Balancing vs. Power Considerations

In a prior post:
"Topic: Pony motor speed question"
Bnesson made a link to some very interesting relationships about starting current versus armature speed. I want to thank him again. I read the whole site and my only dissatisfaction was that many of the terms in the equations were not defined, and as a mechanical engineer, their meaning was not self evident. This is not Mr. Nelsons fault. That information should be very useful for anyone building a converter with a starter motor.

I am interested in the same quality of information relating to optomizing capacitor sizing on a converter. Is any similar thing available that relates the sizing of the run and the power factor capacitors to the power removed from a converter at a given load? I am familiar with two similar rules of thumb (1,2,3 and 60/40) but I am curious about the math that is behind them.

Thank You,
John Gill

John, first of all, I'll try to remedy the confusion in the equations. Some of the stuff I posted was in image format, which I just copied from another site without editing.

I concur with you completely that any quantity or equation discussed on this forum should be able to be defined, in terms satisfying either a physics professor or engineer. That way, much confusion is avoided.

For the capacitor question, one way (not the only way) to look at the current passed to a pure resistor in an AC circuit is:

I = 2*pi*f*e*c
where:
I = current in amps
f = frequency in hertz
e = voltage in volts
c = capacitance in farads (not the usual mfd)

Click to expand...

Let's consider a 5 hp idler. My rule of thumb says 15 mfd/hp as one run cap, across one hot leg (source) to the generated leg (receiver).

That three phase idler will likely have a nameplate rated current of about 10 amps. Now how do you go about getting the 10 amps to be output on the generated leg?

Solving the above equation for c,

c = I / (2*pi*f*e)
c = 10 amps / (2*3.1416*60hz*240volts)
c = 110 X 10^-6, or 110 microfarads.

My rule of thumb says 15 mfd/hp X 5 hp = 75 mfd. But that's just the rule for the bigger run cap. The smaller which goes across the other hot leg for balance, by my rule of thumb, is 9 mfd/hp X 5 hp = 45 mfd.

Adding the two, 75 + 45 = 120.

Very close to the equation value, which is given in any elementary EE text.

This stuff really does prove what they teach you in high school and college physics and engineering has practical utility for designs, such as we forum members do on PM.

Neato and well said.

Your cut and paste was OK. I followed the link under relations, and the terms were not defined there either. I read the whole lesson and I think that you probably took the best stuff.

It's interesting using the inductive current as a resistive current and figguring the capacitance needed to generate it at a voltage.

What I am really interested in is how the 60/40 or 1/2/3 is arrived at. In a recent post
"Topic: Hey bnelson what is the best way to balance voltage?" You responded.

"In general, voltage balance is more important for an unloaded or very lightly loaded converter idler. Amperage is more important if the converter is under a load of 40% or more of its theoretical capacity."

I assume that this has to do more with possible over currents (heat) in a heavily loaded system and with insulation break down in a lightly loaded system.

The 1/2/3 split could also have to do with generating the proper 120 degree phase relationship and not have anything to do with the load on the converter at all.

I am certainly confused.

I don't remember if my general purpose electrical texts had any motor models in them. I also don't know if I could find them.

"What I am really interested in is how the 60/40 or 1/2/3 is arrived at."

From my perspective, these relations were derived empirically, based upon perceived dynamic performance of machines under load, and over a significant number of installations, often by different persons, using essentially the same technique, where these persons shared their machine performance data.

60/40 works well for a machine, such as a lathe, which is operated predominately in one direction, say, forward, and is occasionally "plug reversed".

Where a machine is likely to be operated in either direction, and with equal probability, then 50/50 is preferred.

(Off the top of my head, the machine which would likely be best suited to 50/50 operation is the special case of a small lathe, such as a Hardinge TL, "toolroom lathe", which has a FWD-BRAKE-REV lever, activating a drum type switch for reversing and a mechanical brake for braking, and with which it is desired to operate the machine as a self-converter, that is, using no separate RPC idler, with FWD for threading, followed by BRAKEing to, say, a shoulder, followed by REV, back the starting position, and then immediately initiating the next thread chasing pass).

