What's new
What's new

Overload tripping on rotary converter


New member
Well I've spend a significant amount of time and money building a rotary converter, and I've got issues I'm not understanding.
I'm an engineer, but not electrical.


This is the Jim Hanrahan design I got from the sticky here.

Brand new 20 hp Baldor 3450 rpm motor.
WEG 50 amp contactors CWB50-11-30D15
WEG overload RW67-5D3-U057.
50 amp breaker in my single phase distribution panel.

1st problem:
The idler would run for a about 5 mins, on its own, with no other equipment being run from the system. The overload would then trip. It was initially set at 45 amps.
I then turned up the overload to 50 amps, and ran the system again with no external load. The idler ran for about 20 mins, then the overload tripped.
My clip-on ammeter is not working, so I can't measure the current in the legs, but I'm not getting why the overload would trip with no external loads (no other machines connected).

2nd problem/concern that I don't understand.
When running, I'm finding the following voltages:

Leg 1 to ground 122 volts.
Leg 2 to ground 121 volts.
Leg 3 to ground 236 volts - this is the generated leg.

Leg 1 to leg 2 243 volts
Leg 2 to leg 3 270 volts
Leg 1 to leg 3 254 volts.

Does it make sense that the generated leg to ground would be 236 volts ?

I'm using the following capacitance:
Start caps between T1 & T3 = 1200 mFD.
Run caps between L1 & T3, also L2 & T3 =270 mFD.

I've checked my circuits very closely with the diagram and don't see anything wrong.

So, why is the overload tripping with no external load on the system ?
And is it correct to have 236 volts on the generated leg to ground ?

Thanks for any guidance.
IEC (European) motor overloads trip on sustained current imbalance; where one phase is carrying significantly more or less than the others.

This obviously doesn't work for rotary converters.

I'm not sure what the recommended way of dealing with this is. Likely rewire so that L3 does not pass through the overload, and either L1 or L2 passes through twice (series), so that all three overload heaters see similar currents.


New member
The wild leg will sometimes run at those voltages. Not to big of deal except it will take out any computerized equipment. When I purchased my first computerized wheel balanced over 30 years ago they told me to put the wild leg of the adda-phase on a non computerized leg. Your breaker that is tripping isn't large enough. Google or Bing amps draw of a 20 hp and you will see, even though it isn't under a load it is still pulling plenty of power.


Active member
I'm no engineer or electrician, but I am working on my phase a matic 20hp rpc today, mfr recommends 125 amp breaker and #3 mag contactor (125 amp), in previous application I did get by with a 70A breaker. The generated leg at 236v may be correct, mine is reading 243v. Does any of this apply to your unit, I don't know.


New member
Good info guys, thanks.

Went and got a new clip-on ammeter, these results:

L1, L2 single phase supply both at 7.6 amps.

Taking measurements at the overload:
Leg 1 30 amps
Leg 2 19 amps
Leg 3 45 amps.

I have the O/L set at 50 amps. The system runs for about 20 mins, with no external load, then the O/L trips. Really don't understand why it takes time before it trips.
The idler motor stays warm, not hot.

The overload is adjustable up to 70 amps, I believe. Should I increase the value on the O/L ?
I'm using #8 wire for all the power lines.

Appreciate any help.
Thermal overloads are thermal; they heat up slowly. This means that you can overload the motor for a short period of time, as long as it then has some time to cool down. The temperature in the overload attempts to mimic the heat buildup in the motor; at least somewhat. If the phase imbalance/overload is not corrected, you would expect it to trip in minutes to hours.

Weg Catalog: https://static.weg.net/medias/downloadcenter/hcd/ha1/WEG-thermal-overload-relays-RW-50070227-en.pdf

Page 5 indicates that the trip currents are quite a bit faster if one phase is unloaded. More complex curves are not published, but I would expect two phases loaded to ~50% and one to 90% might trip the phase loss protection - I think this is a similar situation to what you see if a delta-connected motor loses a winding.

You could simply increase the trip rating, although I would tend to caution against it. An overload or breaker without a differential trip would be my preference; e.g. a K-curve MCB.


The high generated leg voltage to ground/neutral is correct. Maybe not the exact voltage, since your generated leg is higher than usual. Generally, voltages to ground or neutral are not important, and the important ones are line-to-line.

The neutral on the system is the center point on the 240V supply. It is NOT a 3 phase neutral, which would be about 139 volts from all lines.

The RPC creates "wild leg / stinger / farm / lighting tap" 3 phase, because of the way it generates the 3 phase based on the single phase 240V. The line voltages are fine, but the voltages to neutral are different from "normal" 3 phase.

Thermal breakers are indeed heat operated. That means that if they are in a warm place, their rating is less, and they may open below the nominal rated current. No idea where yours are, could they be hotter than expected? They claim to be compensated up to 60C, no idea how that tracks.

If they are directly connected to the contactor, as is common, then hot contacts on the contactor can cause extra heating. Wire size less than suggested may also cause extra heating.

The high voltage on the generated leg is associated with higher motor currents. The 20HP motor would be expected to draw about 54A per UL at full load. It probably would idle at around 25 to 30A.

Your 45A is high, and is probably associated with a larger than needed "balance" capacitor, which causes the high voltage.

Try reducing the value of the balance capacitors, which will reduce the motor currents and also the current imbalance. The imbalance may be causing part of the tripping problem due to the "single phasing detector" in the overload..


New member

This shows the panel I built. Overload on the right.

The Jim Hanrahan design suggests 12 to 16 mFDs per hp for the run capacitors, or balance caps.
I used 3 x 90 for each leg; so 270 per leg, or 13.5 mFD per hp. (In my original post I misquoted the size of the caps).

I removed 1 cap from each leg and got the following currents at idle, with no load:

Leg 1 - 25 amps, (was 30)
Leg 2 - 19 amps, (was 19)
Leg 3 - 25 amps, (was 45) this is the generated leg.

My voltages are more stable now, no external load, 242 v, 243 v, 249 volts.

I ran it for about 30 mins, with no external load, and the motor was cool to touch.

So I'm really pleased, and thanks for the excellent responses. I guess these values are OK,I don't really know. When I get my iron-worker or other load running, I'll take some more measurements.

I was thinking, if I could have welded it, or put in the lathe, I'd have been fine; but these electrical issues get really complicated !

Thanks again for the help !


The values are OK when you get the results you want.

It's fairly normal to set the generated leg a little high, because compensating with capacitors is not super-stable, and the voltage will fall under load. However, "a little" means to a value that is in the normal range of voltage, maybe 250V for a 240V line.


Active member
Run caps between L1 & T3, also L2 & T3 =270 mFD.

Every converter I built never had those two run capacitors equal. One capacitor is always about 55% in value compared to the second one.
What counts is the line to line voltage readings when the converter idler and machine motor are running together.

BTW, your post brought up advertisements for movies "Overload" and "Breaker Morant".:drink: