Phase converter balancing

Try this one by Jim Hanrahan:

Mike in BC

Excellent article, Jim covers the whole subject well. But only a couple of paragraphs on balancing. I'm almost sure Fitch had an excellent guide just for balancing. Google search gets thousands of hits with fitch and any subject I could think of.

Karl

Here are some posts and reposts that I have saved from Fitch on rotary converters. The simplest stuff is at the end and the sections are seperated with strings of ****

Cheers,

Kelley

Fitch R. Williams wrote:

Back in 1997 I posted the following two posts on measurements taken on my converter. for those who are interested, they are re posted here in chronological order:

------- First post ----------

I took some more data this afternoon on the two configurations mentioned before and the bare motor. Motivated by the discussion in another thread and its follow up (Help Electrical Mavens) or something like that I was trying to measure several things, the response time to turn on the lathe (shorter is better), the minimum voltage turning on the lathe (higher is better), the minimum voltage reversing the lathe (Higher is better), the current into the third phase of the idling mill (more is better), the stall current into the locked mill motor (more is better), and the output impedance of the converter (less is better) as a function of the capacitor configuration. I made up a simple brake for the mill (didn't want to use the built in one) and used the "MAX" current function of the FLUKE 36 clamp on ammeter and the record voltage function of the FLUKE 97scope meter. The add-on mill brake wore out just as I finished so I made the final measurement using the built in brake.

The output impedance seems to be adjustable in the low current range (less than

7A on a 5hp idler) by using capacitors to boost the generated phase. At some point, the inherent impedance characteristic of the size of rotary converter comes into play. One speculates that larger idlers will have lower characteristic output impedances for lots of reasons. On my little hand sketched graph, the impedance goes flat at ~5 ohms at 2A for the bare idler (no caps), at about 4.5A for the balanced configuration, and at about 7A for the unbalanced one.

The start times are the time from the last peak before the voltage dip to the first peak at the new loaded line voltage measured by running the scope at .5sec/div in single shot mode, and then using the cursor function to measure the delta t.

My reading of the data is that the configuration optimized to balance the idling converter is not as good as the unbalanced one that was optimized (using Rick Zuppan's approach plus a power factor compensation cap) for balancing the idling lathe.

For those that have not been following this series of posts, the configurations are quickly summarized as follows:

Bare idler motor (same in all cases) Gould 5hp 3ph 1725 rpm, 220/440V motor wired in the low voltage configuration. Cp = run cap in parallel with start cap. Cs = run cap opposite, Cp (line to third phase), Cpf = power factor correction capacitor. There were no capacitors on the bare motor.

Balanced Configuration: Cp = 60uF, Cs = 50uF, Cpf = 50uF Unbalanced: Cp = 110uF, Cs = 100uF, Cpf = 12.5uF

The results follow:

If you change your display font to courier this table will be easier to read.

Parameter Bare Idler Balanced Unbalanced

Third Phase Voc 223.7V 239.9V 252.6V

Current/Voltage Into .7A/223.2V 2.4A/234.5V

4.1A/243.4V Third Phase of Idling Mill

Mill Stall I/V 18.1A/140.8V 19.5A/141.3V

21.4A/148.1V

Lathe Start Time 1.3sec 0.9sec 0.68sec

Lathe Start Vac Min 142V 144V 148V

Lathe Rev Vac Min 125V 130V 133V

Zout @ 2A (Ohms) 5 2 1.7

Zout @ 5A (Ohms) 6 5 3

Zout @ 10A 6 5.5 5.2

Zout @ 15A 6 5.5 5+

Zout @ Stall 4.6 5.1 4.9

I've about run out of things to measure except power factor and efficiency. I will tackel that when I get the time. The unbalanced configuration is looking a lot like it is the "final" design. Since I bought the enclosure when I thought the balanced one was "it" I need to get a bit clever about packaging.

--------------------- end of first post --------

Power factor and power consumption are covered in the next post. I have not measured efficiency directly yet - need a load motor dyno to do that. The efficiency of interest is single phase input to load motor shaft output.

There were several posts in between these - but these two capture most of the data.

------- Second post --------

I created an Excel spread sheet and printed it to provide a format in which to record the data. After the numbers were entered I had it do the calculations and printed to a comma seperated ascii file. This views like a sober table if you have your browser font set to courier or some other monospace font. If not, it views like a drunk table.

