Re: [time-nuts] GPSDO won't lock: OCXO aged out?
At 01:27 PM 1/12/2021, paul swed wrote:
Granted at 5V the varicaps may be at minimum. Use a 2K R. That should be
pretty safe.
Current limited bench supply though 4.7 k and an inline DMM to
monitor current. Made quick-n-dirty measurements of EFC voltage, EFC
current, and output frequency.
It looks promising. Dead-bug op-amp, gain of 2, powered by
unregulated input supply (12-15 VDC)? That should get me enough EFC
range for a few years, assuming the loop can cope with double the EFC
sensitivity.
1 V 4 uA 9.999 991 MHz
2 V 9 uA 9.999 992 MHz
3 V 15 uA 9.999 994 MHa
4 V 27 uA 9.999 995 MHz
5 V 41 uA 9.999 996 MHz
6 V 56 uA 9.999 997 MHz
7 V 70 uA 9.999 998 MHz
8 V 85 uA 9.999 999 MHz
9 V 100 uA 10.000 001 MHz
10 V 115 uA 10.000 001 MHz
11 V 129 uA 10.000 002 MHz
12 V 144 uA 10.000 003 MHz
13 V 159 uA 10.000 004 MHz
14 V 173 uA 10.000 005 MHz
15 V 188 uA 10.000 006 MHz
16 V 203 uA 10.000 006 MHz
17 V 217 uA 10.000 007 MHz
18 V 232 uA 10.000 007 MHz
19 V 247 uA 10.000 008 MHz
20 V 261 uA 10.000 008 MHz
--
newell N5TNL
Thats fantastic to see. Yes you have it. Use an opamp to insert an offset
of 6-7V so the DAC itself still uses its normal range of 0-5V. Opamp output
would be 6-11V or something. You just may have the unit for a while longer.
Regards
Paul
WB8TSL
On Tue, Jan 12, 2021 at 8:19 PM Scott Newell newell+timenuts@n5tnl.com
wrote:
At 01:27 PM 1/12/2021, paul swed wrote:
Granted at 5V the varicaps may be at minimum. Use a 2K R. That should be
pretty safe.
Current limited bench supply though 4.7 k and an inline DMM to
monitor current. Made quick-n-dirty measurements of EFC voltage, EFC
current, and output frequency.
It looks promising. Dead-bug op-amp, gain of 2, powered by
unregulated input supply (12-15 VDC)? That should get me enough EFC
range for a few years, assuming the loop can cope with double the EFC
sensitivity.
1 V 4 uA 9.999 991 MHz
2 V 9 uA 9.999 992 MHz
3 V 15 uA 9.999 994 MHa
4 V 27 uA 9.999 995 MHz
5 V 41 uA 9.999 996 MHz
6 V 56 uA 9.999 997 MHz
7 V 70 uA 9.999 998 MHz
8 V 85 uA 9.999 999 MHz
9 V 100 uA 10.000 001 MHz
10 V 115 uA 10.000 001 MHz
11 V 129 uA 10.000 002 MHz
12 V 144 uA 10.000 003 MHz
13 V 159 uA 10.000 004 MHz
14 V 173 uA 10.000 005 MHz
15 V 188 uA 10.000 006 MHz
16 V 203 uA 10.000 006 MHz
17 V 217 uA 10.000 007 MHz
18 V 232 uA 10.000 007 MHz
19 V 247 uA 10.000 008 MHz
20 V 261 uA 10.000 008 MHz
--
newell N5TNL
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
Scott, it looks like you have already proven the tune voltage can go
much higher than specified, and sufficient to get in-range for 10 MHz
ideal. The currents noted show there's probably a pull-down resistor on
the tune line - maybe 75-100 k. The reverse leakage of the varicap diode
should be in the nA range cold, and even hot at oven temperature, I'd
think only up to a few uA or so. You can get a better idea of what's in
there by doing the same test immediately after cold power-up.
You can avoid any issues about doubling the EFC sensitivity by adding a
fixed DC offset instead of higher gain. The DC reference added should of
course be stable and clean, and so should the opamp, the associated
resistors, and its power supply, so you won't have to worry about PSRR.
Ed
One thing I forgot to mention, is that the EFC sensitivity isn't
necessarily the same in this new range as it was before. In other words,
having more gain may not be a big deal, or even beneficial. You're way
past the steeper dC/dV slope anyway, so the actual effect of the varicap
on the net frequency may be much less in the 5-20 Vt region than in the
original 0-5V range. So, it may be OK or even better with more gain. You
can estimate the new tuning sensitivity from your experiments, and see
how it shapes up compared to the likely original characteristic.
Restoring it to "original" would then involve proper gain and offset
adjustments to land close enough.
In my experiments with Si junctions (not just varicaps), I've seen that
once you get past a few volts reverse bias, the slope changes
dramatically (flatter), so passing the 5 V range as are doing could make
a big difference, and more gain may be OK. But, don't overdo it - most
varicap diodes I've seen will avalanche break down (not destructive if
current limited) around 30-50 V.
Ed
Quick update: tried using an AD780 reference + op-amp to offset the
EFC voltage. Worked great, until I slid everything back into the
enclosure and it was tuning the OCXO too high!
After several disassembly/mod/reassembly passes (all tuning too high
once boxed up), I punted and went back to the simple gain stage. It's
been locked at TFOM 2 all day, but it does seem very sensitive to air
currents and temperature changes. I'm starting to think the oven
setpoint may be wrong, too. (How could it lock 1 month ago with EFC
<5V, but now it needs 8.5V to be on frequency when it's outside the
enclosure? Something's fishy there.)
