Dang those silly trains. A lot of scientists have trains of thought. I am pretty sure I have crazy clown cars of thought.

It sounds like a ferroresonant transformer would be a good investment here. If I don't find one of those, I presume 10% variation would pretty much capture the kinds of fluctuations I'll see in standard AC current?

Also, I just procured a spool of 26 ga magnet wire from a local stereo repair shop. MUAH HA HA!!

 

Scratch that. I now have 32 ga wire. *maniacal giggle*

 

As the wire diameter decreases, so does the allowed current. Here's a copper wire table generated from a script I wrote; the columns on the far right help you estimate what the maximum current should be (AAC's software should really be using a fixed font for code blocks):

       Diameter     Res/length    Mass/length    Area          Amps
 AWG  mils    mm   ohm/ft  ohm/m  ft/lb  m/kg    mm^2    60 C   75 C   90 C
----  ----- ------ ------  -----  -----  -----  -----   -----  -----  -----
   4  204.3  5.189   248u   815u   7.91   5.31   21.2      70     87    100
   6  162.0  4.115   395u  1.30m   12.6   8.45   13.3      53     67     77
   8  128.5  3.264   628u  2.06m   20.0   13.4   8.37      40     51     58
  10  101.9  2.588   999u  3.28m   31.8   21.4   5.26      31     38     44
  12  80.81  2.053  1.59m  5.21m   50.5   34.0   3.31      20     20     20
  14  64.08  1.628  2.53m  8.29m   80.4   54.0   2.08      15     15     15
  16  50.82  1.291  4.02m  13.2m    128   85.9   1.31     9.3    9.3    9.3
  18  40.30  1.024  6.38m  20.9m    203    137   823m     5.8    5.8    5.8
  20  31.96  .8118  10.2m  33.3m    323    217   518m     3.7    3.7    3.7
  22  25.35  .6438  16.1m  53.0m    514    345   326m     2.3    2.3    2.3
  24  20.10  .5106  25.7m  84.2m    817    549   205m     1.4    1.4    1.4
  26  15.94  .4049  40.8m   134m  1.30k    873   129m    910m   910m   910m
  28  12.64  .3211  64.9m   213m  2.06k  1.39k  81.0m    570m   570m   570m
  30  10.03  .2546   103m   339m  3.28k  2.21k  50.9m    360m   360m   360m
  32  7.950  .2019   164m   538m  5.22k  3.51k  32.0m    230m   230m   230m
  34  6.305  .1601   261m   856m  8.30k  5.58k  20.1m    140m   140m   140m
  36  5.000  .1270   415m   1.36  13.2k  8.87k  12.7m     90m    90m    90m
  38  3.965  .1007   660m   2.16  21.0k  14.1k  7.97m     56m    56m    56m
  40  3.145 .07987   1.05   3.44  33.4k  22.4k  5.01m     35m    35m    35m

      Amps = maximum current for <= 3 wires with insul temp rating in deg C.
    Ambient T <= 30 C.  (From regression of NEC-allowed current densities.)

You have to be careful about using too much current or the wire will heat up, melt the insulation, and cause a short in the coil which would cause even more current to flow, more heating, etc.

By the way, it's a very good idea to put a fuse in the circuit somewhere to head off such a problem. A fuse is much cheaper than a ruined experiment or a fire in the house.

 

I was definitely planning on putting 2 fuses in this circuit: one for the transformer, and one for the voltmeter. I plan to use enough winds of wire to give me about 0.1 amps and 1.2 W in the primary circuit.

I'm a bit confused about the meaning of the amps column in your chart, though I understand the concept of melting insulation, increased current flow, and fire potential. Can you explain the Amps column a little more, maybe putting it in context with my coil design in the paragraph above this one?

Thanks!

 

The current columns come from an extrapolation of the NEC-allowed current densities for house wiring. (If you're interested in the details, look at resistor.zip here.) Current density is the current in A divided by the cross-sectional wire area in mm2. You'll find as many different recommendations for allowed current density as my daughter has shoes. Part of the reason of so many recommendations is that it's not a simple thing -- how hot things get also depend on the environment the wire is in (if there are lots of other wires, for example, more heat per unit volume is generated and convective cooling is less effective). It's really a heat transfer problem in disguise. So those numbers are purely guidelines. The different columns are based on the rated temperature resistance of the insulation.

Clearly, a wire insulation rated for 90 C can get hotter than a 60 C wire before problems happen. However, the table says they're all equal. That's a limitation of the NEC data and the extrapolation I made. If I was doing a design like you are, I'd measure the temperature rise of the wire. This is easy to do with my $10 Harbor Freight IR thermometer (this little thing is one of the best purchases I have ever made -- I use it all the time). You have to wait for things to come to steady state. (See the BNC_connector_power.pdf file at the above link for an example where this was done.) If you had an RMS current meter, you could use a lamp dimmer to control the power to your coil. Or, better, use a Variac.

