[R-1051] Frequency Standard: Voltage Regulator History Research

[David Wise]

I'm aware of five different Frequency Standards.
Here they are, with salient features, in the order I think they were designed.
I'd like to hear comments, corrections, addtions, disagreement, whatever.

Reference designators are in the context of NAVELEX 0967-LP-970-9010.

1. Early R-1051 plain (Navships 94841A).  58189 (GD), P/N 666231-006, 6625-00-078-4718
    Slug-tuned coils.
    Oscillator runs off +28V, with a shunt regulator.
    Oscillator/rectifier oven control, running off +28V, with a shunt regulator for the critical parts.
    COMP-INT-EXT mode switch.  No 5MHz output in INT mode.
    All discrete transistors.

2. R-1051B (Navships 0967-970-9010).
    Fixed coils with trimmer capacitors.
    Oscillator runs off +20V, with a 10V shunt regulator.  470 ohm series resistor, 21mA.
    DC amplifier oven control, running off +10V and +20V except 28V heater.
    COMP-INT-EXT mode switch.  No 5MHz output in INT mode.
    All discrete transistors.
    When the main MODE is STBY or tuning motors are operating,
    the oven is powered down!

3. Late R-1051 plain (NAVELEX 0976-LP-970-9010).
    Fixed coils with trimmer capacitors.
    +17V preregulator, running off +28V.  ("+27V" is a typo.)
    Oscillator runs off +17V, with a 10V shunt regulator.  330 ohm series resistor, 21mA.
    DC amplifier oven control, running off +10V and +17V except 28V heater.
    COMP-INT-EXT mode switch.  No 5MHz output in INT mode.
    All discrete transistors.
    Thanks to the preregulator, this type does not exhibit Type 2's tune/standby glitch,
    and line regulation of oven temperature is better than Type 4.

4. My R-1051 plain.  06845 (Bendix - Communications Div - bought by Raytheon), P/N 4013399-0701
    Fixed coils with trimmer capacitors.
    Oscillator runs off +28V, with a 10V shunt regulator.  820 ohm series resistor (22mA)
    which burns the board because it's mounted too close.
    DC amplifier oven control, running off +10V and +28V.
    COMP/INT-EXT(NORMAL)-EXT(OVEN STBY) mode switch.  5MHz output available
    in all modes but still interrupted during STBY and tuning.
    All discrete transistors.
    Since the oven controller does not use +20V, this type does not exhibit Type 2's
    tune/standby glitch, but oven temperature varies slightly with mains voltage due to
    lack of preregulation on R24/R23.

5. R-1051H. 98738 (Stewart-Warner), P/N 01A228785-01.
    Fixed coils with trimmer capacitors.
    +15V preregulator, running off +28V.
    Oscillator runs off +15V, with a 10V shunt regulator.  270 ohm series resistor, 19mA.
    DC amplifier oven control, running off +10V and +15V except 28V heater.
    COMP/INT-EXT(NORMAL)-EXT(OVEN STBY) mode switch.  5MHz output available
    in all modes but still interrupted during STBY and tuning.
    Automatic fallback - if external standard fails, automatically switch to internal
    oscillator and turn on oven if not already on.
    TTL IC's.
    Thanks to the preregulator, this type does not exhibit Type 2's tune/standby glitch,
    and line regulation of oven temperature is better than Type 4.

When the standard was redesigned for the R-1051B, a huge mistake was made.
Although it was cost-efficient to run the whole module (except the 28V heater)
off regulated +20V, this rail is not powered all the time.
As a result, if the receiver is in Standby mode, or even while it's just busy moving
to a new MC band, the oven is cut off, causing the receiver to be off frequency,
maybe a lot, when it resumes operation.

The +17V (mismarked as "+27V") preregulator in NAVELEX 0976-LP-970-9010
was added to eliminate this glitch.  Since +28V is available full-time in all modes
except OFF, so is +17V.

The schematic is badly misdrawn.
Here's a corrected description.  +28V enters the A2A5 module through J1 pin 3.
>From there it goes to oven heater A2A5A3R1, and to preregulator assembly A4.
In A4, it goes to Q1 collector and the high side of R1.  R1 and CR1 form an 18V
shunt regulator, with C2 bypassing RF noise.  With its base elevated to 18V, Q1
outputs a stable potential of about 17V on its emitter.  This rail supplies power to
oscillator/oven control assembly A2A5A1, via J2 pin 4.  Inside A1, the +17V rail
powers driver transistor Q7 via R24, and the 10V shunt regulator via R9.
The shunt regulator powers everything else in A1.  As in the other Frequency
Standard types, switched, regulated +20V enters A2A5 through J1 pin 1, to power
multiplier/divider assembly A2.

Although NAVELEX 0976-LP-970-9010 is for the R-1051 plain, it has a later date
than Navships 0976-970-9010.  I believe that after Navships 0976-970-9010, they
decided to have a separate manual for each model instead of one manual with
confusing overlays of addenda.  To this end, they took Navships 0976-970-9010,
stripped out all the B-specific info, and called the result NAVELEX 0976-LP-970-9010.
Then they took Navships 0976-970-9010, removed the plain-specific info, and called it
NAVELEX 0976-LP-427-4010.  (I can't confirm this since I don't have this manual.)
During this time frame, the frequency standard was redesigned to add the preregulator.
Since it appears in the plain-model's NAVELEX, I believe it was assembled into
late-production plains.  I could be wrong about this historical detail, but it doesn't
matter to our radios, which have whatever was on the shelf last time it was serviced.

My own radio came with what I call Type 4.  It doesn't have a preregulator like
Type 3, but it still avoids the oven glitch, by powering driver transistor Q7
from +28V instead of +20V.  Unfortunately, this introduces a small problem
with line regulation, because via feedback resistor R23, rising or falling mains
voltage biases the controller to a higher or lower temperature setpoint.

Type 4 introduces a different change, in the mode switch.  All standards
have a three-position switch, but in the older ones the positions are COMP,
INT, and EXT.  The oven runs all the time, and to monitor the internal oscillator
you must switch it to COMP.  There are only three differences between INT and
COMP.  First, the lamp driver is only powered in COMP.  Second, the output
amplifier is only powered in COMP and EXT.  Third, the output amplifier's
input is grounded in INT.  Maybe they were trying to avoid interference between
local-crystal 5MHz emissions and the external source, but either it wasn't the
problem they thought it was, or the Type 4/5's ability to power down the oven
was deemed a worthwhile trade.

In Type 4 and the R-1051H's Type 5, the positions are
COMP/INT, EXT(NORMAL), and EXT(OVEN STBY).
In COMP/INT, local 5MHz is available at J24 on the rear panel.
(In the EXT positions, J24 reflects external 5MHz as before.)
In EXT(OVEN STBY), the oven is on, and in EXT(NORMAL), it's off,
saving up to 10W of power and heat, plus extending component life,
if the receiver is running off an external standard as it usually did.

Because Types 4 and 5 output local 5MHz when synced to the
internal crystal, they can theoretically act as an external standard for
another receiver, but it's not for serious use, because it's interrupted
during tuning and front panel STBY mode, even in the R-1051H.  I think
it was just for alignment.

Regards,
Dave Wise
SWL in Hillsboro, Oregon