RPC run capacitors

I am considering adding some run capacitors to my self starting RPC. I am reading Jim Hanrahan's article at

formatting link
and I am confused by something. I understand how self starting RPCwould start with one cap between one leg 1 and leg 3 (the generated one). That's how mine is wired. Jim makes a point that it works, but makes unbalanced voltage.

But why would it start is capacitors are connected between 1-3 AND

2-3, like in this picture:

formatting link
I cannot see how it would create assymmetric fields needed to spin up the motor. Is that because capacitances across legs 1-3 are greater than capacitance between leg 2-3?

I could try to use run caps at run time and start caps at start time. In fact, I won a time delay relay for $9 on ebay yesterday, so I could set the RPC to start on start cap (both caps between leg 1-3) and then reconnect the same caps to become run caps, one between 1-3 and another between 2-3. Same TDR could, then, turn out output current aftet time delay, allowing the RPC to spin up and switch to the run mode.

As you can see, I am quite confused, but am willing to experiment. I have 4 unused Furnas 75 A contactors that I can wire, with the time delay relay, to do just about anything.

Idler: 10 HP

Capacitors: 92 mF each, 535 VAC rated, oil filled. I have 5 total, and use 2 for the starting leg, so three are unused.

Reply to
Loading thread data ...

You are correct saying that it's because the capacitance between 1 + 3 is higher than between 2 + 3. My RPC, 10HP 440V 50Hz motor, has 40uF between 2 + 3, 60uF between 1 + 3 and

500uF switched by a start circuit between 1 + 3. The 500uF is a bit too much, it starts very quickly (around 1/3 second) so I could probably drop that to around 200uF. Martin
Reply to
Martin Whybrow

Thank you Martin.

I have five 92 mF caps total.

What I have now is 2 capacitors connected between leg 1 and 3.

It seems that about the only thing that I need, then, wire it as follows:

  1. 1 cap (92 mF) between legs 1 and 2, permanently connected.

  1. 1 cap (92 mF) between 1 and 3, permanently connected.

  2. 2-3 more caps (180-270 mF) between legs 1 and 3, connected at start time and switched off by the time delay relay (TDR), say 4 seconds after startup.

  1. (optional) add another contactor on output side that would only switch output on when the TDR actuates and switches off the start caps.

Theat would get me an RPC that is:

- properly balanced across the range of output HP

- starts quickly

- only turns on when balanced power is output and full rotational speed is achieved, that is, when good quality 3 phase power is available.

Does it make sense?

Reply to

A 3-phase motor will start on single-phase current because the (3rd leg, middle leg, whatever) has phase-shifted current applied to it. That leg in a 3-phase motor operated on single-phase, has phase-shifted current applied via the start cap. That current is "sufficiently out of phase" with the current in the main winding to provide a starting field, if you will, for the main field to operate against, thus causing start-up. Remember, we are talking about a single-phase motor. After the starting interval, the motor will continue to operate as a single phase machine.

The start cap, if left permanently in place, forms a "self starting" RPC. A rather large amount of capacitance is required for starting any motor, be it single-phase or 3-phase, when starting on single-phase current. Therefore, a 3-phase motor running with an over-large starting capacitance, permanently in place, is likely to exhibit severly unbalanced running voltages. This is why serious RPC builders always seperate the starting and running functions.

Assymetric fields are necessary for starting, as above. The "running current" flow paths in a RPC are quite complex and are also assynmetric even in a tuned, voltage balanced RPC because of direction of rotation, among other things. Suffice it to say, the current flow in a RPC and its load requires "steerage" (think series resonance) in order for the fundamental single-phase running current to be guided into paths that emulate 3-phase conditions. Remember, you are still dealing with single-phase power. A RPC does not "generate" 3-phase power - it merely performs adjustments in a fundamentally single-phase scenario which emulates 3-phase.

By all means do "use run caps at run time and start caps at start time". Forget about reconnecting start caps and using them as run caps. Either use a simple push-button switch to temporarily connect the start caps or a NC potential relay that senses 3rd leg voltage to open up the start circuit. Then leave the start circuit alone. Period.

