By Timothy Raneyâ€¦Bald Engineer Guy with Glasses
This time, weâ€™ll finish building the electromagnet and perform an initial verification test, thus finishing this series.
The Final Chapter â€“ it is the End. After another brief hiatus, weâ€™ll finish building the monumental rotatable electromagnet. So, itâ€™s finally time to conduct the initial verification testing.
Initial Performance Verification Testing. With the coils completed except for the binding posts, I assembled the magnet and decided to perform a preliminary verification test. I ensured both coils were connected in series correctly â€“ windings in the same direction. I connected the electromagnet to a BK PrecisionÂ® regulated DC power supply (Model #1666). The power supplyâ€™s integral meters measured the current and potential. The gaussmeter probe was held within the geometric center of the magnetâ€™s 0.5â€ wide gap with a burette clamp.The first trials showed the magnet could produce ~730 gauss flux with a 0.15A current. The power supply achieved its maximum ~21-volt output given the 137W load resistance of the two series-connected coils. This result was just â€œOkâ€.
Then I remembered noticing one magnet core was not flush with the coil form disc.The resulting air gap added to the magnetic circuitâ€™s reluctance â€“ the flux path only went through the steel mounting bolt compared to the core joined to the yoke without an air gap. â€œReluctanceâ€ is analogous to resistance in an electrical circuit. The solution was adding a 0.054â€ thick steel washer to fill the space and function as a better flux path. Since inexpensive mild steel washers vary somewhat in thickness, I found one that would suffice. No magic and I didnâ€™t need to slice-off a washer from a piece of bar stock. Further testing then showed a better result - 835 gauss at 0.15A. Ambient temperature was 20C.
Power Supply Details. The next series of trials used a higher potential power supply assembled from a diode bridge comprised of four (4) TM65 silicon rectifiers. These rectifiers have a 600 peak inverse voltage rating with a 1.25A maximum current. These ratings were satisfactory for the intended purpose. The input for the bridge was a 5A variable transformer. Connected to the bridgeâ€™s DC output was a 6000mF electrolytic filter capacitor with a maximum 250VDC operating potential. Um, let us now pause for a moment and talk about capacitor safety. Weâ€™re not talking about the itty bitty surface mount capacitors of a few micro-farads (mF). Weâ€™re talking about capacitors large and heavy enough to smash your smartphone into smithereens â€“ crunch. Itâ€™s important to remember these capacitors can retain a significant electrical charge â€“ enough to shock or even kill.
Therefore, here are some safeguards. Two precautions included a 2-watt 30KW â€œbleeder resistorâ€ connected across the capacitorâ€™s terminals and having an insulated shorting bar nearby. Selecting the proper value for this resistor depends on the desired time constant and itâ€™s another exercise where we can apply the exceedingly useful Ohmâ€™s law. Another precaution was leaving the single pole switch between the magnet and power supply closed. I varied the potential or slowly reduced the power to zero with the variable transformer. By using this technique, the capacitor discharges into the electromagnet given the time constant for the circuit. This method also prevents the occurrence of damaging transients that occur when switching highly inductive circuits under load. And with the meters in the circuit, we can watch the potential go to zero too. Though I still like to use the shorting bar. I always make sure the circuit is dead, dead, dead. Itâ€™s better than the alternative â€“ much better.
Back to the Experiment. With all circuit connections checked, I applied potentials from 75 to 190-volts with the corresponding currents from 0.5A to about 1.25A. Magnetic flux ranged from 2500 to 3920 gauss. A good result. If you remember the flux calculations in part-13, the estimate was 5310 gauss. We might think the relationship between the calculated and actual magnetic flux does not agree very well. However, the calculation ignored any flux losses in the circuit that occur in actual practice. The equation also estimates the maximum flux we can expect within the ferrous metal used as the magnetâ€™s core. Moreover, we used the estimated wire turns. So, if we recalculate the flux using the actual turns in the coil (3828), the new estimate is 4810 gauss given B=km0NI as discussed in part-13. If we compare this value with the 3545 gauss achieved at 1A (143V), this result then differs from the measured value by 26%.Under the circumstances described herein, this is a pretty good outcome.
