Capacitance required to balance RPC

Hello, I'm fairly new to the site and have been reading up on making a RPC.
There are many plans out there for 5hp and up idlers. I'm using a 3hp 3 phase to power a 2hp Bridgeport Series 1 CNC. I'm going with cheap and easy as I'm not looking to make parts on this machine. I just need it to run so I can sell this one. I have another identical machine that will get a full rebuild. I don't plan on needing the RPC for that one.


This is all for the joy of learning something new, so please don't be going off on me about wasting my time with junk! It's my time! I work full time managing a gage lab and enjoy making things work in my spare time. Cars, Jeeps, Bikes (street & dirt) 4 wheelers, Personal Watercraft, Arduino projects. Now CNC! Starting with an RPC.

With that out of the way I'll get around to my project.
I'm leaning toward the Fitch design.
I have a Jacobs 3hp 3 phase motor wired for the low voltage setting.
First I thought I would need a base reading so; one leg to 3-9, and the other to 1-7, with 2-8 to be the generated leg. Rope start the motor and throw on the power.

generated leg to Black = 209V, to Red = 198V, to Neutral = 166


Measuring the other two legs with it running read the stock incoming voltages of B/R = 233 and 113/119 to neutral.


Since this is such a small idler I planned to use a SPP6 Hard Start kit as the potential relay and start cap. SPP6 Hard Start Capacitor for Supco Relay 1/2HP-1HP | eBay
The run capacitors have me stumped. I've read that anything over 250V will work (preferably oil filled) but the amount/value of microfarads required to balance each leg would be helpful. All the balancing posts I've read are a little vague on how much cap is needed to raise a certain voltage. There's no store round here to just go buy caps, and I don't want to buy a pile of various sizes and "hunt & peck" to find what works.

Does anyone have the values or the formula to balance these readings?

The power running to my barn is 2 hot legs and a ground as neutral, with a separate ground rod staked at the barn. I was told this is common, but so is DEATH!
I read of the control transformer trick to get a real neutral that's not also the ground and I made one up. Is this a valid method? Or should I look into running a Neutral to the barn?


Thanks in advance for any helpful information. I take constructive criticism as helpful, especially concerning high voltage!

Bump it up!
Anyone put together a 3Hp RPC? What run caps did it take?

it will depend on the idler..... There is not really a "universal generic" default value.

You may not even need the balance capacitors, depending on the nature of the load. An oversized idler has less need, as the voltage drop is typically less with a big idler.

In some cases, it may actually be better to use a boost transformer instead of balance caps. That more directly addresses what you really want..

JST said:

it will depend on the idler..... There is not really a "universal generic" default value.

You may not even need the balance capacitors, depending on the nature of the load. An oversized idler has less need, as the voltage drop is typically less with a big idler.

In some cases, it may actually be better to use a boost transformer instead of balance caps. That more directly addresses what you really want..

Click to expand...

At first RPC build, I played with balance caps until I collected quite a number of them. I had jumper wires galore!.

Slowly, I started removing the caps, the latent danger, even with bleed resisters, concerned me. I fond nothing amiss with the absence of "balance caps". Start caps remained even if only re-purposed run caps. The motors I have would start themselves with a nudge from a foot.

Now, with the advent of VFDs, I only keep one RPC, it's a three HP GE sleeve bearing motor that sits on the floor behind the mill. The mill has a two speed spindle motor, a Horizontal shaft motor, a table feed motor and a coolant pump. I thought that combination would be too much for my simple brain to fit a VFD.
OH! Plus the same RPC powers the Wade lathe with PLUGGING for spindle reversal. Ouch!

There are NO run caps in this set up. Starting of the idler is by a "Phase-o-matic" kit built "phase converter" which is nothing but a start cap and a starting relay (voltage senstive, AKA Steveco)

The set up works PERFECTLY, and has for more than 20 years. Well, I can't say perfectly. Sometimes, when the mill spindle motor is set on low speed, and nothing else is energized. I need to hold the spindle OFF button for the count of two before the relay drops.

I get around this by energizing the table feed motor, or the coolant pump, or both.

An expedient to capacitors for balanced power from a RPC is added motors!

Thanks for responding.

JST said:

In some cases, it may actually be better to use a boost transformer instead of balance caps.

Click to expand...

Now you got me thinking I may not need to balance the input to this old CNC. I can't find a print for this model, but the power goes directly into a three winding transformer that knocks it down to 80VDC and 24VDC per leg. Then straight to three more transformers with no values labeled on them.
These old Bridgeport Textron CNCs have some serious power management in them, lots of transformers and big blue VDC caps.

