US power grid frequency tests.

[H. Carl Ott]

Of possible interest. A lot of my clocks use the grid as a reference.

http://goo.gl/KhgtQ

carl
--------------------------------------------------------
Henry Carl Ott   N2RVQ    hcar...@gmail.com

[Adam Jacobs]

I was just logging in to mention this as well.. uh oh.

http://www.google.com/hostednews/ap/article/ALeqM5giHrMC9wYlOzOkUg9wNC2jVKugkw?docId=371623ab59694aef9f0a02fe83faca8a

I wonder if I will be designing crystal controlled 60hz AC signals to bodge into my clocks sometime soon. :(

--
You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To post to this group, send an email to neoni...@googlegroups.com.
To unsubscribe from this group, send email to neonixie-l+...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/neonixie-l?hl=en-GB.

[Nicholas Stock]

Sounds like that's going to screw with my nixieneon clocks....;-(

Sent from my iPhone

[Joe Croft]

This does suck in a lot of ways.

I like how they say it will be a lot harder to control not with things like
wind and solar. I don't see why, I suspect most of these will end up going to
or starting out as DC then being converted to 60HZ. If they could do this with
massive turbines and physical generators, you would think they could easily do
it electronically.

-joe

[Jonathan Peakall]

Glad I live off the grid and don't use the mains for timing my clocks. I
won't notice a thing...

Jonathan

[jb-electronics]

Personally, I do not understand what the benefit is of using the mains
frequency. I always use a 4.194304MHz Quarz and so far all my clocks'
accuracies are very satisfactory.

Best regards,
Jens

[Instrument Resources of America]

The mains "frequency", at least here in the U.S. is always being
corrected for small variances that occur in the frequency during the
day. Therefore the overall error is none over a years time. Clocks are
therefore always accurate and never need resetting, unless of course
there is a power failure, or when changing fron standard time to day
light saving time, and vice versa. Ira

[JohnK]

Maybe you missed this earlier piece of info:-
http://www.google.com/hostednews/ap/article/ALeqM5giHrMC9wYlOzOkUg9wNC2jVKugkw?docId=371623ab59694aef9f0a02fe83faca8a

john k.

----- Original Message -----
From: "Instrument Resources of America" <IRACO...@HUGHES.NET>

> The mains "frequency", at least here in the U.S. is always being
> corrected for small variances that occur in the frequency during the
> day. Therefore the overall error is none over a years time. Clocks are
> therefore always accurate and never need resetting, unless of course
> there is a power failure, or when changing fron standard time to day
> light saving time, and vice versa. Ira
>
> On 6/25/2011 11:01 AM, jb-electronics wrote:
>> Personally, I do not understand what the benefit is of using the mains
>> frequency. I always use a 4.194304MHz Quarz and so far all my clocks'
>> accuracies are very satisfactory.

...clip....

[Instrument Resources of America]

I agree. I wonder what is REALLY behind this?? Something smells
rotten with this.!!! Ira.

[Terry Kennedy]

On Jun 25, 5:04 pm, Instrument Resources of America

<IRACOSA...@HUGHES.NET> wrote:
> I agree.  I wonder what is REALLY behind this??   Something smells
> rotten with this.!!!    Ira.

Indeed. Normally the 60Hz runs low during the day and makes it up at
night. The frequency drops due to large industrial loads during the
day.

I don't understand what the issue is with correcting the long-term
average frequency at night is - it isn't like the plants are running
un-monitored and they'd need to bring people in - they have to be
monitored anyway.

With the increasing use of large DC transmission lines to transmit
billable power (as opposed to the between-region DC links which mainly
serve as starting power for a blacked-out region), the inverter
frequency doesn't need to be tied to spinning generators and won't
drop as load increases (though since it has to be frequency-locked to
the rest of the region's generation, it has to track frequency
changes). As an example, the 3100MW Path 65 DC link.

[Instrument Resources of America]

Hello John,
YES!! I did read it , and for the life of me I can NOT figure out
what the REAL TRUTH is behind doing this. My comment was meant to be
past tense. Thanks, Ira.

[H. Carl Ott]

On Sat, Jun 25, 2011 at 5:11 PM, Instrument Resources of America
<IRACO...@hughes.net> wrote:
> Hello John,
>    YES!!  I did read it , and for the life of me I can NOT figure out what
> the REAL TRUTH is behind doing this. My comment was meant to be past tense.
>  Thanks,  Ira.
>

I think it's a conspiracy by the quartz crystal manufacturers.

