My new home only has 208 3-phase grid service.
I'm about to try a factory updated ST2500XR.
The Zantrex I-net site briefly addresses this
possibility by suggesting use of a "buck-boost
transformer." These are very neat. An example
is a transformer with a 240 volt primary winding and a 24 volt secondary winding. By
bridging the secondary winding to the primary
winding, depending on which way you connect
them, you can have 240 + 24 = 264 volts or
240 - 24 = 208 volts. Fair enough, so that
resolves matching ST2500 voltage output to
my grid connection. And it's even pretty
efficient. Typical transformer efficiency
for these units seems to be around 97%, but
that 3% loss only applies to 24/208 = 11.5%,
so the total system conversion loss is only
about .35% or about a third of one percent.
But that Zantrex suggestion does NOT address
phase matching. In single phase systems,
aren't the two legs of 240 volt feeds 180
degrees out of phase with each other?
Whereas each leg of my 208 3-phase feed is
120 degrees out of phase compared to the
other two. Neutral to any leg of either
system is about 120 volts, but the 120
degree phase relationship prevents any
3-phase leg's voltage peak from time-aligning
with any other's peak, thereby causing the
lower voltage. Here's my concern. [img]images/smiles/icon_confused.gif[/img] If I use
a buck-boost transformer to create a 240
volt connection to which I can attach the
ST2500, it seems to me that even though the
voltage will be 240 and the frequency will
be 60 cps, there will be twice as much time
(60 degrees + 60 degrees)compared to the
other section of the cycle (60 degrees).
That's not like the output curve shape of
any inverter I've ever seen. I know it runs
electric motors and electronic power
supplies well enough that people don't
complain about it. But how efficiently will
an ST2500 drive this asymetrically shaped
cyclical load curve? The fact that the
Zantrex write-up on this does not mention
this concern causes me to worry and
speculate that the writer may not have
considered this issue. I hope my attempt to
explain this makes sense. Does anyone actually KNOW how efficiently an ST2500
series inverter can load into this asymetically shaped curve grid connection?
Is a Zantrex representative reading this?
John
Help help, I'm caught in the kelp. [img]images/smiles/icon_smile.gif[/img]

you drive only one of the 208 phases , so its a 208 to 240 converter or a standard isolation transformer 1 to 1.15 ratio

residental power is 240 single phase that is center tapped at the pole , to give 120/120

Get one with at least a 5kva rating , there have been problems with sites using smaller ones. Smaller than this and the impeadance becomes an issue.

My concern was that the 60 cps cycling curve might be asymetrical. This first consideration was the result of non-rigorous examination. But the more I think about it, more I'm convinced that the intesection of two sine waves cannot produce anything but a difference curve which is also symetrical. So this may simply be a non-issue.

But the approximate 3% loss rate for power flowing through a transformer would apply fully if the primary to secondary winding ratio were 208/240 as compared to the approximately one third of one percent loss if the pass thourgh power were 208 x 100% plus 24 x .97 I still think it's much more efficient to only pass the difference value through the transformer, thereby limiting transformer loss to that portion of the total.
Again, I hope this seems clear, but without graphics, communicating some kinds of concerns seems difficult at best.
Thanks for your consideration,
John

I'll try this again ...

Three phase (commerical power) is three sine waves , each 120 degrees out of phase ... so unlike 240 single phase , which appears as 120 volts 180 degrees out of phase.

You can't feed two legs of "208", 120/60 degrees into an inverter , the wave form is all distorted and you get frequency faults , even if you feed it thru a buck/boost transformer

what you need to do is take ONE 120 vac leg , feed it thru a 1:2 transformer and that connects to the suntie

Henry

Hi John,

You actually have 3 choices of driving a 208V 4 wire 3 phase Y service from a 240V single phase inverter.

1. Drive one 120V leg with a 2:1 240/120V isolation transformer.

2. Drive 2 legs, i.e. L1 and L2, from a 1.15:1 240/208V isolation transformer.

3. Drive 2 legs with an auto transformer. A standard 24V buck/boost transformer will only get you 232V from a 208V line but others are available with a true 240/208 ratio. Using the 24V buck boost transformer might give a little more unexpected disconnects if you have low line voltage.


