The 1 PPS oddity, a quandary

[Token]

Since at least as far back as January of 2018 a 1 pulse per second odd signal has periodically been seen up and down the HF bands. This is a signal that sounds very much like a time station, ticking one time per second, changing frequency sometimes swiftly, sometimes slowly. Naturally, it has been reported on several forums. Some of the places it has been discussed:

https://forums.qrz.com/index.php?threads/time-signal-intruder.614510/
https://www.hfunderground.com/board/index.php/topic,41025.0.html
https://forums.radioreference.com/u...s-ham-40-meter-band-01-jan-2018-2238-utc.html

Various plotting methods have been applied to the signals, most results seem to point to the area around Chicago, Illinois.

Most of the reports seem to be in the ham bands, however that might just be a factor of how many people watch the ham bands at any given moment. To the best of my knowledge, the frequency ranges that have been observed and reported have been:

6320 to 6510 kHz
7095 to 7300 kHz
8100 to 8400 kHz
10100 to 10150 kHz
12200 to 12450 kHz
14060 to 14350 kHz
18065 to 18160 kHz
21200 to 21300 kHz

You can see it has at least 3 full bands that are well outside the ham bands, and a few frequencies just outside ham bands. There absolutely may be more ranges outside ham bands that are simply not being reported because no one has noticed. The reports of operation inside ham bands vastly outnumber the reports of operations outside ham bands, however I am not sure if this is an actual trend, or if it is just a result of the source of data (mostly hams). I know watching the signal today it seemed to spend as much time outside the ham bands as inside.

Various sources have tried, or tentatively, connected these transmissions with High Frequency Trading (HFT) experiments. Some specific HFT experimental licenses have been suggested as related to these transmissions, the following license are sometimes quoted:
WI2XNX
WJ2XGD
WH2XWU
WH2XVO

However, none of these license cover all of the frequencies observed. In fact, all of the authorized frequencies for all of those licenses combined still do not cover all of the frequencies observed.

Further, the HFT application of radio links is in a quest to gain advantages of milliseconds, if not microseconds, on trading transactions. So even if the pulses observed did contain data (and they don't appear to) they would update that data only once per second. It seems, to me, an unlikely fit.

It has been suggested that these pulses have nothing to do with the data of HFT itself, but are just tools to measure or model propagation delays for future HFT ventures. While certainly possible (the signal does look like a propagation / scatter experiment to me) the frequency ranges observed are pretty narrow, and such measurements generally do not require stepping across such narrow frequency ranges. Is moving 10 kHz every 2 or 3 minutes really going to tell you much about the real time propagation variations between the frequencies?

So, what is this 1 PPS oddity? I really don't know. But the more I look at it the less convinced I am that it has anything to do with HFT experiments, or any of the license so far identified. Can anyone convince me differently?

T!

 

[prcguy]

Is the signal just an unmodulated carrier that goes on and off about once per second? I've found what looks like an artifact of a building alarm system that does this throughout the 40m amateur band. I tracked one of them down to a random building in an industrial park with large warehouses. I had one close to my work that was also in an industrial park.

A friend just installed a remote controlled HF radio in, you guessed it, an industrial park and I see the tell tale once per second on off keying on several frequencies across the 40m band. This place is in the high desert of So Cal where the other sites I found were more in the suburban areas of Los Angeles.

Since at least as far back as January of 2018 a 1 pulse per second odd signal has periodically been seen up and down the HF bands. This is a signal that sounds very much like a time station, ticking one time per second, changing frequency sometimes swiftly, sometimes slowly. Naturally, it has been reported on several forums. Some of the places it has been discussed:

https://forums.qrz.com/index.php?threads/time-signal-intruder.614510/
https://www.hfunderground.com/board/index.php/topic,41025.0.html
https://forums.radioreference.com/u...s-ham-40-meter-band-01-jan-2018-2238-utc.html

Various plotting methods have been applied to the signals, most results seem to point to the area around Chicago, Illinois.

