Loading ...
Sorry, an error occurred while loading the content.
 

Re: Indoor (ferrite-) loop antenna test

Expand Messages
  • dl4yhf
    Hello Walter and group, The cascode amplifier has a high-Z input at low frequencies. But a test to determine the input capacitance showed a strange result
    Message 1 of 8 , Sep 1, 2009
      Hello Walter and group,

      The cascode amplifier has a high-Z input at low frequencies. But a test to determine the input capacitance showed a strange result (maybe a wrong measurement):
      I connected the amplifier to a variable AF generator through a 680 kOhm series resistor; then slowly increased the frequency (starting at a few hundred Hz) until the amp's output voltage dropped by 3 dB. To my amazement, this already happened at about 1.1 kHz, which means an input capacitance -caused by the JFET- near 200 pF. Sounds a bit high, because the cascode configuration (almost no AC voltage at the drain) should mostly eliminate the effect of the 15 pF drain-gate capacity (no Miller-effect which would otherwise multiply Cdg by the amplifier's voltage gain); and leaving only the 75 pF between gate and source. I will try some negative feedback (resistor between output and input) and see if this flattens the frequency response, or if a grounded-gate amplifier works better. But for the moment, it already looks reasonably flat in the spectra mentioned below. For a real check I need to take the antenna to a quieter place, and see if the amplifier needs more gain (or an even lower noise figure).

      Latest spectrum and spectrogram are here:

      http://www.qsl.net/dl4yhf/vlf_rdf/latest_vlf_spectra.htm#vlf_loop_indoors

      You can see that without the notch filter, the signal is terrible.
      If this is of any interest, I can leave the stream run for a few minutes without any filtering so you can get an idea how bad the hum (and similar stuff) really is.

      Cheers,
      Wolf .

      --- In VLF_Group@yahoogroups.com, "karula4711" <fedderwi@...> wrote:
      >
      > Hi Wolf
      > > The amplifier is a JFET cascode
      >
      > Is it a low iimpedance input or a hi Z input?
      > Eventually a silly question cause everybody tells you about
      > measuring I from the loop :-)
      >
      > And please post a spectrum and a waterfall diagram of your kitchen
      > table setup somewhere. I find it always interesting to get a feeling
      > for 'normal' noise environments at other places.
      >
      > Viele Grüße
      > Walter
      >
    • Peter Schmalkoke
      Hi Wolf, maybe the cascode configuration is still responsible for an increased input capacitance: The 2nd FET, which is connected to the first FET s drain,
      Message 2 of 8 , Sep 1, 2009
        Hi Wolf,

        maybe the cascode configuration is still responsible for an increased
        input capacitance: The 2nd FET, which is connected to the first FET's
        drain, causes a low drain-source voltage at the input FET, which will
        then conduct less current than under "normal" conditions. This might
        be compensated for by a low gate-source voltage at FET 1. But since it
        is a JFET, that gate's junction capacitance would then be higher than
        usual. So I would suggest to increase the gate voltage at the second
        FET and adjust the drain current with the resistor in the source lead
        of the input FET.

        Regards, Peter S.


        dl4yhf wrote:
        >
        > The cascode amplifier has a high-Z input at low frequencies. But a test
        > to determine the input capacitance showed a strange result (maybe a
        > wrong measurement):
        > I connected the amplifier to a variable AF generator through a 680 kOhm
        > series resistor; then slowly increased the frequency (starting at a few
        > hundred Hz) until the amp's output voltage dropped by 3 dB. To my
        > amazement, this already happened at about 1.1 kHz, which means an input
        > capacitance -caused by the JFET- near 200 pF. Sounds a bit high, because
        > the cascode configuration (almost no AC voltage at the drain) should
        > mostly eliminate the effect of the 15 pF drain-gate capacity (no
        > Miller-effect which would otherwise multiply Cdg by the amplifier's
        > voltage gain); and leaving only the 75 pF between gate and source.
      • dl4yhf
        Hi Peter, Thanks for the suggesion .. ... Confirmed, the first FET has a drain-source voltage of only approx. 4 volts, same as the 2nd FET (with grounded
        Message 3 of 8 , Sep 1, 2009
          Hi Peter,

          Thanks for the suggesion ..
          > maybe the cascode configuration is still responsible for an increased
          > input capacitance: The 2nd FET, which is connected to the first FET's
          > drain, causes a low drain-source voltage at the input FET, which will
          > then conduct less current than under "normal" conditions.
          Confirmed, the first FET has a drain-source voltage of only approx. 4 volts, same as the 2nd FET (with grounded gate), and the 3rd FET (which acts as a constant current source, at least for AC). There's a graph in the datasheet for the input capacitance vs drain-source voltage which shows approx. 80 pF at Vds = 4 V, but "measured at 1 MHz" - mmmh, why do they measure the capacitances of an audio transistor at 1 MHz ? Does the measuring frequency have a significant effect on the measured capacity (Ciss, Crss) at all ?

          I also played around with different drain currents (from 2 to 8 mA) to see the effect on the input noise density (it varied from 0.9 nV/sqrt(Hz) to somewhere around 0.65 nV/sqrt(Hz)), but the amplifier's input capacitance was not affected by the drain current.

          > So I would suggest to increase the gate voltage at the second
          > FET and adjust the drain current with the resistor in the source lead of the input FET.

          Ok, I can simply do this by increasing the supply voltage because the drain-source voltages of all three FETs are roughly 1/3 rd of Vcc.

          Enough hardware playing for today. VLF6 is streaming again, and for those who'd like to hear the UNFILTERED signal from the "indoor loop", I will turn off the automatic notch between 19:00 and 19:30 now:

          http://abelian.org/vlf/detail.php?stream=vlf6

          Cheers,
          Wolf .
        • dl4yhf
          Ooops.. ... Forgot to mention: Times in UTC. In other words, NOW. Cheers, Wolf .
          Message 4 of 8 , Sep 1, 2009
            Ooops..

            > I will turn off the automatic notch between 19:00 and 19:30 .

            Forgot to mention: Times in UTC. In other words, NOW.


            Cheers,
            Wolf .
          Your message has been successfully submitted and would be delivered to recipients shortly.