All three ways are legal since the STXR is internally isolated but the third choice is most likely the best bang for the buck. Phasing and current distortion is not a problem with any of the choices. Much stranger transformer configurations are commonly used to generate 6 phase and even 12 phase power.

Ron Schroeder
Brookhaven National Lab

Ron

I don't understand how you can do #2 or #3

the 208 side , each leg is 120 degrees out of phase , on Suntie side must be 180 degrees out of phase ...

Please explain

<BLOCKQUOTE><font size="1" face="Verdana, Helvetica, sans-serif">quote:</font><HR>Originally posted by Solar Guppy:
I don't understand how you can do #2 or #3

the 208 side , each leg is 120 degrees out of phase , on Suntie side must be 180 degrees out of phase ...

Please explain<HR></BLOCKQUOTE>

But out of phase from what? The Suntie doesn't know what the original phase is, it just knows what it sees. And since the AC input section is isolated from ground, all it can know is that one side is 240VAC (340VDC pk) higher or lower than the other one. If it sees that 240 volt sinusoid it will be happy - even if it's 10, 20, 180 or 359 degrees out of phase with some other reference signal.

As an example, you can hook up a 240VAC appliance to two legs of a 208V 3 phase circuit. It will see a lower voltage, but as long as it's OK with that it will operate normally. All it sees is a 208VAC sine wave.

<BLOCKQUOTE><font size="1" face="Verdana, Helvetica, sans-serif">quote

Thats the issue , and I have seen this an issue with people trying this ...

when you sum to sine waves that are not 180 degrees out of phase , the result is NOT A SINE WAVE.

Thats the reason , the meter reads 208 , the peaks are not 180 degrees out of phase

The Suntie can only generate / follow a sine wave. , draw it out on paper and you'll understand what I talking about.

By the way , this is from debugging frequency fault problems with Sunties and commerical sites ... I know this doesnt work in the real world , you need a 120/240 transformer , its the only way it will work

[QUOTE]Originally posted by Solar Guppy:
[B]Thats the issue , and I have seen this an issue with people trying this ...

when you sum to sine waves that are not 180 degrees out of phase , the result is NOT A SINE WAVE.

Yes it is.

What you have are three sine waves 120 degrees apart. With respect to their common ground, each is 120V. Thus, the voltage between any two legs is the difference between their voltages to ground.

The formula you need is:

Sin(a)-Sin(b)=2Cos((a+b)/2)Sin((a-b)/2)

Given that b=a+120 you have

V=120(2Cos((a+a+120)/2)Sin((a-a-120)/2)
or
V=240Cos(a+60)Sin(-60)
But Sin(-60) is a constant whose value is
-0.8660. (-(Sqrt(3)/2)
Thus V=-208Cos(a+60)
Since Cos(x)=-Sin(x-90)
V=208Sin(a+60-90) or V=208Sin(a-30)
Thus, all that is needed is a transformer to reduce the inverter's 240V output to 208V to match the grid voltage.

Robert

<BLOCKQUOTE><font size="1" face="Verdana, Helvetica, sans-serif">quote:</font><HR>Originally posted by Solar Guppy:
when you sum to sine waves that are not 180 degrees out of phase , the result is NOT A SINE WAVE.<HR></BLOCKQUOTE>

Sure it is, it's just a smaller sine wave. If the two voltages are 180 degrees out of phase with each other, you get maximum possible voltage (2X peak voltage.) If they are in phase, then you get zero differential voltage. If they are 120 degrees out of phase, you get (2 x sin (120/2)) * Vin or 208 volts. If they're 90 degrees out of phase you get 170 volts.

One of the cool things about sine wave power is that the difference between two sine waves of the same frequency is also a sine, just offset and at a different amplitude. You can see this on a scope that will do an (A-B) trace.