[AMPS] more resistor measurements

[John Lyles]

Dang it, now you got my curiosity about this resistor thing...

I measured 3W metal oxide film resistors from DigiKey, cause I happen to
have some. These values were measured on the Hp4193A vector Z meter, with
about 3/16 inch leads on each side to the fixture. Sorry to you metric (SI)
readers. I measured 5.1, 10, 15, 20, 30, and 51 Ohms. If someone sends me a
higher value or different brand, I could add it to my data. This is the
kind of stuff I wish Digikey, Mouser, Newark, Allied, Farnell, and all
should publish in their catalogs, instead of saying "non-inductive design".
That's pure BS, sure they are not wirewounds, but certainly NOT non
inductive.

I put together the data in a 6 page spreadsheet, with the formulas to
convert impedance |Z| and phase (A) to Rs and Ls, which should be the model
for a resistor I think. Rs = |Z| x cos A,  and Ls = |Z| sin A /(2 x pi x F)

If you want a copy, it is in Excel worksheet format, for Office 98 (from a
Mac). However, I can also send it in the older Excel formats. In this
worksheet are 6 plots of the Resistance and Inductance for each resistor,
for every 5 Mhz from 1 to 100.  The lowest value had the highest
inductance. But as they increased, the L always showed a dip at 5 Mhz. And
were higher at 1 Mhz. I am suspect that it is a systematic error of the
4193A probe residual L and R. They are published and in the range of the
low measurements. Also, I could probably look up the error specs and find
that it is worse off at low impedances, like 5 to 10%. So that could
explain why 5.1 Ohms is SOOOO different. Any other explanations to hear?

For those who cannot bring up the spreadsheet or don't care, but want to
know how it compared to N5JA or W8JI's measurements (different techniques),
here is a summary of my Rs/Ls in that order (Ohm/nH):
              5.1 Ohms    10              15                20
30                   51
1 Mhz     5.6/60     10.8/27     15.6/26      20.4/28.3       30.8/25.7
50.9/28.3
10         5.7/59      10.8/24    15.6/23.9    20.4/23.9       30.7/22.2
50.7/19.7
30         6.13/58    11/24.6     15.7/24.4   20.6/24.2       31/22.8
51.1/19.4
50         6.6/57.4   11/24.2     15.9/24.2   20.7/23.8       31.2/22.4
51.3/18.9
70         7.1/57.1   11.1/24.1  16/24.2      20.9/23.8       31.5/22.6
51.4/18.7
100       8.2/56.8   11.3/23.9   16.1/23.7   21.1/23.4      31.6/22
51.6/18.2

The reason it is nice to sweep the whole range for each R, is to observe
that it is indeed flat and doesn't exhibit some resonances, in this case
all behave as inductive resistors up to 100 Mhz. You can also see the
variation (mostly increasing) of the high freq resistance. Wish I could go
higher but....

By the way, this compares to the ballpark values posted yesterday, and
these are the 3W smaller resistors.

Please let me know if you want the data/plots emailed. Also, I haven't
forgotten the folks that inquired for a paper copy of the old Philips
article on pondermotive bending of filaments when arcs occur. It's just
that I've been slightly swamped with getting an amplifier finished and
trying to get one up on old Murphy.... (impossible)

I remember doing an experiment in Radio Engineering Lab at Va Tech, under
Dr. Herb Krauss (he cowrote the book Solid State Radio Engineering), in
which we found that 100 Ohm carbon comp resistors were the most purely
resistive resistors. Now I cannot remember if it was 1/2, 1 or 2 Watt! But
we just built a power resistor ourselves (2 weeks ago), and wanted a really
low L or C 0.1 Ohm resistor, that can handle about 35 Amperes at 5% duty
factor at 5600 KHz. We started with Cesiwid (globar/carborundum types),
with 10 each of 1 Ohm units in parallel. They were not good, very
inductive. We settled on buying 100 real 1/2 watt carbon comp resistors at
10 Ohms each, and soldering them between two pieces of copper punched at
0.1 inch spacing like vectorboard. It was time consuming to build, but it
is one sturdy and fairly good resistor. I realize that I could have located
some power chip resistors to do this, and soldered them in parallel, but
the leaded R's (which cost me almost 50 cents each!) seemed to be fine and
quick to get. Didn't even consider the Metox film types, as we prototyped
it with carbon comps. Wonder how the 3W guys would have worked?

John
K5PRO



--
FAQ on WWW:               http://www.contesting.com/ampfaq.html
Submissions:              amps@contesting.com
Administrative requests:  amps-REQUEST@contesting.com
Problems:                 owner-amps@contesting.com
Search:                   http://www.contesting.com/km9p/search.htm