TSSP: List Archives

From: Paul
Date: Sun, 03 Sep 2000 07:58:22 +0100
Subject: Re: [TSSP] Mystery of the missing loss

Further results from Terry,

With a pristine foil ground plane:

Bare Coil        Coil at 1"   Coil at 12.5"
f1 = 148.3kHz     Q = 57.7      Q = 61.7
f3 = 347.87kHz    Q = 33.3
f5 = 513.76kHz    Q = 20.3

45 inch Torroid
f1 = 97.93kHz     Q = 80.9      Q = 97.3
f3 = 322kHz       Q = 36.6
f5 = 491.2kHz     Q = 23.8

Thanks again Terry. These figures, together with your measurement
on the effect of a shorted loop around the coil, give me a total
of 9 numbers with which a model must simultaneously agree with.

Some comments,

Terrell W. Fritz wrote:

> So moving the coil further away from the eddy currents, or
> whatever, of the plane did not have a "dramatic" effect on Q...

Perhaps not dramatic, but significant I think - a 20% improvement.

> The coil's Q seems to vary much depending on hard to control
> (at least in one's basment) factors.  Perhaps if some "standard"
> evironment or setup could be extablished that would go well with
> the math models, we could take Q reading more consistently.

I disagree. I think your basement setup is rather good - it's
reasonably stable and well defined. OK, the individual readings
are bound to wobble a bit, but a clear pattern does emerge. 

> I fear the best expermenters trying to reproduce the same Q
> measurements could get fairly large errors depending on a number
> of little factors.  I would guess that +-20% accuracy is all
> that could be expected.

I'll be a little bit more optimistic and go for +/- 5%. Let me
throw in a quick input impedance comparison with Mark's small,
elevated coil, which I'm still working on,

    Zin meas    Re(Zin) model
f1   64.0 ohms     67.9 ohms
f3  118.1 ohms    119.5 ohms
f5  162.3 ohms    168.7 ohms

Terry, thanks to your efforts we've got a pretty good set of
measurements against which to test models of eddy current losses
in the ground plane. I've now got all the formulae gathered
together and can set about incorporating an eddy current loss
estimate into the program. That'll keep me busy for a week or
so, meanwhile I'll refrain from suggesting any more experiments!

Bryan Steinbach wrote:

> Perhaps a better choice for a gnd might be sheet metal on a
> hard floor.

That may be a good way to go in some situations. Accept the
large eddy currents but make sure they circulate in a high Q
ground plane. The advantage - the circulating currents will
raise their own B field which will tend to cancel the coils'
B field which would otherwise extend further into the ground
and perhaps encounter a higher loss material - this is the
'screening can' approach.

Regards All,
--
Paul Nicholson,
Manchester, UK.
--


Maintainer Paul Nicholson, paul@abelian.demon.co.uk.