Shelby Ennis, W8WN

NOTE - This is a preprint of the article published in QST, November 2000, pp. 28-32. It is not identical, but very similar. This copy was placed on the Web to allow a number of other operators to check it before it went to print. While not identical, it should contain all of the information in that article. There are no graphics with this copy, but illustrations similar to the ones found in QST can be found on various pages of the W8WN Web site. Also, there is no bibliography in this copy, but it can be found at the end of the Rich-Text Format version.


"Your mission, Mr. Phelps, if you choose to accept it, is to complete a contact on 144 MHz over a 600-1000 mile (960-1610 km) path in less than 20 minutes every morning of the year, meanwhile developing techniques so that others can do the same thing. As always, if you fail, the Secretary will disavow any knowledge of you or your activities. It is suggested that you recruit a top-notch team for this operation. This tape will self-destruct in 15 seconds. Good luck."

Report: For Eyes Only:

"Mr. Secretary - After a slow start, we are now approaching the goal. Between January 1999 and April 2000, 172 out of 215 144 MHz schedules were completed with KØXP for an 80% completion rate, usually within 15 minutes, over an 813-mile (1308 km) path, which included schedules during the poorest time of the year. During the period of February through April 2000, 21 of 22 attempts were also completed with K1JT on 144 MHz, and 11 of 11 were completed on 50 MHz over that 650 mile (1050 km) path. Details to follow."

While the above might sound like an opening scene from "Mission Impossible," it actually is a brief summary of what has been happening during the past three years. 144 MHz contacts with these stations (and between a number of others) have become so routine that we are now more surprised when a schedule is not successfully completed. Satellites are not used, nor some exotic mode of propagation. Rather, it is simply a continuation of what Hams developed back in the 1950ís - meteor scatter.

A brief history of the time of meteor scatter

Meteor scatter operation began in 1953 when Paul Wilson, W4HHK (in western Tennessee) and Ross Bateman, W4AO (NE Virginia) kept hearing bursts of signal while trying to work during a widespread tropospheric opening. They, along with W2UK, W5RCI, W2NLY, W2AZL, W1HDQ, W1FZJ, and others were soon running tests, establishing how communication could be done using this mode., With the publication of two articles in QST by Walt Bain, W4LTU,, meteor scatter soon became a popular mode for making contacts beyond the normal extended-tropo range. During the annual Perseids meteor shower, even though stations were well spread out, enough were active that QRM became a problem at times. (Everyone was crystal controlled, and there was no way to use a "calling frequency" in those days). Using various means of keying and two- or three-speed reel-to-reel tape recorders, some managed to operate at speeds up to 100 wpm, slowing the tape for copying.

Most of the time the "pings" were few and short. As SSB operation became more common on VHF in the 1970ís, North American Hams saw an opportunity to push across a greater amount of information in the same amount of time. Using SSB, it was possible to exchange information as rapidly as even the best operators, using multiple speed tape recorders, could on CW.

But SSB operation still required pings of one second or longer to be really useful. On 144 MHz, these seldom occurred except during the peaks of major meteor showers. (On 50 MHz, where pings are longer, SSB MS contacts can be made routinely by well equipped stations.) So nearly all North American 144-MHz MS operation occurred only during the "big three" meteor showers each year (the August Perseids, December Geminids and January Quadrantids).

Meanwhile, the Europeans had a different idea. SSB meteor scatter operation was still much too slow to utilize the barrage of little meteors that constantly strike the earthís protective atmosphere. Their pings, while numerous, are usually weak and very short, being caused by the underdense ionization of the tiny particles. The European idea was to transmit CW at much higher speeds using an electronic keyer, record incoming pings on a modified audio cassette recorder, then slow the pings down to a readable speed. This was a brilliant solution. Using this method, speeds up to 400 wpm (or 2000 lpm; lpm, letters per minute; lpm=wpm x 5) or faster were quickly achieved. European operators soon were routinely making contacts every morning of the year, and logging dozens during meteor showers.