Most of the reports seem to be in the ham bands, however that might just be a factor of how many people watch the ham bands at any given moment. To the best of my knowledge, the frequency ranges that have been observed and reported have been:

6320 to 6510 kHz
7095 to 7300 kHz
8100 to 8400 kHz
10100 to 10150 kHz
12200 to 12450 kHz
14060 to 14350 kHz
18065 to 18160 kHz
21200 to 21300 kHz

You can see it has at least 3 full bands that are well outside the ham bands, and a few frequencies just outside ham bands. There absolutely may be more ranges outside ham bands that are simply not being reported because no one has noticed. The reports of operation inside ham bands vastly outnumber the reports of operations outside ham bands, however I am not sure if this is an actual trend, or if it is just a result of the source of data (mostly hams). I know watching the signal today it seemed to spend as much time outside the ham bands as inside.

Various sources have tried, or tentatively, connected these transmissions with High Frequency Trading (HFT) experiments. Some specific HFT experimental licenses have been suggested as related to these transmissions, the following license are sometimes quoted:
WI2XNX
WJ2XGD
WH2XWU
WH2XVO

However, none of these license cover all of the frequencies observed. In fact, all of the authorized frequencies for all of those licenses combined still do not cover all of the frequencies observed.

Further, the HFT application of radio links is in a quest to gain advantages of milliseconds, if not microseconds, on trading transactions. So even if the pulses observed did contain data (and they don't appear to) they would update that data only once per second. It seems, to me, an unlikely fit.

It has been suggested that these pulses have nothing to do with the data of HFT itself, but are just tools to measure or model propagation delays for future HFT ventures. While certainly possible (the signal does look like a propagation / scatter experiment to me) the frequency ranges observed are pretty narrow, and such measurements generally do not require stepping across such narrow frequency ranges. Is moving 10 kHz every 2 or 3 minutes really going to tell you much about the real time propagation variations between the frequencies?

So, what is this 1 PPS oddity? I really don't know. But the more I look at it the less convinced I am that it has anything to do with HFT experiments, or any of the license so far identified. Can anyone convince me differently?

T!

 

[Token]

Is the signal just an unmodulated carrier that goes on and off about once per second? I've found what looks like an artifact of a building alarm system that does this throughout the 40m amateur band. I tracked one of them down to a random building in an industrial park with large warehouses. I had one close to my work that was also in an industrial park.

A friend just installed a remote controlled HF radio in, you guessed it, an industrial park and I see the tell tale once per second on off keying on several frequencies across the 40m band. This place is in the high desert of So Cal where the other sites I found were more in the suburban areas of Los Angeles.

This is a pulsed, unmodulated CW, signal. The pulse lengths appear to be multiples of 8 msec (today was all 8 msec, a few months ago was 16 msec, I have heard of, but not received myself, other multiples of 8 msec). The pulses are sent exactly (as close as can be determined) once per second and apparently in sync with GPS / UTC time zero. By this I mean the signal comes on for 8 msec of unmodulated CW, turns off, and comes back on 0.992 seconds later, rinse and repeat. The pulse has a sharp rise and fall time, and a flat top, it is nicely rectangular.

The signals are heard across the US easily and also occasionally in Europe.

Various methods of geolocation have established the source is somewhere near Chicago, IL.

This signal is frequency stable, it picks a freq and that is what it uses for a while, then it changes to a new freq after variable times. The frequencies almost always end in 0 or 5 kHz. The known ranges of operation are as my first post in this thread, but it would not surprise me if there were not more bands that have not yet been discovered or reported.

This is almost certainly not an unintentional signal. There has been speculation that it is some kind of ionospheric test or measurement tool, and possibly associated with the HFT licensees I mentioned, although I am not really convinced of the later.

T!

 

[RFI-EMI-GUY]

This is a pulsed, unmodulated CW, signal. The pulse lengths appear to be multiples of 8 msec (today was all 8 msec, a few months ago was 16 msec, I have heard of, but not received myself, other multiples of 8 msec). The pulses are sent exactly (as close as can be determined) once per second and apparently in sync with GPS / UTC time zero. By this I mean the signal comes on for 8 msec of unmodulated CW, turns off, and comes back on 0.992 seconds later, rinse and repeat. The pulse has a sharp rise and fall time, and a flat top, it is nicely rectangular.

The signals are heard across the US easily and also occasionally in Europe.

Various methods of geolocation have established the source is somewhere near Chicago, IL.

This signal is frequency stable, it picks a freq and that is what it uses for a while, then it changes to a new freq after variable times. The frequencies almost always end in 0 or 5 kHz. The known ranges of operation are as my first post in this thread, but it would not surprise me if there were not more bands that have not yet been discovered or reported.

