OK, thanks Bill. That article is just what I was looking for. For an AC capacitor, would you do the process polarized one way, then flip the leads and re-do the same? Or is there no advantage to doing that? Certainly not as convenient to set up a 120V DC through a rectifier as it is to just direct wire to the AC.
Good luck, wherever you’re going. I’ll fuss about with the setup this weekend. Looking at all the things I need to put together, that could take a while, so you may not be missing much, here.
There’s enough space as long as I don’t let the baseplate whip around and strike the bedways of the lathe.
There are boards on the lathe ways to protect them until I start to run it.
BTW,
This mounting is premature because I also need to re-connect the wires to parallel-Wye so that there’s no chance of generating 480V! This is a dual-voltage motor rated at 230V/460V and it’s currently set up for 460V.
Quick update. I did some tests on the weekend, but I have a lot more to do before I get the whole picture. Some quick lessons:
With the wiring still in 460V/1200RPM, I tested 45, 270, and 315 uF capacitors. Cut-in where the capacitors start to excite the winding varies with capacitance. I can extrapolate that to get the generator to self-excite at 150 RPM, I’ll need about 600 uF capacitors.
I tested capacitors in a Delta bank. I measured voltage line-to-line, and I measured current on the lines and in the capacitors. There seem to be high circulating currents, even when there is no load.
I did make up some 30-ohm loads and connected them in Star to the output of the capacitors. This made a noticeable difference in the mechanical load driving the generator but results in only a small about of useful electrical power. In one test setup I measured the following:
1096 W Mechanical power input
1014 VA Line power
379 VA Load power
559 VA Capacitor power
So this is only 35% efficient. Of course I am probably far away from an optimal point. I’m just using materials that I have available. I’ll need to test more capacitors to know what value gives the best power output, and likewise I need to find out the ideal size of the resistor bank before I’m sure it’s able to dissipate a lot of energy.
I should add that the loads are simple resistors, so I trust that I can use “379 VA” as equivalent to “379 W” for comparison to “1069 W” mechanical power to derive the efficiency.
I can’t do that with the line load or the capacitor load; both in VA. Actually I will need to do a lot more work before I can really determine anything useful with those figures. But I measured them and I’ll keep them just in case I can, later.
Something that may be interesting to try.
If your motor is wye connected, connect the capacitors across the 480 Volt connections. T1 T2 T3.
And connect a test load across the 240 volt connection. T4-T7, T5-T8, T6-T9.
Wye connection identification:
Continuity across one group of three wires.
Continuity across three groups of two wires.
Delta connection identification:
Continuity across three groups of three wires.
Past few weekends, had loads of fun.
So far I’ve only tested a fraction of the possible combinations I can try.
The results I first posted about were pretty far away from the optimum. I also had unbalanced phases, for no good reason, and it was really hard to make any sense of the current in the different phases.
I don’t have a way to measure power factor, yet. I only have VA to work with, not the reactive vectors. I believe this can be done with an oscilloscope.
One thing that is predictable is the frequency at which the field begins to form, which varies with the capacitor being used.
These speed ranges are too high for a 8-ft to 12-ft size of turbine that I would likely build for a generator this size. I would prefer a cut-in speed around 100 RPM, but that indicates I need about 800 uF capacitors in a Delta. I am hoping that when I re-wire the motor/generator in parallel-star that this will drop to a lower range of speeds.
Edit: I should add that I’ve noticed the cube-root power in the curve-fit above. I haven’t figured out what it means, yet. But it’s so darn close to -1/3 that it really can’t be anything else.
There is always the red-neck method of PF calculation. grin
Tools needed:
Capacitor,
Clamp on ammeter,
Drafting table.
Compass.
For the motor under test:
Connect a set of capacitors temporarily.
With the clamp-meter measure:
Line current, (Unknown phase angle)
Capacitor current, (Assume 90 degrees leading)
Motor current. (Unknown phase angle)
Pick a suitable scale and lay out and draw a vertical vector representing the capacitor current.
Set the compass to represent the line current and from the top of the vertical vector lay out and draw a suitable arc.
Set the compass to represent the motor current and from the bottom of the vertical vector lay out and draw a suitable arc.
Draw vectors from the intersection of the arcs to the top and to the bottom of the vertical vector.