Attaching the Binding Posts. I know all too well. With this paragraph heading, you probably canâ€™t wait to read further. I finished attaching a binding post to each coil assembly. The binding posts are the banana plug type mounted on a small phenolic cylinder and epoxied to the coil form disc on the outboard sides. Each input coil lead was soldered to a ring connector and then attached to the binding post screw.Afterwards, I applied 5-minute epoxy to each binding post and held them in place with a rubber band. When using epoxy, too much clamping force will squeeze-out the epoxy and result in a weak joint.It is usually sufficient to apply enough pressure to spread the epoxy, force out any air bubbles and hold the part in place.Not using too much epoxy is important too â€“ itâ€™s a waste and often allows the part to slide around. It just might end up where you donâ€™t want it â€“ if you donâ€™t pay attention. Sounds like something a parent would say.
Lastly, after much thought, the binding post solution was what I chose. Without changing the design late in its development, I think itâ€™s a good solution, but Iâ€™m not sure how well the posts will hold up in service. I did prepare the surfaces by scratching them with 220-grit abrasive paper, followed by cleaning with denatured alcohol. So, time will tell. Surface preparation is very important in applying epoxy and any other adhesive. I suppose a possible exception is gluing stuff on construction paper with white glue in kindergarten. Though weâ€™re talking about engineered adhesives here â€“ take the time to do the surface preparation properly. Have I talked enough about surface preparation yet? No? Youâ€™re right. I will also apply epoxy to the lead wires were they exit the coil assembly â€“ I donâ€™t want to snag them on something and break the wire. I would cry if that happened. Well, maybe not literally.
Data and Results. Now I suppose this is the real beginning of the end. With the electromagnet completed and sitting on the lab bench in all its unadulterated glory, it was time to do the final verification test. I wonâ€™t keep you in suspense. Drum roll, please. Yes, it worked fine. No issues, except for the magnet coreâ€™s ~170 gauss residual magnetism. As Mr. James Hannon implied some weeks ago, I should have studied the magnetic properties of this iron alloy closer. Weâ€™ll place that mistake in the â€œyou knew better, but didnâ€™t listenâ€ pile. However, I did make the magnetic cores removable.Â Later, I can machine new cores from annealed low carbon steel. Though demagnetizing the electromagnet with 144VAC did reduce the residual magnetism to ~51 gauss.
Maximum Flux Test. Another trial determined the maximum flux, coincidentally called the â€œmaximum flux testâ€. What a surprise. The magnet achieved 4080 gauss at 1.44A (204V). This trial was quick â€“ I did not want to ruin the 1.25A rated bridge rectifier. Ambient temperature was 22C. However, this test did show the magnetâ€™s capabilities with a 0.5â€ wide gap. Sure, I did this test for my own satisfaction, but its results do answer a commonly asked question. All the other values are shown in the table above and corresponding graph below. These values essentially describe the magnetâ€™s performance given currents from the milliamperes to amperes. In future experiments, I will measure the flux at different gap widths preparatory to using the magnet for magnetic anisotropy work.
Conclusion. Perhaps we all learned something useful from this series on building a rotatable electromagnet. My intent was to document the effort in enough detail to guide my efforts and show others basic skills and techniques used in electromagnet design and construction. You will remember this project is largely a consequence of earlier magnetic anisotropy experiments where the need for a rotatable electromagnet was made evident by the rather poor experimental results largely due to my own ignorance. So, I have at least elevated my knowledge a degree or two. I also wanted to fill a gap between two electromagnet extremes, i.e., an electric wire wrapped around a nail and a large laboratory type costing thousands of dollars.
Therefore, my other purpose was to show how we can build an electromagnet somewhere between these two extremes. This project was not written as a â€œcookbookâ€ with specific steps you must follow. What worked for me (or didnâ€™t work) may not work for you. We learned a little theory along the way too. Lastly, I hope documenting my errors will leave you the freedom to make your own creative mistakes. Thatâ€™s it. Keep experimenting and have fun.
I hope youâ€™ve enjoyed this lingering, though dazzlingly unremarkable series of building a rotatable electromagnet. This is the last installment of this awe-inspiring and equally prosaic project. Please shut the door on the way out.Â