A bit of explanation for any first time viewers or those unfamiliar with this series of posts on the 5hp rotary converter project.

This is a table of measurements related to the single phase 220V input to a 5hp rotary converter evaluated with a number of capacitor combinations and loads. The converter idler motor is a Gould 5hp 1725 rpm "Y" wound

208-220/480V induction motor. A number of posts have been made detailing other measurements associated with these converter combinations. The purpose of this most recent in a series of measurement sets is to explore the effect of capacitor configuration on power factor, and idler power consumption.

A brief tour of the table and definition of variables. CP is the run capacitor in parallel with the start capacitor (connected between one single phase line and the third phase). Note that CP values of 125uF and larger make this particular converter self starting if there are no other run capacitors. CS is the other run capacitor which is between the other single phase line and the third phase. Cpf is the power factor correction capacitor when one is used. All capacitor values are expressed in microfarads.

V(Sh) is the rms voltage in volts developed across the current shunt (resistor). It is multiplied by the Shunt constant of 159.59 to convert it to amps. V(LL) is the converter input line to line voltage in volts.

Ph is the phase angle between voltage and current. The positive value means voltage is leading current which indicates that the input is in all cases at least slightly inductive. Pf is the cosine of angle Ph.

W is the actual power input to the system in watts which is calculated as:

W = 159.59 * V(Sh) * V(LL) * Pf

The instrument used was a FLUKE 97 scope with 10:1 probes except for the idling converter measurement on the last two cases where a 1:1 probe was used on channel B (current measurement) to increase the size of the wave form and improve the accuracy of the phase angle measurements. The current shunt is .006266 ohms as measured on a very accurate Valahalla micro-ohmmeter. The scope calculated rms values of the voltages were used. The scope determined phase angle measurements were used. Because of some variability in the measurements, the phase angle variations were studied and the data from the best "marker" placement (closest to zero reference crossing) was used.

Power and Power Factor Measurements

Shunt constant = 159.591 Description V(Sh) V(LL) Ph W Pf

CP = 0 Idler 0.078 232 70 988 0.34 Idler + L(N) 0.116 232 70 1,469 0.34 Idler + L(1500) 0.116 232 69 1,539 0.36 Idler + L(1500) + M(223 0.148 232 69 1,964 0.36 Idler + M(stall) 0.420 224 35 12,299 0.82

CP = 50uF Idler 0.060 232 77 500 0.22 Idler + L(N) 0.096 232 77 800 0.22 Idler + L(1500) 0.096 232 69 1,274 0.36 Idler + L(1500) + M(223 0.132 232 69 1,751 0.36 Idler + M(stall) 0.408 224 51 9,179 0.63

CP = 100uF Idler 0.044 232 69 584 0.36 Idler + L(N) 0.076 232 69 1,008 0.36 Idler + L(1500) 0.080 232 60 1,481 0.50 Idler + L(1500) + M(223 0.116 232 60 2,147 0.50 Idler + M(stall) 0.380 224 43 9,935 0.73

CP = 125uF Idler 0.038 232 61 682 0.48 Idler + L(N) 0.068 232 69 902 0.36 Idler + L(1500) 0.072 232 51 1,678 0.63 Idler + L(1500) + M(223 0.104 232 60 1,925 0.50 Idler + M(stall) 0.380 224 43 9,935 0.73

CP = 160uF Idler 0.030 232 51 699 0.63 Idler + L(N) 0.056 232 60 1,037 0.50 Idler + L(1500) 0.062 232 44 1,651 0.72 Idler + L(1500) + M(223 0.096 232 51 2,237 0.63 Idler + M(stall) 0.360 224 43 9,412 0.73

CP = 60uF CS=50 Cpf=50 Balanced as an Idler AKA "Balanced Config." Idler 0.012 232 9 439 0.99 Idler + L(N) 0.040 232 60 741 0.50 Idler + L(1500) 0.048 232 44 1,278 0.72 Idler + L(1500) + M(223 0.080 232 51 1,864 0.63 Idler + M(stall) 0.360 224 43 9,412 0.73

CP = 110uF CS=50 Cpf=12.5 Balanced lathe currents AKA "Prefered" Idler 0.016 232 26 532 0.90 Idler + L(N) 0.046 232 60 852 0.50 Idler + L(1500) 0.060 232 35 1,820 0.82 Idler + L(1500) + M(223 0.088 232 53 1,961 0.60 Idler + M(stall) 0.384 224 43 10,040 0.73

Some notes:

Of interest to me is the 17% lower power consumption of the converter which was balanced as an idler when compared to the "prefered configuration" which was optimized to balance the lathe currents. The total rms current taken by the scope from the waveform was 1.91A for the idling balanced converter. It is cooler and maybe quieter (although I have no objective way to evaluate "quiet").

