[R-390] Comments on ER Issue 208 AGC modification

[Charles Steinmetz]

David identified a hopelessly muddled (and, thereby, mistaken) paragraph 
in my original post and offered a correction with specific instructions 
for rewiring the AGC switch (S107) to accomplish what he correctly 
inferred that I intended.  Responding to David's post, Larry offered an 
alternative and somewhat easier re-wiring proposal.  My abundant thanks 
to both of them.

David and Larry both managed to understand what I meant.  The key point 
is to rewire the AGC switch to choose among three, grounded capacitors 
to prevent the large AGC voltage step on switching.  This is easy to do, 
and, as David and Larry show, there are several ways to organize it.  [I 
have not traced the switch connections for either of their proposed 
fixes, but they are both careful workers so I'm confident their 
proposals will work as advertised.  I also have not compared their 
suggestions to my notes ("They're around here somewhere..."), so I'm not 
sure which, if either, is exactly what I do.]

In my original "Comments," I did not intend to provide step-by-step 
instructions for curing the "moment of silence."  In this vein, I am now 
drafting a revised, PDF version, and will replace the muddled paragraph 
with the following:

"CURING THE MOMENT OF SILENCE:  To cure this -- and only this -- you can 
very easily modify the AGC circuit by removing the capacitance 
multiplier tube (V506A) from the circuit and re-wiring the switch to 
select three plastic film capacitors in sequence, each of them grounded. 
  To produce the same AGC release behavior as a stock R390A, one would 
use 0.1uF, 2uF, and 20uF for Fast, Medium, and Slow AGC, respectively. 
[In my view, the "Slow AGC" capacitor should really be much larger, from 
50 to 100uF -- I frequently use 56 or 68uF as a practical compromise.]"

While I did not originally intend to provide step-by-step instructions 
for curing the "moment of silence," I see no reason not to include 
David's and Larry's contributions in the PDF version with appropriate 
attributions and thanks, if I have their permission (please let me know 
offline).

One further note:  While the "cure" prevents the massive DC transient on 
the AGC line and the associated "moment of silence," the "Medium AGC" 
and "Slow AGC" capacitors can still be left with residual charge (and, 
therefore, non-zero voltage) when switching to one of the other 
capacitors.  It is good engineering practice to include bleed resistors 
from each of these two capacitors to the AGC buss, so that the voltages 
on the capacitors are maintained near the current AGC voltage and even 
these much smaller voltage transients on the AGC line are minimized. 
The bleed resistors must be very large, with time constants of a minute 
or so.

Larry asks if I know who originated the "two diode" AGC mod.  I do not 
know who thought of it or did it first.  It seemed obvious to me, and I 
have been doing it myself since the early 1960s.  I mentioned it 
subsequently to other R390A tinkerers and found that some of them had 
also thought of it on their own.  By the mid-1960s, quite a few folks 
were doing it.

Best regards,

Charles


Larry wrote:

> Dave, Good catch on Charles' unclear fix to the 'moment of silence'.  I
> like his idea, though.
>
> Your proposed implementation of it will certainly work, but I think the
> following procedure is a little easier:
> 1. Remove wire from switch S107 term 8 and securely insulate the end of the
> wire.
> 2. Remove both wires off of C551 and connect them together. This provides
> direct path from V506A grid to S107 wiper.
> 3. Remove ground from S107 term 9 and connect 2 mfd cap to it and other end
> of cap to ground.
> 4. Connect 20 mfd cap to S107 term 8 and other end of cap to ground.
> Note: If an IF deck is installed that does not have the C551 bypass on it,
> it will not hurt anything, although the AGC switch will not work correctly.
>
> What do you think?
>
> Regards, Larry


David wrote:

>> Charles, I liked your article.  High bit rate.
>>
>> One nitpick.  Regarding the Moment of Silence, you write
>>
>> ++++++++++++++
>> To cure this -- and only this -- you can very easily modify the AGC
>> circuit by adding a 20uF film capacitor to ground and permanently
>> grounding the end of C551 that is away from C548, so that the AGC switch
>> adds C551 in parallel with C548 for Medium AGC and the new 20uF film
>> capacitor in parallel with C548 for Slow AGC.
>> ++++++++++++++
>>
>> You must be misremembering, because grounding the end of C551 away
>> from C548 would give you MED speed when FAST is selected.  I suppose
>> you actually did this (since all C551's are leaky nowadays):
>>
>> 1. Disconnect and abandon the wires at S107 terminals 7 and 8;
>> 2. Move ground wire from S107 terminal 9 to terminal 7;
>> 3. Add new 2uF or 2.2uF C1102 between S107 terminal 9 and S102 terminal 3
>> or 5;
>> 4. Add new 20uF or 22uF C1103 between S107 terminal 8 and S102 terminal 3
>> or 5.
>>
>> This does away with the capacitance multiplier, and grounds C1102 for MED
>> and C1103 for SLOW.
>> (C551 is decommissioned.)  C1102 and C1103 only see AGC voltage, not B+ .
>>
>> This is the simplest approach, and it's cheap too - Mouser has a 22uF 63V
>> film cap for $5 and it's about 1" x 1.5" x 0.5", an easy fit on the front
>> panel.
>> Even though I'm put off by the brute-force aspect, I say that your approach
>> is superior to mine, because it doesn't require a new S107.
>>
>> But Collins did not have small cheap 20uF film caps in 1950, so my approach
>> ("Ending the Moment of Silence", which goes the opposite way by multiplying
>> a small cap for MED) is still historically interesting as a way they could
>> have
>> done it had they put in the effort to do it right.
>>
>> Regards,
>> Dave Wise
>>