About HP10544A
Hi Bruce
A more efficient buffer like a white emitter follower using 2 x npn + 1
x pnp would be best.
If you want I can post such a circuit using BC546 + BC556 or similar.
Yes if you have that kind of schematic I'd like to test it. It won't
take much time to do breadboard testing. I have BC547's and BC557's
ready in the shelf, all E12 resistor values are available but capasitors
and coils are not so easy; of course any standard values like 100n, 10n,
22...33p etc. are available at a moment. Coils are most difficult
however LCR meter is available but I don't have any cores that might be
necessary to make coils for this purpose?
This buffer needs to be located with 5cm or less of the OCXO connector
unless you use capacitance cancellation techniques.
Ok.. there was a "bottom PCB" used all my previous tests, which itself
has about 5 cm. trace before coaxial connector. I don't know where this
PCB is originally used. It included 12V regulator and some filtering.
But maybe i could get the signal directly from output pins without using
a coax at all, with just a pair of very short wires and take the OCXO as
close as breadboard as possible.
First you need to decide what output level you want:
+7dBm? +13dBm?
Many instuments seems to want as much as 1V input for external
reference. That seems odd because the level is so high that it will
cause a 10 MHz spurious peak easily. So I think the +13 dBm should be
enough. There will be some kind of distribution amplifier in final
design and I have no idea about it's gain (not selected any yet). But
let's assume it will have gain of 1:1.
The how far do you need to suppress the 2nd and 3rd harmonics?
There's no specification for these. But if we think about the
application which is to clean the LPRO's output signal it would be nice
to have at least same kind of harmonics performance that LRPO itself
has. It has some kind of lower grade crystal inside but I have no idea
maybe there are some filtering because the spectrum is fine:
http://www.amigazone.fi/files/gpsdo/LPRO-3.PNG
Some spurious on 50 MHz and below 150 MHz still present.
You need to set some actual specs.
I think good spec is that it has to clean the LPRO's output, not get it
worse. I know it's main purpose is rubidium noise cleaning, which is not
seen in these spectrum measurements at all and I even don't have the
cleaning loop yet.
I'm trying to pass on step by step so first thing is to find out how to
generate a good 10 MHz signal and next thing will be that (digital?)
phaselock needed to lock that on rubidium at suitable timing constant.
Then there will be another stage to control LPRO's C-field with GPS,
some diagnostics, power supplies, distribution amplifier etc... Lot's of
work to do. Infact I had no idea how hard it is to get just a clean 10
Mhz...
Attenuating them by 20dB is easy if you need to suppress them much
further it gets more difficult.
Yes that's the point I'm afraid with 10544.
Don't use a high Q bandpass filter as its phase shift tempco will be
relatively high.
What do you think if I could find FM radio IF filter somewhere (10.7
MHz) and tune it to 10 MHz? I don't know if there's tunable filters
available anymore because any new ones seems to be ceramic. But maybe
from some old radio could have tunable one (wishful thinking in Finland,
thanks to "recycling" of the electronics).
--
73s!
Esa
OH4KJU
Esa Heikkinen wrote:
Hi Bruce
A more efficient buffer like a white emitter follower using 2 x npn + 1
x pnp would be best.
If you want I can post such a circuit using BC546 + BC556 or similar.
Yes if you have that kind of schematic I'd like to test it. It won't
take much time to do breadboard testing. I have BC547's and BC557's
ready in the shelf, all E12 resistor values are available but capasitors
and coils are not so easy; of course any standard values like 100n, 10n,
22...33p etc. are available at a moment. Coils are most difficult
however LCR meter is available but I don't have any cores that might be
necessary to make coils for this purpose?
I'll see if I can produce a schematic as a pdf file.
I presume you can also obtain red LEDs?
Can you obtain parts from Farnell or RS Components?
This buffer needs to be located with 5cm or less of the OCXO connector
unless you use capacitance cancellation techniques.
Ok.. there was a "bottom PCB" used all my previous tests, which itself
has about 5 cm. trace before coaxial connector. I don't know where this
PCB is originally used. It included 12V regulator and some filtering.
But maybe i could get the signal directly from output pins without using
a coax at all, with just a pair of very short wires and take the OCXO as
close as breadboard as possible.