But humor aside, I also can't quite figure what the reasoning is
behind this one is.
Somebody somewhere is probable figuring out a way to save a couple of
bucks, but it's weird one for sure.

-carl

[Charles MacDonald]

On 11-06-25 02:01 PM, jb-electronics wrote:
> Personally, I do not understand what the benefit is of using the mains
> frequency. I always use a 4.194304MHz Quarz and so far all my clocks'
> accuracies are very satisfactory.

The benefit WAS that the Hydro Utilities always averaged out the
frequency to avoid long turn drift, so that the clock would stay on time
for months. It is hard (but not impossible) to get a crystal source
that will stay to the same level of accuracy.

Now of course the "way forward" is to use GPS or WWVL methods, or even NTP.

--
Charles MacDonald Stittsville Ontario
cm...@zeusprune.ca Just Beyond the Fringe
http://www.TelecomOttawa.net/~cmacd/
No Microsoft Products were used in sending this e-mail.

[neutron spin]

I think the original idea of grid frequency maintenance is when clocks were purely mechanical before the digital era.  They of course were just synchronous A.C. motors that were terribly inaccurate without the synchonization to the power grid's frequency.  This also was before GPS and sophisticated devices to reference to.  The only concern to electric utilities today is synchronizing the generators when bringing them online to avoid motoring...but that does not require any great frequency tolerance since it is all relative....

[Instrument Resources of America]

A LOT of today's 'digital' clocks are still locked up to the power line
frequency, which is still the quickest, easiest, most in expensive way
of maintaining accuracy. Some are locked to either GPS or NIST 60 khz
broadcasts. Ira

> frequency tolerance since it is all relative.... --

> You received this message because you are subscribed to the Google
> Groups "neonixie-l" group.

> To view this discussion on the web, visit
> https://groups.google.com/d/msg/neonixie-l/-/AHMxZj9OJvcJ.

[neutron spin]

Yes ...that was a clever and (cheap) method of creating a relatively accurate time base but with modern technology there really is no reason to keep this method.  Of course existing designs that use the grid's frequency are going to have to live with it.  Similar to going from analog TV broadcasting to digital made many TV's useless unless converters were employed.  I suppose someone could capitalize on this situation by offering to convert the devices from line time bases to crystal, GPS or WWVB systems. 

[Wayne de Geere III]

My mother in law handed me an article today about this from their local paper and asked if my "funny tube clocks" are still going to work. WIth the exception of my MOD6 clocks, I'm all GPS over here but I'd be a bit bummed that the masters of 60Hz have unilaterally decided to screw with that "to see if it affects anyone" which was the gist of the article published. After 80+ years of having a national 60Hz timebase, it seems somewhat odd to go and change that "just because" -- did anyone find anything on the net about why they're doing this and for what hoped benefit?

> --
> You received this message because you are subscribed to the Google Groups "neonixie-l" group.

[Wayne de Geere III]

Now we're all going to need to GPS or use an atomic time standard to discipline our mechanical flip clocks like http://leapsecond.com/pages/atomic-nixie/

[J Forbes]

We have about eight clocks here that we use to tell time at home
regularly....they ALL are based on line frequency!

Maybe we need to get a mast protest going, have everyone you know mail
at least one line frequency based clock to Joe McClelland, head of
electric reliability for the Federal Energy Regulatory Commission.

I wonder how many thousands (or millions) of clocks it would take for
him to get the message?

Jim

[neutron spin]

It's a conspiracy between Elm electronics the FERC.

http://www.elmelectronics.com/ebench.html#Oscillators

[Larry]

I recently completed a Kabtronics Nixie clock that uses line
frequency. Now I'm going to have to add a 60Hz generator to it.
Thanks for the link.

[neutron spin]

You are quite welcome. The chip is actually a microcontroller that
coded to produce the 60, 50 or 1 Hz signals. The 60 Hz version will
flip back and forth between the grid freq and the micro. I think it
is a Microchip MCU but not sure. Simple way of ensuring the clock
always has a clock signal. The color burst crystal is still subject
to minor drift but should function well I assume. I am guessing
perhaps around 10 to 20 PPM drift.

> >http://www.elmelectronics.com/ebench.html#Oscillators- Hide quoted text -
>
> - Show quoted text -

[Adam Jacobs]

It's a clever idea, shouldn't be too hard to implement in an equivalent
AVR chip (Tiny12) or whatever. At $10 a pop, given the number of clocks
I have, it'd definitely be worth writing some simple code. I was already
planning on doing the 60hz generator uC code, the colorburst crystal
surprises me, though. Is the colorburst crystal higher accuracy than a
standard cheap 8mhz or whatever? The only thing I know of colorburst
crystals from is that they are popular in 80m homemade CW rigs for
obvious reasons. I also like the sync to line frequency option. I might
add a switch, since they are initially only running this experiment for
a year. If the year ends and they discontinue the experiment, I could
flip the switch and go back to line frequency.