This is almost certainly not an unintentional signal. There has been speculation that it is some kind of ionospheric test or measurement tool, and possibly associated with the HFT licensees I mentioned, although I am not really convinced of the later.

T!

There are indeed some companies in Chicago experimenting with high speed data links for stock trading. One is using HF bands. This could be some sort of beacon used to measure propagation delay.

Sent from my SM-T350 using Tapatalk

 

[prcguy]

I double checked my friends new remote HF rig in the CA high desert and one of the signals has about a two second period with 50% duty cycle, about one second on, one second off and it repeats. There are several right now in the lower end of 40m between 7.0 and about 7.178Mhz.

I have personally DF'd similar signals to warehouse type buildings, so these are different than what you described.

This is a pulsed, unmodulated CW, signal. The pulse lengths appear to be multiples of 8 msec (today was all 8 msec, a few months ago was 16 msec, I have heard of, but not received myself, other multiples of 8 msec). The pulses are sent exactly (as close as can be determined) once per second and apparently in sync with GPS / UTC time zero. By this I mean the signal comes on for 8 msec of unmodulated CW, turns off, and comes back on 0.992 seconds later, rinse and repeat. The pulse has a sharp rise and fall time, and a flat top, it is nicely rectangular.

The signals are heard across the US easily and also occasionally in Europe.

Various methods of geolocation have established the source is somewhere near Chicago, IL.

This signal is frequency stable, it picks a freq and that is what it uses for a while, then it changes to a new freq after variable times. The frequencies almost always end in 0 or 5 kHz. The known ranges of operation are as my first post in this thread, but it would not surprise me if there were not more bands that have not yet been discovered or reported.

This is almost certainly not an unintentional signal. There has been speculation that it is some kind of ionospheric test or measurement tool, and possibly associated with the HFT licensees I mentioned, although I am not really convinced of the later.

T!

 

[Boombox]

Dumb question time: Why would a high frequency trading network use HF? Isn't the internet more secure, and just as fast (if not faster, depending on propagation characteristics)?

What would be the point?

 

[RFI-EMI-GUY]

Dumb question time: Why would a high frequency trading network use HF? Isn't the internet more secure, and just as fast (if not faster, depending on propagation characteristics)?

What would be the point?

The internet is not secure at all and is the slowest form of propagation. Airwaves go through free space and travel at the speed of light through the atmosphere.

While fiber optic and copper cables have a slow velocity factor and microwave links require repeaters and filters that slow propagation, HF travels unimpeded .

Sent from my SM-T350 using Tapatalk

 

[Token]

What RFI-EMI-Guy said. Radio propagation is faster than the Internet ever will be (unless the net is over RF), even when the same communications standards / techniques are used.

Not only do fiber / copper cable operations suffer the velocity factor issue, there is always overhead (time) to any type of packetized communication. So the internet takes a slow medium and makes it slower yet.

Typical multimode fiber has a velocity factor of about 0.65, and single mode fiber can be slower yet. Using this number, over a 100 km link the RF link would take about .00033 seconds to travel the distance, while the fiber link would take 0.000505 sec to cover the same distance. Or the RF would achieve the same link ~170 microseconds faster. Or the RF link would be 35% faster than the fiber link.

This speed difference matters to the HFT folks.

T!

 

[Boombox]

So what will they do when the propagation is down? Time is money in stock trading, after all.

 

[Token]

So what will they do when the propagation is down? Time is money in stock trading, after all.

Worst case they will have to resort to more traditional techniques, such as the Internet or whatever they have been using before this, during such times. However such instances should be few and far between.

With a well engineered system, intelligent selection of frequencies, transmitter / receiver / antenna site, transmitter power, antenna pattern, and digital techniques that meet the communications speed goals with robust performance at low SNRs, such times of no propagation should be extremely limited. Probably talking about short duration events like solar flares / geomagnetic storms only.

During such times the HFT folks would have to revert to less advantageous links or systems.

But for the rest of the time they can enjoy their speed advantage over folks who don't lease time on such systems.

T!

 

[RFI-EMI-GUY]

I hope they have researched long delayed echoes! It would be a shame if they churn a stock up and then buy it over and over again at the inflated price, on a delayed echoe!



Sent from my SM-T350 using Tapatalk