[GPS_Standard] [Re: Xtal Oven coming up to temperature from cold?(Dick)

[Bert, VE2ZAZ]

Hi again Dick,

Your circuit would work, but I think you will be spending too much work on
this startup circuit. The best thing to do with an ovenized oscillator is to
leave it on all the time. I have a Un-interruptible Power Supply to provide
power in case of an outage.

As for the +5V regulators, I think you plan is a bit overkill. If you really
want to improve your stability, use a separate 78L05 for the PIC only. For a
0 to 5V tuning range, slow supply variations on the integrator will have no
effect on the output voltage. This is not quite true for a -5V to +5V range
using my circuit, but running the XO around 0V (center point) actually
minimizes the impact of voltage variations on the tuning voltage. Jacques,
VE2AZX demonstrates this on his website. http://www.geocities.com/ve2_azx .

You should definitely run the GPS receiver on a separate 7805 regulator
though, as it may draw up to 200mA. As for the meter, I don't see a problem
as long as the current for full scale deflection much less than 1mA. You
would probably be better off using a digital display with a high input
impedance. You'll get a better resolution. These can be found on eBay for
~10$.

By the way, DAC mid-point is hex 0200, not 0020...probably just a typo from
your part.

Good luck!

Bert, VE2ZAZ

----- Original Message -----
From: "Dick" <dick at zs6ro.co.za>
To: <GPS_Standard at mailman.qth.net>
Sent: Saturday, December 16, 2006 12:27 PM
Subject: [GPS_Standard] [Re: Xtal Oven coming up to temperature from
cold?(Dick)


Thanks for the information Bert  - what you say makes sense (of course) and
I have already smacked myself across the side of the head - figuratively of
course <grin> .. I had read and re-read the docs on this project but I
obviously didn't digest the part about the time taken for the error-voltage
steps <sigh> ...

I figure then that if I just leave things, approximately 11 steps or so
would be wasted effort, then by that time the oven is at the correct
temperature and then the PIC will have to do another 11 odd steps in the
opposite direction before it declares that the frequency is close enough on
frequency to lock !! ... That's what - almost an hour of useless work ... So
after much thought I think I will take advantage of my 18 - 24 volts to the
oven and fit a comparator to take the reset pin of the PIC high as the
voltage approaches 23 volts ... I have a 5MHz xtal so I assume my 'one step'
would be 10x8 and not 10x16 ?? ... If I make 'S' double what you suggest for
10 MHz, should I then get the same 'count' with my 5 MHz xtal ?? .. Thus get
back some accuracy, taking into account that I have a high-stab unit !!

I think that this is even more elegant than my first choice of using a fixed
timer. Say I switched the supply off five minutes earlier and back on again,
 then the oven will still be warm and the 18 - 24 volts supply will be
closer to 24 volts, and should take less time to get to it's  'sweet spot',
thus less delay in things happening !!

Note that you had this problem Peter, I wonder what you actually discovered
? - could you possibly describe the symptoms ?.. What sort of end-to-end
frequency range do you have - 1Hz or 10Hz or possibly even more ??

>From further reading of the documents and findings on this project, I get
the distinct impression that there may be problems from having a single 5
volt supply feeding everything .. I intend to use a 78L05 for the PIC and
input IC's .. Another 78L05 just for the Integrator (the OPA chip) ..  A
further 5 volt regulator for the divider and output chips (10, 5, 1 etc MHz)
and another 5 volt regulator for the GPS receiver .. I figure it may be an
overkill but really these regulators cost very little including the various
associated capacitors for each regulator ...

Another idea I'm contemplating is to use a smallish edge type meter (the
scale is 4cm wide) .. I want to calibrate this so that full-scale will
indicate  5 volts  .. Then attach this to the error-voltage line .. If I
adjust things up correctly then midway on the meter should represent the
sweet spot of the moment !!! ...  As aging and temperature etc takes it's
toll on the system, I should see the meter varying according to the
error-voltage .. I have the parts - its not as if I have to buy anything ..
I wonder if I would have to use an op-amp to get a high impedance to not
load the OPA chips output ??  ..

Bert, I assume that the idea when calibrating the 0 - 5 volt error-voltage,
is that 2.5 volts is the obvious point to set up ?? .. (hex 0020 I think you
mentioned in the docs) ... That then gives me roughly a half cycle either
way for frequency compensation .. The oven I have is quite a high-stab one -
a Motorola model KXN1132AA 5MHz ... (Please Bert, don't let me smack my head
again - I'm starting to get a figuratively sore headache !!) ..


73, Dick ZS6RO
Johannesburg
South Africa
http://www.zs6ro.co.za



______________________________________________________________
GPS_Standard mailing list
Home: http://mailman.qth.net/mailman/listinfo/gps_standard
Help: http://mailman.qth.net/mmmain.htm
Post: mailto:GPS_Standard at mailman.qth.net