Hi

The spectrum you show in the middle is the correct one. It shows the 10 MHz output and the sub-harmonics related to the 5 MHz crystal oscillator. The other two are a bit broken. One appears to have a stage oscillating. The other is more interesting. It has a stage oscillating that is injection locked to the 5 MHz crystal oscillator. I suspect more than one capacitor went bad in these OCXO's. Somebody may have sold Morion a bad reel of bypass caps.

Bob

On Feb 25, 2012, at 4:01 PM, Roberto Barrios rbarrioss@msn.com wrote:

MHz

locked to

in

Interesting.  The MV89As are very stable, but they have a lot of AM noise
and they don't meet their PN specs (about -150 dBc/Hz at 10 kHz rather than
-155).  I've tested a few and they all look the same in that respect.
Sounds like replacing all of the SMT caps may be a good move.

-- john

Hi,
I have two of Morion MV89A, one is prominently marked (RoHS 5).
I assume that means lead-less solder. Pure tin solder is much more
prone to
crystallised joints which break later.
Are these the dodgy units ?

cheers, Neville Michie

On 26/02/2012, at 11:02 AM, John Miles wrote:

Hi

Are the specs at 5 MHz or at 10? They should be 6 db worse at 10 than at 5.

Bob

On Feb 25, 2012, at 7:02 PM, "John Miles" jmiles@pop.net wrote:

Hi

Hit send to soon...

The AM noise could easily be another bypass issue.

Bob

On Feb 25, 2012, at 7:02 PM, "John Miles" jmiles@pop.net wrote:

Now that I'm looking back at the data sheet posted with the auction, they
were indeed specified at 5 MHz.  But even after the 6 dB penalty, their spec
limits versus their real-world numbers are pretty marginal.

Morion actually announced this particular oscillator at IFCS 2002, so
there's a paper of sorts:
http://www.ke5fx.com/Morion_MV89A_IEEE_IFCS_2002.pdf .  Figure 4 was taken
from a 10 MHz part, but all three of the ones I've tested look like the
attached.  Same as figure 4 until about 200 Hz... then they never go any
lower.

-- john

Hi

The paper mentions HC-40 crystals. They probably have the crystal drive current cranked down to optimize short term stability at 1 sec and longer. The phase noise floor takes a back seat.

The other interesting thing in the paper is that the short term numbers shown aren't into parts in the 13th.

Bob

On Feb 25, 2012, at 9:11 PM, "John Miles" jmiles@pop.net wrote:

On 02/26/2012 05:20 PM, Bob Camp wrote:

Come to think of it, I never saw a good explanation as to why the
crystal drive current really has this effect. I can only seem to recall
hand-waving and experience show stuff. Could be I read it but forgot
about it as I don't design crystal oscillators, only tend to use the
once that others have designed and built for me.

Cheers,
Magnus

Is this the paper you are speaking of?
Google:  Optimization of Drive-Level in High Stability Low-Noise OCXOs

Thomas Knox

The ADEV plot in figure 3 doesn't exactly show the oscillator in its best
light.  The right side of the 'bathtub' is well beyond 1000 seconds with
these parts, if you actually let them warm up before you measure them.

So the person who generated the plot either didn't allow sufficient warmup
time, or they used a low-quality reference.  The Ch A versus Ch B amplitudes
are almost 3 dB apart, so they probably were not using another MV89A.

-- john

That's not the one Bob (and I) was talking about, but there are some
interesting contradictions between the paper you mentioned:

http://www.magicxtal.com/articles/Optimization%20of%20Drive-Level%20in%20Hig
h%20Stability%20Low-Noise%20OCXOs.PDF

and Rakon's article:

http://www.rakon.com/Products/Public%20Documents/Whitepapers/PHASE%20NOISE%2
0IN%20CRYSTAL%20OSCILLATORS.pdf

The Rakon author states that the flicker and white PM regions, including the
broadband floor, are influenced by factors other than the crystal (Figure
9).

Meanwhile, the Russian authors of the first paper state repeatedly that the
broadband noise floor is a function of the crystal current, and that
close-in noise isn't affected by crystal current.

It's obvious enough that high levels of crystal current are bad for
long-term stability and aging, but it's not as easy to see why high crystal
current in itself should rule out a good broadband noise floor.  There are a
lot of assertions in the literature, especially in the magicxtal.com paper,
that are backed up by measurements but not by solid analysis.

-- john

I wonder how Wenzel looks at drive level, they seem to have their products dialed in.

Thomas Knox

Hi

Most people put the best of the batch in the paper. It is indeed odd that the very experienced authors of the paper did not do so. With low drive a 5 MHz HC-40 should get into parts in the thirteenth over much of the bathtub region.

Bob

On Feb 26, 2012, at 3:22 PM, "John Miles" jmiles@pop.net wrote:

Hi Bob,

I wonder what the other oscillators that others are receiving look like in
the SA. Replacing all the capacitors would take a serious effort, not sure
if it would be justified.

Can I ask how can you know all that by looking at the spectrum plot?

Thank you,
Roberto EB4EQA

-----Original Message-----
From: time-nuts-bounces@febo.com [mailto:time-nuts-bounces@febo.com] On
Behalf Of Bob Camp
Sent: sábado, 25 de febrero de 2012 23:55
To: Discussion of precise time and frequency measurement
Subject: Re: [time-nuts] Morion MV89A Repair

Hi

The spectrum you show in the middle is the correct one. It shows the 10 MHz
output and the sub-harmonics related to the 5 MHz crystal oscillator. The
other two are a bit broken. One appears to have a stage oscillating. The
other is more interesting. It has a stage oscillating that is injection
locked to the 5 MHz crystal oscillator. I suspect more than one capacitor
went bad in these OCXO's. Somebody may have sold Morion a bad reel of bypass
caps.

Bob

On Feb 25, 2012, at 4:01 PM, Roberto Barrios rbarrioss@msn.com wrote:

reasons for this?

time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.