Capacitors in a Rotary Phase Converter
Ok. Here is one for you electrical wizards.
I have a set of plans for building a R.P.C. which shows two capacitors. One for starting and one for running. the values that are given in the plans are a little confusing. I am going to be using a 5 hp motor for the R.P.C, and will be pulling a 3 hp load for my mill.
the plans show that both capacitors need to be 240V. However, In the McMaster-Carr online catalog, I can only find motor run capacitors in either 370 VAC. or 440 VAC.
the plans also show only one run cap, and show it wired from the 240V hot lead to the cap, and then back to the generating lead of the three phase R.P.C. motor.
In some of the posts I have read, some have used multiple caps to balance the voltage. In my limited knowledge, it seems like a good idea to me. My question is what size caps do I need for both starting, and running, and what is the best way to wire them in so that I don't end up letting the smoke out. i would like to have it so that the three phase R.P.C. motor will start on its own. If need be, I can put in a kicker motor for starting, but would like to keep it simple, (so to speak).
Starting and running an RPC are two quite different processes.
Let's take the running condition first.
Ideally, an RPC should be balanced, that is Cab and Ccb (b representing the manufactured phase), therefore the capacitor from L1, which is also the A phase, to B and the capacitor from L2, which is also the C phase, to B should be the same values, thereby creating a balanced condition.
As these capacitors are always in the circuit they MUST be rated 1.414 * 1.1 = 1.56 times the applied line voltage, and for our purposes 370 volts is correct.
Now, a balanced converter cannot self-start; in order to start a very imbalanced condition must be created within the idler.
For this purpose, a starting capacitor must be temporarily placed across one of the run capacitors (it really doesn't matter which at this point as the RPC is balanced), and this capacitor should be an electrolytic capacitor rated for the applied line voltage, and for our purposes 250 volts is correct.
If the values have been properly selected, the RPC should start in two seconds or less and the B phase voltage should be within 10 percent of the applied line voltage.
So much for a balanced converter.
Fitch Williams has shown that the transient performance of an RPC is best when there is a slight imbalance, often about 60 percent/40 percent.
In this case, the starting capacitor should be placed across the larger of the two running capacitors.
Finally, the power factor may be improved by adding a capacitor Cpf between A and C, but this should be done last, after the RPC is meeting all performance expectations. This capacitor should also be rated 370 volts.
As usual, per NEC, the capacitors should be after the disconnecting means and before the idler overload protection, assuming the idler is protected from overloads.
Thank you Peter for that excellent explanation. Soooo. I see from the catalog that there are a lot of different microfarad ratings to choose from. For the start cap, do I want to go for the highest rating (270-324)? and for the run cap, there are ratings from 2mf to 50mf. what would you recomend?
Alas, one cannot say without actually trying it.
The motors which are employed in home-built RPCs usually fall into these classifications:
1) pre-NEMA motors (roughly, pre-1952),
2) NEMA motors,
3) late NEMA motors, and
4) high-efficiency motors (roughly post-1999).
As the motors improved their efficiencies, the requirements for building RPCs with them also significantly changed.
It is usually best to have a variety of capacitors available and to "tune" the RPC to achieve +/- 10 percent voltage no-load to full-load and to start within two seconds or less.
If you buy your capacitors from a source like Burden in Lincoln, NE, you will get good capacitors at a very reasonable price:
There are, of course, other surplus sellers, but I have had very good luck with Surplus Center.
Motors used in commercially available RPCs are very often modified to achieve what is claimed to be a "high resistance" rotor. These are very often just off-the-shelf motors which are modified by trimming some of the aluminum off of the ends of the rotor, thereby increasing the rotor's resistance, which is otherwise quite low.
For a load motor, you want a low-resistance rotor; for an RPC, you MAY want a higher resistance rotor.
thanks again Peter.
