Best counter setting for ADEV?
Hi All,
thanks to John's superb free Timelab software, I tried ADEV for the
first time, after having been mostly interested in phase noise yet.
I learned that ADEV sensitivity is limited by the 500 ps resolution of
the counter that leads to a noise floor of 5E-10 at 1 sec, going down to
5E-11 at 100 sec etc.
Actually, the dynamic range allows only for meaningful results beyond a
few hundred seconds. Anything below 100 sec is simply below the system
noise floor.
Note that the measurement noise floor appears to be determined
exclusively by the resolution of the counter time interval mode, which,
for a HP 53131A is 500 ps.
A 53132A with 200 ps resolution should produce a noise floor of 2E-10 at
1 sec. Likewise, a 20 ps counter (5370A, SR620) should get you to 2E-11
at 1 sec.
Here is an example that shows what I mean: http://www.ke5fx.com/rb.htm.
Add limit lines to the diagram as discussed above to see what I mean.
Even a 5370A or SR620 would not be good enough to measure the 5065A
below 100 sec in that example, since anything below a straight line
through the points 2E-11 / 1 sec and 2E-14 / 1000 sec would be below the
measurement limit.
I was wondering why the resolution in TI mode is so much limited, since
I never had any problems measuring 5 or 10 MHz frequencies with up to 12
figures on that counter.
As a quic 'n dirty test, I tried to measure ADEV with my 53131A in
frequency mode using a gate time of 5.1 s for max resolution, and found
the noise floor had shifted down to 3...4E-12 at 10 sec, and going
further down to 5E-13 at 1000 sec from where it was equal to the TI mode
noise floor.
So, what is the best method to use for ADEV? What instruments and setups
are you using, and what works best for you? How to get the max out of
the given instruments?
Adrian
On 10/01/2012 10:31 PM, Adrian wrote:
Hi All,
thanks to John's superb free Timelab software, I tried ADEV for the
first time, after having been mostly interested in phase noise yet.
I learned that ADEV sensitivity is limited by the 500 ps resolution of
the counter that leads to a noise floor of 5E-10 at 1 sec, going down to
5E-11 at 100 sec etc.
Actually, the dynamic range allows only for meaningful results beyond a
few hundred seconds. Anything below 100 sec is simply below the system
noise floor.
Note that the measurement noise floor appears to be determined
exclusively by the resolution of the counter time interval mode, which,
for a HP 53131A is 500 ps.
A 53132A with 200 ps resolution should produce a noise floor of 2E-10 at
1 sec. Likewise, a 20 ps counter (5370A, SR620) should get you to 2E-11
at 1 sec.
Here is an example that shows what I mean: http://www.ke5fx.com/rb.htm.
Add limit lines to the diagram as discussed above to see what I mean.
Even a 5370A or SR620 would not be good enough to measure the 5065A
below 100 sec in that example, since anything below a straight line
through the points 2E-11 / 1 sec and 2E-14 / 1000 sec would be below the
measurement limit.
I was wondering why the resolution in TI mode is so much limited, since
I never had any problems measuring 5 or 10 MHz frequencies with up to 12
figures on that counter.
As a quic 'n dirty test, I tried to measure ADEV with my 53131A in
frequency mode using a gate time of 5.1 s for max resolution, and found
the noise floor had shifted down to 3...4E-12 at 10 sec, and going
further down to 5E-13 at 1000 sec from where it was equal to the TI mode
noise floor.
So, what is the best method to use for ADEV? What instruments and setups
are you using, and what works best for you? How to get the max out of
the given instruments?
Ahem, you should be careful with certain modes of the 53131/53132
counters, amongst others. As they do internal smoothing they do a form
of filtering which gives you artificially lower numbers than the actual
numbers. We have covered this several times and from many angles.
Stick to the TI mode for correct numbers.
Cheers,
Magnus
I was wondering why the resolution in TI mode is so much limited, since
I never had any problems measuring 5 or 10 MHz frequencies with up to 12
figures on that counter.
