[ARC5] What's going on in this small-signal RF amplifier?

[Leslie Smith]

Hi Bob:

  You wrote: <A few comments:>
Me:  I appreciate your reply!!

  <1. My preferred 'spice' program is LTSpice (now being supported by
Analog Devices). However it only runs (apparently) on Windows or Apple
devices. Oh, well.>
Me:  Useful info.  I saw the name 'LT' Spice and assumed it may be
GNU-related, but now I know.  I run a version of Puppy Linux, based on
Ubuntu.  I'm sure some-one on the Puppy Forum will point me in the right
direction.

  <2. Your 'load' (I assume that is a ferrite transformer with 5:1 turns
ratio) sort of indicates the RF load is about 220 ohms * 25 or about 5K ohm>
Me:  Yes, FT37-43 core.  After that, I follow your logic.  I wound an
alternate transformer.  Same FT37-43 core.  5t in the collector circuit,
link to load: 2t.  I did this because I believe the load Z was too high as
you said.  Result when lower Z load was in circuit:  Less distortion on the
collector (bec of lower inductance) but not the 'nice' sine I'm looking
for.   But still (apparent) cut-off.

More comment:  I'm waiting for you to tell me, "You're crazy!  Why have you
got the 2200 ohm resistor in the base circuit?"  Ah, yes.  I observed the
base voltage when collector signal distored.   I calculated the effective
parallel AC load of the 4k7+10k bias resistor(s).  The 2k2 resistor is
meant to attenuate the signal to the point where I see no distortion.
Close, but not quite there yet.  Need a little more attenuation.  But the
sketch has the words, "As built", and what you see is what I have.  The 2k2
is in place, and it's shown on the sketch.

I anticipate your next Q.  "Surely it would be easier to remove turns from
the input link!"  Answer:  I did that, but at that point I had only 1 turn
on a toroid.  Nothing to hold the link in place.  So I added one more
turn.  That's why I have 2 turns.  I haven't show the oscillator.  It's a
two-device negative resistance oscillator that delivers a sine wave over a
few hundreds of kHz.  Good amplitude stability.

You:  <That seems high - the device is set for about 5 mA quiescent (i.e.
Vb - Vbe or about 4 v - 0.7v or about 3.3 V (quite close to what you see)
divided by 620'ish ohms) so theoretically you have a collector voltage
swing downwards of about 12V - Ve or about 7V and then you would clip.>
Me:  That's how I figure it too.  Also, I need to say that the device is
S9018.   I have a bottle full of them.  Not high gain, but good F(t).

You:  <And your results indicate that you are clipping some place - the
base will swing up but it won't swing down or maybe it runs out of steam
when it swings down - the swing down on the base also coincides with
collector going up.  At this point, I am going to go off and ponder this a
bit while the turkey is cooking.>
Me:  I'm going to sit tight while the turkey is cooking.  My thinking
followed yours until I wound a transformer with 5+2 turns.  Of course, if
you suggest I'll wind another.  Say 3t+3t.  I have a handful of TF37-43
cores I bought because ... we'll ya'know I'm just a little stoopid.  Stuff
I build always has some problem and it's just handy to have a handful of
cores to wind alternate coils ...

In conclusion:  I can't figure out what's going on.  I can easily replace
the transformer with a resistor.  I could even solder in a series of
resistors staring with a ridiculously low value (e.g. 220 or 330 ohms).
That would give a gain of 220/56 or 330/56 approximately.  Generally I
avoid resistive loads in RF circuits because the signal on the collector is
assymetric.

Thanks for your comment.  I have no idea what's going on here and this is
an excellent chance to learn something.    Final final musing.  Maybe I
should feed the beast with +18 or +24 volts (just to see what happens).  On
second thoughts:  Just looked at the spec sheet.  I think I should replace
the 9018 with a 2N2222 or BC547.  At this frequency - the '547 will work.
Generally I use the 9018 on 6m.  I'll do that later today.

Leslie
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.qth.net/pipermail/arc5/attachments/20221125/b3c8a126/attachment-0001.html>