Lab electromagnet from scratch / Projects / Forums

Forums

4hv.org :: Forums :: Projects

« Previous topic | Next topic »

Lab electromagnet from scratch

previous 1 2

Move Thread

LAN_403

Ash Small

Mon Jun 01 2015, 10:48PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

This is a really interesting project, Rich. When I joined this forum I had plans to build an electro-static accelerator, but these electro-magnet threads of yours have persuaded me that electro-magnets are a viable alternative to hundreds of kV.

I'm still not convinced that a conventional amateur cyclotron is viable for my ideas, but using electromagnets certainly looks viable in some configuration from your observations

klugesmith

Sat Jul 04 2015, 11:39PM

Registered Member #2099 Joined: Wed Apr 29 2009, 12:22AM
Location: Los Altos, California
Posts: 1714

Thanks, Ash.
Soon after last post, I thought I'd blown something up. New +/- 15 A analog meter would not stay past 7; motor would suddenly slow down and meter pointers fall back a little. Disconnected the main inductive load, fearing that momentary drops in control voltage (e.g. from scratchy potentiometer) had "braked" the load & pushed energy back into the DC power supply. Soon even the little motor would not turn. Output voltage was about 0.05 V & unresponsive to control dial.

During the next few weeks I looked into buying another servo amp, and considered seeing if my quad audio amp chip (paralleled for 100 W into 1 ohm) could drive DC. Decided to first open "broken" amp and check H-bridge MOSFET voltages. Those all appeared to be in order. Ultimate conclusion: I had run afoul of amplifier current limit settings (fast & slow) and had made a control knob with intermittent connections.

With a new self-powered and filtered control knob, I got power satisfaction on the 4th of July. For loads of around 1 ohm, the control knob goes from -15 A to +15 A. Here's the new, simplified configuration in schematic form:

and for real:

The servo amp is shown converting DC 24.2 V 2.5 A to 7.19 V 7.5 A. Wooden clothespin secures a temporary 15 A electrical connection, with no hazard from far end of a conductive clip lead. Magnet top plate has been replaced by a stiffly-supported crescent wrench.

Electromagnets are naturally more power-efficient as they get bigger, a trend apparent even at 3 inch pole diameter. The wrench demo was able to keep it up at 1 ampere.

As the drive was turned down little by little after that, we read 313 mA before the still-magnetized wrench fell off.

Next step is to start measuring magnetic flux. Don't really need a DC current sensor with ground-referenced output, if we read the analog meter and manually tabulate the data.

klugesmith

Sat Jul 11 2015, 12:16AM

Registered Member #2099 Joined: Wed Apr 29 2009, 12:22AM
Location: Los Altos, California
Posts: 1714

Finally got to measure some magnetic properties of my "plain old steel".
First, restored the magnetic circuit to a simple gapless configuration. Here the top and bottom plates have much less cross-sectional area than the round pole pieces (27.8 vs 45.6 cm^2), so they will saturate first. My quick-and-dirty sense coil is four turns of wire, conservatively placed about as far as possible from the forcing coil (whose decoration is left over from Halloween 2013).

Fluxmeters depend on an electrical conductor going around the magnetic flux of interest. It's analogous to measuring electric current by putting a ferromagnetic core around the place of interest. Both are noninvasive, and can take readings on solid bars or empty space. Fluxmeters measure changes of flux; a change of 1 weber (at any speed) generates 1 volt-second per turn in a sense coil. This unit has an analog voltage integrator with a reset button, a very sensitive offset-adjustment knob, and a digital display.

To get flux change from a gapped magnet, one can reset the integrator while sense coil is at the place of interest, then rapidly move the sense coil away from the strong field. With my electromagnet, I could have run demagnetizing cycles and then set the fluxmeter zero. This time, started at 10 amperes and set fluxmeter output to half of known peak-to-peak value, using a slow motion control kluge. (6 volts through 10 kÎ©, applied in parallel with sense coil.)

One effect stood out which I had never seen while fluxmetering transformers. After each current step that caused a large flux change, but not similar current steps with small flux change, I had to wait for the least significant digit to stop changing. This has got to be an eddy current / skin effect thing. Coil current lags voltage a little. Average flux (esp. the last percent) lags current a lot. It will be fun to see that in a quantitative way.

Ash Small

Sat Jul 11 2015, 12:59AM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

The bottom two graphs look very familiar, they look identical to a typical hysteresis curve.

previous 1 2

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob