| |
G8MNY > TECH 25.12.13 20:47l 127 Lines 6783 Bytes #999 (0) @ WW
BID : 49457_GB7CIP
Read: GUEST
Subj: Off Air Lock for Ref Osc.
Path: IZ3LSV<I0OJJ<GB7CIP
Sent: 131225/1513Z @:GB7CIP.#32.GBR.EU #:49457 [Caterham Surrey GBR] $:49457_GB
From: G8MNY@GB7CIP.#32.GBR.EU
To : TECH@WW
By G8MNY (Updated Apr 13)
(8 Bit ASCII graphics use code page 437 or 850, Terminal Font)
Here was my old method when we had analoge TV of locking the master osc in a
PLL signal generator to 625 line TV syncs. (BBC Ceasium ref was better than
1/10^9 off air), obtained from TV or video Rx with video output. This was much
better method than other off air sources as there is less fading, no AM or PSK
on carier data etc.
NB. Digital TV sources are NOT accurate as they only use a local xtal timebase.
In the UK only the BBC have accurate syncs, CH4 is quite accurate, ITV is poor,
& five is no use as it's Tx are fed from a digital sat feed!
MY DESIGN
This design will work with any 1MHz clock that has (or can have) a varicap to
trim its frequency. So it can easily be used on a frequency counter if a 1MHz
clock can be derived or extracted. It will greatly improve the accuracy &
temperature drift & aging problems of simple Xtal timebases.
FEATURES
Automatic switching to passive free run mode with no applied Video.
Reasonable immunity to TV sync noise & ghosts etc.
Slow PLL Lock up has a "speed up" circuit.
THE CIRCUIT
220u + +12V SYNC PRESENT SWITCHED 12V
ÚÄ´ÃÄÄÄÄÁÂÄÄÄÂÄÄ¿ ÚÄÄÄÂÄÄÄÄÄÄÂÄÄÄÄÄÄÂÄÄÄÄÄÄ¿ +12V
_³_ ³ ³ ³e ³ ³ ³ ³ D3_³_ ³
³ 22k \ /T2 ³ PRESET _³_ \_/ 220k
T1 ³ ³ ÄÂÄ ³ 4k7 /_\ ³ SPEED ³
OFF AIR u1 ³/e ÃÄÄÄÄÁÄ´ÃÄÄ´ ³ ³D1 10K UP ³ Control
VIDEO >´ÃÄÂÄ´ 100k + 2u ³8 1k ³ ³ PHASE Úij>³ÄÄ´ To Master
INPUT ³ ³ ³\ ³ 15p ÚÄÄÄÄÄ¿3 ³ ³ TPoÄ´ COMP ³ D4ÃÄÄÂÄ> Osc
³ ³ ÃÄÄÂÁÄ´ÃÄ¿ ³ ³ÄÄ)Ä220kÄ´ ÃÄÄÂÄ47kÄ´ ³ ³ Varicap
³ ³ ³ ³ ³ ³ 555 ³ ³ ³ ³/ ³ Ãij<³ÄÄ´ ³
75ê ³ ³ ³ 220k ³ 7³ÄÄ´ ÃÄÄÂÄ´T3 === ³ D5³ ³+
³ ³ ³ === ÃÄÄijCV ³ 12k _³_ ³ ³\e ³22n ÃÄÄ1MÄÄ´ === Tant
³ 1M 10k ³1n5 ³ 5³ 2,6³ÄÄ´ D2\_/ ³ ³ ³ +³ ³ ³100uF
TERM / ³ ³ ³ === ÀÄÄÄÄÄÙ === ³ 56k ³ ³ === 220k ³ FILTER
SW ³ ³ ³ ³ ³1n ³1 3n3³ 3k* ³ ³ ³ Tant³2u ³ ³
ÚÄÁÄÄÄÁÄÄÄÁÄÄÁÄÄÄÄÄÁÄÄÄÄÁÄÄÄÄÄÄÄÁÄÄÄÄÄÄÁÄÄ)ÄÄÄÁÄÄÁÄÄÄÄÄÁÄÄÄÄÄÄÁÄÄÁÄ
_³_ SYNC AIR LOCKED ³ 10:1 ATTEN
STRIPPER 15,625Hz 15,625 ³
Hz ³
e.g. 64uS ³ 5ÚÄÄÄ+5VÄÄÄ¿5
T1-2 any small PNP HFE 100, ³ ÚÄÄÄÄÄÄ¿1 ÚÄÄÄÄÄÄ¿1 TTL
T3 NPN HFE 100. Àij74LS93³ÄÄij74LS93³ÄÄÄ<1MHz
D1 is a germanium/Schottky. 11ÀÄÄÄÄÄÄÙ 11ÀÄÄÄÄÄÄÙ
D2-5 are common 1N4148. ³2,3,10 ³2,3,10
ÚÁÄÄÄÄÄÄÄÄÄÄÁÄ
_³_ DIVIDE BY 64
HOW IT WORKS
The input can be terminated or not onto 75R. T1 strips the syncs of the video,
which are filtered almost to a saw tooth with the 1n5. The average DC level of
T1 collector changes from high to low with the presents of video & this turns
on slow acting T2 to power up & enable the rest of the circuit.
