[time-nuts] 5061A Insufficient Cs Beam Current Modulation (long)

[Michael Ulbrich]

Dear time-nuts,

I've had the chance to work on a 5061A over the last couple of days. I
guess it's a fairly old unit with "singing" oven controller and 105 5
MHz crystal oven assembly. Last digits from serial number are "0964" -
not sure if that indicates the date of mfg. It definitely has been
tampered with before as evident from some missing screws, a module
cover  and pulled coaxial interconnects. All considered it's in a quite
miserable state, but anyway - I couldn't resist taking the opportunity
and started fiddling with the unit.

Basic checks revealed a standard unit (no options) with a standard beam
tube 05061-6077 and a complete set of modules. Probability that it's not
the original CBT is quite high since the screws on the hold-down straps
were loose and the oven controller taps for transformer T4 were not
wired according to the tube label (that I actually found out much
later). Uhm, and - most obvious - the -2500V connection did not fit
since both plug and chassis socket are of the same - female - type ...
but a carefully placed clip lead actually worked ok here.

The owner had told me that the unit must have been stored for years,
maybe even more than a decade ...? So I was positively surprised when I
found out that it pumped down the cesium tube within a day and the unit
stays on since with the Cs oven enabled and an ion current about 2 ticks
from zero. OSC oven and Cs oven meter readings come down within 30
minutes from powering up and stay within reasonable limits. 5 MHz
reading is present and rises until crystal oven has reached its
operating temperature.

Initially I could get only a faint reading in the MULT position of the
circuit check switch. Frequencies and levels from the Multiplier A3 and
Synthesizer A1 are ok although levels seem a tad on the low side. 137 Hz
side bands of 90 MHz are visible on an 89441 vector analyzer if MOD is
switched on. As a final resort I started to tweak the pots in the A4
harmonic generator and got the MULT reading up to about 20 on the meter
scale. That seemed sufficient for the moment.

Next step "BEAM I" - again no deflection on the meter scale. Using an
external known standard I set the internal crystal OSC close to 5 MHz
and (literally) grabbed the magnifying glass to watch the meter needle
while tuning the coarse frequency adjust - and there it finally was: a
tiny movement showing a minuscule peak of about one tick deflection with
barely noticeable adjacent peaks of half that size! Hahaa ... here we
go! It took quite some time (and manual reading) to find the "Beam I
Meter" adjust accessible from the front panel. With that I was able to
bring up the resonant peak meter reading to 35 or so, with clearly
visible valleys and lower peaks on both sides of the main resonance.

Now on to "OPER" and closing the loop with 137 Hz modulation on. No
deflection in position 2nd harmonic. The manual led me to AC amplifier
A7 which was thoroughly checked and found without fault. No 2nd harmonic
at test output A7J2 and no fundamental at A7J6. Switching A7
amplification to HI and cranking up loop gain does help in a certain but
not satisfactory way. Now I get a 2nd harmonic reading in the
20-somethings but this obviously is just noise since the meter reading
isn't steady and it does not depend on modulation being on or off. A7J6
shows a narrow-band filtered signal with strong components around 137 Hz
but these also are mostly of random nature as excited by the amplified
CBT and amplifier noise. Moving off the main resonant peak (MOD off,
Loop open) at 250 by +/-50 to either 200 or 300 on the 10e-10 frequency
dial does not change a lot. Only if the 10e-10 dial is fully turned CW
or CCW using the full range of +/-250 the 137 Hz component at A7J6
becomes stronger but still on a very noisy background.

Funny thing is that the unit actually "tracks" (I wouldn't call it
"locked", though ...) the changes on the 10e-10 dial and develops a
control voltage for steering the main osc. Even Alarm goes off and the
Continuous Operation light comes on and (mostly) stays on, but because
of the cranked up loop gain it e.g. doesn't detect if modulation is
switched off, since the noisy narrow band signals (137 + 274 Hz)
obviously are somehow able to trick the Logic Assembly A14 to see a
"locked" operating condition.

Now (finally, thanks a lot for your patience ...) onto my question:
Isn't it so that a CBT with clearly discernible resonant peaks and
valleys should mandatorily deliver a sufficiently modulated beam
current? - provided input signals to and harmonic generator A4 itself
are within specs. Or might there be a condition internal to the tube
which will show "normal" resonant behavior (not strong but sufficient
IMHO) and at the same time prevent the beam current from being
modulated? My impression is, that the tube is not overly noisy, since
the outputs A7J2 and A7J6 do not show excessive noise from the tube but
only from gain fully cranked up. But maybe I'm wrong here?!

I did check the beam current "peak to valley" ratio according to the
manual and found with a 10 megohm DC voltmeter from A7J1 to ground: 66
mV peak and 22 mV nearest valley. The peak value should correspond to a
peak current of .66e-8 A which is bit lower than the .8e-8 A given as an
end of life indication, but OTOH the manual says the tube may well work
beyond these EOL specs albeit with a somewhat degraded performance.

I also did a complete realignment of the A4 harmonic generator according
to the manual instructions. All alignment steps could be carried out
with the expected results and finally a beam current near saturation
(plateau with dip) could be established.

My apologies for bothering the list with a such a long post, but since
being a mostly lurking member for some years now I learned that nearly
nothing could be more on-topic than bringing an old Cs beam frequency
standard back to life ... ;-)

Thanks in advance ... Michael U.