It occurs, you are attempting to use components that may not be appropriate, or the best way, just becasue you got them on the cheap. No amount of aimless, and possibly dangerous experimentation, with the wrong things can necessarily force success.

Bob Swinney

Reply to
Robert Swinney


Makes full sense.

That's fine. I must note though, that I want to use the same kind 92 mF 535 VAC rated oil filled caps for both starting and running, just because I already have them and they are built like tanks.

See a paragraph near the end of my post where I discuss run caps vs. start caps.

I could use 2 for dedicated starting, and 2 more for running.

Or, I could use 2 for starting, and then reconnect one of them to another leg pair, with a relay. That could possibly save some space in the RPC as these caps are rather large. That would also leave me with more caps for future projects. Surely, if that's dangerous, I may choose not to do it. I just would like to understand the issues involved.

At least I have something working now.

As far as I understand, capacitors that can be used as "run capacitors" are superior to capacitors that can only be used as "start capacitors" in every possible way, EXCEPT that run capacitors are much more expensive. Caps that are sold as start capacitors cannot be left plugged in permanently and could explode. They also lose their capacity over time. Run caps do not do that. Mine are run capacitors and can be plugged in indefinitely.

My caps are 535 VAC rated, oil filled, continuous duty 92 mF capacitors.

You can see their pictures at

formatting link
formatting link
So, I am not afraid of using them for essentially lighter, starting duty.

ALl five cost me $5 ($1 apiece), I could never hope go get anything like that for similar price on ebay. Someone said that they could go for $40 apiece, which I do not quite believe.

Reply to

With five equal value oil filled capacitors your options are fairly limited but you can still finish up wth a perfectly good converter.

One thing to remember is that capacitors of this type have extremely low internal series resistance (milliohms)- that's why you get a sizable bang if you short circuit a charged one! The same thing happens when you connect a charged capacitor to an uncharged one - VERY high peak currents flow as the charge voltage equalises on the two capacitors. The capacitors are not greatly bothered by this treatment but it's very unkind to the switch contacts or relay contacts used to parallel connect two capacitors if there is a substantial voltage difference at the instant of connection.

Brute force oversizing of the switching contacts can give reasonable contact life but it's much neater to avoid the problem by using separate start and run capacitors that are never parallel connected. This is easily done with your capacitor collection.

Use three parallel connected as your start capacitor. Only in circuit for the few seconds needed for the idler to run up to speed.

You now have a bit unbalanced but perfectly usable converter system. Ideally the load motor should not be switched in until the idler is up to speed. However, if the load motor is initially running light, this start capacitance is probably enough for a simultaneous idler and load start.

Your START switching should be a changeover contact which EITHER connects the start capacitor OR the run capacitor across L1 and L2.

For the run capacitor, try one capacitor or two capacitors series connected to halve the effective value - whichever gives the best voltage balance with full load on the load motor.

Do not place a capacitor across L2 L3. With the values you have available this would do more harm than good.


Reply to

An excellent point.

Yep, makes full sense.

Now, based on common sense, unlike CONNECTING capacitors, DISCONNECTING them should not produce any big sparks.

I have a plan for a start circuit with a separate contactor. The sequence of events is as follows.

  1. I turh the ON switch.
  2. The contactor that connects starting caps engages.
  3. When it engages, it actuates the SECOND contactor that supplies primary 240V voltage to the electric motor.
  4. Everything starts spinning.
  5. After a few seconds, a time delay relay that I won on ebay today (NO and NC), will open a pair of contacts and DISCONNECT the starting caps.

I think that I know how to wire this properly so that all of this happens, and yet everything shuts down when I turn the on/off switch to off. I have a mental picture. Basically both contactors would have a neutral wired directly to neutral, but the 110V line for signal would go through the switch, then to 3rd poles of second contactor, and to the 3rd pole of the first contactor but through a time delay relay. then from the switched side of both contactors to the hot switch terminal of the second contactor.

