggunn

PE (Electrical), NABCEP certified
Location
Austin, TX, USA
Occupation
Electrical Engineer - Photovoltaic Systems
Likely the route I would have chosen from the get go. I cannot understand why PV designers plan a system for an AC voltage other than that of the service or an existing SDS.
I try not to do that but sometimes it is unavoidable. It has happened to me a couple of times.
 
1. despite what is shown on page 28 of their installation manual, all of the connections indicated are not actually necessary.

2. Given that they do show a delta configuration on the utility-side as shown, then I believe it would fall under the NEC requirements for a "seperately-derived system", which would necessitate a typical GEC & grounding electrode on the inverter-side.

3. Regarding your suggestion of three 1P xfmrs, I'm a little unclear how this will be more effecient than one 3P xfmr? Physical space is a concern, and it seems like three 1P xfmrs will take up quite a bit more room.

To share a little more detail, my main concerns are the following:

4. - Wye/Wye xfmrs aren't very common, and I wonder if it would be more cost-effective to design using a delta/wye xfmr?

5. - Between the step-down transformer and the POI, the utility requires both a disconnect and an inline utiliity meter. The meter itself requires a neutral connection, so I have to bring the 3P/4W from the POI to at least the meter. It just seems strange to pull a neutral all the way to the meter, then not continue it the final 5ft to the step-down xfmr, only to re-establish a neutral again on the inverter-side of the xfmr.

6.- It seems cleaner to simply run 3P/4W on both sides of the xfmr, and configure it Wye/Wye. If I do that, though, the question I have is whether to bond the two neutrals or not. If I do, then I don't believe it qualifies as a "separately-derived system", and question whether I should establish a new GEC on the inverter-side of the xfmr. If I don't bond them, then I believe a new GEC would be required.

7. My inclination is to do the following: 1x28kW + 1x23kW inverter combined in a panelboard, 480/277V Wye xfmr primary --> 208V Delta xfmr secondary, 3P/3W to the meter housing, meter housing, 3P/4W to the POI. Bond and ground the 480/277V-side of the xfmr per the rules for separately-derived systems. Does this make sense?

1. Well, that isn't terribly helpful! Other companies do that too, I always have to wonder why?

2. I'm reading that thick double black line between wye and delta on page 28 as the "POI", if POI means PCC/service connection point. Point Of Interconnection then? So if that wye section on page 28 is technically (hypothetically) showing the 208Y/120V service...you'd want the 480Y/277 inverters to connect to that (the inverters would be another separately derived system). If you added step down xfmrs for the inverters to the page 28 diagram, they'd still be connecting to those 208Y/120V lines and going to grid thru the green neutral/ground setup?
I don't get why you'd want to use anything other than 208Y/120V for the step down side of your inverters to grid xfmr(s)....you COULD because it's another separate system, but when both grid and load are 208Y/120V...I'd think you'd simply want to stay with that.

3. I was suggesting 3 one phase because while 208/120 to 480/277 three phase step UP xfmrs are common, the reverse are not- but single phase 277V to 120V step DOWN xfmrs are, and 3 of those, wye on both sides, gives you the 480/277 to 208/120 step down you are after (you aren't using the 480 L-L from the inverters- I see connecting the inverter ouput L-L to get 480V as an over-complication.)

4. To reference number 3- it may in fact be that cost of 3 one phase xfmrs are = to or even less than 1 three phase of the same kVA- you really want to look around online, then call the xfmr maker you like and ask for the nearest distributors- I've got 3 or 4 xfmr sellers within easy driving distance- no shipping and the prices are a bit lower too usually.

5. Right- the LV neutral from POCO is also connected to loads, correct? So you can't just "skip it" for 5 feet.

6. You want a second opinion here FOR SURE, but I say you don't bond them. The PV inverters want a neutral, that PV neutral goes to ground on PV-480/277V side of xfmr(s), and the load/grid 208/120V side of same xfmr(s), you attach that neutral connection to the load/grid neutral.
I have a question- how would a PV system qualify as NOT separately derived?

7. Wouldn't using a 208 delta on step down side cause issues with 208Y/120 service and/or load? Because you can't or shouldn't go L-N for 120V from 208V delta.
 
quote_icon.png
Originally Posted by Smart $

Likely the route I would have chosen from the get go. I cannot understand why PV designers plan a system for an AC voltage other than that of the service or an existing SDS.


I try not to do that but sometimes it is unavoidable. It has happened to me a couple of times.

Sorry about reposting this from the other 480/277 to 208/120 thread that is going on now, but...
Distances and therefore lengths of wire and therefore costs and AC and/or DC losses factor in.

The above matters even more because the DC wire sizes (combiner box(es) and DC inverter inputs) and the AC inverter outputs are all limited by the inverter maker- 2 AWG for input on DC side and 2 AWG on the AC output for some of them...that makes a lot of difference.
Some 480/277V inverters are limited to 6 AWG max wire size for AC ouput, but there's still less losses with them than the same exact kW of 208/120 inverters with their 2 AWG AC output wire limit.

Edit: the only way I've found to easily get "numbers" for those DC/AC losses due to wire runs is with Helioscope software, btw.
---
repost-

Ok, about using 480/277 instead of 208/120 inverters... for instance...

Say you have a (PV) panel to inverter DC run that is fairly long, 200 feet or so.
Then you have a fairly large building in a direct line between panels and PCC/service point.
You can go around the building OR straight through it.

Other option is:
Single phase 208/120 inverters, wired for 3 phase vs. 480/277 3 phase inverters.

Going around the building, with 208/120 inverters, you are using considerably more DC wire, at lower voltage- the AC wires may be shorter, but...the DC losses add up. Possibly not the best option overall.
Going thru the building with 208/120 (say 100 feet or so of AC). the *AC losses* add up.
Going around the building with 480/277, same amount of wire, less DC losses.
Same going thru the building- less AC losses over 100 feet at 480/277.

That's just using 100 feet as an "example", but isn't there a certain number there where it does matter?
 
Status
Not open for further replies.
Top