On Tue, 09 Oct 2007 02:16:37 -0000, Frnak McKenney
wrote:

Am I asking too much?

Hi Frnak,

Judging by the questions and responses, I would have to say "Yes."

To this point you haven't exactly demonstrated you have a problem,
just a complaint of a signal of poor quality to a human's perception.
For the clock itself, that complaint is arguably weak.

Let's just examine the evidence for the problem: There is none!

You have a clock that has 100mS resolution, and yet you have never
said how much it is off. 100mS? 1S? 10S? 1 minute? 1 Hour? All,
or any part of any of these metrics?

As Reggie would have chimed in at this point "If you can't measure it
and express it with a quantifiable, then you don't know anything."

Of course, your only source of accurate information is the one you are
suggesting has a problem. It probably doesn't have a problem, but
then how does one use this source's accuracy to check itself? You
would need a second clock to check it, and we would be hearing your
complaint in stereo.

I've calibrated time standards to the nearest 100nS and it is
accomplished at one sitting, no need for total connectivity such as
you might imagine (unless the clock you have is especially crappy).

Your clock has a resolution of 0.1 second. There are roughly 1
million ticks of the display in a day. A simple XTAL oscillator at 10
MHz would exhibit 50ppm stability and in a day wander up to 0.5
second. The next day it might wander back, the day following it might
slip below by 0.5 second. You would be hard pressed to confirm this
with over the air matching to the strike of the WWV gong - except if
the clock is especially crappy (and it could be). The same XTAL might
also exhibit an absolute error of 50ppm and accumulate time error.
This would be far more noticeable over the course of a week (you could
confirm the error by listening to time announcements - but you have
been silent to this issue).

These worst case errors all presume that the internal circuitry cannot
over the course of 24 hours manage to pull out one of 1400
synchronizing opportunities to phase lock out the error. These
circuits are generally optimized to accomplish just this (they work
fine in watches with a 60KHz signal after all). Your clock may be
especially crappy (but that is unlikely).

The clock synchronizing circuits don't have to listen to the bandwidth
of noise you hear, the speaker is for your convenience, not the
clock's. I am sure that it works fine with only 1 LED lit - this is
not a case of "can you hear me now?"

73's
Richard Clark, KB7QHC

On Wed, 10 Oct 2007 10:19:49 -0700, Richard Clark wrote:
On Tue, 09 Oct 2007 02:16:37 -0000, Frnak McKenney
wrote:

Am I asking too much?

Hi Frnak,

Judging by the questions and responses, I would have to say "Yes."

To this point you haven't exactly demonstrated you have a problem,
just a complaint of a signal of poor quality to a human's perception.
For the clock itself, that complaint is arguably weak.

Let's just examine the evidence for the problem: There is none!

You have a clock that has 100mS resolution, and yet you have never
said how much it is off. 100mS? 1S? 10S? 1 minute? 1 Hour? All,
or any part of any of these metrics?

It's hard to tell exactly how much the clock is off by.

Every time the power hiccups, or I have to move the MAC-II, or power
down the outlet the display switches to something like this (best
viewed with a fixed-width font):

_ _
_ _ /_ /_ /_ /_
/ / /_/ /_ _/ /_ /_

and it stays that way for weeks. Or months.

As Reggie would have chimed in at this point "If you can't measure it
and express it with a quantifiable, then you don't know anything."

Given the extent of my ignorance concerning 'most everything, that
seems likely. grin!

Of course, your only source of accurate information is the one you are
suggesting has a problem. It probably doesn't have a problem, but
then how does one use this source's accuracy to check itself? You
would need a second clock to check it, and we would be hearing your
complaint in stereo.

Hm... I don't _think_ so. At least, I havent heard of any plans for
a High-Def upgrade to Usenet lately, but with Congress currently in
session I suppose anything is possible.

As for testing the clock's accuracy, you're right about needing a
second source ("Qui custodiet ipsos custodes?" or something like
that? grin). On the other hand, as long as the digits are
flashing by, I'm happy to "just trust them".

I've calibrated time standards to the nearest 100nS and it is
accomplished at one sitting, no need for total connectivity such as
you might imagine (unless the clock you have is especially crappy).