  1. When the NO contacts on the time delay relay close, it would actuate the third contactor that would allow power to be output.

Sorry if I am speaking in strange language, I am a computer programmer. It should work, logically.

read on...


right. it's much nicer to have something working and add stuff to it, than build a "dream system" without testing to only realize that something went amiss. That's how I do my programming too.

Looks like one run cap per side should be good, based on Jim Hanrahan's writings. I'll see.

Hm, why is that so?

Anyway, my main question is: if I do all this (as I outlined in steps

1-6), will I get a Mercedes Benz of phase converters? Or am I wasting my time for load up to and under 5 HP? i
Reply to


It would be rude for me to agree that you'd be wasting your time by sophisticating this RPC. I can tell you that nothing you are likely to

*ever* use in your house will be able to operate any better because of *anything you do beyond just spinning up that 10 HP idler.


Reply to
Jerry Martes

Thanks Jerry... I think that you are right... It is fun to play with logic and relays etc, it is like computer programming. I'll try to think of more fruitful things to do.

Reply to

This is an extract from an earlier post which goes some way to explaining the peculiarities of "balancing" capacitors.


A converter of this type is basically a capacitor/inductor phase shift system which produces an open vee 3 phase system. This phase shifter is a series resonant circuit and when it is set up to give the

60 deg phase shift it is working a long way below its natural resonant frequency. 60 deg is of course the correct phase angle between the two legs of an open vee system.

The motor(s) is the inductor in the system and unfortunately the apparent inductance of the motor changes with rotor speed. For any particular rotor speed greater than about 90% of synchronous speed (the lower limit varies a bit with motor type) it is possible to choose a capacitor combination which produces a pretty close approximation to balanced 3 phase at the motor terminals.

For near the full load rated speed of the motor, large run capacitance is needed with most or all of it as a single capacitor feeding the phantom phase from supply live. At light load the speed of the rotor rises and if the capacitor value is chosen to achieve the right phase angle the phantom phase voltage will be excessive. This could be corrected by feeding the capacitor from a lower voltage single phase source but this would mean feeding it from an auto transformer across the supply.

It is much simpler (and of course everybody does this) to use two capacitors arranged as a voltage divider to simultaneously achieve the correct phase angle and phase voltage.

The effective capacitanceof the two capacitors connected in series across the supply is the sum of the capacitances because the source impedance of the supply is zero and this effectively parallels the two capacitors.

Because the they also act as a voltage divider, this sum capacitance is effectively fed from a voltage of supply voltage times C1/(C1+C2) where C1 is the top capacitor and C2 is connected phantom phase to neutral.

Because it looks nicely symmetrical there seems to be a tendency to believe that C1 and C2 should be equal and any inequality in their optimum value must result from some strange second order effect. This is NOT true. There is nothing magic about equal C1 and C2. It simply results in a capacitor of value C1+C2 fed from half the supply voltage. At this low effective supply voltage it is only possible to get close to balanced operation at no load or light loads which enable the rotor to operate close to synchronous speed. As the load increases with consequent slowing of the rotor speed the total capacitance needs to increase with both more in C1 and less in C2. By the time full load is reached the optimum value for C2 is usually zero.

These effects are very noticeable if you're using a single motor on a variable load up to near rated full load power and some compromise necessary. The saving grace is that industrial motors are surprisingly tolerant of reasonable overvoltage when operating at light loads so the trick is to size the capacitors for at or near full loads and to accept some overvoltaqe at light loads. This increases the motor losses at light load but the total motor losses still remain below the losses at rated full load so temperature rise is acceptable.


In your case, while a small additional capacitor across L2 L3 might yield better phase balance for some load conditions, the 92uF you have available is far too large and would do more harm than good.

I fully support Jerry Martes comment to the effect that fine tuning rarely achieves significant practical benefit.

If you've a nice big idler and have solved the starting problem, careful choice of balancing capacitors may give you a nice warm feeling but you're unlikely to notice much practical difference in the overall performance!


Reply to

PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.