Capturing NIST signal with a Propeller?
Rsadeika
Posts: 3,837
Has anybody set up a Propeller to capture a signal from an NIST station? I wonder how much something like that would cost in materials, of course personal time is always for free for your own projects. Now I know that you could purchase a GPS receiver with an antenna and capture just time and date from that, but that is a ~$50.00 dollar purchase. Can it be done for less than $50.00 using a Propeller?
The reason I am interested in this is because I am experimenting with data logging that would have accurate and consistent readings. At the moment I am using an RPi to do that, which has an WiFi attachment, but the unit does not have its own internal clock so it relies on the internet connection. Well last night, in my area, we had a power outage, a couple of minutes of down time, plus the internet was out for about four hours. Those two combinations made my RPi useless in terms of providing the correct time. So I am looking for a way of providing an accurate, auto updatable time signal. Anybody have any workable ideas?
Ray
The reason I am interested in this is because I am experimenting with data logging that would have accurate and consistent readings. At the moment I am using an RPi to do that, which has an WiFi attachment, but the unit does not have its own internal clock so it relies on the internet connection. Well last night, in my area, we had a power outage, a couple of minutes of down time, plus the internet was out for about four hours. Those two combinations made my RPi useless in terms of providing the correct time. So I am looking for a way of providing an accurate, auto updatable time signal. Anybody have any workable ideas?
Ray
Comments
How about an extremely accurate TXCO RTC like the DS3231 etc + an ESP8266 WiFi for the occasional update maybe, all to be had for less than ten bucks.
Ray
Those things are NOT Atomic Clocks even if they try to call them that, just because somewhere the signal is calibrated against an Atomic clock as are GPS signals etc. I'm not sure what you are needing this ultrafine precision for. Aren't the DS3231s extremely stable? They are based on trimmed and calibrated MEMS technology and the time signal is local, there, and ready all the "time".
Ray
NIST covers a very limited area of the US and the coverage depends upon the ionosphere etc. Besides the ferrite rod is bulky and the whole idea of time signals is to get a stable and precise time, isn't it? Now if you have a extremely stable TXCO clock built in on your board then why would you ever bother with a WWVB receiver decoding NIST signals? Sure, there was a time when that was probably the best way and if you wanted to build such a clock just for the fun of it or didn't know any better then go for it. I know what I would do.
Besides, this comes back down to the application as to what is really required. Is it accurate to the millisecond time-stamping? Or perhaps it's really to the tenth of a second or more.
The more expensive ones are more performing, but for clock sync this is overkill. Just don't expect the same navigation performace or time to fix.
Another alternative with Spinneret is cited here:
http://forums.parallax.com/showthread.php/151794-NIST-NTP-Code-for-Spinneret
I guess a NTP solution for a serial to wify solution is reasonable too.
Massimo
Ray
I'm implementing SNTP protocol on my Prop network servers so you don't need a P2 or RPi for this. However I would design a system that was stable in itself and when data is collected off that then the system's time would also be timestamped by the collector. It's an easy thing in software to compensate for an offset.
http://www.mouser.it/ProductDetail/Maestro-Wireless-Solutions/A2235-H/?qs=sGAEpiMZZMsjLMBIknjmkinf3subEiP4byj7tRuoHmI%3d
Unless you are in a basement you should get the time, even with no fixing on position.
It can be assembled also on a home made board using toner transfert (I did it with the discontinued A1035H module and it's doable).
Massimo
I imagine there are other boards like this one around.
Why can't people groove with that - or at least not be jackwagons?
WWVB covers all of the "continental 48", Mexico, and Canada south of the Hudson Bay (excl. Newfoundland).
http://tf.nist.gov/stations/wwvbcoverage.htm
WWV and WWVH, on shortwave, have greater coverage areas (potentially world-wide).
They use audible tones for top of the hour and so on, but the time of day data is coded using 100 Hz (0s and 1s corresponding to 100 Hz on-time.)
Fundamentally, if you have a general-coverage receiver, with a decent signal and audio into a low-pass filter (or not) then you could time the length of the 100Hz occurrences and the rest is academic.
You can get more info googling "WWV".
Beyond this, I'm not "sharing" design info or code.
-Phil
I used it for caribration of rtc.
In Japan, there are 2-point(Fukushima:40kHz and Kyushu:60kHz) as transmitter.
These cover all over Japan(from Okinawa to Hokkaido).
Of course, time code is different with USA.
Now Phil you should know better than some who jump to conclusions. If the guy was building a NIST receiver then it would be about that. But the OP wants an accurate timestamp and we forumistas in general are not so blinked that we can only answer in the constrained terms of an OP's question. How many threads have gone on and on and on because they focus way too tightly instead of stepping back and asking the right question? So I am asking questions rather than assuming the old lady wanted to cross the busy road when in fact she was waiting for a cab!
The OP may not be into hardware so presenting some options can only help.
http://www.ebay.com/itm/High-Sensitivity-WWVB-Atomic-Clock-Receiver-Module-Fort-Collins-60KHz-Receiver-/231367674457?pt=UK_BOI_Electrical_Components_Supplies_ET&hash=item35de968a59
http://www.ebay.com/itm/WWVB-Atomic-Clock-Receiver-Module-Fort-Collins-60KHz-Receiver-/331354144562?pt=UK_BOI_Electrical_Components_Supplies_ET&hash=item4d263eff32
Both sellers are in the UK but it looks like they're selling the US version of the receivers.
Edit: Apparently this site sells components for to receive these clock signals.