-Adam

[neutron spin]

The crystals are cheap these days whether a colorburst or other
generic types. My guess is they picked that frequency that they are
readily available and they wrote the code using that number to make it
difficult to clone...lol... but it was probably just a choice when
they developed the device. I am not connected to this company but am
just a user of some of their products. I do not mind paying them for
their time and effort to produce these cool devices. I bought a few
for around 8 bucks each and used them for a few projects.

> >>>http://www.elmelectronics.com/ebench.html#Oscillators-Hide quoted text -
> >> - Show quoted text -- Hide quoted text -

[J Forbes]

I wonder how I'm gonna get that neat flapper mechanical clock to keep
time after the experiment starts? Or my scope clock? or my TTL nixie
clock? or my old Cathode Corner nixie clock? or my cheezy 70s LED
alarm clock? or the clock in our kitchen, on the electric range? etc.

The battery powered Mickey Mouse clock in the garage should be fine.
Also my Nixie watches. And my cell phone will probably keep up ok too.

Jim

[neutron spin]

Personally I rely on my Heliochronometer or sundial....needs no
batteries and lasts a long time....

[neutron spin]

Personally I am going to use my old standby..... the Heliochronometer

On 27 June, 16:03, J Forbes <jforbnos...@selectric.org> wrote:

[Adam Jacobs]

Ya, I certainly don't mind supporting other people's work. It's a
time/cost tradeoff in my case, though. I have probably 10-12 clocks that
I'd like to modify.. for $100ish dollars, it is worth it to me to put in
the hour of work. If it was less money (or more difficult) then I
wouldn't mind it at all. For example, I don't begrudge Moses the $15-$20
he gets for his extremely full featured clock chips.

-Adam

[Charles MacDonald]

On 11-06-27 02:21 PM, Adam Jacobs wrote:
> the colorburst crystal surprises me, though. Is the colorburst crystal
> higher accuracy than a standard cheap 8mhz or whatever? The only thing I
> know of colorburst crystals from is that they are popular in 80m
> homemade CW rigs for obvious reasons.

I always figured that they were used as they were about the chepest
crystals around. Every NTSC TV set needed one so they were made in the
Zillions. Don't know if the Digital TV sets use anything like them.

look at http://www.futurlec.com/ICCrystalsMain.shtml and the TV ones are
30 cents while the rest are 75 cents.

Actually as they were working against a reference at 15725 Khz
repetition, they probably did not need to be too accurate.

[threeneurons]

http://www.elmelectronics.com/ebench.html#Oscillators

| Larry <unmitigated_f...@earthlink.net> wrote:
| I recently completed a Kabtronics Nixie clock that uses line
| frequency.  Now I'm going to have to add a 60Hz generator to it.
| Thanks for the link.

If you can change the input source (3.58MHz xtal), or tweak it, you
might just want to leave things alone.

There are 525,960 minutes in a year, on average. 525,600 for a normal
year, and 527,040 for a leap year. Being off 20 minutes in a year
works out to ~38ppm (East Coast). 8 minutes comes out to 15ppm (West
Coast). A typical crystal has 30 to 50 ppm accuracy, or between 15 to
26 minutes off, in a year. You'll get no improvement with a common
crystal. You might just as well just stick with the line sync, and
just occasionally hit that minute button, to correct the time.

That module won't help you unless you swap out that xtal, and tie a
TCXO to the clock input. Of course, you'll need to match the
frequency, or write your own uC code. A TCXO has an initial accuracy
in the 1ppm to 2.5ppm territory. Even when extra errors, such as aging
coming into play, you're still a lot better off than using a cheap
xtal.

I actually use to sell a coded uC (a tiny12 in fact) on eBay, that
outputed 50Hz, 60Hz, and 1Hz. Sold it for $5 each. Maybe I should just
post the source code, to screw with that guy. Code is pretty trivial.