(This is a very special case, and I have provided a schematic diagram of this machine).

TL Converter.jpg

(Select the specified file as Yahoo! doesn't allow direct links to its files sections. Sadly, you may be asked for your Yahoo! Groups password for access. I will shortly move this file to a more generally accessible area).

I agree with Peter that most likely these ratios were found by trying out various configurations, and seeing which gave the best performance.

Except for the balanced 50:50 case, I can see no mathematical relationship or physical law behind a 60:40 ratio.

Which brings up a point about motor design. Early electrical engineers often tried out variations in windings topology. They rapidly discarded what didn't work, and kept and improved upon what did.

There probably weren't too many early innovators like Tesla, who remarked that he did not need to try out his concept of the induction motor in a lab; he simply did the field experiment in his head, and it worked.

I suspect in many cases, we actually figured out WHY something worked, perhaps even discovered a new principle, after inventing something which was found to work.

"In general, voltage balance is more important for an unloaded or very lightly loaded converter idler. Amperage is more important if the converter is under a load of 40% or more of its theoretical capacity."

Click to expand...

I don't really know why this is true. I've found there's a correlation in that if I balance an unloaded idler by voltage, then connect it to a load, the currents are pretty well balanced; however under VERY heavy loads, I often need to add additional caps, which produce pretty wild open circuit voltages on the idler.

For me, if it works (gives good, robust machine performance), I use it, whether I understand it or not.

Believe it or not, this makes sense to me. In my crude way of thinking, a rotary phase converter looks a lot like a rotating center tapped autotransformer with some capacitors strapped across it. Based on the mechanical analog, it has always seemed to me that a 50/50 split in the capacitor sizing should give the best oscilation between the capacitors and the voltage source. The generated leg would be closest to the center tap potential and not biased. I never really considered the possibility of stopping or starting frequently. Tapping on a bridgeport type mill would be another good example.

Peter, When you get that schematic moved, please make a post. I am interested in sseeing it.

Thank you,
John Gill

"Except for the balanced 50:50 case, I can see no mathematical relationship or physical law behind a 60:40 ratio."

But, we have empirical evidence, which is how we got to the 60/40 case from the otherwise "obvious" 50/50 case, in the first place.

Consider an RPC where the capacitance is 60/40, and the start cap is placed across the 60 percent run cap. Such a starting cap would usually be about 600 percent, or six times the value of the total of the run caps, one of which it is placed in parallel with.

OK, the motor starts with 660 percent across the A-B phases and 40 percent across the C-B phases.

660:40 (16.5) is "significantly imbalanced" (with respect to a perfect 50/50 balance), so the motor is caused to rotate, and it starts.

(Note the direction of rotation of the RPC).

Next, the start cap is removed from the circuit and the RPC is now operating at 60/40.

60:40 (1.5) is not "significantly imbalanced", in fact, it is pretty close to "balanced", so the motor continues to run. And pretty quietly, too. For purposes of this post, I will call this the "overbalanced" case.

Stop the RPC.

Now, consider another possibility, namely that the start cap is placed across the 40 percent cap.

OK, the motor starts with 60 percent across the A-B phases and 640 percent across the C-B phases.

640:60 (10.67) is still "significantly imbalanced" (with respect to a perfect 50/50 balance), so the motor is caused to rotate, and it starts.

(Note the direction of rotation, it should be opposite that observed earlier).

Next, the start cap is removed from the circuit and the RPC is now operating at 40/60.

40:60 (0.67) is not "significantly imbalanced", in fact, it is pretty close to "balanced", or so it seems, so the motor continues to run. For purposes of this post, I will call this the "underbalanced" case.

However, the motor is rotating opposite that which it was in the previous case, yet the placement of the 60 percent and 40 percent caps is the same as that in the previous case.

Hence, the RPC may (probably will) run roughly, and it doesn't provide the best possible load motor dynamic performance.