Voltmeters are high-input resistance devices, so they don't really need to be protected with a fuse.

 

My circuit-mounted voltmeter has a 1 amp maximum, and the coil will have some resistors in line at some point, so I see what you're saying about the fuse issue. That's one less thing to install (though I already cut the hole in my project board. Meh.)

I actually have access to an IR camera. I could set the scale for copper and shoot the whole coil to look for hot spots.

I was planning to wrap the coil with gaffer's tape once I wind the magnet wire around it to keep the wires in place. Based on your information above I'm not sure this is a good idea. Any thoughts on that?

In addition, I should note that this coil is not intended to be used for more than maybe 5 minutes at a time. Not sure if that matters. Of course, some twit could leave it on indefinitely.

 

I should note: I could be that twit.

Also, just discovered series parallel resistor array calculations. Looks like I could use a bunch of 1 watt resistors to do what I need for my nearly 2 watt system. This is completely awesome, however I'm too tired to make my brain do math right now. Maybe tomorrow.

 

I don't understand what you mean by "circuit-mounted voltmeter has a 1 amp maximum" -- voltmeters measure voltage, not current.

If you use an IR camera (or some other device that measures radiation), do NOT set it to copper, as you'll get wrong readings (unless you're measuring bare copper). Metals have low emissivities (0.02 to 0.2 or so, IIRC) and plastics, enamel, and non-metallic things often have emissivities in the 0.8-0.95 range. It's a good bet for plastics to use 0.95.

Instead of tape, I'd just secure it with some cheap plastic tie-wraps -- 5 or 6 around the circumference. You can tape it later once you know the temperature rise isn't significant, but realize the tape will reduce the heat transfer, making things warmer.

I'm sure if we were to measure twitness, you'd probably score around absolute zero compared to me. I've done tons of dumb things -- just ask my wife!

 

I'm looking forward to doing more dumb things. Hopefully, none of them injure or kill me. I still have about 40 years left to find out.

The voltmeter says it's rated for 1 amp. I presume it's because the circuit powers the voltmeter. I had considered putting a shunt resistor in line with it anyway. Maybe that isn't very important? I assumed it was.

As for the IR camera. That makes sense. The coating is plastic. Silly me.

Also, I just got about 100 feet of wire wound on the coil. It looks pretty good. I like the idea of the zip ties. I happen to have tons of those around because, well, I drive a VW.

Now, here's what I didn't really think out too well. What's the best way to connect the coil terminal wires to the rest of the circuit? I know I'll need to sand off the varnish to get good contact. The coil is wrapped around a wooden frame, and the coil wire is run around the circumference (about 6 feet) about 100 times as tightly packed as I could get it (600 feet total wire). The resulting tails stick out of one corner, one on each edge of the frame.

Here's what I was thinking:
- get 2 wood screws and 4 washers (all brass)
- clean off about 1/2 inch of insulation from one tail of the magnet wire and also from some flexible, insulated wire
-twist the two together
-put two washers on the screw, squish the wires between the washers, and insert the screw into the edge of the coil frame to secure it.
-repeat for the other tail on the same corner so that both wires come off one corner of the square coil


Which brings me to yet another question: I need the wires that go from the coil to the circuit box to be flexible enough to be bent around. I also heard that, to minimize interference, I should twist together the wires that come off the coil. I have some old ethernet cable with twisted pairs of wires in it.

Do you think I could leave those wires in the exterior sheath and pick one twisted pair to hook to the coil (there are 8 wires in the bundle)? Or should I strip out a twisted pair and use them bare? They are 24 AWG stranded wire, which seems like a pretty good choice for this circuit setup unless I'm completely mistaken.

Thanks for your patience.

 

I've got probably 10-20 years left unless my wife kills me first.

First, there are a couple of ways to get the insulation off the wires (I'll assume you have what's called magnet wire, so it looks like a varnished coating). If you're lucky, it's an insulation that can be melted with a blob of solder on the end of a soldering iron. This makes for quick work because it tins the wire too. The other method is sanding or scraping with a penknife.

As to making the connection, there are numerous choices. I prefer soldered joints with something like vinyl tape over them because they're reliable. However, if the wires get moved a lot, solder isn't the first choice because it tends to make the copper a bit more brittle, although this is typically more important for smaller wires (it used to be a big deal 30 years ago in copper metallization in ICs). Probably the fastest and certainly good enough is a small wire nut. You can find them at the hardware store and they have the advantage that you can take the joint apart again. They can be a little tricky for small wires, but 24 gauge should be OK. But take some sample wires to the store and make sure they work as intended before buying them. Normally wire nuts tend to be used with 12-16 or 18 gauge wire.

