tales of the rogawski coil?

• tales of the rogawski coil?
  Posted by CopperCone on 18 Mar, 2018 15:55
• Anyone ever build one of these before?
  
  Here is a nice design:
  http://www.ee.bgu.ac.il/~kushnero/belles-lettres/Rogowskoil.pdf
  You can make it from an old coaxial cable.
  
  Seems to be alot of options for these guys, seems that high bandwidth = lower current and vise versa.
  Also alot of options on how to do the loops.
• #1 Reply
  Posted by coppice on 18 Mar, 2018 16:06
• Rogowski coils are widely used for high current measurements in grid applications. They can be problematic for small currents, as they are hard to screen from surrounding fields. The key positives are that no core means no core to saturate and no core to limit bandwidth.
  
  I've never seen anyone use coaxial cable to make a Rogowski coil. In the slides it looks like they are just using it as a flexible former.
• #2 Reply
  Posted by CopperCone on 18 Mar, 2018 16:28
• some designs will use the inner wire, basically you wind the coil up, and return the current through the middle of the coaxial cable, and terminate the other end of the cable on the shield, so you can use a regular BNC connector for it without having to mold some kinda mechanical fixture for a coaxial connector.
  
  The routing arrangement has some differences in the pickup behavior of it I think.
• #3 Reply
  Posted by ahbushnell on 19 Mar, 2018 00:59
• We use Rogawski coils in pulsed power all the time.  High currents high bandwidth.  We use coaxial cable as the coil form.  Peal back the braid and wind the coil on the insulation and connect the other end of the coil to the center conductor.  We also add a shield of copper foil on the out side and connect to the braid to block efields.  You need to leave a slot in the foil to allow the field to reach the coil.  use tap to insulate to you don't put a shorted turn around the coil.  If you don't use the coax method still bring the wire back along the coil so you don't add a big loop that can pick up external fields. 
  
  Remember the Rogawski coil can not measure DC. 
  
  IEEE paper
  http://ieeexplore.ieee.org/document/4084317/
  you can find at a library if you don't have an IEEE account
  
• #4 Reply
  Posted by Quarlo Klobrigney on 19 Mar, 2018 03:06
• Looks like new design fodder for the Free Energy or Audiophool people.
• #5 Reply
  Posted by CopperCone on 19 Mar, 2018 15:02
• They are more meant to measure things like capacitor dicharge currents and pulsed loads... Igbt, power mos... I dont think it really has much use in audio unless your are measuring some kinda super massive speaker
• #6 Reply
  Posted by coppice on 19 Mar, 2018 15:10
• Quote from: Quarlo Klobrigney on 19 Mar, 2018 03:06

  Looks like new design fodder for the Free Energy or Audiophool people.
  

  Yeah, the very latest thing from a guy who published his work on current sensing in 1912, and died in 1947. 
• #7 Reply
  Posted by free_electron on 19 Mar, 2018 17:03
• I made several of those as a flexible pcb . simply wrap them around the wire to sense.
• #8 Reply
  Posted by CopperCone on 26 Mar, 2018 20:44
• My current problem is shielding this fucking thing.
  
  I made a coil, used 11nF mica cap and series/parallel resistor like in the powerpoint I linked.
  
  I made a fast capacitor bank, consisting of 1.2uF @ 2000V using foil capacitors.
  
  When I discharge it with a reasonable voltage, like 200V, the coil basically triggers just being near the spark. In fact, it triggers when I touch one end of the capacitor with my pliers (without creating a spark!). I noticed the capacitor has alot of pickup, so I shielded it with some copper tape. The coil is not in the circuit.
  
  I was told that coil symmetry is the #1 factor I should be concerned about here.
  
  I will try again with a better wound coil I suppose. I don't know exactly how to shield it though. I saw in the powerpoint that you need to terminate the shield according to the coaxial properties. I thought I could slit open a vinyl tube, coat it with copper tape/conductive epoxy, then curve it around the coil, to get known stable shield dimensions.
  