Limited? The question is not so much counting the figures but asking if the how much the figures count.
As a quic 'n dirty test, I tried to measure ADEV with my 53131A in
frequency mode using a gate time of 5.1 s for max resolution, and found
the noise floor had shifted down to 3...4E-12 at 10 sec, and going
further down to 5E-13 at 1000 sec from where it was equal to the TI mode
noise floor.
In frequency mode, the hp 53131A/53132A counters use a clever internal averaging mode. It's mentioned in the manual. The readings it reports are a highly oversampled mean frequency. This is nothing to complain about, really. Typically, with a frequency counter you are only interested in a smoothed averaged result. And these hp and other modern frequency counters do this quite well.
So, what is the best method to use for ADEV? What instruments and setups
are you using, and what works best for you? How to get the max out of
the given instruments?
An ADEV measurement is rather different from average frequency measurement. ADEV tries to tell you the variance, the deviations from the average frequency, as a function of tau. It's purpose is to measure the noise, not make the noise go away. So the more a frequency counter averages (in order to give you a smooth average frequency) the more it is actually suppressing the very variations that you are trying to measure. In order to gain "precision" these frequency counters are removing part of the "variance"; ADEV is completely ignores precision and is only concerned with the variance.
This is why one-shot phase meters or time interval counters give a more pure view of oscillator performance. Yes, they tend to report a little to a lot more noise -- but that's because there is more noise. An oversampling frequency counter takes the liberty to average away as much noise as it possibly can, to suppress the short-term variations, and present just a single value as the one true answer.
/tvb
As you've found, time interval counters by themselves can't measure
Allan Deviation to the levels required for today's precision oscillators.
Here are three google searches - either a general search or search the
Time Nuts archive at www.febo.com. They will get you started in the
topic of Allan Deviation measurement.
"single mixer" -"dual mixer" "allan deviation"
"Tight PLL"
DMTD oscillator
These describe the three main methods of measuring Allan Deviation. To
oversimplify, they increase the time-domain noise in a way that allows
measurements that don't require such small time interval resolutions
(i.e.. less than 1 picosecond). They each have strong and weak points.
The single mixer method is probably the easiest, but requires a
reference oscillator that is comparable in quality to the device under
test and is offset in frequency by typically 1-100 Hz.
The Tight PLL method was discussed a few years ago on this list but
hasn't really caught on much even though it appears to be as good as the
other methods.
The DMTD (Dual Mixer Time Difference) method is the standard for
comparing two oscillators that are running at the same frequency. It
also requires an offset frequency oscillator, but it doesn't have to be
as good as the devices under test because it's noise is cancelled out.
It has the most complicated hardware setup of the three methods.
All of these methods require extreme attention to detail to get top
performance.
Note: Once you start down this rabbit hole, there is no turning back.
You have been warned! :-)
Ed
On 10/1/2012 2:31 PM, Adrian wrote:
Hi All,
thanks to John's superb free Timelab software, I tried ADEV for the
first time, after having been mostly interested in phase noise yet.
I learned that ADEV sensitivity is limited by the 500 ps resolution of
the counter that leads to a noise floor of 5E-10 at 1 sec, going down
to 5E-11 at 100 sec etc.
Actually, the dynamic range allows only for meaningful results beyond
a few hundred seconds. Anything below 100 sec is simply below the
system noise floor.
Note that the measurement noise floor appears to be determined
exclusively by the resolution of the counter time interval mode,
which, for a HP 53131A is 500 ps.
A 53132A with 200 ps resolution should produce a noise floor of 2E-10
at 1 sec. Likewise, a 20 ps counter (5370A, SR620) should get you to
2E-11 at 1 sec.
Here is an example that shows what I mean: http://www.ke5fx.com/rb.htm.
Add limit lines to the diagram as discussed above to see what I mean.