Another approach would be to use a crimp-on solderless connector. These are also fast and you can buy connectors that have mating spade connectors that let you plug the wires together, yet take them apart. You'll want the pink or red size or smaller ones if you can get them.

Your wood screw approach should work OK too. Galvanized steel is cheaper and will work just as well. I like hex head sheet metal screws better than wood screws.

The Ethernet cable could be a good choice. As I recall, the wires in them are 24 gauge solid copper and there are 4 pairs. Connect the 4 wires from one set of twists together and the other 4 together and solder each group into a single mass. Then you have a heavier gauge wire so the resistance will be less in your connection wire.

Only doctors have patience. I have hives.

 

Holy BLEEEEEEP!!

I did it! I hooked up a fuse, the transformer (12 V), my 600 feet of 32 ga magnet wire, and my modified ethernet cable, and I get about 360 mG in the middle of the coil. The magnet wire stayed cool to the touch, and the transformer didn't make funny smells.

My calculations found that the amps I was kicking out on the coil were below the temperature where the wire would get hot by about 50%.

To top that off, I figured out series-parallel resistors (I'm pretty sure), and I'll be testing that concept tomorrow.

I thought you'd find this site pretty useful too:
http://demonstrations.wolfram.com/search.html?query=helmholtz

I played with one of the Wolfram models for a square Helmholtz, and it was not bad for modeling what I wanted ultimately tried in my circuit. There seems to be a ton of amazing stuff on that site for the math-impaired (such as myself).

I'll spend a little time testing the circuit directly before bed, but I feel like I'm over the hump on the design. Of course, once I figure this out, I need to design the centering bracket to keep the meter centered on all three axes, and then test it in different backgrounds of magnetic field. But that seems to be the least of my worries.

WHEW.

PS. High-five me! Thanks for all your help!

 

Also, I was able to confirm I'm running about 13.6 VAC. Need to get a more sensitive ammeter tho. Was able to find the sweet spot for all three axial coils in my EMF meter too. Discovered the wood floor gives a different field than my cast iron work bench (no shocker there. Pretty neat.

 

Congratulations! You know, of course, that this might just be the beginning of a hopeless downward spiral. Oh, I need a new ammeter. Gee, I can't read that voltage well enough; maybe I need one of those fancy $1500 voltmeters. Uh oh, it's changing and I can't see what's going on -- now I need an oscilloscope. Oh, here's a nice one for $1000. Then you'll find you need lots of accessories... Pretty soon, you've sold your children into white slavery and hocked your household belongings to satisfy the lust for more stuff. After that, it goes downhill...

 

Pretty soon, you've sold your children into white slavery and hocked your household belongings to satisfy the lust for more stuff. After that, it goes downhill...

You then wake up one day and realize that you have far more stuff than you have room for! It's decision time. The stuff,..... or one of the kids gotta go...maybe two!

 

So, you're saying it's like owning a VW?
LOL.

 

UGH. Now I'm having trouble rigging the rotary and power switch. Will have to figure out how they work (bought them used) and diagram/take a photo of the cranky-pantsed thing. Tomorrow.

 

The common (Wiper) of rotary switches are typically offset from the polls to identify them.

 

I think I may be struggling with issues of polarity and current direction. Not sure how to separate my three resistor arrays from each other so that they do not form individual loops of circuit between the different switch positions. Might have a rotary switch that contacts two terminals at the same time instead of just selecting one. Maybe I'm over-thinking it. Wouldn't be the first time.

I can't well explain it without a photo or diagram tho, so will have to do some documentation to share the concept. Meh.

 

Might have a rotary switch that contacts two terminals at the same time instead of just selecting one.

Yup, rotary switches can come in all shapes and configurations. What you're describing is not uncommon but I'm not sure if it's appropriate for your application. Are you going to post a schematic of what you're attempting? If two polls are always making contact with the wiper you will always have two resistors switched in. How many positions, besides off does this switch have?

 

It has 4 plus the input terminal. I unsoldered my gear and tested it yesterday to confirm that the switch only selects one position at a time. I was doing my continuity tests in the wrong location, it turns out. After I re-brained myself, I could see what I did wrong.

Will definitely post a schematic or photos later. Hopefully today. Got derailed yesterday. I'd like to wire and test this thing by the end of the weekend so I can use my bench for other giant projects. Stay tuned!