[neutron spin]

Yes you could do that. Perhaps you may have a great idea and make the
code open source so all of us clock nuts can benefit. Using a
DS3231 with the MCU also would be a fix but more hardware and higher
cost. It would be quite easy to write the code and connect the DS3231
to the MCU and jigger the code to give you the required 60 or 1 Hz
signal....but....is it worth it?...perhaps if one had a real
collectable old TTL clock and wanted to use it to tell time...I would
just like firing it up and watching the pretty numbers...lol...:)

[neutron spin]

One more thing those Elm modules were meant to be just a short term
fix when the power gid went away or for battery operated
situations....not a long term time base...regards...

On 27 June, 20:09, threeneurons <threeneur...@yahoo.com> wrote:

[neutron spin]

Also many of the cheaply designed nixie clocks just use a cheap 4 Mhz
or whatever the design uses and these are usually rated at 10 to 20
PPM. Color burst crystals can be found with similar tolerances....so
what's the difference unless you are using temperature compensated
oscillators????..I don't see your point but we must compare apples to
oranges I suppose...lol...:)

On 27 June, 20:09, threeneurons <threeneur...@yahoo.com> wrote:

[Adam Jacobs]

The DS3231 is a combination RTC & TCXO.. If your clock already keeps track of the time (I'm betting it DOES), then you don't need the RTC part. The DS32khz is a 32khz TCXO (Costs about $10.. but they will sample them).

-Adam

[Adam Jacobs]

Yep, that's what I was thinking.. If the line frequency gets bad enough, I'll add little TCXO 60khz oscillators to the front of my clocks and cut the trace that used to get the 60khz from the AC.
This is a very well reasoned and clearly thought out illustration, and it demonstrates nothing quite so much as that a regular crystal is a terribly inaccurate way to run a clock. If you post the source code, I'd definitely use it. :D

-Adam

On Mon, Jun 27, 2011 at 5:09 PM, threeneurons <threen...@yahoo.com> wrote:

If you can change the input source (3.58MHz xtal), or tweak it, you
might just want to leave things alone.

There are 525,960 minutes in a year, on average. 525,600 for a normal
year, and 527,040 for a leap year. Being off 20 minutes in a year
works out to ~38ppm (East Coast). 8 minutes comes out to 15ppm (West
Coast). A typical crystal has 30 to 50 ppm accuracy, or between 15 to
26 minutes off, in a year. You'll get no improvement with a common
crystal. You might just as well just stick with the line sync, and
just occasionally hit that minute button, to correct the time.

That module won't help you unless you swap out that xtal, and tie a
TCXO to the clock input. Of course, you'll need to match the
frequency, or write your own uC code. A TCXO has an initial accuracy
in the 1ppm to 2.5ppm territory. Even when extra errors, such as aging
coming into play, you're still a lot better off than using a cheap
xtal.

I actually use to sell a coded uC (a tiny12 in fact) on eBay, that
outputed 50Hz, 60Hz, and 1Hz. Sold it for $5 each. Maybe I should just
post the source code, to screw with that guy. Code is pretty trivial.

[David Forbes]

On 6/27/11 5:09 PM, threeneurons wrote:

> A typical crystal has 30 to 50 ppm accuracy, or between 15 to
> 26 minutes off, in a year. You'll get no improvement with a common
> crystal. You might just as well just stick with the line sync, and
> just occasionally hit that minute button, to correct the time.

I use cheap crystals for my clocks and watches. But I get 2 PPM accuracy
by adjusting them with a variable capacitor. The capacitor costs 50 cents.

I did have to buy an HP counter, the 5245L Nixie tube model, at surplus
for $20 to have a suitable test instrument. Now and then it needs
servicing, so I have a parts box on hand whose oven blew out.

I also had to work out the proper settings and offset to adjust the
Nixie watch crystal, since it runs at 32768 Hz and is rather sensitive
to the capacitance of the scope probe I use to measure the frequency,
even when I probe it on the chip side of the 510k ohm drive-reduction
resistor.

--
David Forbes, Tucson AZ

[H. Carl Ott]

I think you could put a 32768 xtal, a trim cap, and an 8 pin AVR (my
preference) onto a very small pcb for about dollar or two.
The advantage of using a micro is that could also add a simple
calibrate function.

For people who like to solder. (but don't like to code).

http://www.8085projects.info/post/60Hz-signal-circuit-using-32768Hz-SJT-crystal.aspx

-carl

[neutron spin]

Great idea...sometimes the simple designs are the most
effective....the only drawback of course for any of these crystals is
the temperature effect on the crystal and drift. We are not dealing
with NBS traceable standards here but if you a real stickler....I have
some Cesium....just kidding...the NRC would frown on
that!...regards...