Would it damage anything to run the machine with a low voltage condition as described above?

The low leg will be "lazy"..... not supplying full current. So the others will have to pick up the slack. At low power, that will not be an issue, but could be if the machine is asked to produce full power for long. It is a sort of "single phasing" problem.

The internal transformer is not a direct fix for that, As generally used, it just repeats the voltages that come in.

IF you can re-set the transformer to different voltages per winding, in small increments of a few volts, THEN it may be possible to "balance" the load with the internal transformer. You want to be careful though, you do not want to forget what is what and get into trouble later if moving the machine to a different power source, etc.

Ok, one of the E.E. techs at work told me low voltage would start drawing more current, and I don't want to start popping stuff.
I checked with the previous owner of these machines, and he had a balanced/tuned RPC.

I don't know enough to start changing the winding of a transformer. I have a MOT to play with but I haven't cut into it yet. I need it to heat a swing-arm bolt on a 4-wheeler that's seized.

So I'll try to get some caps to start with.
I'm just tossed as to get 15Uf or 50Uf to start with, and how many.

Any suggestions?

Low voltage does pull more current if talking about a motor that has a certain load on it. Power has to come into the motor in order to be turned into shaft power. So if the voltage is lower, the current has to be higher. If the motor was already fully loaded to max at normal voltage, then if the voltage drops, the current has to go up. If the motor was already at full load current, it will be in current overload. And, 10% extra current means over 20% more heating. That can fry motors.

CNNC often has a DC bus and servos with drives running off the DC. Rectifiers act differently than motors. the highest voltage phase tends to charge the caps to a higher voltage than the others, and then the others, since they are lower voltage, only charge the bus during the time their voltage is higher than the DC bus, which is less time than the highest voltage phase.. So the highest voltage phase does more work and draws more current, working those diodes harder than the others, etc.

If the previous guy ran OK with an RPC and balance caps, I wonder how hard he ran the machines. If you don't run them hard, there may be no issue.

If you are not going to run at full power, then I'd not fool with the transformer etc. You can go ahead and try to balance the RPC and just run the thing.

I found a local HVAC supply house that carries caps in this range.

After some trial and error I ended up with 2 - 50Uf on Black L1 to the generated leg, and a 25Uf on Red L2 to the generated leg with equal voltage of 247. (And some spare caps, not bad for less than $30)

The current has me a little puzzled. Black = 9.35, Red = 5.05, Generated = 11.35! The motor is rated 9.6 at 240V.
I tried adding different Uf between the L1 - L2 as the Cpf with almost no change so I left that out.


The Bridgeport was under power for about an hour with the spindle running and Y - Z axis moving just fine. X axis moved a little bit then the stepper just humms while the readout displays movement. The gib is loosened and ways are clean, so I'm thinking it may be power related.

Thoughts?

Oh, I'll post v/a readings of while it's running later.

BDGiven said:

I found a local HVAC supply house that carries caps in this range.

After some trial and error I ended up with 2 - 50Uf on Black L1 to the generated leg, and a 25Uf on Red L2 to the generated leg with equal voltage of 247. (And some spare caps, not bad for less than $30)

The current has me a little puzzled. Black = 9.35, Red = 5.05, Generated = 11.35! The motor is rated 9.6 at 240V.
I tried adding different Uf between the L1 - L2 as the Cpf with almost no change so I left that out.


The Bridgeport was under power for about an hour with the spindle running and Y - Z axis moving just fine. X axis moved a little bit then the stepper just humms while the readout displays movement. The gib is loosened and ways are clean, so I'm thinking it may be power related.

Thoughts?

Oh, I'll post v/a readings of while it's running later.

Click to expand...

I don't know what context the man told you the lower voltage increases current to the motor, this is only true when all three phases are low voltage and when the load is higher than the optimal voltage for that load. so for instance, an induction motor at no load, the line amps will decrease until the voltage drops to on the order of 20-30% of nameplate volts. at 75% load the optimal voltage is about 70-90% of nameplate voltage. at full load, anything less than nameplate volts will increase current.

In the case of an induction motor rpc, the generated leg is going to produce less current into the driven motor because its voltage is lower and lagging behind where it should be. but at a given load, the total amps into the motor will usually be higher because the third phase is weak. adding capacitors works up to a point but to find the optimal value probably requires a watt meter as well. simply adding capacitors until the third leg amps into the load are equal to the other two phases may not be the optimal solution.

