[time-nuts] Vectron Type 229 Crystal Oscillator

C. Turner turner at ussc.com
Mon Jun 25 20:54:34 UTC 2012


FWIW,  I've re-crystalled a number of those 100 MHz-area '229 and 
similar Vectron oscillators and have had pretty good results and I 
thought that I'd make a comment or two.

The '229's that I've seen that operate in the 100 MHz area use TO-5 
mount crystals operating at 1/2 the output frequency (e.g. about 50 MHz) 
and it turns out that International Crystal does, indeed, have the 
formula in their files.  The only thing that really needs to be 
specified is the operating temperature which I've always seen written on 
a small paper label attached to the crystal.  As it turns out, there's a 
potentiometer inside that will adjust the oven operating temp so that 
could, in theory, be matched to the turnover temperature of the 
crystal.  The output of the oscillator has an active doubler and I've 
been able to tune them from 96 to 108 MHz just by compressing/expanding 
the air-core inductors achieving an output in the +5 to +10 dBm range.

Before attempting to open the case, take ohmmeter readings between all 
of the bottom pins and ground:  Heating the bottom plate for removal one 
can sometimes unsolder some of the components and this will provide a 
quick check to verify that things are OK after the plate is removed.  As 
noted in a previous email, there's a simple Zener-type regulator for the 
tuning voltage reference and a few bypass capacitors - nothing 
particularly critical, but it is very easy to mess this wiring up.  
Fortunately, all of the wiring that I've seen is in PTFE insulation so 
nothing melts!

The trickiest part by far is getting the can open with minimum damage, 
but the way to do this is to make a small jig (or use the original 
mounting bracket - removing rubber grommets temporarily) that attaches 
to the four screws so that one can pull up evenly on the "bottom" 
(facing upwards) with the body of the oscillator firmly clamped in a 
vise with wood.  Before heating, remove the tuning screw to avoid 
pressure buildup (and the subsequent "pop" when the solder seal lets go 
burps hot metal) and it's recommended that one use an awl or screwdriver 
and make marks on the bottom plate and side so that they may be matched 
up again upon reassembly.

The trickiest part is to patiently use a propane torch (set on low) or a 
powerful head gun and evenly apply heat on all sides, always keeping the 
flame moving:  Having more than one torch/heat gun (and a helper or two) 
is always useful at this point.  Eventually, one will be able to gently 
rock the bottom plate and it will lift up and off.  The foam inside is 
always brittle and is likely to break - but just be careful and save the 
pieces, noting their orientation.  Once the bottom is off it may be 
necessary, using a long screwdriver, to carefully push out the "guts" 
through the oscillator adjustment hole.  At this point, it would be a 
good idea to make a quick drawing and note which-color wire goes where 
on the connector on the base of the unit as they tend to break off after 
not too many flexes as well as how the wires were routed through the 
foam and to the base.  This would also be the time to re-check the 
resistance-to-ground readings of the connections at the feed-through and 
carefully inspect the circuity wired to it and re-solder as necessary 
and then powering the unit up again to verify that it still works as before.

The foam is held together using what looks like polyester tape inside is 
a metal enclosure containing the circuitry which is often stuck to the 
foam, sometimes requiring a bit of persuasion and/or a thin knife to 
un-stick it.  On the units that I've seen, nubs of the circuit board 
protrude from at least one end of the small metal box and blobs of RTV 
hold this together, so removing this and a small amount of prying will 
disassemble the case, revealing the circuit board.  When this is done, 
pay close attention to a small, round piece of paper (which may or may 
not be stuck to the TO-5 crystal) and on it will be the temperature for 
which the original crystal was cut:  Specify this temperature (or 
something close to it) when ordering a new crystal.  Also, as soon as 
one opens things up, draw/note which wire goes where on the board in 
case one of these breaks off, too...

The first one of these that I re-worked I didn't pay much attention and 
simply guessed that the crystal formula was a 7th OT, series resonant, 
and when I got the crystal and installed it I was getting a weak output 
from the unit on completely the wrong frequency and it was then that I'd 
realized that there was a frequency doubler: Re-working the oscillator 
for the correct overtone saved the day (and my not having to order 
another $50 crystal!) and when I've ordered since, I've made it a point 
to work with ICM and find their already-existing formulae and have since 
avoided this problem! Somewhere in ICM's computer is my version of the 
formula which is the standard '229 formula but with slightly lesser 
specs (accuracy, stability) to save a few dollars.

In replacing the crystal, I simply made note of the original lead dress 
and installed the new one in a like manner, pushed the case back 
together and then RTV'd (using the "Silicone II" formula - the stuff 
that does NOT smell of vinegar) the nubs of circuit board to hold it in 
place.  The next day I put it back in the foam, carefully pushed it back 
into the case - verifying that the tuning access lined up with the hole 
- and the press-fitting the bottom back in. When reassembling, I've 
usually not bothered re-taping the foam as the wires and foam have 
pretty much "set" into the desired shape - and I didn't have any of that 
polyester tape, anyway:  I would guess that polyimide ("Kapton") tape 
with a silicone-based adhesive would work as well if it were necessary.  
Almost inevitably, one will break the "base" piece of foam and I simply 
fit the two pieces together as I'm reassembling.

At this point I apply power and walk away for an hour or so, coming back 
later and noting that the current has dropped with the oven's coming to 
temperature.  I've then attached a 10-20 turn, 10k-20k pot across the 
tuning lines (ground, tuning V+ which is in the 5-7 volt area and the 
tuning line itself), setting it to mid-rotation and then adjusting the 
mechanical tuning for the intended frequency.  I then let it "burn in" 
for a day or two so that I'm satisfied that it will work as it is 
supposed to (and seems pretty stable) before resoldering using a propane 
torch on low heat.

Of these units, two have gone in 10 GHz beacons.  The first of these, 
for the K7RJ beacon, used my re-worked oscillator (e.g. the 
"on-frequency" oscillator instead of the half-frequency) which feeds a 
microwave "brick" while the second beacon uses the modified ICM formula 
with the intended half-frequency crystal with the "brick" feeding a 
modified DB6NT beacon module - both beacons using the tuning line for 
FSCW keying.  Both of these '229's have been found to be pretty good 
performers and in the first several weeks the 10 GHz frequency drifted 
10-15 kHz and then gradually slowed down as the ovens and their 
components baked and the still "green" crystal aged a bit, etc. and 
during this time, as long as it was powered-up and warm, I left the 
tuning access screw off to allow moisture and volatiles (mostly from the 
RTV) to escape.  By the time a year of beacon service had passed, the 
drift/aging rate of the oscillators leveled off and they are staying 
within a few hundred Hz of their intended frequency over a wide range of 
ambient temperatures despite the fact that they are single-oven. Now, 
they seem to be moving only a kHz or less in long-term drift over the 
course of a year.

FWIW, a video describing some of the "guts" of the K7RJ beacon may be 
found here:  http://www.ka7oei.com/microwave.html

In theory, one could tweak the oven to match the turnover temperature of 
the crystal and get even better stability, but I've never done that - 
I've simply ordered a crystal for as close to the marked temperature of 
the original as I could.

I've also taken apart some Vectron '225's and they seem to be very 
different internally as they would appear to multiply a 5-ish MHz 
glass-case crystal up to the 100 MHz-range frequency.  I've not 
re-worked one of those and have not checked with ICM to see if they have 
their crystal formula on hand.

73,

Clint
KA7OEI





More information about the Time-nuts_lists.febo.com mailing list