That PCB or equivalent is typically used in counters that use the 10544A.
That would simplify things considerably as you shouldn't need to tune
out any capacitance.
First you need to decide what output level you want:
+7dBm? +13dBm?
Many instuments seems to want as much as 1V input for external
reference. That seems odd because the level is so high that it will
cause a 10 MHz spurious peak easily. So I think the +13 dBm should be
enough. There will be some kind of distribution amplifier in final
design and I have no idea about it's gain (not selected any yet). But
let's assume it will have gain of 1:1.
The how far do you need to suppress the 2nd and 3rd harmonics?
There's no specification for these. But if we think about the
application which is to clean the LPRO's output signal it would be nice
to have at least same kind of harmonics performance that LRPO itself
has. It has some kind of lower grade crystal inside but I have no idea
maybe there are some filtering because the spectrum is fine:
http://www.amigazone.fi/files/gpsdo/LPRO-3.PNG
Some spurious on 50 MHz and below 150 MHz still present.
You need to set some actual specs.
I think good spec is that it has to clean the LPRO's output, not get it
worse. I know it's main purpose is rubidium noise cleaning, which is not
seen in these spectrum measurements at all and I even don't have the
cleaning loop yet.
Can easily reduce the distortion down to 50dB below the carrier.
80dB below the carrier is possible but requires a lot of effort and
attention to detail.
I'm trying to pass on step by step so first thing is to find out how to
generate a good 10 MHz signal and next thing will be that (digital?)
phaselock needed to lock that on rubidium at suitable timing constant.
Then there will be another stage to control LPRO's C-field with GPS,
some diagnostics, power supplies, distribution amplifier etc... Lot's of
work to do. Infact I had no idea how hard it is to get just a clean 10
Mhz...
Attenuating them by 20dB is easy if you need to suppress them much
further it gets more difficult.
Yes that's the point I'm afraid with 10544.
Don't use a high Q bandpass filter as its phase shift tempco will be
relatively high.
What do you think if I could find FM radio IF filter somewhere (10.7
MHz) and tune it to 10 MHz? I don't know if there's tunable filters
available anymore because any new ones seems to be ceramic. But maybe
from some old radio could have tunable one (wishful thinking in Finland,
thanks to "recycling" of the electronics).
Using a 10MHz bandpass filter isn't usually a good idea unless the
temperature doesn't change too rapidly and the filtered output is phase
locked to the stable frequency source.
The phase shift tempco of the bandpass filter then isnt too critical as
slow changes in the filter phase shift will be corrected by the phase
lock loop.
Even so you don't want too high a Q or the filtered signal amplitude
will change too much over temperature.
Bruce
I'll see if I can produce a schematic as a pdf file.
I presume you can also obtain red LEDs?
LEDs are available ready in shelf...
Can you obtain parts from Farnell or RS Components?
Yes I can. One design goal for the whole system is that all parts should
be easily availale somewhere like Farnell, ELFA etc. distributors
selling small amounts.
--
73s!
Esa
OH4KJU
Esa Heikkinen wrote:
I'll see if I can produce a schematic as a pdf file.
I presume you can also obtain red LEDs?
LEDs are available ready in shelf...
Can you obtain parts from Farnell or RS Components?
Yes I can. One design goal for the whole system is that all parts should
be easily availale somewhere like Farnell, ELFA etc. distributors
selling small amounts.
Esa
White emitter follower circuit attached.
The LED ensures a temperature stable collector current for the npn
common emitter device
The 47 ohms in series with the output ensures stability with a
capacitive load.
Distortion of this circuit is lower than that of the 10544A.
Some work is required - it should be OK but you will need to check that
the circuit doesn't oscillate at VHF - can only go so far with spice.
Bruce
White emitter follower circuit attached.
Ok and thanks, Bruce!
I'll test that later. Now it's family time.. :)
Distortion of this circuit is lower than that of the 10544A.
Ok it would then give the actual output spectrum of HP10544. Would you
do the test with or without the 1000 ohm. termination before C1? The
datasheet recommends the 1k termination but how with this kind of amplifier?
--
73s!
