RPC Theory

[rwm]

I have been reading about RPCs and trying to understand how they work. Most explanations have mis-statements or gloss over how they make true 3 phase power. For example, several sources say 'they take single phase power and add a generated phase to get 3 phase power.' So basically 1+1= 3. Since when? I realize the standard 240V power is "split phase" but it is not 2 phase.
I guess my fundamental misunderstanding is how does the RPC take the single phase L1 and L2 lines and offset these by 120 deg? Especially since these appear to be wired directly to two legs of the 3 phase in most diagrams?
I know we have a lot of people here who are very knowledgeable on this. Please help!


You can see in this diagram that L1 is connected to T1 directly and L2 is connected to T2. By definition these phases are 180 deg apart?

 

[rwm]

Well at least one discussion says they do not make 3 phase power. They make this:

I don't know if this is correct?

 

[benmychree]

You neglect to illustrate the magic part; the capacitor,

 

[Eddyde]

One of the best explanations I have heard; Think of the idler as a rotary transformer, not a "generator".

 

[Chipper5783]

Like RWM Robert says, they do not make true 3 phase power. The graph Rob provided shows you the voltage between any two legs (L1-L2, L1-L3, L2-L3). You can’t change the voltage phase relationship - that is baked into the grid. However, you can push around when the current flows, compared to the voltage. You can make the current lead or lag the voltage - that is called the “Power Factor”. That is what the motor windings do (inductance - where the peak current will lag, or follow the peak voltage) and the capacitors (add capacitance - where the peak current will lead the peak in the voltage). It is the amount of current flowing in a given conductor that provides the magnetic field (more current = a stronger magnetic field). The result with a well designed RPC is that you get a pretty good separation of the current in each leg, thus you will get a rotating magnetic field in the stator of a connected 3 phase motor.

It still isn’t true 3 phase power. Like was said, it is more of a transformer. Works fine.

 

[rabler]

Well at least one discussion says they do not make 3 phase power. They make this:
View attachment 498773
I don't know if this is correct?

I've dug into this myself. I went so far as to build a digital system to analyze the voltage, current, phase relationships and noise coming from my rotary phase converter. For the technically minded, that involved A/D conversion of both voltage and current of all three phases, oversampled, and used FFT's to get accurate data.

A simple rotary phase converter can produce accurate 3 phase power for one unchanging (static) load. The problem arises in that there are three possible complex loads (in electrical engineering speak, impedances), one across each phase. An "impedance" has a resistive load value, which changes the magnitude of the voltage/current relationship, and a phase angle, which changes the offset of the sinusoid (angle) between voltage and current, also known as power factor. So two values per phase, times three phases, six values that have to all be "right" to get "perfect" three phase. I say perfect because even grid 3 phase isn't mathematically perfect. The problem is loads change, and a simple three phase converter doesn't have a way to compensate. Typically with a RPC you measure the voltage/current produced at a "typical" load and try to tune it to that load. My goal with a digital system was initially to dynamically analyze the load and "tune" it on the fly. That kind of tuning is similar to what VFD's do. I might get back to that at some point, for now I log the performance of the system, sampled IIRC, 16 times per cycle of each current and voltage wave.

One thing that makes it simpler is if the load on all three phases change similarly (or identically). A three phase motor does pretty much present identical loads to all three phases, which is why RPC's work OK (but not "perfect) for motors but not for general three phase power where you may have some single phase loads scattered across the phases. The utility supplied phase is pretty stable (which is why you want to use it for the control circuitry on your machine), but loads on that phase can still have effects the generated leg. (I'm using the terminology that a "phase" is between two wires, a leg is one wire).

Three phase motors are not particulary fussy customers, so they do OK, with less than perfect three phase, although there is some loss of performance (power) and increased heating in the motor. For most machine shop uses this doesn't matter, machines are not running at 100% of rated capacity continously so there is enough margin to cover the lack of perfect 3 phase.

That's about the best I can do as a hand waving explanation. Otherwise you fall into "here is the mathematical model". Which is how engineering would cover it. I'm retired, so I don't do that anymore

 

[rabler]

they take single phase power and add a generated phase to get 3 phase power.' So basically 1+1= 3.