Draw a horizontal line from the point of intersection of the arcs to the vertical vector.
Now you have the picture that may be worth many words.
The horizontal line represents the real current and is proportional to Watts.
The upper inclined vector represents the actual line current and is proportional to the line VA and to the corrected PF.
The lower inclined vector represents the motor current and is proportional to the motor VA and to the uncorrected PF.
You may measure your phase angles directly or you may choose to scale the values and calculate the phase angles arithmetically.
I am sure that you will find a way to adapt this method to your test setup.
Hope it helps.
Disclaimer. This method disregards the effect of any internal discharge resistor of the capacitor.
Justification:
The charge time of the capacitor is in the order of 1/4 cycle or 4 milliseconds.
The discharge resistor, if required must reduce the charge to less than 50 Volts in under 1 minute.
That is a maximum time ratio of over 14,000:1
The discharge time may be less.
Consider a discharge time of 6 seconds. The time ratio is still over 1400:1
I am willing to disregard any error introduced by ignoring the effect of the discharge capacitor.
Shred observation.
Industrial PF correction capacitors are rated in KVAR, it makes the selection and application easier.
From the KVAR, we can calculate back and determine the impedance of the caps. This will be important later in the explanation.
The KVARs are rated at a specific voltage, generally 480 Volts.
Similar to a heater, at half voltage the Watts or KVARs are 1/4.
But this is dependant on the impedance of the capacitors.
The impedance of a capacitor is frequency related. When you double the speed, the impedance is halved. This effect will add one or two powers to the exponent.
But you have the capacitors connected to an inductive winding.
The impedance of the winding is also frequency dependant, but the relationship is not linear due to the combination of inductive reactance (frequency dependant) and resistance (not frequency dependant).
While the impedance of a capacitor drops with increasing frequency, the impedance of the induction rises with increasing frequency.
I suspect that the power of the exponent may vary with the X:R ratio of the winding from motor to motor.
I hope that this helps explain the effect that you are seeing.
The cut-in speed for a given capacitance doubles when using parallel-star motor connections. Where the 370uF capacitors made it cut-in at 200 RPM when the motor was wired in series-star, with parallel-star the cut-in is about 360 RPM. This is many times too high to match a set of rotor blades of a size necessary to run as a wind turbine.
I’ll keep investigating this motor configuration only as long as there’s some possibility that it might be more efficient or stable than the series-star wiring. Otherwise there won’t be much benefit. Either by handling the cut-in with extra controls (switching relay) or by more expensive capacitors, the goal is not to make the generator more expensive to use.
I may be going to Calgary sometime next week. If so, I may be able to drop off the bucket of caps on my way home.
(I may not get a chance to go home on the way to Calgary.)
Not positive but a good possibility.
If you won’t be home, I can leave the bucket in front of your shop and be on my way.
Hi Bill,
I am still staying with my parents thru next week, so the timing is bad. I’d rather be there when you drop by to at least say hello and show you the apparatus.
Let’s plan the rest by phone or e-mail, for simplicity.
I’ve continued my testing - just haven’t posted much about it lately. Still thinking about the results a lot. It’s very interesting to try out ideas.
I tried something like that yesterday. Frankly, I mis-remembered what you’d suggested, and tried putting the Capacitors across the 240 connections, and the load stayed on the T1-2-3 connections. When I started up the lathe to drive it - absolutely nothing happened. It ran like it was completely unloaded, whether the load was switched in or not. It wasn’t even like there was a capacitor load. I could measure no current any where. Shut it down, and switched the setup back to the way it was (all capacitors and loads on the 480V outputs). And still nothing happened. It had stopped working altogether!
After many hours pondering my mistake and what could have changed, I realized that I had managed to completely demagnetize the armature. Not even a whiff of remanent magnetism to start the induction cycle. This morning I connected a 12V battery across two of the output lines, held it there for 10 seconds, then started it up and Voila! Back to normal. Haha, a little lesson learned about how the self-induction cycle begins.