The motor that I have for the R.P.C. is an old General Electric, made in Schenectady N.Y. I'm sure that it is pre-1952. i jotted down some info from the name plate, and maybe you can help me date it a little closer. the model # is 5K254D136, Type code is "H", Frame is 254,rpm is 1735, Ser# is 6325265. She is old and heavy, but has the classic look of an old electric motor which for some reason appeals to me.
Can you give me some idea as to where to start with the microfarad ratings on the caps? On the run caps, does the one want to be 60 or 40 percent of the other to get it out of balance? should I go with the big 270-324 Mf for the start cap? I'm a little lost, and any suggestions you can give me would be greatly appreciated.
Alas, there are no hard and fast rules with RPCs.
One COULD generalize and state that early motors generally require LESS capacitance than late motors.
You may want to investigate RPC topologies using Steveco 90-66 potential relays ...
RPC Topologies Using Potential Relays
... which I authored some time ago.
thanks Peter for all of your help. i guess i have some experimenting to do.
The Fitch Williams design has been made in a variety of capacities.
Obviously, the 5 HP (or less) is the easiest as the Steveco 90-66 potential relay has a built-in contactor for the starting capacitor and it is rated 5 HP (the General Electric potential relays are rated 3 HP).
So, I would suggest taking a F.W. design which approximates your situation and start from there.
It is a fact of life that every RPC is necessarily different.
thanks again peter. I did look up the Fitch Williams design, and found a couple of others that also gave some good ideas on where to start on determining what size capaciters to use. Your help was greatly appreciated.
I've seen Fitch's slight imbalance 60-40 suggestion before. This begs the question: which line gets 60% and which gets 40%? Obviously L1 and L2 are differentiated in an RPC design with a start capacitor, but what about in a design which uses a timed pony motor? In that case, maybe imbalanced doesn't make sense?
You are correct unless it improves voltage balance to have different values. RPCs require trial and error.
Originally Posted by metalmagpie
Ok folks, I got a bunch of caps from the surplus store, and my cabinet is almost ready to start putting rverything in place. The question that crossed my mind is: since i will be using run caps at different values to balance the phases, and will be using more than one cap on each phase, do I wire them in series or parallel?
well I did some more reading on the subject, and in case anyone else needs to know, when you are combining two or more capacitors to get a given value, they are to be wired together in parallel. Thanks for all of the helpful replies. I plan on installing everything in the cabinet as soon as the paint is dry.
"... do I wire them in series or parallel?"
If you have selected capacitors of the recommended ratings (370 volts for 240 volt RPCs), then you would wire these in parallel.
The only time you would wire these in series would be to obtain a higher voltage rating, such as would be required for a 480 volt RPC, in which case you need capacitors rated 740 volts, which is not anywhere near a standard rating (600 volts is usually the maximum).
My dad used to build and sell RPC and static PC. Be aware that if you fry the relay contacts together and the start caps stay online sometimes they will go BOOM, they are not made for continuous duty.
I'm not sure if you could use some kind of solid state relay maybe triggered by the voltage sensing relay or not. When properly set up the normal relay will work great for a long long time, but when you are playing around getting things working you can weld the contacts on the relay together.
For just a RPC just a timer relay might work just fine too, this could be used to actuate a contacter to cut out the start caps....it would work much like the setups that start a motor in one fashion and run it another (star vs delta ?).
Instead of a relay, I am going to be using a drum switch. I will have it so that if I place the lever in one position, it will energise the start caps to start the motor, and as soon as it comes up to speed, i will flip the lever to the run position.
I built my RPC according to the plans, but it started so hard that it shook the cabinet. I took out half of the start capacitance, and it now starts very easily with no jolt. Regards, Clark
One issue with that, if your power goes out with the RPC running, when power comes back on you have a bad deal going on there, might be better to use a motor starter and a push button or other deadman switch for the start caps, that way if you lose power it resets.
Originally Posted by loggerhogger
good point Willbird, but where I live, power outages are very rare, and if it does happen, I am pretty anal about making sure that everything is turned off before leaving the shop. Ieven throw the breakers on the bigger machines when not in use.