The identification of 60/40 as the "ideal" was made by Fitch Williams after numerous tests, with deviations from the "obvious" 50/50 case, upon load motor dynamic performance, especially including those cases where the load motor was "plug reversed", which perhaps represents one of the most difficult loads for an RPC to handle.

The best numerical example I can come up with which succinctly demonstrates this "ideal" is to note that the ratio was greater than one (is was 1.5, hence "overbalanced") in the first case, whereas it was less than one (it was 0.67, hence "underbalanced") in the second case.

Summary: a "slight imbalance" was observed to result in the best possible load motor dynamic performance, and it was an "overbalance" which most contributed to that best possible performance. IOW, a "slight imbalance" is "ideal", but only if that imbalance results in a ratio greater than one, hence an "overbalance".

Thank you Peter. It looks like you could just wire it into the existing machine!

Hi,

I'm new here. I saw my name in a post about rotary phase converters. Something about 60/40 being ideal.

Was that refering to capacitor ratios in an RPC?

Fitch

Hi Fitch,

Us r.c.m. refugees would enjoy a shop update... in a different thread perhaps.

Bob Powell

"I'm new here."

Welcome, Fitch.

PracticalMachinists and r.c.m (AKA, "the Wreck") refugees, all, Fitch is one of the few who could be called a legend in his own time.

Please welcome him!

Peter.

Fitch, so glad that you popped up here on this forum.

Yes, I have quoted your design several times, as an example of superb balance of three phase output.

Your design uses far higher capacitance between phases A and C, than most of us are accustomed to. This capacitance is for power factor correction, and most of the converters I build don't use it at all, since balanced run caps between phases A-B and C-B produce a high power factor converter.

We would be most interested in learning why the large power factor correction caps make for such a good balance. From what I've been able to read, you are the inventor of this idea.

This forum is pretty lively and we get into new designs for special situations. Your thoughts and contributions will be most appreciated.

Fitch *who*?



Hope your plans in PA are working out apace!

Regards - Jim

Thanks for the warm welcome. Good to see some old friends here.

I told Bob I'd post my converter balancing procedure here. I'll do it in a new thread.

I've not been active in a forum like this. I need to figure out how to upload a jpg schematic to go with it. The procedure is in a MS-WORD document - hopefully I can cut and paste it into this "system".

Fitch

Fitch,

I'd be glad to post what you sent me by email if you want.

I use Photobucket to host images (jpg, gif, etc) to this board. Unfortunately they do not host Word docs. Also, some members have Macs on PM, and cannot open Word docs. I've gotten around this by converting to Adobe pdf, which everyone likes.

However, to post an actual pdf file to this website, you have to email it to Don Thomas and ask him to host it; he'll give you the URL link when he does.

Let me know if you'd like me to convert your Word doc to pdf; I can email it to you.

bnelson wrote "I'd be glad to post what you sent me by email if you want."

Thanks! That would be great. I don't have any place to host pictures on line, and probably won't until after I get our new house and my new shop built.

Fitch

Fitch, I'll start a new thread and post what you sent me in the next few days.

"I use Photobucket to host images (jpg, gif, etc) to this board."

So do I, insofar as JPEGs are concerned.

I develop the JPEGs I submit here at "High" or "Maximum" resolution, and at 150 dpi, as most of my original art is fine-line, or are small, illustrative extractions from other sources.

I let Photoshop create a more compact version for submission to Photobucket by reducing the resolution to "Medium" when invoking the final, "save copy as" function.

Also, I scale JPEGs to about 3" wide, also using Photoshop, immediately prior to invoking the "save" function.

I believe this enhances the on-screen impact, while preserving usability should a user want to download the image for off-line purposes.

I use GraphicConverter to convert GIF and other proprietary formats to JPEGs, which everyone can use.

What is:
"r.c.m (AKA, "the Wreck")"

John Gill

"r.c.m (AKA, 'the Wreck')"

The USENET group rec.crafts.metalworking.