Even though the modification to an audio cassette recorder was simple, this style of meteor scatter operation (now called HSCW or HSMS - high speed CW or high speed meteor scatter) did not cross the Atlantic. North American Hams somehow felt that SSB was superior to any type of CW. Several operators attempted to create interest in HSCW in the Western Hemisphere, but few if any contacts were made. Itís difficult to make a contact when youíre the only one operating!

HSCW MS finally crosses the Atlantic

HSCW in North America started almost by accident when, in May 1997, Steve, KOØU (now KØXP) and W8WN learned that each liked CW MS operation. At this same time, DL3JINís SBMS ("Sound Blaster Meteor Scatter") receiving program, which allowed a computer to emulate a variable-speed tape recorder, was discovered. Several schedules were run at speeds up to 80 wpm using a programmable keyer or OH5IYís "MS-Soft" program. W8WN was using SBMS to assist with receiving, while KOØU copied by ear at speed. (SBMS is an excellent program, yet it never caught on either in Europe or North America, for reasons unknown.)

Early in August of that year, 9A4GL, Tihomir Heidelburg, a college student in Croatia, released the first Beta version of his HSCW receiving program, MS_DSP ("Meteor Scatter using DSP"). It was not as well developed as DL3JINís program, but it had several additional features that showed promise. E-mail messages began to fly back and forth across the Atlantic as Tihomir sent us version after version of MSDSP to test, eventually adding nearly every feature that we requested. Before the peak of the Perseids that year, a test version with transmit capability was available, and speeds jumped to 2000 lpm (400 wpm). Things suddenly became interesting! Other stations learned of the HSCW experiments and began to join the fun. Routine speeds soon were up to 4000 lpm (800 wpm), with one contact between Valerie, WD8KVD (visiting in EM77), and KOØU (FN42), having been made at 8600 lpm (1720 wpm), the highest speed then possible. The next year (1998), while again visiting in Kentucky for Christmas, WD8KVD and KOØU made a contact on Christmas Day at the unheard-of speed of 16,600 lpm (3320 wpm)!

Some operators had trouble using the DOS MSDSP program, as many were familiar only with Windows. In 1999 Tihomir released his first Windows version, WinMSDSP 2000. With more features and capabilities, it was quickly downloaded by a number of VHF operators around the world.

During 1997 and 1998, a group had also been testing various techniques for HSCW MS operation. It soon became apparent that the procedures in use for slow CW or SSB meteor scatter were inadequate for HSCW. The Europeans, with their 20 years of HSCW experience, had developed many additional ideas. However, some of their practices were different from ours, and itís hard to suddenly change 40 years of operating experience. Many of the schedules between W8WN and KOØU were devoted to testing various procedures, speeds, techniques, and equipment settings. This is one reason that the percentage of completed contacts did not increase as rapidly as would be expected from the increasing speeds. (For more data on these test schedules, see the "Archive News" page.)

Characteristics of HSCW MS operation

HSCW MS operation in North America is now established, with speeds of 4000-10,000 lpm (800-2000 wpm) routinely used. (All information here pertains to 144 MHz operation unless otherwise specified, as this is the most-used band for all types of MS operation.) Most operation is by schedule, with few routine CQís except during the annual North American HSMS Contest, which coincides with the Southern Hemisphere Eta Aquarids meteor shower during the first week of May. This is because the number of North American stations capable of HSCW operation is still too few to provide many random contacts. (The HSCW calling frequency is 144.100 in North America.) Most schedules on this side of the Atlantic are made using the "MS Rocks Live!" real-time Web page (often referred to as "Hot Rocks"), or via the HSCW Reflector. A station equipped for weak signal operation on 144 MHz (150 watts or more, 16-element Yagi, and a decent VHF weak-signal location) has a good chance of completing a contact nearly any morning of the year using the underdense pings of sporadic meteors, if someone at a suitable distance is available for a schedule.