Check out the white OSC trace in the LH screenshot--that can't be good!
Hoping to limp by until I can find either a drop-in 5V OCXO or time
to adapt a 12V OCXO into the box.
thanks!
newell N5TNL
Hi
One of the options is that the EFC has an issue.
Another option is that the oven control circuit has a problem.
Either one could put the center frequency someplace odd. If it’s the heater control
(or heater set point) then indeed, the temperature sensitivity will go way up.
Most of the OCXO’s off these boards are a sub $20 sort of item on eBay. How good
are they? Who knows ….
Bob
On Jan 15, 2021, at 8:18 PM, Scott Newell newell+timenuts@n5tnl.com wrote:
Quick update: tried using an AD780 reference + op-amp to offset the EFC voltage. Worked great, until I slid everything back into the enclosure and it was tuning the OCXO too high!
After several disassembly/mod/reassembly passes (all tuning too high once boxed up), I punted and went back to the simple gain stage. It's been locked at TFOM 2 all day, but it does seem very sensitive to air currents and temperature changes. I'm starting to think the oven setpoint may be wrong, too. (How could it lock 1 month ago with EFC <5V, but now it needs 8.5V to be on frequency when it's outside the enclosure? Something's fishy there.)
Check out the white OSC trace in the LH screenshot--that can't be good!
Hoping to limp by until I can find either a drop-in 5V OCXO or time to adapt a 12V OCXO into the box.
thanks!
newell N5TNL<uccm_gain_2_11h.gif>_______________________________________________
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
Not sure how hard it might be to get a thermocouple into the oven. That
would give you a hint.
Regards
Paul
On Fri, Jan 15, 2021 at 8:51 PM Bob kb8tq kb8tq@n1k.org wrote:
Hi
One of the options is that the EFC has an issue.
Another option is that the oven control circuit has a problem.
Either one could put the center frequency someplace odd. If it’s the
heater control
(or heater set point) then indeed, the temperature sensitivity will go way
up.
Most of the OCXO’s off these boards are a sub $20 sort of item on eBay.
How good
are they? Who knows ….
Bob
On Jan 15, 2021, at 8:18 PM, Scott Newell newell+timenuts@n5tnl.com
wrote:
Quick update: tried using an AD780 reference + op-amp to offset the EFC
voltage. Worked great, until I slid everything back into the enclosure and
it was tuning the OCXO too high!
After several disassembly/mod/reassembly passes (all tuning too high
once boxed up), I punted and went back to the simple gain stage. It's been
locked at TFOM 2 all day, but it does seem very sensitive to air currents
and temperature changes. I'm starting to think the oven setpoint may be
wrong, too. (How could it lock 1 month ago with EFC <5V, but now it needs
8.5V to be on frequency when it's outside the enclosure? Something's fishy
there.)
Check out the white OSC trace in the LH screenshot--that can't be good!
Hoping to limp by until I can find either a drop-in 5V OCXO or time to
adapt a 12V OCXO into the box.
thanks!
newell
N5TNL<uccm_gain_2_11h.gif>_______________________________________________
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to
and follow the instructions there.
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
At 07:49 PM 1/15/2021, Bob kb8tq wrote:
then indeed, the temperature sensitivity will go
way up. Most of the OCXOâs off these boards
are a sub $20 sort of item on eBay. How good are they? Who
Any suggestions for a SC-cut 5V OCXO? Searching
for "Trimble 73090" or "73090 OCXO" just brings
up the ~$80 GPSDO boards. Nearly everything 5V
I'm seeing is square wave, but maybe that doesn't
matter in this box? (Or maybe the 73090 is square
wave; I'll check next time I have it open.)
--
newell N5TNL
If the gpsdo output pps signal does not drift, but the gps pps signal drifts slowly, it is bound that after a period of time, the gpsdo pps signal will have a very large beat in order to follow the gps. How to solve this problem? In tens of nanosconds, I see that the pps of the gps are always drifting.
| |
李
|
|
邮箱bd4kc@126.com
|
签名由 网易邮箱大师 定制
On 01/16/2021 12:46, Scott Newell wrote:
At 07:49 PM 1/15/2021, Bob kb8tq wrote:
then indeed, the temperature sensitivity will go
way up. Most of the OCXOâs off these boards
are a sub $20 sort of item on eBay. How good are they? Who
Any suggestions for a SC-cut 5V OCXO? Searching
for "Trimble 73090" or "73090 OCXO" just brings
up the ~$80 GPSDO boards. Nearly everything 5V
I'm seeing is square wave, but maybe that doesn't
matter in this box? (Or maybe the 73090 is square
wave; I'll check next time I have it open.)
--
newell N5TNL
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
Hi
Simple answer is to do a general search on eBay for “OCXO” then take a
look at the pictures. The descriptions often are a bit off ….
Bob
On Jan 15, 2021, at 11:46 PM, Scott Newell newell+timenuts@n5tnl.com wrote:
At 07:49 PM 1/15/2021, Bob kb8tq wrote:
then indeed, the temperature sensitivity will go way up. Most of the OCXOâs off these boards are a sub $20 sort of item on eBay. How good are they? Who
Any suggestions for a SC-cut 5V OCXO? Searching for "Trimble 73090" or "73090 OCXO" just brings up the ~$80 GPSDO boards. Nearly everything 5V I'm seeing is square wave, but maybe that doesn't matter in this box? (Or maybe the 73090 is square wave; I'll check next time I have it open.)
--
newell N5TNL
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.