High Voltage Magnetics: Chokes / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

High Voltage Magnetics: Chokes

1 2 3 4 next

Move Thread

LAN_403

sngecko

Fri Nov 26 2010, 11:20PM

Registered Member #3447 Joined: Fri Nov 26 2010, 11:10PM
Location: North Jersey
Posts: 97

I'm building a 20kVDC linear power supply out of big iron (for now; maybe SMPS later) and would like to build a 200mH inductor as part of an LC filter. I have a 2uF, 25kv capacitor, so I figure that a low pass frequency of around 250Hz would give me a 200mH inductor. The power supply need only provide 160mA max current, but...

The problem is that I don't know where to get a laminated iron core with windows big enough to allow enough turns of insulated wire to get to 200mH. Has anyone here built one of these up?

Thanks.
Scott.

PS - This PSU is going to be used, along with an active current regulator and optical resonator, to operate a >100W CO2 laser.

radiotech

Sat Nov 27 2010, 05:26AM

Registered Member #2463 Joined: Wed Nov 11 2009, 03:49AM
Location:
Posts: 1546

How about finding a used 20 mA iron core neon sign transformer
.Plenty of window in that to make a 200 mHy choke out of
insulated wire, which will have only ~60 volts across it at 60 hZ
at 160 mA, so your problem will just be winding to core and not
turn to turn or end to end. If you run it at 250 hZ that voltage
rises ro ~250 VRMS. The high voltage wire need only be to connect
the choke to external components. your 20kV insulation is just coil
to core,

monokel

Sat Nov 27 2010, 09:07AM

Registered Member #2981 Joined: Thu Jul 08 2010, 01:47PM
Location: Germany
Posts: 35

If currents with frequencies up to about 250 kHz run through an iron cored coil, there may be large core losses. But you can build an air cored coil for little money:
If you want to build an air coil with a specified inductance and wire diameter with as few as possible wire length, you have to use the Brooks coil geometry:
http://www.nessengr.com/techdata/brooks/brooks.html
If the coil width(height, parameter c in the image) is 1.54931 centimeters and the number of turns is N=2250 the inductance will be 200 mH. The wire will have a diameter of 0.31915 millimeters in this configuration. The current/wire area quotient will be 2.5 A/millimeter^2 (assuming a current of 0.2 A). The length of wire will be 329 meters; the mass of copper will be 0.234 kg, you will get it for few dollars. Of course the copper resistance (17.4 Ohms) will be a bit higher compared to an iron cored coil, but this should be no problem if you have HV and small currents (if the current is 0.2 A, there is a loss of 0.7 Watts at the coil).

monokel

Proud Mary

Sat Nov 27 2010, 09:49AM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

You might also consider a linear inductor made by filling a length of plastic - or glass - tube with as many ferrite rods as it will contain, and then winding a single layer coil around it. This is not so efficient as a windowed laminated core, or ferrite toroid, but has the advanatge of an even potential gradient between the two widely separated ends. HV RF power chokes are sometimes made in this way, while smaller RF chokes can be made with a test tube containing a single ferrite rod of suitable size.

monokel

Sat Nov 27 2010, 11:49AM

Registered Member #2981 Joined: Thu Jul 08 2010, 01:47PM
Location: Germany
Posts: 35

Proud Mary wrote ...

You might also consider a linear inductor made by filling a length of plastic - or glass - tube with as many ferrite rods as it will contain, and then winding a single layer coil around it.

But you need a very long core if you don't want the ferrite core to go into saturation. Sometime ago I found formulas to calculate the inductance of a linear coil. Assume the core is 8 centimeters long and the diameter is 1 cm. Then you need about 1900 single layer windings to get an inductance of 0.2 H. This leads to a wire diameter of 0.042 millimeters; this is extremely thin. In this configuration the B field in the core will be about 0.3 T at a current of about 0.16 A (RMS).

Now I try to make the core longer: If the core is 30 centimeters long and the diameter is 1.0 centimeter, you need about 1517 windings for an inductance of 0.2 H. Thus the wire diameter is 0.2 millimeters; this is a more reasonable thickness for a current of 0.2 A (RMS). The maximal B field in the core will be about 0.37 T at a current of 0.2 A.

The core for this will be more expensive than the copper for an air cored coil with the same inductance.

I have made diagrams which help designing linear inductors if the core should not go into saturation. But they are in German. Maybe I'll translate them and post them soon here.

Proud Mary

Sat Nov 27 2010, 12:34PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

monokel wrote ...

But you need a very long core if you don't want the ferrite core to go into saturation. Sometime ago I found formulas to calculate the inductance of a linear coil. Assume the core is 8 centimeters long and the diameter is 1 cm. Then you need about 2041 single layer windings to get an inductance of 0.2 H. This leads to a wire diameter of 0.04 millimeters; this is extremely thin. In this configuration the B field in the core will be about 0.3 T at a current of about 0.15 A.

Now I try to make the core longer: If the core is 30 centimeters long and the diameter is 1.0 centimeter, you need about 1517 windings for an inductance of 0.2 H. Thus the wire diameter is 0.2 millimeters; this is a more reasonable thickness for a current of 0.2 A. The B field in the core will be about 0.3 T at a current of 0.2 A.

The core for this will be more expensive than the copper for an air cored coil with the same inductance.

I have made diagrams which help designing linear inductors if the core should not go into saturation. But they are in German. Maybe I'll translate them and post them soon here.

I found using bundles of cheap Russian 200mm x 10mm MF ferrite rods effective across a range of applications. Cheap ferrite slabs can also be stacked with gaps to form rectangular cores. I'm not surprised that your experiment with a single core was not successful.

monokel

Sat Nov 27 2010, 01:05PM

Registered Member #2981 Joined: Thu Jul 08 2010, 01:47PM
Location: Germany
Posts: 35

Proud Mary wrote ...

I'm not surprised that your experiment with a single core was not successful.

I have not made any experiments; I've calculated the inductance and the B field in the core. If you build the coil in the way I described, experiments will be successful. If you use 200 mm long cored instead of 300 mm, the inductance will be less.
Of course, if core saturation is allowed, you can build the linear coil smaller.

monokel

Sat Nov 27 2010, 01:07PM

Registered Member #2981 Joined: Thu Jul 08 2010, 01:47PM
Location: Germany
Posts: 35

-- deleted --

sngecko

Sat Nov 27 2010, 01:20PM

Registered Member #3447 Joined: Fri Nov 26 2010, 11:10PM
Location: North Jersey
Posts: 97

I like the simplicity of the single layer designs you've mentioned. I had only been thinking of NST cores.

Would it be reasonable to use a ~250mm x 10mm diameter glass tube filled with ferrite rods (potted in paraffin, perhaps?)? I see hoards of ferrite rods on eBay that should serve the purpose.

monokel

Sat Nov 27 2010, 01:27PM

Registered Member #2981 Joined: Thu Jul 08 2010, 01:47PM
Location: Germany
Posts: 35

sngecko wrote ...

Would it be reasonable to use a ~250mm x 10mm diameter glass tube filled with ferrite rods (potted in paraffin, perhaps?)? I see hoards of ferrite rods on eBay that should serve the purpose.

With these dimensions, you need about 1550 single layer windings for an inductance of 0.2 H. Thus the wire diameter will be about 0.16 mm; this is quiet small. Assuming an RMS current of 0.160 A, the max. B field amplitude in the core will be about 0.37 T.
Is the current of 0.2 A realistic ? Of course this depends on the way the coil is connected and on the voltages and frequency occuring at the coil. If the current was smaller, you could make the coil smaller, of course.

1 2 3 4 next

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