I don't have the ability to print waveforms from this scope (optical coupled cable is missing). That could change shortly after the first of the year when I am supposed to get the chance to use a TEK THS720 with PC connect cable and SW for a short time. However I made a note that the current wave form was looking noticably non sinusoidal and "pinched" at the peaks in the case where CS =

160uF, I might recreate it and take a picture of it with my digital camera.

Based on data taken during previous testing the increasing power consumption as CP is increased is due to the increase in current flowing in the third phase of the idler motor. While it is reactive current, it still dissipates power as an IR drop in the third phase windings. The output voltage of the third phase increases markedly with increases in CP. 125uF or 130uF is about as far as I am comfortable with (153+volts).

The differences in power consumption between the last two configurations is easily explained by the significnatly lower circulating currents in the idler in the balanced configuration.

I will be adding a few more cases to this table shortly after the first of the year after I pick up a buck/boost converter to try in place of CS. I am currently expecting to try it with CP values of 25, 50, 75, and 100 uF.

Note that the power factor of unloaded induction motors is worse (lower) than when they are loaded and working hard. This is because the magnitizing current is such a larg part of the total current when the motor is idling and the real part grows as the motor is loaded and does "real" work.

------------

There is more, but that is probably more than enough for now. I keep thinking I will put this in the Metal Working News page - roundtuit needed!

Fitch In So. Cal.

**************************************************************************************************

Newsgroups: rec.crafts.metalwork>f anyone is interested, I can email you a spreadsheet I put

If you add capacitors so that when the converter is idling the voltage of the two generated phases is 108% to 110% of the incoming line voltage, and then add capacitance across int incoming line to minimize the measured line current you may find that it works even better. That will give you nearly balanced voltages when the load is connected, but will not cause an idling over voltage that will damage the run caps (assuming they are 370V caps).

I would be interested in a copy of your data if you care to E-mail it.

Thanks Fitch In So. Cal.

*******************************************************************************************

Newsgroups: rec.crafts.metalwork>Not only am I late to this thread, but I'm not a theoretical physicist

There are three opportunities to add capacitors to a rotary converter. Some definitions:

L1 and L2 are the incoming 240V single phase leads.

Cp is the run cap that is in parallel with the start cap. Assuming the start cap is between Line 1 (L1) and the third phase lead, then Cp is also connected between L1 and the third phase lead.

Cs is the other run cap and is connected between L2 and the third phase lead.

Cpf is the power factor correction capacitor and is connected between L1 and L2.

You select Cp and Cs so that the idling converter has voltages from L1 and L2 to the third phase lead that are equal and at or up to 108% more than the incoming line voltage between L1 and L2. To do this, you only need a volt meter and a selection of run caps.

Once the two run caps are selected, you need a current measuring device to select Cpf. Assuming a clamp on ammeter is available, clamp that around one of the incoming lines, L21 or L2, "before" the junction where the run caps connect. Start adding Cpf in increments. After each addition (capacitors are connected in parallel), read the line current. As capacitance is added, the line current will decrease to a minimum, and then start back up. After the first measurement indicating an increase in current, remove the most recently added capacitor and call it quits.

The rule of thumb is to add Cpf until the idling converter line current is minimized.

Hope this adds more clarification than confusion.

Fitch In So. Cal.

*********************************************************************************************

Also, be aware the Hanrahan rotary is a self starting device. Thus the excessive amount of start capacitance would make it difficult to voltage balance ala the Fitch method.

I find this message posted by Fitch to be the easiest tuning instructions for a novice to follow.

Mike (the other Mike in BC, Hi Mike!)

It seems to me that balancing is getting ahead of the situation--why won't it start is the first question.

Steve Smith

Karl Townsend wrote:

Many thanks - just what I was looking for.

Also, Thanks Bob Swinney; for the information you sent by email.

FWIW, the motor being used as the phase converter was WAY to small. Switched to a five horse and it took right off. I left the guy a copy of Fitch and Bob's information so he could learn the fine art of converter tuning.

Karl