The equipment needed for HSCW MS operation is the same as is found in a typical VHF shack; a multi-mode transceiver, amplifier, horizontal beam antenna, and computer running Windows 95/ 98. (There are a few other methods of operating HSCW MS besides using MSDSP or a modified cassette recorder. For more on these other methods, go to any of the HSCW Web sites and check there, then follow the links). If you can work out well on auroras, tropo, and other weak-signal modes of propagation, you should be able to do well on HSCW MS. For distances greater than about 1250 miles (2000 km), a good location and high antenna are needed. However, for distances under 800 miles, the optimum take-off angle begins to rise. Thus, a low antenna (in the clear) is quite usable for medium distances of 600-1200 miles (960-1900 km). A quiet location is a definite asset, obviously. No modifications are needed to a standard SSB rig. Keying is done by an injected audio tone (as is done with many of the digital modes), and standard SSB filters work fine up to about 10,000 lpm (2000 wpm). The emission type is designated J2A. This method produces keying that is indistinguishable from on/off keying of the main carrier, and is the same method used by many rigs to produce CW. Using standard SSB filters, the bandwidth of an HSCW signal is about the same as that of a voice transmission. For more technical information, see the numerous papers on the several HSCW Web sites13 and the lists of resources in the article by Jim McMasters, KD5BUR (now KM5PO).

HSCW MS operation is decidedly different from either slow CW or SSB meteor scatter operation. Both slow CW and SSB operation require overdense bursts or good (i.e., strong and long) underdense pings to complete a contact. SSB and slow CW operators hope for specular reflections from heavily ionized trails instead of "the abominable Ďping,'" as one writer put it.6 Unfortunately, these overdense bursts are seldom observed except during major shower periods. HSCW relies on those very short (less than 1 second), weak pings scattered from underdense trains. Pings of this type are often produced by the "sporadic" meteors that bombard the earth constantly. Most of them are not fragments from the asteroid belt but are particles from dust trails left by ancient comets. Now they are widely distributed and no longer dense enough to produce recognizable showers. Their number may vary heavily day by day and even minute by minute. Most of these particles are tiny, no larger than grains of sand, or less. But because of their extreme speed, the ionization they produce as they enter the atmosphere is often enough to scatter or even refract a radio wave.6,

Even though the duration of these pings is usually very short, the number of tiny pings available on many mornings may surprise you. It has been estimated that if you have a 5% chance of completing a contact on SSB, you have a 95% chance on HSCW. Of the approximately 215 completed HSCW contacts between W8WN and KOØU/KØXP, no more than 10 contacts could have been accomplished on SSB. Nearly all of those would have been during showers, when SSB MS comes into its own.

Meteor scatter (using any mode) is difficult at distances shorter than about 500 miles (800 km) or greater than about 1400 miles (2250 km) due to the height of the meteor trails, antenna characteristics, the scattering mechanism, etc.16 For communication at distances under 500 miles, back- or side-scatter or elevated antenna headings are usually needed. At distances beyond 1400 miles, a good location and high power are necessary. While difficult, both are possible during showers.

Because HSCW relies entirely on the tiny amount of scattered signal from underdense meteor trains, the typical sporadic "ping" will be weak and very short. In contrast, the intense ionization of an overdense burst is much more efficient at refracting a VHF signal. However, as noted previously, these overdense bursts are rare and cannot be counted on except possibly during the peak of a major shower. Higher output power is more important for utilization of the underdense pings than it is for the overdense bursts. So, more power to Ďya!

Does this mean that a 150-watt station cannot use sporadic underdense pings? In order to see what could be done, arrangements were made with our daughter, WD8KVD, to operate portable from her home near Duluth, Minnesota during a visit in July 1999. An IC-706 MkII drove a 150-watt amplifier, and an 11-element Yagi was mounted on a telescoping paint pole with a maximum height of about 20 feet (6 meters). The location was only fair for VHF. A used laptop computer ran either the Windows or DOS version of MSDSP. The purpose was to see how difficult this type of operation would be, and also to provide a new grid (EN46) for the operators. Compared with the kilowatt and large array at home in Kentucky, contacts were more difficult, of course; but a number of fellows were able to add a new grid to their logs. One morning was spent beside the road on a nearby hilltop. Even though the location was better, the meteors did not cooperate, and only one contact was made that day. However, many mosquitoes (the state bird of Minnesota) had an extra meal.

The portable operation (indoors only) was repeated at Christmas, 1999, from our sonís home near Clio, Michigan (EN83). Using the same equipment but with very flat terrain, contacts between 500 and 1000 miles were easily accomplished. (Details, with a number of photos from both operations, are available on the W8WN Web site.)