The 555 15,625Hz (64uS) astable osc is fed with only a few mV of the recovered
syncs on its CV pin 5, this is an unusual way of controlling a 555 & is used to
make the astable lock up only over a narrow range, so it free wheels over noisy
& missing syncs e.g. during the frame sync pulse.
The 555 output is compared in T3 with the local 1MHz clock after that has been
divided by 64 (8x8 to give 64uS) in 2x 74LS93. Diode D2 & 'select on test' 3k*
are to temperature compensate T3 so there is no change in phase sensitivity.
D1 & D3 isolate the varicap control output when there is no video.
T3 output is lightly filtered on its collector (TP) so that the phase errors
can be seen on an oscilloscope for setting up. The final varicap control line
is attenuated & heavily filtered to remove all traces of TV variables & clock
pulse etc. This would cause long lock up times of several minutes, so an out of
lock speed up diodes bypass the filter & attenuator circuit when the control
voltage is more than 600mV out.
On applying video, the two 15,625Hz phases will be well out of phase & the
control voltage will be > ñ 0.5V off the normal varicap voltage, & diodes D4/5
will conduct & bypass the 10:1 attenuator/filter, speeding up the lock. When in
lock the attenuator insures only a small amount of heavily filtered control is
applied.
The master oscillator varicap feed has 2 220k Rs across the 12V, (this could be
a single 100k from 5V rail) so that when there is no video the osc free runs at
the normal frequency. Some accurate osc modules have internal thermistors for
thermal tracking the Xtal with a varicap, this circuit is able to drive that
load without the need for these Rs.
TESTING
Lock the scope to incoming video "line". Scope the filtered output syncs on T1
collector. Check that T2 is turned on with video being present, & check that it
turns off & the varicap voltage returns to the normal setting with no video.
Initially set up the 555 astable preset so that pin 3 output square wave is in
line lock with the video.
Check the divider outputs are OK & produce a 15,625Hz square wave. Using a high
Z probe (1/10) on T3 collector, check that there is a level somewhere between
the power rails (it will have some line & frame sync content on it), this DC
level indicates lock up. Now adjust the scope timebase to see a 1/25 Sec & show
2 video frame syncs, & accurately set the 555 preset for minimum disturbance
during both odd & even frame syncs.
Use your osc/counter for some accurate work, e.g. listen 70cms SSB carrier (osc
harmonic) or is a counter measure 70cms frequency. Remove video & adjust your
master osc's trimmer for the same frequency. This should leave the non-locked
accuracy as near to the off air as possible, but still with drift etc.
LOCK INDICATION
Not shown here, but I use a couple of transistors to sample the two 15,625Hz
phases again, but with slight CR windowing so that they do not turn on when the
unit is properly locked. This then has a long time constant to stop indicating
lock, until everything has settled down. This feeds a tri-state LED transistor
switch circuit to indicate NORMAL (OFF) TV SYNCS (RED) & TV LOCK (GREEN).
See also my "Calibrating Frequency", "198kHz Off Air Standard", "Comparing Off
Air Freq Standards", "Crystal Drift Compensation" & "Simple Crystal Oven" buls.
Why Don't U send an interesting bul?
73 de G8MNY @ GB7CIP
Read previous mail | Read next mail
| |