Your clock has a resolution of 0.1 second. There are roughly 1
million ticks of the display in a day. A simple XTAL oscillator at 10
MHz would exhibit 50ppm stability and in a day wander up to 0.5
second. The next day it might wander back, the day following it might
slip below by 0.5 second.

If I read the MAC-II manual correctly, each time it "connects to
WWV" (gets a recognizable signal) it calculates and saves an
adjustment value. The front panel has two LEDs labelled "TRIM UP"
and "TRIM DN" to indicate how well it's doing.

... You would be hard pressed to confirm this
with over the air matching to the strike of the WWV gong - except if
the clock is especially crappy (and it could be). The same XTAL might
also exhibit an absolute error of 50ppm and accumulate time error.
This would be far more noticeable over the course of a week (you could
confirm the error by listening to time announcements - but you have
been silent to this issue).

True. And, while I'm sure the _WWV_ announcer hasn't been silent,
_I_ haven't heard anything comprehendable from him/her/it out of my
MAC-II's speaker at any point in the past few weeks.

These worst case errors all presume that the internal circuitry cannot
over the course of 24 hours manage to pull out one of 1400
synchronizing opportunities to phase lock out the error. These
circuits are generally optimized to accomplish just this (they work
fine in watches with a 60KHz signal after all). Your clock may be
especially crappy (but that is unlikely).

Based on the feedback from other posters, it's likely a consequence
of 10MHz propagation. A VLF RF signal like 60KHz reportedly does a
much better job of getting a readable signal to a wide area.

The clock synchronizing circuits don't have to listen to the bandwidth
of noise you hear, the speaker is for your convenience, not the
clock's. I am sure that it works fine with only 1 LED lit - this is
not a case of "can you hear me now?"

No, but (assuming you're subbing for WWV grin!) it would be nice
to know I was going to get a readable message from "you" more than
once every couple of months. (Why do I hear the echo of my parents'
frustration during my colege days? grin!)


Thanks for the feedback. I admit I hadn't thought that much about
the accuracy of the MAC-II; I'm afraid I've been too caught up in
simply trying to get digits instead of "error text" on the display.


Frank
--
"A good traveller has no fixed plans and is not intent on
arriving." -- Lao Tzu (570-490 BC)
--
Frank McKenney, McKenney Associates
Richmond, Virginia / (804) 320-4887
Munged E-mail: frank uscore mckenney ayut minds pring dawt cahm (y'all)

On Thu, 11 Oct 2007 12:12:14 -0000, Frnak McKenney
wrote:

_ _
_ _ /_ /_ /_ /_
/ / /_/ /_ _/ /_ /_

and it stays that way for weeks. Or months.


A whip antenna should be able to sort out WWV for at least one of 1400
synchronizing events in a day. This may be a problem of too much
antenna at one time - and a nearby lightning event at that same time.
Your front end got fried out.

73's
Richard Clark, KB7QHC

On Thu, 11 Oct 2007 07:51:55 -0700, Richard Clark wrote:
On Thu, 11 Oct 2007 12:12:14 -0000, Frnak McKenney
wrote:
_ _
_ _ /_ /_ /_ /_
/ / /_/ /_ _/ /_ /_

and it stays that way for weeks. Or months.

A whip antenna should be able to sort out WWV for at least one of 1400
synchronizing events in a day. This may be a problem of too much
antenna at one time - and a nearby lightning event at that same time.
Your front end got fried out.

If you're reading these posts in the same order I'm posting, you'll
have already read my good news:

It's Working!

(Oh. That's right -- this is Usenet. Y'all can't here the
"doooonb"-ing from down the hall. Well, take my word for it -- or
even the MAC-II's display! grin)

This doesn't prove that the MAC-II's J-FET RF amplifier hasn't been
degraded through an... er, "very wideband RF overload" grin!, but
I think it does say that "completely fried" is unlikely (which
pleases me greatly! grin).

My earlier MAC-I had a nice whip built in, and could select the
strongest signal among (IIRC) 5, 10, and 15MHz, but it had the same
problems getting an "acceptable" WWV signal most of the time.


Frank
--
"Whatever you can do, or dream you can, begin it.
Boldness has genius, power and magic in it" -- Goethe
--
Frank McKenney, McKenney Associates
Richmond, Virginia / (804) 320-4887
Munged E-mail: frank uscore mckenney ayut minds pring dawt cahm (y'all)

Frnak McKenney wrote:





Hm... I don't _think_ so. At least, I havent heard of any plans for
a High-Def upgrade to Usenet lately, but with Congress currently in
session I suppose anything is possible.