The amps you are measuring of 9.5, 5, 11.35 are nearly 2-3 times as high as they should be for a 2-3 hp motor at no load. where did you measure these amps at? do they include any current flowing through the capacitors?

if you remove the capacitors and use an rpc to generate the third leg, you will have no amps on the third leg once the driven motor is up to speed if it has no load and is similar in size to the rpc.

50uF is 4.65 amps of reactive current at 247 volts btw.


but regarding the servo drives, if the incoming 3 phase voltage is simply rectified to dc, then it will work on single phase, same as any vfd will but there are thermal limitations which will inspire the capacitors or rectifiers to fail (unlikely). And dc ripple limitations which may impact performance depending on how well the dc ripple is removed from the pwm loops. (not unlike the 60 and 120hz hum from an old guitar amp with failing capacitors) and this 120hz ripple will only show up when trying to max out the torque at high velocity in a program.

Here's my record of values starting with the Lincoln AC motor TEFC, 182T Frame, 1750 RPM, 3 Hp, 3 Ph, 8 INS, Max Amp 40, 200/400 Volt Amp 9.6/4.8
Wired to the low voltage setting.
Motor wire to L1 Black (B), L2 Red R, Generated G, Ground/Neutral N.
3-9=B, 1-7=R, 2-8=G

1st pull start (I only have a cheap volt meter)
G-B= 209, G-R=198, G-N=166, Incoming B-R/R-B=233, B-N=119, R-N=113

Add 50Uf to B-G
G-B=234, G-R=217, G-N=191, B-R/R-B=237, B-N=121, R-N=115

Add 25Uf to R-G. (Bought a Clamp Meter)
G-B=233, G-R=228, G-N=197, B-R/R-B=237, B-N=121, R-N=115
Current Measured at the wires coming out of the 3ph idler motor (RPC).
B current=7.2, R current=6.4, G current= 5.3

Add 25Uf to the 50Uf on B-G, and change the 25Uf to 50 Uf on the R-G.
Connections between capacitors is made with jumpers to the existing cap.
B===Cap1===Cap2=open
G===Cap1===Cap2=open
Not;
B===Cap1===Cap2===G
Readings: G-B=237, G-R=251, G-N=214, B-r/R-b=235, B-N=120, R-N=114, B current=6.8, R current=7.05, G current=10.66

Change B-G to 2 50Uf caps, and R-G to 1 25Uf cap.
G-B=247, G-R=247, G-N=219, B-R/R-B=233, B-N=120, R-N=112, B current=9.35, R current=5.05, G current=11.35

Notice the generated leg to neutral value rapid rise. What's up with that?

I plan to swap the leg wires to the mill to see if a different axis is affected, because I noticed voltage dropped to about 213 and 203 measured B-G and R-G with the mill running.
Should I add more to the lowest leg to try to make it more balanced while running a load?

As you can tell; I really don't know what I'm doing!
That's why I'm here!

i presume you have 120/240v single phase, but your readings are consistently measuring 6 volts difference from one side to the other? while this may not be a problem it could be in the future if you have a bad neutral.

Secondly, this 200/400 volt motor that you are using as an idler is that a 50hz motor? if so that's fine. if its a 60hz motor then it is running saturated. 200v motors do exist for 60hz they are designed for 120/208 three phase, not 240v three phase.


anyhow, the generated leg to neutral should read 208 for your application assuming 120/240v single phase input, and yes adding capacitors causes that voltage to climb rapidly, that's the whole point.

your initial voltage of G-B= 209, G-R=198, G-N=166, Incoming B-R/R-B=233, B-N=119, R-N=113 seems really low for a 3 hp motor.


if you're certain this is a 1750 rpm motor then using it as an rpc at 60hz 240 explains the high amps and the low volts with no capacitors.

johansen said:

i presume you have 120/240v single phase, but your readings are consistently measuring 6 volts difference from one side to the other? while this may not be a problem it could be in the future if you have a bad neutral.

Secondly, this 200/400 volt motor that you are using as an idler is that a 50hz motor? if so that's fine. if its a 60hz motor then it is running saturated. 200v motors do exist for 60hz they are designed for 120/208 three phase, not 240v three phase.


anyhow, the generated leg to neutral should read 208 for your application assuming 120/240v single phase input, and yes adding capacitors causes that voltage to climb rapidly, that's the whole point.

your initial voltage of G-B= 209, G-R=198, G-N=166, Incoming B-R/R-B=233, B-N=119, R-N=113 seems really low for a 3 hp motor.


if you're certain this is a 1750 rpm motor then using it as an rpc at 60hz 240 explains the high amps and the low volts with no capacitors.