  What should I focus on here?
• #9 Reply
  Posted by CopperCone on 27 Mar, 2018 00:30
• also, why does it spike when I touch it with a insulated pair of pliers on one terminal without a discharge? I got a big pulse waveform from doing that (with the capacitors disconnected from the ground referenced PSU).
• #10 Reply
  Posted by dmills on 27 Mar, 2018 10:09
• Rogowski coils sense dI/dt so really fast events produce large amplitudes at the output, charging up your pliers from a film cap sounds likely to count as a fast event to me. To see the current waveform integrating the output is required (either in some preamp electronics or with your scopes math mode).
  
  Note that you need to wrap the coil right around the target conductor to capture all the flux (and to close the loop to minimise sensitivity to nearby external fields, physically bigger then it needs to be is not a figure of merit.
  
  I made mine on a thin plastic tube of the type used as a spout on a can of air duster, chucked it in the lathe (poked the end into a collet in the tailstock to keep it straight) wound the coil then fed the return down the centre of the tube, covered it in glue lined heatshrink to keep the coil in place then screened with copper tape and kapton, that bit was a faff mainly because of the need to slit the foil.
  
  Regards, Dan.
• #11 Reply
  Posted by CopperCone on 27 Mar, 2018 13:56
• How much rejection does it have though? The coil can be laying like several inches away from the capacitor bank, without any wires in the loop and the signal it picks up looks the same and has the same amplitude values as when it is over the conducting wire.
  
  I will make one with the center conductor looped back as you describe. I believe it is said to aid in rejection.
  
  I removed the capacitor from my circuit because i thought it might be a source of pickup but this is not the case.
  
  
  I read from one manufacturer that a way they shield them from large dvdt transients is by using the sheild as a return for the coil (rather then routing it inside) but i have yet to find and construction details or formulas about this design.
• #12 Reply
  Posted by CopperCone on 27 Mar, 2018 15:55
• Also dan, did you terminate the shield or connect it directly to ground? 
  
  I see in the pdf i linked they recommend terminating the shield from both ends
• #13 Reply
  Posted by CopperCone on 29 Mar, 2018 16:09
• Do you guys think the loop back through the center conductor will give better performance? I am short on coppper tape and idk if i should make a new rogawski coil that has the compensating turn or if i should shield the one i have.
  
  The compensating turn does not do anything for immunity to high electric fields right?
  
  I would like to use lme600 cable but its a bit thick. All the cables i have are super wimpy, and i think simply putting a loopback wire through a tube without a centering former wol lead to erratic performance
• #14 Reply
  Posted by mikerj on 01 Apr, 2018 12:19
• The uniformity of the winding in a Rogowski coil is very important for the refection of external fields. 
  
  The company I used to work for used Rogowski coils for measuring the traction current harmonics in electric locomotives.  These were powered from a DC third rail, but used electronic drives to power 3 phase motors, so switching noise could be considerable and could interfere with the track circuit block signalling.  They were so sensitive that temperature gradients causing uneven expansion of the former that the coils were wound on could degrade external field rejection.
• #15 Reply
  Posted by CopperCone on 22 Jun, 2018 16:06
• What kind of mechanical specifications would be good togo for? I am angry at this bullshit.
  
  I am imagining i can get good results with a tiny knife file and some gauge/angle blocks i have to make a helix marking along with a caliper. I also have some machinist contours and i have an idea how tomake something precise withouta 5 axis cnc machine using magnification.i think you just need to work real slow and relax your eyes so youdont get hamfisted bullshit
  
  I feel likw this god damn coil is taking me hostage
• #16 Reply
  Posted by coppice on 22 Jun, 2018 17:09
• Quote from: CopperCone on 22 Jun, 2018 16:06

  What kind of mechanical specifications would be good togo for? I am angry at this bullshit.
  
  I am imagining i can get good results with a tiny knife file and some gauge/angle blocks i have to make a helix marking along with a caliper. I also have some machinist contours and i have an idea how tomake something precise withouta 5 axis cnc machine using magnification.i think you just need to work real slow and relax your eyes so youdont get hamfisted bullshit
  
  I feel likw this god damn coil is taking me hostage
  

  Many engineers are attracted to the apparent simplicity and low material cost of Rogowski coils, but most people fail to achieve their goals and end up using a different sensor. You are in good company.
• #17 Reply
  Posted by CopperCone on 22 Jun, 2018 23:45
• You can't do any real research into high impulse current without one. All I have is one of those freon filled probes but its useless if I keep picking up the voltage signal with my current probe.
  