> http://www.8085projects.info/post/60Hz-signal-circuit-using-32768Hz-S...
>
> -carl

[H. Carl Ott]

On Tue, Jun 28, 2011 at 11:37 AM, neutron spin <mrst...@charter.net> wrote:
> Great idea...sometimes the simple designs are the most
> effective....the only drawback of course for any of these crystals is
> the temperature effect on the crystal and drift.  We are not dealing
> with NBS traceable standards here but if you a real stickler....I have
> some Cesium....just kidding...the NRC would frown on
> that!...regards...
>

Yeah, temperature might be a bit of a problem. Hopefully the temp
inside a clock is reasonably stable over time.

But if the point is just to try to get accuracy close to that
currently existing via the mains then maybe we don't need temperature
& aging compensation.
Can anybody translate these charts into a PPM error figure?

http://www.leapsecond.com/pages/mains/

> On 28 June, 07:47, "H. Carl Ott" <hcarl...@gmail.com> wrote:
>> I think you could put a 32768 xtal, a trim cap, and an 8 pin AVR (my
>> preference) onto a very small pcb for about dollar or two.
>>  The advantage of using a micro is that could also add a simple
>> calibrate function.
>>
>> For people who like to solder. (but don't like to code).
>>
>> http://www.8085projects.info/post/60Hz-signal-circuit-using-32768Hz-S...
>>
>> -carl
>

> --

[neutron spin]

32.768KHz has a nice property that it oscillates 2^15 times per second
which makes the timer math easy...I made a calc using the average
tolerance of these crystals and came out to an accuracy of around
0.002 % accuracy or better - about 2 seconds a day. This of course
does not account for temperature drift....wow...are we splitting hairs
or seconds?...lol...regards...

On 28 June, 12:24, "H. Carl Ott" <hcarl...@gmail.com> wrote:

> > --- Hide quoted text -

[H. Carl Ott]

On Tue, Jun 28, 2011 at 1:03 PM, neutron spin <mrst...@charter.net> wrote:
> 32.768KHz has a nice property that it oscillates 2^15 times per second
> which makes the timer math easy..

A little trickier dividing 32768 down to 60hz though.
Anybody got a simple way to do that in a micro?

Going to have short and long cycles, that's okay as long as they
eventually add up to 60hz.

[David Forbes]

On 6/28/11 10:03 AM, neutron spin wrote:
> 32.768KHz has a nice property that it oscillates 2^15 times per second
> which makes the timer math easy...I made a calc using the average
> tolerance of these crystals and came out to an accuracy of around
> 0.002 % accuracy or better - about 2 seconds a day. This of course
> does not account for temperature drift....wow...are we splitting hairs
> or seconds?...lol...regards...
>

The standard 32768 Hz watch crystal has a parabolic
frequency/temperature curve, which tops out in frequency at 25C (77F)
and gets slower in either hotter or colder weather.

Once adjusted, it's good to a couple PPM at ~20C to 30C (68F to 86F).
One PPM is 30 seconds a year, half a second per week.

I highly recommend adjusting a crystal if you're going to use it for
timekeeping. The hairs you can split with it are much finer.

[David Forbes]

Sure, you can do that. The 60 pulses won't all be the same length, but
they'll all be there. The counter in the target clock doesn't need a
pure 60 Hz tone, it just needs 60 rising edges per second.

[neutron spin]

Yes you are correct...for long term accuracy you need to calibrate it
to a reliable frequency standard. If all you want is to generate 60
Hz the a good reference is Fairchild's Application Note 42025. There
is some good info there. I use the Maxim DS32kHz oscillator on many
of my projects and it is quite accurate for a clock time base.

Perhaps our wonderful government will keep the utilities doing what
they are doing today....If I received anything other than an accurate
60 Hz supply from them then I will insist on some form of rebate.
Time to break out my strip chart recorder and monitor my
supply...regards...

[Jonathan Peakall]

I wonder if an adaptation of my GPS time system mught work. I have a central
clock that reads the GPS and then sends it to the other clocks via RF. What
about a central clock that sends out a 60/50 hZ beat via RF? RF receivers
are fairly cheap, and then you only have to calibrate one xtal or whatever
you use. You should be able to bodger it into many exsisting clocks.

Sure am glad that I went the GPS route now! All my clocks will remain
synched, do DST etc.

Jonathan

[Adam Jacobs]

Are the RF receivers cheaper than the GPS receivers at this point?
Sparkfun sells some pretty small GPS receivers for about $25.. Course,
if you're into WWVB, digikey has a WWVB receiver & antenna for $10..
although for my clock, I needed to replace the original tuned loop
antenna with a larger one.

-Adam

[Grahame Marsh]

This clock

http://www.sgitheach.org.uk/atomic.html

has two timekeepers - one from the Rb oscillator and one from mains
frequency. I set the clocks using a GPS receiver. I can display the
time difference between the two clocks and idly watch the difference
wobble backwards a forwards. The greatest difference I have seen is
about 2 seconds. As commented earlier, frequency (in the UK Mainland)
dips during the day and then runs higher overnight to compensate,
usually but not always.

cheers Grahame
Highlands Scotland

[Jonathan Peakall]

Way cheaper. Check these out.

http://www.google.com/products/catalog?rlz=1T4GGLL_enUS328US328&q=sparkfun+rf&um=1&ie=UTF-8&tbm=shop&cid=11913633787621706539&sa=X&ei=MCgKTtCoL4e6sQOqwrGcCQ&ved=0CCcQ8wIwAw#

[Mich...@aol.com]

Very good price, but that isn't a GPS/WWVB.

 

I guess you could attach that to another and connect to a WWVB/GPS for remote location of the receivers instead of a long antenna wire.

 

Maybe have it connected to your garage door so the clock can open/close the door depending on the time of day.  Or make sure the front door is closed after a certain time at night.  :)

 

Michail

[Adam Jacobs]

Still, that's a pretty neat piece of kit. I might have to find an excuse to put one in my next project.

-Adam

[Jonathan Peakall]

What I meant was not a GPS/WWVB thing, but that one could generate a 60Hz signal and then TX it via RF, with each clock haveing a RF reciever to RX the beat. Won't set the time/date but could be a less expensive way of giving clocks that need a TTL/CMOS 60Hz beat.

[Adam Jacobs]

Make sure to buy the 315mhz version instead of the 433mhz.. 433mhz is smack in the middle of the 70cm band and is used for linking repeaters. Also, unless you're a licensed amateur, transmitting on 433mhz is limited to part 15.231 which authorizes "Intermittent Control Signals and/or Periodic Transmissions". I think it is safe to say that a 60hz pulse transmission would be much too high of a duty cycle to be characterized this way.

-Adam

[Brian P. Poi]

Digikey sells a 3.93216 MHz crystal; that divides into 60 Hz 2^16 times,
which is easy to do if the microcontroller has a 16-bit timer. Or even
with a plain logic chip or two.

Brian Poi

[Adam Jacobs]

I think that the appeal of the 32.768khz oscillator is that the DS32khz
TCXO from Maxim uses this frequency. The jury is out on if even a well
calibrated crystal will be more accurate than the coming line frequency
changes (we'll see how bad they are... )

-Adam

[taylorjpt]

I did this for fun a while back. It uses buttons to trim the add/
delete pulses per time block... if the average temperature is fairly
constant then the average clock speed is as well. It has a 0.125ppm
trim resolution with a +/-32ppm trim range.

http://www.tayloredge.com/reference/Circuits/ClockDivider2/SimpleClock_PIC12F675.jpg
http://www.tayloredge.com/reference/Circuits/ClockDivider2

jt

[neutron spin]

I just knew you would do something like that! Now I will make a small
temperature controlled oven for the crystal and we be happy!.....You
make the best nixie power supplies BTW...regards...

On 29 June, 16:51, taylorjpt <j...@tayloredge.com> wrote:
> I did this for fun a while back.  It uses buttons to trim the add/
> delete pulses per time block... if the average temperature is fairly
> constant then the average clock speed is as well.  It has a 0.125ppm
> trim resolution with a +/-32ppm trim range.
>

> http://www.tayloredge.com/reference/Circuits/ClockDivider2/SimpleCloc...http://www.tayloredge.com/reference/Circuits/ClockDivider2
>
> jt

[James]

On Jun 24, 6:10 pm, "H. Carl Ott" <hcarl...@gmail.com> wrote:
>  Of possible interest. A lot of my clocks use the grid as a reference.
>
>  http://goo.gl/KhgtQ
>
>  carl
>



What are they smoking? The vast majority of line powered clocks use
the grid frequency for timing, both analog and digital. It's great
because I can set all the clocks and they all stay exactly in sync
with one another. With the deviations mentioned in the article this
could cause serious problems, and for what gain? The frequency
compensation could be fully automated, I'd be shocked if it isn't
already.