Click to expand...

It is a 60 Hz motor. What Do you suggest? Can this Work?

Further above, you mentioned a 240V motor, which I thought was the idler.... maybe that was the load motor?

Using a 200V motor in a 240V situation is not going to work well. I agree that the currents are going to be distorted by the overvoltage condition, and you just need to use a 240V motor for the idler.

An unloaded motor usually draws about 30 to 50% of full load current, depending on design issues, etc. As an idler, it may draw a little extra, since just one phase is being powered.

ANY normal motor used as an idler motor is going to produce low volts on the generated leg. The "output voltage" is really just the "back EMF" of the motor. That HAS TO BE lower than the actual line volts, or the motor would not draw any power. So the generated voltage starts out low, and sags farther under load.

The whole idea of the "balance" capacitors, is to cancel out some of the motor inductance that causes voltage drop, and possibly to get close to resonance with the motor inductance, which will raise the voltage. But the resonance condition is affected by load, so the boost effect is damped down by the load motor. A sufficient capacitor to raise the voltage under load, may produce a very high voltage at low load, or a lighter load, which is not desirable.

One can also boost the generated leg by a special idler motor that is wound to produce more generated leg voltage, OR by a boost transformer that raises the output voltage to equal, or slightly exceed the incoming line volts. Those approaches do not have the disadvantages of the capacitors, but are rarely used.

An expedient to capacitors for balanced power from a RPC is added motors!

Click to expand...

Yea, I did this.

I used a 5HP static phase converter (which is really just a bunch of caps) to start one motor. As more motor is added to the circuit and spinning, the voltage became more and more balanced.

BDGiven said:

...
generated leg to Black = 209V, to Red = 198V, to Neutral = 166
...

Click to expand...

Your initial numbers shown above.

1) do not bother measuring line-to-neutral. Does not matter.

2) balance as shown is within 10 percent. Leave it at that.

3) do not bother measuring line currents with a clamp on meter on the manufactured leg - this will be confusing for you.

4) read the Fitch Williams instructions again.

5) for a slightly larger idler motor you will see even better balance w/o capacitors added.

If the machine runs fine, you are done - don't over-think this. More meters does not mean better.

An old thread comes to life again!

However, it may be helpful even if old.

Voltage balance does look good. From that standpoint it might appear fine.

Current is a better way to assess final balance, because voltage does not ensure current, and current runs motors. If the generated leg is not supplying current, it is providing no torque and you are not getting the benefit of 3 phase.

Despite Jim Rozen's "ODS" (Oersted Derangement Syndrome), the clamp-on meter is actually a perfectly good method of determining if the motor is actually drawing current, and how well balanced the currents are. If it shows current, the current is flowing. The amp meter does not show actual phase, but that is not too important, it can show if phase is good enough.

The voltage balance actually can be confusing, because it does not show you if you are getting any actual power to the machine motor.

The two things affecting the measured voltage are voltage, and phase. The correct voltage, at the wrong phase, will not correctly provide power. A phase error can show a low voltage even if the actual voltage is good. Also, the voltage needs to be high enough to overcome the "back EMF" (generated voltage) of the load motor.

In general, the phase will be OK with NO capacitors. A motor inherently develops back EMF in the correct phase, due to it's physical construction.

Once you add capacitors, you start to affect phase. In order to raise voltage, you have to add sufficient capacitance to approach resonance with the effective motor winding inductance. As the voltage is raised, the phase changes also, along the normal resonance curve. Usually, this is not an issue, but the effect exists. Often capacitors are split, some connected between the generated leg (phase "B") and phase "A", others between it and phase "C", in an attempt to correct phase.

Any motor develops a back EMF when energized and spinning. That is going to be around 90% or so of the applied voltage, but will differ with different motors. The difference between that back EMF and the applied voltage is the "effective" voltage that actually drives current through the motor. So, since the idler and the load motor are both generating roughly similar back EMF, you would not expect much current flow on the generated leg if you just connect them.* That's the reason for capacitors etc to raise ("balance") voltage.

At idle, the current flow is not too important. It is good if it exists, but under load is where you need it.

