[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: AC Resistance of wires - was 8 kHz Tesla Coil



Original poster: "Gerry  Reynolds" <gerryreynolds@xxxxxxxxxxxxx>

Hi Kurt,

The Medhurst resistance tables assume "high frequency" and that the proximitry effect is "fully saturated". The table was also developed for coils with a small number of turns (30-50). The proximitry effect may be overestimated.

Gerry R

Original poster: "Kurt Schraner" <k.schraner@xxxxxxxxxxx>

Hi Bart,Gerry,

Medhurst's empirical Method for HF-Resistance of solenoids is implemented in my TC-Plan Excel:
http://home.datacomm.ch/k.schraner/TCplan.zip
The relevant sheets are "Skineff." and "Proxi-eff.", whose results are
transferred to "TC calc."' cells E34:E38. In the light of more up to date,
more general methods, Medhurst may soon become of lesser relevance, but I've tried to compare results from it, with Q of some of my coils, estimated from bandwidth measurements at Fres. A summary Excel of the results can be had at:
http://home.datacomm.ch/k.schraner/Qsec2.xls
or here (view in fixed width font, sorry for the possible mess!):


     TC-Secondary Q-Estimation by Medhurst's Empirical Method
      for some of my Coils

     Geometric Data
     Secondary Diameter Length Aspect L/D Turns
     Name      [cm]     [cm]     [ - ]    [ - ]
     Sk-Seibt   4      150.3     37.58     4187
     Sk-UBTT  12.54     60.3      4.81     1680
     Sk-5cm    5.11     41.1      8.04      934
     Sk-12cm  12.12     58.5      4.83      894
     Sk-Long  16.1     140.2      8.71     1976
     Sk-20cm  20.52     66.8      3.26      943
     Sk-38cm  38.4      44        1.15      346
     Sk-B&W *)40.13    177        4.41      813
     Sk-B&W *)40.13    177        4.41      813

     Geometric Data (cont.)
     Secondary Wire-Copper  Pitch Comment
     Name        [mm]       [mm]
     Sk-Seibt   0.319       0.359 Big error unexplained!
     Sk-UBTT    0.319       0.359 Test-signal via low-Z amp.
     Sk-5cm     0.4         0.434 no experiment
     Sk-12cm    0.6         0.654 Test-signal via low-Z amp.
     Sk-Long    0.64        0.71  Test-signal via low-Z amp.
     Sk-20cm    0.63        0.708 no experiment
     Sk-38cm    1           1.2   cardoard/cotton insul. wire
     Sk-B&W *) 1.268        2.177 Test-signal via low-Z amp.
     Sk-B&W *) 1.268        2.177 if R Sig-Gen=136 Ohm


Electrical Data Secondary at1kHz Fres,bare DC Resist. Wire straigt Name Lexp[mH] exp.[kHz] exp.[Ohm] with Skineff. Sk-Seibt 18.13 366.11 114 240.82 Sk-UBTT 66.59 205.4 147.8 147.86 Sk-5cm 5.203 923.8 21.7 n.a. Sk-12cm 18.2 402.8 21 27.61 Sk-Long 67.7 147.4 55.2 61.05 Sk-20cm 49 202.7 34.8 40.5 Sk-38cm 28.12 217.95 9.2 16.32 Sk-B&W *)54.67 119 14.5 22.3 Sk-B&W *)54.67 119 14.5+136 22.3+136

     Electrical Data (cont.)
     Secondary Rac-calc. Qexp     Qmedh. Qerror
     Name    with Proxy.E  [ - ]  [ - ]    [%]
     Sk-Seibt    698.9     156.5  70.8    54.8%
     Sk-UBTT     440.65    211.8  186     12.2%
     Sk-5cm       n.a.      n.a.   n.a.    n.a.
     Sk-12cm     91.01     269    270     -0.4%
     Sk-Long     190.6     164.4  168.4   -2.4%
     Sk-20cm     128.0      n.a.  478.3    n.a.
     Sk-38cm     49.79     104.3  803.7  -670.6%
     Sk-B&W *)   44.02     141.7  576    -306.5%
     Sk-B&W *) 44.02+136   141.7  141.7    0.0%


The method seems to underestimate Q of my small diameter coils, and to (heavily) overestimate the Q of the big diameter coils. What seems obvious is the dominant proximity effect over skineffect for usual TC frequencies, as mentioned before on the relevant treads. Maybe dielectric loss, not considered in the calculations, come to play here...

*) For the bandwidth measurements it seems important using a low-Z signal
source (--> Terry's low-Z amp), which was not the case for "Sk-B&W".
Otherwise the generators internal resistance will reduce measured Q value.
Anyway, bandwidth measurements seem prone to higher error, 'cause of
amplitude reading on the scope and frequency precision reading from the
sig-gen.

Well ... my 2cents for the interesting threads about coil Q. If I see coming out a method estimating Q with perhaps 10% error, I'd readily adapt it!

Best regards
Kurt