[neutron spin]

In fact this is the controller I may use:

http://www.w6pql.com/crystal_oven_controller.htm

On 29 June, 19:53, neutron spin <mrstan...@charter.net> wrote:
> I just knew you would do something like that!  Now I will make a small
> temperature controlled oven for the crystal and we be happy!.....You
> make the best nixie power supplies BTW...regards...
>
> On 29 June, 16:51, taylorjpt <j...@tayloredge.com> wrote:
>
>
>
> > I did this for fun a while back.  It uses buttons to trim the add/
> > delete pulses per time block... if the average temperature is fairly
> > constant then the average clock speed is as well.  It has a 0.125ppm
> > trim resolution with a +/-32ppm trim range.
>
> >http://www.tayloredge.com/reference/Circuits/ClockDivider2/SimpleCloc...
>

> > jt- Hide quoted text -

[neutron spin]

I have always wondered what do you do when there is a power
outage?....Also do the utilities account for any outages or does that
even matter? Yes, Yes I know you just reset the master line clock
after the outage and all is well...regards...

[John Rehwinkel]

> The vast majority of line powered clocks use
> the grid frequency for timing, both analog and digital. It's great
> because I can set all the clocks and they all stay exactly in sync
> with one another.

Oh, they still will. They'll just all be wrong.

- John

[JohnK]

All this talk of line frequency reminds me of when we attached a 'powerful'
audio sig gen to the lecture room clock back in 1968. Got us 10 mins less
boredom.

John K.

----- Original Message -----
Subject: Re: [neonixie-l] Re: US power grid frequency tests.

>> The vast majority of line powered clocks use
>> the grid frequency for timing,

....clip....

[A.J. Franzman]

On Jun 30, 1:01 am, "JohnK" <yend...@internode.on.net> wrote:
> ----- Original Message -----
> Subject: Re: [neonixie-l] Re: US power grid frequency tests.
> >> The vast majority of line powered clocks use
> >> the grid frequency for timing,
>

> All this talk of line frequency reminds me of when we attached a 'powerful'
> audio sig gen to the lecture room clock back in 1968. Got us 10 mins less
> boredom.

I remember reading about something like that, probably more than 25
years ago when I was still in high school. Apparently there was a
certain professor who was so meticulous in preparing and pacing his
lectures, that with seldom even a glance at the clock, he always
managed to have them end right at the end of the class period. Some
ingenious hardware hacker students did as you described and messed
with the prof's head by speeding up and slowing down the clock, just
to watch him get flustered from running out of time or having time
left over at the ends of his presentations. I always wondered just how
it was done, and the size of the hardware that would be required. Of
course today it's fairly trivial to do in a small package, but back
then, I'm not so sure.

A.J.

[David Forbes]

On 6/30/11 8:25 PM, A.J. Franzman wrote:
>> All this talk of line frequency reminds me of when we attached a 'powerful'
>> audio sig gen to the lecture room clock back in 1968. Got us 10 mins less
>> boredom.
>

> I always wondered just how
> it was done, and the size of the hardware that would be required. Of
> course today it's fairly trivial to do in a small package, but back
> then, I'm not so sure.
>
> A.J.
>

I used to own an HP 201B audio signal generator. It was powerful enough
to ring a telephone bell, providing 100V RMS at 20 Hz.

[Wayne de Geere III]

You must be a real phone fan to know that proper ring generation is at 20Hz, I'm impressed!

[David Forbes]

Yes, I am. I have an analog PBX in my house.

I was building a cordless Western Electric 2500 desk set, so I needed to
ring the original bell with a 3.6V battery pack. I ended up rewinding
the coil for lower voltage and using a 555 timer oscillator feeding an
H-bridge driver to make it ring. It sounded exactly like the real thing.

[JohnK]

As to our hardware: the valve/tube audio oscillator could provide watts of
output. I can't remember if we had to add a transformer eg speaker
transformer in reverse. The venue was a government department training
school. That particular lecture room was adjacent to the site technical
equipment room and projection booth. The equipment room housed a couple of
racks of comms, telephone and PA and time-announcer/bell ringer etc. Wasn't
a problem to park the large osc in there. The single pair jumper wire [used
for interconnecting the racks via tag blocks ie mdf, idf etc] was noticable
running between the top of the projection window frame and the clock. It was
an old building and pretty scruffy looking which helped with the lack of
extra camo.

John K.

----- Original Message -----
From: "A.J. Franzman" <a.j.fr...@verizon.net>
I always wondered just how
it was done, and the size of the hardware that would be required. Of
course today it's fairly trivial to do in a small package, but back
then, I'm not so sure.