In case anybody still reading has misunderstood anything: I have no intention or belief that I’m inventing anything here. No overunity discoveries are planned, nor will any be fabricated for the sake of argument. I’m just learning stuff and taking notes as I go. I go back to the books and I can sort-of see how some authors (Gary Johnson for example) have definitely done work like this and understood this better than me, but their didn’t offer enough detail for me to appreciate what really happens. Not as much as just doing it myself. Now that I have, I am going back to the books and realizing “that’s what they meant when they wrote that”.
Here are some detailed results from the tests. I’m collecting current at various points in the circuit, and in each phase to make sure they’re balanced.
For example, with 420 uF of capacitors in Delta, across the leads of the motor in series-Y, I see this:
Starting with the generator running Open circuit:
251 RPM
10% slip (belts and lathe drive motor)
12.6 Hz
142V line voltage
11.3 V / Hz
9.5A line current (2337 VARs)
5.9A capacitor current (1451 VARs)
18.8 Newton-meter torque
497 Watts mechanical power
Same setup, now loaded by just 60 ohms:
221 RPM
16% slip
11.1 Hz frequency
90V line voltage
8.1 V / Hz
4.5A line current (857 VARs)
2.9A capacitor current (447 VARs)
5.5A Load current (281 VARs)
242 Watts on wattmeter
86% load factor
22.3 Newton-meter torque
517 Watts mechanical power
47% mechanical efficiency
And that’s just one of the test runs. I’ve repeated this for various speeds for each capacitor value, and varied the capacitors many times. It’s a big spreadsheet…
One general observation: the V/Hz looks good when lightly loaded. The basic motor is dual voltage, and having it in series-Y makes it 460V at 60Hz. If you consider the rated speed actually has some slip, and the base frequency actually extrapolates to 480 V, then the standard V/Hz for this motor is either 7.7 or 8.0 depending on how it should be defined. Either way, the tests where I have the induction motor loaded, it is maintaining a loaded voltage that corresponds to this ratio rather well. Except in some cases where the speed isn’t high enough to maintain the field and the voltage is about to drop out anyway. It is as important to me to use different speeds on the lathe to investigate the BAD speed ranges as well as the good ones, to know what’s going on in both scenarios.
Some general comments:
Ahhhhhh residual.
Something to try.
Disconnect the load and caps before stopping rotation.
Let the generator coast to a stop with caps and load connected.
While residual is needed in induction generators, it may cause excess in rush current in transformers.
We have had a couple of discussions at an alternate site as to techniques to kill residual in transformers.
The various techniques to kill residual are based on reducing the applied AC voltage so that as the EMF drops to the point where it interacts with the magnetic hysteresis the residual will be left at zero.
You may be a victim of a similar, unintended consequence.
A thought:
What voltage and frequency are you seeing across the caps?
You are looking for VARs.
Volts times Amps(reactive).
A higher voltage on the caps may incur the magic of multiplication.
If you have much less than 480 Volts across the caps, you may be able to use transformers to boost your VARs.
Success will depend on frequency and V/Hz ratios.
Yours
Bill
A comment on slip:
When we talk about slip in an induction motor, we are looking at the difference between the synchronous speed and the actual speed. eg 1800 RPM vs 1760 RPM = 40 RPM slip.
When the 40 RPM is resolved into frequency, it gives the frequency of the rotor circuit.
In your case the slip of interest will be the actual RPM of the induction generator vs the frequency of the generated voltage.
In a motor the slip is closely proportional to the load, up to about 200% to 250% of full load.
In a motor the slip frequency is constant from zero RPM to at least 200% of rated speed.
That is, while the slip frequency varies with loading, the slip frequency for a given load remains constant across a wide speed range.
The real current for a given slip remains closely constant.
The slip frequency of your IG rotor may be of interest.
I’m not currently set up to measure the difference between the lathe motor’s slip and the slipping of the belts in the lathe gearbox. Maybe there is something interesting in there, if I just tried to measure it.
Actually, it should be more useful to measure the frequency in the generator itself.
So far I have been inferring it from the speed of the lathe but Bill you’ve just opened my eyes to the OTHER SLIP going on in this system.
I’ve been missing it all along. I can re-run a few tests to see if I can add some useful measurements.
About re-running tests: I’ve tried this a few times, such as when I got out my wattmeter. Running under seemingly identical conditions did not have exactly identical results. Close… It’s as if the generator is susceptible to temperature in my garage, I think.