As for testing the clock's accuracy, you're right about needing a
second source ("Qui custodiet ipsos custodes?" or something like
that? grin). On the other hand, as long as the digits are
flashing by, I'm happy to "just trust them".


One needs three clocks.. if you have only two, you only know that they
differ. If you have three, you can detect the failed clock, because the
other two read the same time.

Of course, if the different clocks have different accuracies or
reliabilities, that's another story.

For a more detailed treatment of such things, you might want to check
out the "Byzantine General" problem.


Based on the feedback from other posters, it's likely a consequence
of 10MHz propagation. A VLF RF signal like 60KHz reportedly does a
much better job of getting a readable signal to a wide area.

Although in my house (southern California), the 60 kHz signal seems to
fade in and out on a daily basis.

On Thu, 11 Oct 2007 09:52:53 -0700, Jim Lux wrote:
Frnak McKenney wrote:
--snip--
As for testing the clock's accuracy, you're right about needing a
second source ("Qui custodiet ipsos custodes?" or something like
that? grin). On the other hand, as long as the digits are
flashing by, I'm happy to "just trust them".

One needs three clocks.. if you have only two, you only know that they
differ. If you have three, you can detect the failed clock, because the
other two read the same time.

Well, you have a reasonable _presumption_, anyway. grin

In _theory_, one could have two clocks in error and one on time. Or
even all three out of step with (say) the NIST's clock. But
two-out-of-three would be the place to put your money. grin

Of course, if the different clocks have different accuracies or
reliabilities, that's another story.

For a more detailed treatment of such things, you might want to check
out the "Byzantine General" problem.

Faint recollections... "unreliable communication on the
battlefield" category?

Based on the feedback from other posters, it's likely a consequence
of 10MHz propagation. A VLF RF signal like 60KHz reportedly does a
much better job of getting a readable signal to a wide area.

Although in my house (southern California), the 60 kHz signal seems to
fade in and out on a daily basis.

I can believe it. I've seen those LCD "clock and weather" stations
take days to synchronize the time (I assume they're using WWVB).
OTOH, I'm hardly one to talk, since I've let the MAC-II sit
liostening for months at a time without successfully locking-in on
WWV.

Frank
--
Hanlon's Razor: ÿNever attribute to malice that which can
be adequately explained by stupidity.ÿ
--
Frank McKenney, McKenney Associates
Richmond, Virginia / (804) 320-4887
Munged E-mail: frank uscore mckenney ayut minds pring dawt cahm (y'all)

On Oct 11, 5:12 am, Frnak McKenney
wrote:
On Wed, 10 Oct 2007 10:19:49 -0700, Richard Clark wrote:
On Tue, 09 Oct 2007 02:16:37 -0000, Frnak McKenney
wrote:

Am I asking too much?

Hi Frnak,

Judging by the questions and responses, I would have to say "Yes."

To this point you haven't exactly demonstrated you have a problem,
just a complaint of a signal of poor quality to a human's perception.
For the clock itself, that complaint is arguably weak.

Let's just examine the evidence for the problem: There is none!

You have a clock that has 100mS resolution, and yet you have never
said how much it is off. 100mS? 1S? 10S? 1 minute? 1 Hour? All,
or any part of any of these metrics?

It's hard to tell exactly how much the clock is off by.

Every time the power hiccups, or I have to move the MAC-II, or power
down the outlet the display switches to something like this (best
viewed with a fixed-width font):

_ _
_ _ /_ /_ /_ /_
/ / /_/ /_ _/ /_ /_

and it stays that way for weeks. Or months.

As Reggie would have chimed in at this point "If you can't measure it
and express it with a quantifiable, then you don't know anything."

Given the extent of my ignorance concerning 'most everything, that
seems likely. grin!

Of course, your only source of accurate information is the one you are
suggesting has a problem. It probably doesn't have a problem, but
then how does one use this source's accuracy to check itself? You
would need a second clock to check it, and we would be hearing your
complaint in stereo.

Hm... I don't _think_ so. At least, I havent heard of any plans for
a High-Def upgrade to Usenet lately, but with Congress currently in
session I suppose anything is possible.