[JohnK]

Er, are we being sloppy here rounding 17 to 20? Or do the weirder countries
[ie not Australia :-)) ] use 20 ? I thought it was bad enough rounding 16
2/3 [16 and two thirds] to 17 !
My home system used a motorised exchange ringer running at 16 2/3 . Had to
replace the contacts with standard 3000 type as the scrappies always ripped
the platinum ones off.

John K.
[apologies for more drift in subject ]

[James]

Yes, my point was that they are line frequency referenced because they
stay in perfect sync and the long term stability is excellent. The
former will not be the case if the latter changes and I have to resort
to individual timebases.

[neutron spin]

Probably the best method would be to build a WWVB or GPS 60Hz
synchronized clock and designate that one a master and feed the others
with that signal. Another possibility would be to build a small 60 Hz
alternator with a WWVB or GPS synchronized phase locked loop to
control it's frequency and just power all the line frequency clocks
from that source independent of the utilities grid...regards...

[Charles MacDonald]

On 11-07-01 02:43 AM, JohnK wrote:
> Er, are we being sloppy here rounding 17 to 20? Or do the weirder
> countries [ie not Australia :-)) ] use 20 ?

I have always heard of 20Hz. ATT and ITT divided the world up at one
time, with North America being ATT and the rest of the world being ITT.
ITT found that it was easier to negotiate with different countries if
they had a different standard for each country. SO 17Hz may have been
the Aussie way.

--
Charles MacDonald Stittsville Ontario
cm...@zeusprune.ca Just Beyond the Fringe
http://www.TelecomOttawa.net/~cmacd/
No Microsoft Products were used in sending this e-mail.

[Wayne de Geere III]

Very sweet. I collect WECO gear like 550/551/556 cordboards and Pre 1960s dial sets, including the earliest dial subscriber sets like the AA1, 102 and 200 series phones. I had a 1a2 key system in my house from high school until the late 1980s when Panasonic introduced the KXT616 analog PBX. My current favorite phone is a 102 w /subset Bluetooth linked to my iPhone, it even supports outbound pulse dialing through the iPhone.

Sent from my iPad 2 3G

[Terry Kennedy]

On Jul 1, 2:43 am, "JohnK" <yend...@internode.on.net> wrote:
> Er, are we being sloppy here rounding 17 to 20? Or do the weirder countries
> [ie not Australia :-)) ] use 20 ?  I thought it was bad enough rounding 16
> 2/3 [16 and two thirds]  to 17 !

Standard (individual and 2-party) lines used ringers nominally
configured for 20Hz which would actually operate on a relatively wide
range of frequencies. 2-party operation was done by sending ringing
voltage between one wire of the phone pair and ground for the first
subscriber, and from the other wire to ground for the second
subscriber.

More complex party line configurations used frequency-selective
ringers. It was possible to get up to 8 customers on one pair, with 4
selective ringers on each half of the phone pair.

Beyond that, or on equipment that didn't support frequency-selective
ringers, you'd know if the call was for your number based on the ring
pattern. Old equipment normally wasn't upgraded if it didn't already
have selective frequency support, using the logic "the customer is
using a party line to reduce their costs - why should we (the phone
company) spend money to make their life easier?"

When making calls, 4 parties could be identified (ring, tip, tip + 1K,
tip +2.65K) if the switching equipment supported it (see above for why
it often didn't). If the equipment didn't support it, any direct dial
call would go to an intercept operator that would ask you for your
number. And of course, if you were on a non-dial party line, you had
to tell the operator both the number you wanted to call and the number
you were calling from.

[threeneurons]

> > Er, are we being sloppy here rounding 17 to 20? Or do the weirder countries
> > [ie not Australia :-)) ] use 20 ?  I thought it was bad enough rounding 16
> > 2/3 [16 and two thirds]  to 17 !

Just remember the origins of the 1st ring generators:

http://www.flickr.com/photos/34878756@N04/5559552293/
http://www.flickr.com/photos/34878756@N04/5559593001/

You cranked them by hand. Frequency and voltage regulation might be a
bit iffy.

[JohnK]

Yes, I have some magneto [wooden and beakelite] phones too. [And a lovely
tech's intstrument circuits book dated 1914 when they 'rationalised' the
system here.]

After all the time-base talk re the grid and at other times [atomic] I
thought that I had better determine whether the current posters had taken
off their Nixie-clock-foible hats :-)) .

The ringer I mentioned has several sets of cam/contact for specific ring
styles too.
[Dad was telephone guy and was twice given the honour of making the last
call from a closing exchange as part of the 'ceremony'. Still got the
'gold' phone and the 'gold-painted bi-directional selector' given as
memento.

John K.