[Milsurplus] BC-9: NETTED! WoooHOOO!

[Mike Morrow]

Dave wrote:

>The BC-9 is Netted!  WooHoo!
>The solution was indeed to insert a *real*
>RF Choke in the DC shunt feed to the Osc.
>Plate.  C6 is no longer needed (nor has
>any effect) and the external RF chokes 
>on the 120VDC line are eliminated.  
>Here's the revised diagram:
>  https://goo.gl/photos/y6mJsdF9RrYDZzJN9
>All you folks working on replicas, take note.
>
>I'm still impressed with how stable this rig
>is and how little chirp.  Of course, the DC/DC
>converters provide excellent voltage regulation
>at such low load levels.

Very good results indeed!  I figured things would settle down once the stuff connected to the capacitor end of the filter was prevented from affecting the RF on the oscillator plate.  But I thought you'd get a larger frequency offset key open vs. key closed, just due to the gross plate voltage change.

>The receiver and transmitter net to within
>just a couple of  hundred cycles from one
>end of the band to the other.

That works OK for ham purposes.  But in a multi-station net, ideally every station would listen on about the same frequency, and when any one station transmitted, its RF tube frequency would automatically offset presumably about 1000 Hz for an enjoyable note in all the other receivers.  The few hundred hertz offset you have now sounds too small for service use.

I believe that the filter was specifically and deliberately designed to allow circuit changes on its capacitor end to affect in a controlled manner the oscillation frequency differently for key open vs. key closed, creating the desired offset (set by C6) on key closed.  I'm sure the designers knew about RF chokes in addition to low pass filter circuits and would not have chosen the latter if the former would have functioned satisfactorily.  But ALL of that sophisticated original design likely worked properly ONLY with the original loop, tubes, and batteries operating near the original 4600 kHz design frequency.

It surprises me how little the frequency changed when the RF tube's plate voltage went from receive (70 vdc??) to transmit (120 vdc).  How much does the frequency change with, say 90 vdc instead of 120 vdc ... simulating battery discharge?

>And this is consistant across tube types.
>I tested 01A, Triode-connnected 6AQ5
>and VT-25A.  All behaved similarlly.
>Note that, with the loop, the 6AQ5 is
>not the best tube.  It does something 
>at high bias levels that causes the audio
>amps to motor-boat.  Don't know what.
>The VT-25A is the best for both TX and RX.
>And that's with only 6.5 V on the filament
>of a 7.5V tube.  Tight fit but it goes.
>Figures; Can't get a $600 tube, so
>the best one is just $125, LOL.
>
>I will be tinkering with the rig on the big
>copper loop tonight around 3555 KC +/-.
>If you hear a little signal with a little
>wobble, let me know, HI.
>Discovered that the radio might even work
>on 40 meters.  Already pushed it to 6 MC
>and it worked.  Another time.

What tubes are you using in the two AF stages?  It would be neat to make a version that used 1.5 or 3 volt directly-heated cathode tubes in all three sockets.

I wonder what performance change would occur if an RFC were placed in series with the filter that is in the RF tube bias supply.

You have had success and fun proving that the simple 95-year-old BC-9-A functioned usefully and surprisingly well.  Congratulations!  (This would be good material for an article in Electric Radio...seriously!)

Mike / KK5F