As for testing the clock's accuracy, you're right about needing a
second source ("Qui custodiet ipsos custodes?" or something like
that? grin). On the other hand, as long as the digits are
flashing by, I'm happy to "just trust them".

I've calibrated time standards to the nearest 100nS and it is
accomplished at one sitting, no need for total connectivity such as
you might imagine (unless the clock you have is especially crappy).

Your clock has a resolution of 0.1 second. There are roughly 1
million ticks of the display in a day. A simple XTAL oscillator at 10
MHz would exhibit 50ppm stability and in a day wander up to 0.5
second. The next day it might wander back, the day following it might
slip below by 0.5 second.

If I read the MAC-II manual correctly, each time it "connects to
WWV" (gets a recognizable signal) it calculates and saves an
adjustment value. The front panel has two LEDs labelled "TRIM UP"
and "TRIM DN" to indicate how well it's doing.

... You would be hard pressed to confirm this
with over the air matching to the strike of the WWV gong - except if
the clock is especially crappy (and it could be). The same XTAL might
also exhibit an absolute error of 50ppm and accumulate time error.
This would be far more noticeable over the course of a week (you could
confirm the error by listening to time announcements - but you have
been silent to this issue).

True. And, while I'm sure the _WWV_ announcer hasn't been silent,
_I_ haven't heard anything comprehendable from him/her/it out of my
MAC-II's speaker at any point in the past few weeks.

These worst case errors all presume that the internal circuitry cannot
over the course of 24 hours manage to pull out one of 1400
synchronizing opportunities to phase lock out the error. These
circuits are generally optimized to accomplish just this (they work
fine in watches with a 60KHz signal after all). Your clock may be
especially crappy (but that is unlikely).

Based on the feedback from other posters, it's likely a consequence
of 10MHz propagation. A VLF RF signal like 60KHz reportedly does a
much better job of getting a readable signal to a wide area.

The clock synchronizing circuits don't have to listen to the bandwidth
of noise you hear, the speaker is for your convenience, not the
clock's. I am sure that it works fine with only 1 LED lit - this is
not a case of "can you hear me now?"

No, but (assuming you're subbing for WWV grin!) it would be nice
to know I was going to get a readable message from "you" more than
once every couple of months. (Why do I hear the echo of my parents'
frustration during my colege days? grin!)

Thanks for the feedback. I admit I hadn't thought that much about
the accuracy of the MAC-II; I'm afraid I've been too caught up in
simply trying to get digits instead of "error text" on the display.

Frank
--
"A good traveller has no fixed plans and is not intent on
arriving." -- Lao Tzu (570-490 BC)
--
Frank McKenney, McKenney Associates
Richmond, Virginia / (804) 320-4887
Munged E-mail: frank uscore mckenney ayut minds pring dawt cahm (y'all)

Gee, all this trouble you're having getting a good signal from WWV on
10MHz makes me wonder, "why??" I mean, why bother? It must be the
challenge! I'm a bit closer to Ft. Collins, but I wouldn't expect
things to be all that much different, and in any event, the same
antenna I've used for them has worked fine for signals from W1AW, for
frequency measuring tests. That antenna is just a short piece of
wire, maybe five feet long, connected to a signal analyzer's input
port. The signal analyzer's input doesn't even have a particularly
good noise figure. But with it, I get a good enough signal from WWV
to easily track the nocturnal/diurnal frequency shifts that happen as
the path length changes. (The analyzer may not have a great RF front
end, but it has a very stable frequency reference...) Similarly, I
have a portable short wave radio that has an awful front end, and with
just a 3 foot whip antenna, it gets WWV 10MHz fine most of the time.
Obviously, there are times of the day when propagation just doesn't do
it, but over the period of one day, and not during a geomagnetic
storm, the signal is usually available.

All this makes me wonder if the receiver in your clock is OK. I'd
start by looking at that; or at very least, see if a known-working
radio receiver has as much trouble with the signal as the clock seems
to. Given that the clock has a single frequency receiver, even a
pretty simple receiver design should give decent performance.

It's also possible that you have some signal source on nominally 10MHz
nearby, and you hear than instead of WWV. There are soooo many
microprocessors around the average home these days that it's entirely
possible that the source of the trouble is very nearby--but could also
be in a neighbor's house (or car -- or garage -- or ??).