Even a 5370A or SR620 would not be good enough to measure the 5065A
below 100 sec in that example, since anything below a straight line
through the points 2E-11 / 1 sec and 2E-14 / 1000 sec would be below
the measurement limit.
I was wondering why the resolution in TI mode is so much limited,
since I never had any problems measuring 5 or 10 MHz frequencies with
up to 12 figures on that counter.
As a quic 'n dirty test, I tried to measure ADEV with my 53131A in
frequency mode using a gate time of 5.1 s for max resolution, and
found the noise floor had shifted down to 3...4E-12 at 10 sec, and
going further down to 5E-13 at 1000 sec from where it was equal to the
TI mode noise floor.
So, what is the best method to use for ADEV? What instruments and
setups are you using, and what works best for you? How to get the max
out of the given instruments?
Adrian
On 10/01/2012 11:13 PM, Tom Van Baak wrote:
I was wondering why the resolution in TI mode is so much limited, since
I never had any problems measuring 5 or 10 MHz frequencies with up to 12
figures on that counter.
Limited? The question is not so much counting the figures but asking if
the how much the figures count.
As a quic 'n dirty test, I tried to measure ADEV with my 53131A in
frequency mode using a gate time of 5.1 s for max resolution, and found
the noise floor had shifted down to 3...4E-12 at 10 sec, and going
further down to 5E-13 at 1000 sec from where it was equal to the TI mode
noise floor.
In frequency mode, the hp 53131A/53132A counters use a clever internal
averaging mode. It's mentioned in the manual. The readings it reports are
a highly oversampled mean frequency. This is nothing to complain about,
really. Typically, with a frequency counter you are only interested in a
smoothed averaged result. And these hp and other modern frequency counters
do this quite well.
So, what is the best method to use for ADEV? What instruments and setups
are you using, and what works best for you? How to get the max out of
the given instruments?
An ADEV measurement is rather different from average frequency measurement.
ADEV tries to tell you the variance, the deviations from the average
frequency, as a function of tau. It's purpose is to measure the noise, not
make the noise go away.
True, but the measure is also there to show you how credible your
frequency measures are, so what you wrote above does not add up with
what you then later say.
Turns out that ADEV fills two purposes, both giving a reference curve
for oscillators and estimating the precision of it's frequency measure.
If you change your frequency measure, you will need another estimator
for the precision since it needs to include the averaging effect that
your frequency stability method applies. That number would however not
give you the proper ADEV curve.
It's tricky business.
The difference between these ADEV and the deviation of the estimator is
a bias function, which is usually expressed as a multiplicative bias
function for these measures.
So the more a frequency counter averages (in order
to give you a smooth average frequency) the more it is actually suppressing
the very variations that you are trying to measure. In order to gain
"precision" these frequency counters are removing part of the "variance";
ADEV is completely ignores precision and is only concerned with the variance.
Which is true when we talk about the AVAR/ADEV reference. You are
confusing things a little when you talk about ADEV and then say
variance, it's like saying AVAR and deviation.
This is why one-shot phase meters or time interval counters give a more pure
view of oscillator performance. Yes, they tend to report a little to a lot
more noise -- but that's because there is more noise. An oversampling
frequency counter takes the liberty to average away as much noise as it
possibly can, to suppress the short-term variations, and present just a
single value as the one true answer.
It's a matter of learning what the tool does for you and what it doesn't
do for you.
I rarely see the frequency measures with the ADEV next to it or even
better... under it.
Cheers,
Magnus
I was wondering why the resolution in TI mode is so much limited, since
I never had any problems measuring 5 or 10 MHz frequencies with up to 12
figures on that counter.
If you look closely you'll see that the 53132A resolution drops by a factor of ten (to 11 digits per second) when then input is too near 5 or 10 MHz. There's a note on page 185 of the user manual:
http://cp.literature.agilent.com/litweb/pdf/53131-90055.pdf
/tvb