Terminology issue really. A "phase" is the voltage and current between two wires/points. You can't measure voltage or current on a single terminal, voltmeters have two leads. Common terminology, the voltage "at" one point, is relative to a wire stuck into the earth (ground). Single phase has two wires, usually called L1 and L2. You measure the voltage and/or between L1 and L2. By adding one more wire (leg), you can now measure L1-L2, L2-L3, and L3-L1, three separate values, thus 3 phase. Or to look at it the other way around, in a three phase power system with L1, L2, and L3, you can pull single phase off of an combination of two of those, doesn't have to be L1 and L2.

One of the idiosyncrasies of three phase as generated by RPCs is that the generated leg has some greater voltage relative to neutral/ground. You can arbitrarily use one of the two lines in 240V single phase relative to neutral and get 120V. If you use the generated leg, there is no way to get 120V off of that leg while still having anything like 3-phase 240V. This is another reason you don't want to use the generated leg for control circuitry which may reference neutral. This is sometimes confused by saying generated leg has improper voltage. It is correct when referenced to the other two leg, but not to neutral/ground.

 

[rabler]

Here's a simple graphic of an RPC voltage output as 2D picture. Any three phase system can be graphed liked this, just may have to shuffle the points around a bit.

Where L1 and L2 are your line supplied single phase, and the green (N) is the neutral reference. For the generated leg to be 240V to both L1 and L2, then the voltage between generated and neutral will be 207.8 ... volts. A three phase motor doesn't care about the generated to neutral voltage, the neutral voltage isn't used (except as a safety ground to the case frame, which shouldn't be carrying any power unless something goes wrong). But this is where the idea that the generated voltage is "wrong" sometimes comes from, as in your graph above.

There is nowhere in that triangle to put the neutral in a 240V three phase that gives 120V to neutral/ground from all three legs. This is completely independent of using an RPC, it's just a reality of how trigonometry (vector math) works. If you use 208V three phase, you can put the neutral in the "middle" of the triangle and get 120V to all three legs. These variations are where 3 phase systems get complicated and you see all sorts of "standards".

Calling residential single phase power 120V split phase only works when neutral is in the middle of the two lines (L1, L2) as is common in single phase power systems. But the reality is if you hook up the two resulting 120V circuits to an oscilloscope, one will be the inverse of the other. Mathematically a sine wave shifted by 180 degrees is the same as the inverse. So you can say from a math perspective that they are represented by two sine waves at different angles (180 degrees offset). Engineers call that offset the phase angle, or "phase" for short. So in one sense it is two-phase power. But it isn't common power electronics nomenclature to call split phase as two phase. The power company either runs two wires to a customer (single phase) or three wires (3 phase). U.S residential use is to run single phase through a center tap transformer to produce those two phases, but since the power company provides only two wires with one voltage between them, the power company terminology of it being "single phase" is used.

 

[rwm]

Thanks for taking the time to respond in such detail. I am certainly no EE but I will study your reply.
BTW, The point of the graph I posted was not that the generated leg voltage was wrong but that the phases were not 120 deg apart.
I was under the impression that at the power plant, the generators put out 3 phase power (120 deg apart) and in fact, you could measure their voltage and phase relative to ground? It may not matter for a motor, but it that concept physically accurate?

 

[rabler]

I was under the impression that at the power plant, the generators put out 3 phase power (120 deg apart) and in fact, you could measure their voltage and phase relative to ground? It may not matter for a motor, but it that concept physically accurate?

No. The generator may be grounded by a connection, but the generator itself has no inherent reference to ground. You can test this with your small gas generator. True, it is single phase only but the same idea applies. Unless you run a wire from the generator frame (grounding point) to actual earth ground it is floating. A voltmeter measurement from any of the generator outputs to earth ground won’t show real voltage. Nor will a light bulb light if connected from the generator hot to earth ground. Of course if the generator frame is sitting on wet soil it will ground itself. Or it will pick up your home electrical system ground if hooked in to your house.

A really simple equivalent- take a 12v car battery (or any battery) and measure from + to earth ground. Unless the - terminal is hooked up to a ground rod, you won’t measure any voltage. Same from - to ground. It is isolated by the car tires.

Also, relative to ground goes out the window as soon as you pass through an magnetically coupled transformer transformer, of which there are many in the power lines including one on the pole near your house. The only reason that ground matters for home power is your transformer’s center tap is connected to a ground rod. In some situations that connection doesn’t exist. Hospital operating rooms and some marine applications may use free floating (no reference to ground) power to reduce shock hazards. Note that auto-tranformers (buck/boost transformers) have only one winding so they are electrically coupled, not just magnetically.