  I also thought a good method for starting would be to take your ring, glue it to a spin indexer, and make some kind of fixed pivot with a knife edge on it for initial marking, so you rotate it in your spindexer and peck at it with your fixed edge, or even mount a contoured pattern instead of a knife edge so you can scribe your follow line on the core.
  
  You can make a contour pattern by drilling a hole the same diameter as your inner core, cutting it roughly in half, then filing it exactly in half, to make a semicircle, but instead of cutting it in half you would cut it along the axis between two opposite designated points and then file it on a flat. This way you don't really need machine tools.
  
  Trying to think about how to get past the need for a spinindexer though. If its large enough it might be possible to make a precision rotational table using a fixed magnifier, a computer printed diagram and a round, I am wondering if you can use a sheet of glass cut with a glass hole scriber, then cut teeth in the glass using a diamond knife file (very fine for a very fine metric screw), you would need a very perfect circle, I don't think something generated by a hole saw is going to work good enough with how precise everyone wants this thing.
  
  I would have to inspect some common round low thickness objects for roundness. What comes to mind is old CDROMS (they might be well regulated) that are glued together, old record player, or I think for a hobbiest the most accessible accurately sized and bearinged material would be to use an old hard drive, then glue a bunch of spindles together lined up on an axis, print out a marked strip (like for an encoder, glue it to the outside edge, then use a fine point file to make teeth that can interface with a lead screw (now getting the dimension of the tooth groove is difficult, its like making your own gears, I need to research this more... but you can also do something with brass), not sure how to do this again without a broach an a spindxer, all I can say is you need to be a competent and careful with filing and visual inspection, you might need some kind of a go/no go gauge too
  
  But, if your lead screw is plastic or long enough to flex a bit without warping the table, it will act as a break and greatly reduce the amount of precision you need your gear teeth to have, since your optically aligning it like a caliper and you don't need high linearity.
  
  Since your just using it for marking, backlash does not matter, only how ridgid it is, so your top computer printed scale would have to line up with some kind of overlay
  
  This procedure would produce a little table that is capable of very fine rotation if you do it right, by adjustment of a lead screw, so you can mark sub degrees of your torroidal ring very accurately. I would recommend making fixed optical magnifiers so all the work during rotational marking is done under magnification (I plan on using a bench magnifier).
  
  I hope that this will allow for enough symmetry that you get high CMRR without the shield, so you can make a useful measurement on a spark, after you shield it
  
  Figuring out how to do the shield well is the next part.
• #18 Reply
  Posted by CopperCone on 23 Jun, 2018 00:05
• Am I missing some obvious manufacturing method here though? This seems insane. All I can think of is this, a very mobile CNC machine that can cut a helical groove into a torroid, 3d printing or this above described painstaking as hell manual assembly.
  
  I imagine the 3d printer type that uses a laser and rises a former out of the water like the clear terminator skulls would work well here? I don't see a regular 3d printer doing a good enough job because of warping?
  
  I don't know if I want to build a precise hand adjusted rotational table, maybe one made of hard drive plates would later prove useful for aligning optical mirrors for laser experiments, like stuff involving diffraction gratings or high microwave radiation that's well focused? I only say this because harddrive bearings are pretty decent and everything is fairly well machined in them, otherwise this tool is just borderline too complicated for me to build.
  
  
  oh wait, since you want to reroute the wire in internally, this is useless since you can't get a bore hole inside of your torroid.
  
  
  Can anyone quantify how much unbalance is added by routing the wire internally through the middle vs carefully laid out wire that goes strait into a twisted pair without the cancelation loop or whatever you want to call it so a trade off can be made?
  
  I also wonder what the CMRR on those PCB rogawski coils is, like the ones texas instruments have on their website as samples, they are not shielded. Supposedly bandwidth of 1MHz
  
  But same problem ,you can't make them at home, they require at least 3 layers. I can do VIAs myself but not the feedback winding.
  
  I guess I can make a thick one, using a piece of double sided and single sided PCB, if I align my coppersets right..