If you want an accurate clock and get tired of fooling with WWV-10MHz,
and don't want to use WWVB-60kHz, you might consider using a GPS. As
long as you can manage an antenna with a reasonably clear view of the
sky, you should be able to have a clock reliably set to within less
than a second accuracy practically all the time. Or, if you'd like to
be independent of external references, modify your MAC with an oven
oscillator. Oven stabilized crystal oscillators left on for a long
time will almost always settle out to very low drift rates---one part
in 10^8 over a year shouldn't be difficult, in my experience, and GPS
signals can be used to calibrate it occasionally. One part in 10^8 is
about 1/3 of a second per year.

Cheers,
Tom

On Thu, 11 Oct 2007 11:26:03 -0700, K7ITM wrote:
On Oct 11, 5:12 am, Frnak McKenney
wrote:
On Wed, 10 Oct 2007 10:19:49 -0700, Richard Clark wrote:
On Tue, 09 Oct 2007 02:16:37 -0000, Frnak McKenney
wrote:

Am I asking too much?

Hi Frnak,

Judging by the questions and responses, I would have to say "Yes."
--snip--

Gee, all this trouble you're having getting a good signal from WWV on
10MHz makes me wonder, "why??" I mean, why bother? It must be the
challenge! I'm a bit closer to Ft. Collins, but I wouldn't expect
things to be all that much different, and in any event, the same
antenna I've used for them has worked fine for signals from W1AW, for
frequency measuring tests. That antenna is just a short piece of
wire, maybe five feet long, connected to a signal analyzer's input
port. The signal analyzer's input doesn't even have a particularly
good noise figure. But with it, I get a good enough signal from WWV
to easily track the nocturnal/diurnal frequency shifts that happen as
the path length changes. (The analyzer may not have a great RF front
end, but it has a very stable frequency reference...) Similarly, I
have a portable short wave radio that has an awful front end, and with
just a 3 foot whip antenna, it gets WWV 10MHz fine most of the time.
Obviously, there are times of the day when propagation just doesn't do
it, but over the period of one day, and not during a geomagnetic
storm, the signal is usually available.

Hm. Sounds great.

All this makes me wonder if the receiver in your clock is OK. I'd
start by looking at that; or at very least, see if a known-working
radio receiver has as much trouble with the signal as the clock seems
to. Given that the clock has a single frequency receiver, even a
pretty simple receiver design should give decent performance.

The RF section is a one-transistor (J-FET) amplifier followed by a
TDA1072A "Integrated AM receiver" chip. I'm not sure how to go
about comparing that to your equipment.

It's also possible that you have some signal source on nominally 10MHz
nearby, and you hear than instead of WWV. There are soooo many
microprocessors around the average home these days that it's entirely
possible that the source of the trouble is very nearby--but could also
be in a neighbor's house (or car -- or garage -- or ??).

Highly possible, but the noise level "at the speaker" seems to have
fallen off appreciably in the past few days. Also, I think I'm
doing a better job of tuning the antenna; adjusting with no clearly
defined signal is a chicken-and-egg problem, but two days ago I
started hearing a deafening cackle. grin!

If you want an accurate clock and get tired of fooling with WWV-10MHz,
and don't want to use WWVB-60kHz, you might consider using a GPS. As
long as you can manage an antenna with a reasonably clear view of the
sky, you should be able to have a clock reliably set to within less
than a second accuracy practically all the time. Or, if you'd like to
be independent of external references, modify your MAC with an oven
oscillator. Oven stabilized crystal oscillators left on for a long
time will almost always settle out to very low drift rates---one part
in 10^8 over a year shouldn't be difficult, in my experience, and GPS
signals can be used to calibrate it occasionally. One part in 10^8 is
about 1/3 of a second per year.

Yup. GPS would eb the way to go for accuracy... or -- for the
billion-dollar-budget people -- your very own Cesium Clock. grin!


Frank
--
To learn is to change. Learning allows an animal child to
finish the long, slow process of evolution by changing in its
own lifetime. Tiger cubs, eaglets, or babies, nature brings
us all into existence with the ability to learn, and the rest
is up to us. -- Susan McCarthy / Becoming a Tiger
--
Frank McKenney, McKenney Associates
Richmond, Virginia / (804) 320-4887
Munged E-mail: frank uscore mckenney ayut minds pring dawt cahm (y'all)