I must confess, this all started with the thought of the actual Propeller chip being able to capture a signal wave. I thought I remember a post from a very long time ago where the Propeller was used in sending a signal wave or something along those waves. Actually I have not seen very many new Propeller hardware ideas, that is, actual use of the Propeller chip to do something different.
Ray
Phil did come up with a way to use a Propeller as a radio receiver.
Oh, moan & groan. What's this "we" stuff? The Propeller audience?
Here's a little something "from out of left field": You don't have to be able to play a guitar or sing to be a member of the Monkees Fan Club, but you should if you want to be recognized as a musician or part of a pop group.
I got a WWVB reciever, I think it was like $20. The idea is to have a propforth dattalogger run off a supercapacitor charged by a solar cell. And understand that on some occasions, the rig will run flat due to clouds, snow, whatever.
When the prop boots, it nottices that its time stampos not initialized. It listens to the WWVB unttil it gets the time, then adjusts its stamp accordingly. Since records are logged sequentially, we can always determine the time stamps for the uninitalized time records, no matter how long or how often thay occure (unless it rruns flat before it gets time). No additional RTC is needed, the prop code is accurate enough although it does drift fast and slow depending on temperature. Once it figures the offset, it get really close.
Might be handy for data logging in an isolated location. Should be toatlly doable. Haven't actually built it yet, but no reason why it wouldn't work, all the parts are there.
If they work like the Sparkfun module the British modules would require quite a bit of lifting from the Propeller; they just output a presence-or-absence signal which would need to be error checked and interpreted. In that sense they represent about the minimum additional hardware needed to grab the 60 KHz signal out of the air and amplify it so you can work with it with a cihp like the Propeller. The instructions for reading the signal are straightforward -- it's not high speed so could probably all be done in C or Spin without difficulty. But I did read of people having "adventures" with the Sparkfun modules, such as not being operable during the day (develop at night only!) or in one case not being operable while the development PC was running nearby due to interference.
http://www.ebay.com/itm/High-Sensitivity-WWVB-Atomic-Clock-Receiver-Module-Fort-Collins-60KHz-Receiver-/231367674457?pt=UK_BOI_Electrical_Components_Supplies_ET
It is interesting that most Propellers (as far as I know) use 5Mhz crystals which could
be convenient to feed a direct conversion receiver circuit because WWV is on that frequency and harmonically related ones.
It might also be very easy to get WWVB on 60 KHz, if you can get / make a sensitive resonant antenna loop for it.
The format is different on that frequency. (I think there are two binary codes.)
It is my observation that all the NIST time signals seem to come in strong at night but very weak during the day, in New York.
The formats also have changed slightly over the years but usually not enough to make radio controlled clocks obsolete.
At my location I can also always hear CHU from Canada on 3.33 MHz, it talks, beeps, and sends 300 baud digital time.
Edit: I used a prop rather than BS-2. This thread got sort of "anti-hijacked" to
http://forums.parallax.com/showthread.php/158053-NIST-WWVB-WWV-WWVH-amp-Propeller
Since round-the-clock reception isn't likely, some sort of fairly accurate clock is still necessary, I think.
The WWVB would be used to update/correct that clock (as most of the wall-clocks we see out there do.)
Yeah mate, that's what I've been hinting at and in fact those wall clocks probably drifted by several seconds or more until they could get a signal again, and then that signal depends a lot upon interference and position etc. Unless there was an indication on the clock for how long it has been without synch then there is no real way to totally rely upon it other than having that warm and fuzzy psychological peace of mind that it's an "Atomic" clock
Don't get me wrong, I like the idea of a NIST receiver although we don't have anything like that here in Oz, and I think it would be fun to get the Prop to be the whole front end of it as well as the decoder. For me to test that out though I'd have to build my own NIST transmitter, low-power at least, which could be fun too and I think the Prop could handle that at 60kHz. But I think if I built a time server I would just use GPS time signal because it's easy to pick up GPS modules for $10 on ebay. Maybe I will, and then I can use ISM bands or those ESP8266 WiFi modules to distribute the time signals around the area.
Now that the broadcasters are digital and I have been out of the loop for 10 years this might not exist today. It would take someone with a demod and delay scope to look at the signal. Could be more reliable then trying to receive distant HF signals without suitable antennas.
Carry on and ignore this comment...
Since PJ Allen has a thread devoted solely to reception of NIST broadcasts, I'm taking license here to mention yet another GPS module.
This module has a 1pps signal pin to which a wire can easily be soldered. Various comparisons to much more expensive time references has shown this inexpensive module's 1pps signal to be amazingly accurate and consistent, far more so than you'd expect from a cheap HobbyKing board.
Meanwhile, I've never been happy with clocks that used just the NMEA serial data stream. Even clocking on the leading edge of the start bit of the first character of an RMA message, the timing is "all over the place."
Pretty hard to beat 12 ns accuracy for anything close to $17.
While interesting solutions, none of the GPS solutions presented are capable of capturing NIST, so must be disqualified. Although I now have to buy some, and make a non-NIST project, as soon as I finish the NST project, otherwise my workshop will look like erco's without all the working robots.
Regarding clock drift, what IMPACT will a few seconds drift have on any given project, particulary where NIST was selected in the first place? Its not a GPS project (otherwise we would have STARTED with GPS). We got by for thousands of years with "about noon" and "near sunset" level of accuracy for most things, and that hasn't turned out too bad (relatively speaking).
I suggest that if one has an application "off by a few seconds" is fine, we can use NIST to set the clock, and the prop itself to keep time, and it will me more accurate then most folks can manually set in the first place. And it will automatically get the correct time. Now to build it and check it this is true.
Massimo
Ray