Re: AGC circuit to try?

Gordon Gibby

​Allison, what on earth do you do for a living, and for WHOM???


From: <> on behalf of ajparent1 <kb1gmx@...>
Sent: Saturday, April 7, 2018 5:29 PM
Subject: Re: [BITX20] AGC circuit to try?

Rare is an order for just one part here.  Typical order for my lab are 4-20 parts and quantity.
Stock devices like 1n4007 a hundred in a order is common, same for 2n2222A, 2n3906, 2n3904
as they are cheaper when purchased in volume.  I do have the BAP64 in various packages and
several others as well. Even some exotic Avago PIN parts.  I maintain a stock that allows me
to build then optimize.  I get annoyed when I don't have something unless exotic like 10ghz
gain blocks.

When I write for open audiences I tend to use parts I know they can get or easily and effortlessly 
substitute and not require shipping for 5 pieces. 

The junction capacitance is about 15-25PF at -4V depending on vendor. Vishay says 15pf and
Motorola says 20.  For the application that's small enough to not be a factor.  Reason is rare
will it be at zero bias as it is being use in the forward direction.  However, if bias is zero, the
capacitance measured for a few 1n4007s is around 25 pf and compared to the nominal
bypass for that stage it is larger than the emitter resistance used and far higher than the
gain set resistor that was there.  At 30mhz thats a capacitive reactance of around 200
ohms and the resistor was only 10, so gain reduction is still very significant.  

Regarding  "That should probably be R33, not R35"  depends on the version of the schematic
you have.   It is annoying but it is what it is.  For the latest dump of the manual 33 would be it.
When you follow a design for over 12 years to accumulate all the flavors and variants, big disks
are cheap so I keep information going back to the flood.  The Bitx folder is small only 43MB
and history is interesting.  

One of my radios has a twin-Tee attenuator in the front end as image reject DC receivers 
have few places at RF to put variable attenuation in.  It works very well in that place but might be
harder to insert into the (u)BITX board. 

It allows the BITX ( all variants) to have RF agc and on a trial version of a bitx20 about 10 years ago
I have used it in the TX path to allow ALC for power compression to allow increased talk power.
The agc signal in that case was rectified RF out.   IT has another useful characteristic in that any
time the SWR is unacceptable the RF at that point will climb and clamp the TX power.  Its not a 
true SWR protection but it helps.

Its a really neat design for those that like to mod.

FYI that opamp AGC mod was published by W4OP and also a Y0 a while ago as applicable to 
KNQ7A, KN10, and the X1M all of which use a MC1350 for the RX IF and require a 4 to 8V signal
for AGC where 4V is max gain and 8V is min (about -50DB) gain.  It would take a transistor
and a few resistors to invert the signal and apply it to the diodes.  An opamp can do that too
and a quad device would be convenient.  


Join to automatically receive all group messages.