The issue is that at no load, the voltage (from A toB or C to B) should not be extremely high, but it should be at or slightly above the incoming voltage . For 240V, a maximum of about 260V is best. However, the voltage will always be higher at no load than under load, and a "balance" capacitance which is correct for the loaded condition may lead to a very high voltage at no load (due to resonance effects). So it may not be possible to provide a "one size fits all" correction at the RPC. (see below)

The way to determine correct balance is to measure and see that when the machine motor is loaded (doing work) there is current flow in the generated leg, and that the voltage (A-B and B-C) is at a reasonable value. Hopefully that flow is fairly balanced between the three wires to the motor. Voltage should be balanced and appropriate before checking current.

At the least, the current on "B" should flow, and increase with load, just as the current on "A" and "C" does. The better balanced it is between the three wires, the better the motor will perform.

So, first get the voltage corrected with capacitors with no load on the converter. I'd suggest that the generated leg ought to be set high, up to 10% high, relative to the other two wires. Get the voltages balanced to the other two wires, so that the voltage from "B to C" equals that from "B to A". That can be done to some extent by changing how much capacitance is connected "B to A" vs "B to C"

Check the current flow on all three when the load motor (machine motor) is running unloaded. Ideally the currents will be balanced, but that might not happen.

Now apply a load, by running the machine with it doing work. Check that there is more current flowing in the phase"B" wire than at no load. If you can apply a heavier load also, check that current increases again. As usual, it is best if the currents are balanced, but that is not always possible.

If it seems that the voltage sags and current drops under load, then it may be possible to add one or more capacitors, but do that at the machine motor. You want the capacitors to only be in-circuit when the motor is running, since you already have a good "generic" voltage balance with no load, and adding more capacitance is likely to lead to extremely high voltage at no load..

But the key here is to check BOTH voltage and current. Voltage makes sure that nothing is way off, while current verifies that the RPC is actually supplying current (power/torque). And to check them under load, because that is what you want... proper operation under load.

What you want is to have good voltage that supplies proper current across the range of loads. You are trying to make 3 phase that is equal to what the power company would supply.

By just checking voltage or current alone, you can get way "off in the weeds" trying to optimize the converter. You need BOTH to be at reasonable values.

An alternative to the capacitor method is to use a boost transformer to bring the generated leg (phase wire "B") up to proper voltage. That is more expensive, and adds some resistance, but it should work better. Because it does not depend on a resonance effect, the voltage boost is more stable.


* Under load, the machine motor slows, which reduces back EMF, and can allow the generated leg output from the RPC to provide current even if the voltage is low. The question is how much difference it makes, because the slowing is generally not large, only 4 or 5 percent. That alone will not completely correct for an idler voltage that is 10% low.

An amp clamp meter will show very little current on the manufactured leg, until the load motor is actually drawing current. It's an odd effect you would see if you ever did this - an amp-clamp willl show large currents on the load motor utility legs - but those are nearly entirely reactive. The manufactured leg will show very low current, until the motor is loaded. As the load motor does more and more work, the amp-clamp on the manufactured leg will begin to rise until it equals the utililty leg reading.

The current on the utility leg does not appear to change at all as the load increases, what's really going on is the current shifts from reactive to more in-phase. Try it sometime. This is why I suggest that the amp clamp readings on all three load motor leads *will* confound a novice.

Another reason why Fitch's discussion of "how to tune a rotary converter" has absolutely zero mention of inductive clamp meters. He was aware of the pitfalls.

I suspect Fitch would not be in a position to weigh in on this discussion. Unfortunately.

Sigh.....

Again with this? Really?

If you take a look at the standard circuit model for a motor, the reactive portion remains constant aside from a small parasitic effect. The magnetizing current flows through a parallel component, and is dependent only on voltage and mains frequency.

The load current increases with load (surprise) and that increases the power factor simply due to vector sum of the two. There is not a "reduction" of the inductive portion other than what happens as the input voltage is drawn down slightly by external circuit impedances.

If you read what I wrote, you would see why the "B" leg typically draws more current as the load increases.

Please don't bring that amp-clamp fixation into every discussion.

If an amp-clamp could somehow distinguish between, and indicate only the reactive or the "power" current (whichever), it would be a very useful thing, However, just like all the rest of the plain ammeters, it reads total current without separating those two.

You forgot to include the part where Fitch talked about current-balancing a rotary converter.

Try the experiment sometime and you will see why a novice would be somewhat flummoxed by measuring this way. You do have a rotary converter, yes? What sort of voltage balance do you get?