Esa
OH4KJU
-----Original Message-----
From: time-nuts-bounces@febo.com
[mailto:time-nuts-bounces@febo.com] On Behalf Of Bruce Griffiths
Sent: Saturday, March 07, 2009 8:48 PM
To: Discussion of precise time and frequency measurement
Subject: Re: [time-nuts] About HP10544A
Scott Newell wrote:
At 07:41 PM 3/7/2009, Tom Van Baak wrote:
I know late model 5245L counters used the 10544A oscillator.
Perhaps others in the group can provide other examples and
then you
can head off to the online manual sites and grab the info you need
from the PDF files.
My 5345A has a 10544A.
Scott
The 10MHz buffer is merely a cascade of 3 ECL line receivers.
Bruce
Bruce,
How much isolation can on expect to get from that kind of buffer?
Didier
Don't use a high Q bandpass filter as its phase shift tempco will be
relatively high.
What do you think if I could find FM radio IF filter somewhere (10.7
MHz) and tune it to 10 MHz? I don't know if there's tunable filters
available anymore because any new ones seems to be ceramic. But maybe
from some old radio could have tunable one (wishful thinking in Finland,
thanks to "recycling" of the electronics).
Using a 10MHz bandpass filter isn't usually a good idea unless the
temperature doesn't change too rapidly and the filtered output is phase
locked to the stable frequency source.
The phase shift tempco of the bandpass filter then isnt too critical as
slow changes in the filter phase shift will be corrected by the phase
lock loop.
Even so you don't want too high a Q or the filtered signal amplitude
will change too much over temperature.
You should be able to lower the effective Q value by use of external
serial or parallel resistor. A parallel resistor will add some
capacitance, but detuning the frequency should also require some
additional capacitance so it should work fine in combination.
Cheers,
Magnus
Esa Heikkinen wrote:
White emitter follower circuit attached.
Ok and thanks, Bruce!
I'll test that later. Now it's family time.. :)
Distortion of this circuit is lower than that of the 10544A.
Ok it would then give the actual output spectrum of HP10544. Would you
do the test with or without the 1000 ohm. termination before C1? The
datasheet recommends the 1k termination but how with this kind of amplifier?
Esa
The emitter follower output stage of the 10544A should be fine driving
the higher input impedance of this buffer without any external resistor
to bring the load down to 1k.
N.B., since the 10544A has an output coupling capacitor the input
coupling cap shown on the amplifier schematic isn't required.
Bruce
Didier wrote:
-----Original Message-----
From: time-nuts-bounces@febo.com
[mailto:time-nuts-bounces@febo.com] On Behalf Of Bruce Griffiths
Sent: Saturday, March 07, 2009 8:48 PM
To: Discussion of precise time and frequency measurement
Subject: Re: [time-nuts] About HP10544A
Scott Newell wrote:
At 07:41 PM 3/7/2009, Tom Van Baak wrote:
I know late model 5245L counters used the 10544A oscillator.
Perhaps others in the group can provide other examples and
then you
can head off to the online manual sites and grab the info you need
from the PDF files.
My 5345A has a 10544A.
Scott
The 10MHz buffer is merely a cascade of 3 ECL line receivers.
Bruce
Bruce,
How much isolation can on expect to get from that kind of buffer?
Didier
Didier
The 5345A buffer is actually a CC stage feeding a CB stage with a CE
(with series feedback) output stage.
Total reverse isolation will probably be somewhere in the vicinity of
100dB or so.
However coupling via the bias circuitry will reduce this somewhat.
Cascading 3 ECL line receivers (longtailed pair plus output emitter
follower buffers) in a single package as in the 5370A should achieve a
reverse isolation of 60dB or so per stage, but package parasitics and
coupling through the internal bias regulator etc will limit this to a
somewhat lower value.
Bruce
This question is directed at Bruce, but if anyone else has a contribution,
feel free to speak.
Bruce,
What is the best way to measure the reverse isolation of an amplifier
(particularly a buffer amplifier for a 10 MHz reference), when it is
expected to be in the order of 100 dB or more?
Feeding the output with a known signal and measuring at the input with a
spectrum analyzer comes to mind, but I am sure there must be something wrong
with that technique, it sounds too simple.
The presence of a signal at the input (or not) may affect the operating
point of the amplifier, so measuring from output to input without such
signal may not give a true result.
Didier