Since then, K9KNW/MM has completed a number HSCW contacts from his boat, running either a halo or a small beam. During May, 2000, K9KNW completed 28 HSCW contacts while sailing in seven water grids (EL93, FL 03, 13, 14, 15, 23 and 24). His 12-element Yagi was only about 3 meters above the water, limiting his maximum distance to about 1250 miles (2000 km). Stations less than 1100 miles (1775 km) distant found the contacts to be very easy, usually taking less than 20 minutes. Those at greater distances were more difficult. Joe has plans for more trips, including the possibility of a grid-hopping trip with much more time devoted to Amateur Radio.

How fast is fast?

Obviously, the higher the speed of the transmission, the more information can be packed into each ping. MS speeds originally were 25-35 wpm (this is still the standard speed for slow CW operation). A number of the early operators could operate at 50 wpm, copying in their heads. (Yes, Iíve heard some; and if the tapes havenít deteriorated, I still have a few of those pings on 1/4" audio tape.) SSB brought slightly higher speeds to nearly everyone, but with the added requirement of needing to be exactly on frequency to be readable. When the Europeans developed HSCW, routine speeds increased to 2000 lpm (400 wpm) or faster. North American HSCW started at 1500 lpm and quickly went to 4000 lpm (800 wpm). The current version of WinMSDSP is capable of speeds up to 20,000 lpm (4000 wpm). But are these extreme speeds practicable? What is the maximum usable speed?

While a number of contacts have been made at speeds up to 18,600 lpm, the normal limit at this time seems to be about 12,000 lpm (2400 wpm). Using unmodified SSB transmitters and receivers with their standard SSB filters and keying by injecting a keyed audio tone into the mike or data port, speeds up to about 10,000 lpm have proved to be no more difficult than the slower speeds. However, at higher speeds, several problems arise. First, the signal-to-noise ratio seems to become poorer, making very weak pings unusable. Second, the keying begins to sound "soft" and difficult to copy. Using a 2000-Hz injection tone and receiving with a 1500-Hz tone, a dit may not even get a full cycle. Speeds faster than 10,000 lpm are definitely possible, but the QSO efficiency drops - and this defeats the main purpose of HSCW. By using wider filters and higher tones, it is possible that faster speeds could be used. None of the MSDSP test stations had this capability, and no one has had the ambition to purchase new filters or to modify their rigs. So at this time North American speeds are typically 4000-10,000 lpm.

Other bands

Little HSCW MS operation in North America has been attempted on the other VHF bands for several reasons. First, most MS activity today is on 144 MHz. And second, on the other bands itís either difficult or not really needed.

50 MHz - (calling frequency 50.300) - Pings tend to be weaker but longer, while the number of pings is somewhat greater than on 144 MHz. The lower gain antennas and lower power typically used on 6 meters are apparently the reason for the weaker signals. On 6, the pings average about 1 second in length with an occasional ping lasting up to 5 seconds or longer. This is why SSB MS is possible many mornings on 6 meters by well-equipped stations. Also, with Es, F2, tropo, and other modes of propagation available, grids can eventually be worked using these other modes. Surprisingly, HSCW MS has not been as easy on 50 MHz as had been expected because of the weak signals. Most operators have found 144 MHz to be easier. For these reasons, although a number of stations have made contacts on 50 MHz, HSCW is not common on this band.

However, some interesting questions have been raised. It has been observed that the day-to-day rates, and even the minute-by-minute rates, change radically. Thus, comparisons between 50 and 144 MHz are not necessarily valid, even when made by the same pair of stations. Therefore, how would 6 and 2 compare on the same pings? What is needed are a pair of stations at a proper distance, one of whom can transmit on both bands simultaneously while the other receives and records on both. Any volunteers? And has anyone completed a double-hop MS contact on 6?

