Hi Jim,
>> Note that despite this characteristic, you can still use 43
>> material quite well for boradband transformers covering the whole
>> HF range. For example, if you make a bifiliar transformer that will
>> be used to transform from 50 ohms down to 12.5 ohms, you can
>> easily wind enough turns on a small toroid to get 1000uH of total
>> inductance.
> Yes -- BUT -- winding those turns also adds parallel capacitance,
> which forms a parallel resonance with the inductance.
That's right, but in many cases this capacitance is still irrelevant.
Even when it causes a resonance inside the band of interest, which it
often does indeed, this is at such a high level of reactances, that it
is swamped by the low circuit impedance and causes no noticeable effect.
For a practical example, that 50 to 12.5 Ohm transformer with 1000uH
total inductance might need 30 turns or so on a small toroid, and might
have 5pF of stray capacitance. That would put the resonance around 2MHz,
but with reactances of about 14000 ohms in parallel with the 50 ohm
circuit impedance, that's nothing. And if we will use this transformer
at up to 30MHz, then at that frequency the capacitance will totally
dominate over the inductance, but will still be around 1000 ohms, and
thus have almost no effect.
But clearly in each particular case the capacitance and its effects has
to be evaluated. It's easy to run into trouble with it, when the circuit
impedances become higher.
> You can see MEASURED DATA from an excellent lab in Appendix One of
> http://audiosystemsgroup.com/RFI-Ham.pdf
You have lots of good info there.
>> On 160 meters, that will give you a reactance of more than 10
>> kiloohm, and the parallel loss resistance will be even larger, so
>> that this transformer will have next to no loss. And on 10 meters,
>> the inductive reactance will have risen to almost 200 kiloohms,
>> while the equivalent resistance might be as low as 10 kiloohms
>> (don't take this too literally, I haven't checked the data sheet,
>> I'm just giving estimated numbers out of my head). Note that using
>> this transformer as an inductor would be absolutely terrible,
>> because it would have a Q factor of only around 0.05! But in that
>> transformer, the loss caused by the 10 kiloohm resistance in
>> parallel with the 50 Ohm circuit impedance is totally negligible!
>> That's why 43 ferrite does work well in broadband transformers.
> The above analysis is seriously flawed, because it overlook the
> contribution of stray capacitance, as discussed above.
Yes, indeed I overlooked the capacitance! But even so, with the
capacitance dominating over the inductance, the transformer still works
well...
> The Q of the resulting resonant circuits wound on #31 Fair-Rite cores
> is on the order of 0.4 at resonances in the MF and HF range. For #43
> ferrites, the Q is slightly higher. Above resonance, a parallel
> resonant circuit looks like a lossy capacitor. Below resonance, it
> looks like a lossy inductor. At resonance, it looks like a resistor.
I agree.
> In general, you trade permeability for low loss -- that is, low loss
> materials also have lower values of mu.
That's right. And in the situation originally described, a plate choke
wound on a ferrite rod, the permeability of the material isn't very
relevant at all, because the total equivalent permeability of teh entire
path is dominated by the air, through which the flux lines have to
return. This allows obtaining almost the same inductance with a material
having pemeability 10, than with one having 1000! So it's a good case
for using a low permeability material, even as low as what's provided by
powdered iron. The 43 material with its 850 permeability is certainly
out of place here.
> Ferrite materials also vary widely in their resistivity. Some are
> quite good insulators, while others have substantial conductivity. If
> the material you're using is good conductor, it may be important
> that the coil is insulated.
A good point. That's probably the main reason why that recipe mentioned
in this thread has the ferrite rod covered by a teflon tube!
Generally the low permeability ferrites are decent insulators. The high
permeability onces are indecent conductors! :-)
>> And then you have to design the whole thing, considering the AC
>> flux density you will be putting through the ferrite, and the
>> resulting loss. You also need to consider DC flux density, and make
>> sure it won't saturate the ferrite, but this is usually not a
>> problem in such chokes, because the AC flux density limit for
>> acceptable loss is so low that the resulting DC flux density is
>> child's play!
> Not necessarily. If there is enough DC current in the choke, it can
> cause saturation.
In theory, yes, but hardly in the case of a plate choke on a rod. The
instant flux density is roughly proportional to the instant current (it
would be exactly proportional if the magnetic material had a linear
response). So we can compare teh RF current to the DC current in the
choke, and infer the ratio between RF and DC flux density. Now, a
typical amp might have up to 1A DC current, and the RF current might be
anyway from about 30 to 500mA, depending on choke value and frequency.
Considering that a typical magnetic material (ferrite or powdered iron)
can only take an RF flux density of relatively few millitesla, but at
least 300 millitesla at DC, teh result is that indeed very often there
will be no saturation from DC after designing for proper RF performance,
even when no care has been given to what happens at DC. Many a builder
has been lucky with this! But the correct approach, of course, is to
calculate everything, and that includes calculating the DC flux density,
and making sure it causes no problem.
> 300 turns on a core will set up a lot of stray capacitance, that
> shorts out the choke.
That's precisely the point why some people wish to use a magnetic core!
It helps keeping down the turns number.
Of course, a great lot depends on how the turns are arranged. A solenoid
isn't too bad. The hot end should be away from the chassis and other
stuff. Thin wire, with some space between turns, is better than
close-wound thick wire.
> Remember old-fashioned RF chokes wound
> "pancake" style? The purpose of that winding style was to reduce
> stray capacitance.
I think you mean this criss-crossed winding style called "PI winding"? I
have often wondered how to make such a winding at home!
So, I hope you can forgive me my forgetting stray capacitance! ;-)
Manfred.
========================
Visit my hobby homepage!
http://ludens.cl
========================
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|