222 MHz - (no random operating; schedules only) - As the frequency increases, so does the difficulty.6,16 On 222, bursts can be strong but are fewer in number than on 144 MHz. One-hour schedules are typical between well-equipped stations. Little HSCW operation has been done on this band, but more is expected this year. In fact, the first known HSCW MS contact was made on May 2 between N7STU and NØKQY, while this article was being prepared. The next day a second 222 contact was completed, this time between N7STU and KØGU. N7STU was running 450 watts to a 7 wavelength Yagi, while KØGU had only 25 watts to a 22-element Yagi! Because of the few but strong (and longer than expected) "burns" that have been experienced, some of those who have operated MS on this band have wondered whether HSCW or fast-break SSB might be more effective. This depends upon conditions. For the peak of major showers, SSB should be the most effective mode, while HSCW has a great advantage at all other times.

432 MHz - (no random operating; schedules only) - This band is more difficult, of course, though possible. Most 432 MS activity has taken place in Europe (the SM3AKW - UA9FAD contact, at a reported distance of 2141 km, is believed to be the world 432 MS record; the North American record is held by N6RMJ and W7XU at 2036 km). Using large antennas and high power, the Europeans reportedly have been rather successful. Pings are few, as expected, and usually short and weak. However, a number of European operators say that they have been surprised by the length and strength of some of the bursts. Schedules are typically one or two hours during showers. They suggest that you should take advantage of elevating your antenna so that much of the ground noise is in the first null of your antenna pattern. Also, the notes about antenna aiming and using the "hot spot" should be carefully considered because of the much narrower beamwidth of 70 cm long Yagis.6,17

Higher frequencies - These bands are generally considered to be unlikely candidates for MS operation, although 902 MHz should be possible by well equipped stations during major showers. Who will make the first MS contact on this band?

Using WinMSDSP

Although modified cassette recorders and other means are used for HSCW operation, WinMSDSP has become the most common method. A limited run-time version (shareware) can be downloaded from the 9A4GL Web site and other locations. It requires a computer running W95 or W98 and supporting DirectX. (W95 and stripped-down versions of W98 may require the addition or updating of some Microsoft files; lists and URLís can be found in the MSDSP "Manual", also called the "Index" or "Help" file.) WinMSDSP should run with any full-duplex sound board that supports DirectX. Because of the demands on the audio board and the fact that their standards seem to be moving targets, there have been problems reported with a few boards. The "Manual" and a "Problems" paper have suggestions for working around any difficulties. The program is easy to use and requires only a few minutes to learn. However, it has many features. You really need to slowly go through the short Manual at least once as you set it up.

Many of the HSCW Web sites have a few sample pings that can be downloaded to see what different speeds and strengths are like. However, after a few minutes spent playing with the program and learning the main functions, further "practice" of this type is of little value. Now itís time to join the HSCW Reflector and request a sked. (The meteor-scatter reflector is often used in Europe.) There is nothing that can properly simulate actual HSCW operation. Make a set of interface cables (the same as required for PSK31, etc.). Several different interface boxes can be found in the "Manual" and on the W8WN and other Web sites, though they are not absolutely necessary. (See also May 2000 QST, p. 45.)

What can you expect? This depends upon your location, equipment, distance to the other station, his equipment, the time of year, the time of day, and maybe what your dog had for breakfast! For two smaller to medium size stations at a proper distance, you may get only a few pings during a 30-minute morning schedule, or as many as 4 or 5 pings per minute. Conditions can vary greatly. 4000 or 6000 lpm (800 or 1200 wpm) are good first-schedule speeds. (See the MSDSP "Manual", "Problems," and the "Procedures" for initial program settings, etc13.)

Does it work? The Europeans knock our socks off with the number of MS contacts they routinely make, because they run primarily HSCW (except during the peaks of major showers, when SSB comes into its own), while North America is stuck with SSB. Also, Europe has many more stations on HSCW (and on VHF DX generally) than there are in North America. The biggest disadvantage to HSCW in the Western Hemisphere is the same as with the other weak-signal modes - the "white noise" syndrome. There just are not enough stations on the band for many contacts.

Things to watch for

Once you start running schedules on HSCW, you may immediately notice some things that you had previously wondered about (assuming, of course, that you have been active on MS and other weak-signal DX work previously).

The first is how radically the number of pings varies month by month, day by day, and even minute by minute. The best time of year for sporadic meteors is the July-January period, with February-May being the poorest. Using HSCW means that it is possible to complete contacts almost any day of the year, though certain periods may be easier than others. The number of sporadic meteors reaches a maximum around 6 a.m. local time because the morning side of the earth is facing toward the direction of its orbital travel, thus sweeping up even slow meteors. Around 6 p.m. local time your location is now on the trailing side, so only those meteors fast enough to overtake the earth will be captured. Thus, MS is much easier in the morning due to the larger number of meteors entering the atmosphere. The exception to this may be during a meteor shower. However, do not make the mistake of attempting to use shower meteors when its radiant is still below your horizon. (Remember that there are also several daylight showers - listings are found on and also linked from the HSCW/MS "Hot News" Web page.)

If you have operated MS during one of the large peaks of a recent shower, you could hardly keep from noticing that even strong, long-duration signals were sometimes difficult to copy, especially on 144.200 SSB. The strengths appeared to vary greatly, and sometimes stations would seem to be vying for your receiverís attention, almost as in the "capture" of an FM signal. This is caused by several different phenomena and is seldom seen on an underdense ping, but is common on overdense bursts - see the various references for more information.

The first thing you may notice on a good schedule is how many pings there sometimes are, but that most are very, very short. On 144 MHz, a one-second or longer ping is the exception. Most are shorter than this, and many are much shorter. A lot are only 1/10th of a second or less, and may be so short that they are difficult to either hear or to see on the MSDSP screen. Because HSCW requires a ping of 1/10th second or longer even at 10,000 lpm (2000 wpm), only 1/5 to 1/2 of the pings are likely to contain usable information. But why are so many more pings heard when using HSCW than when just monitoring 144.200? Primarily because there is so much more usable signal available with HSCW.

Other interesting things to watch for are Doppler shift, doubles, ionospheric scatter, and - who knows? Doppler shift is not often observed on these tiny underdense pings. However, if you get Doppler on one ping during a schedule, you are more likely to have Doppler on another ping or two. Why? Another phenomenon you may notice is that on certain days (especially during certain showers) the pings may seem to occasionally come in pairs. At first glance, this would appear to mean that some meteors are traveling together, separated by a second or so. Meteor scientists have long contended that this is only a statistical fluke. However, Hams have noticed it for many years, and visual observers have recently been reporting it. The jury may still be out on this question. (It will be interesting to see what ideas the dust-trail Leonid predictions of Asher and McNaught eventually bring to this idea.) Finally, when two EME-class stations have had HSCW schedules, traces of ionospheric scatter have been reported on a number of occasions. So when youíre doing this type of operating, remember to be alert for more than just completing a contact or working a new grid. HSCW MS is much easier than other types of MS operation and has the advantage of visually displaying the pings, so you can more easily carry out other observations. MSDSP also gives you the ability to save any particularly interesting pings for later study. (If it all becomes too easy and you want a real challenge, see Maj. O. R. Disasterís collection of the works of that great wireless pioneer, Owa Taboo Byam.)

And finally....

If you are serious about VHF DX, you almost certainly have a multi-mode rig with an amplifier, a decent antenna, and a computer. Donít let the "CW" in the HSCW MS scare you away, for you can slow the received signal down to any reasonable speed. (In fact, using the Cool Edit audio editing program, several European Hams who could not copy CW at all by ear have successfully operated HSCW! This is also possible by using some of the new features of WinMSDSP, although it is not really advised and actually is much more difficult). But try HSCW MS - you may be surprised.

Thanks to all of those who have helped with this article and to those who have assisted getting HSCW MS started over here. Special thanks go to my wife Lora, WD8LPN, who has assisted with so much and put up with everything; to Val, WD8KVD, and to Steve and Alisca for the use of their homes when operating portable in Minnesota and Michigan; to Steve Harrison, KØXP, for three years of schedules while we tested all of the things listed above and for help with this article; to Maarten, W1FIG, and Joe, K1JT, for their tail-ending, giving a chance for another type of operating; to the MSDSP Alpha-test group for all of their work with the many versions of 9A4GLís program; to Ilkka, OH5IY, whose multi-part MS-Soft program is used by nearly every MS operator around the world; to Peter, DL3JIN, and Tihomir, 9A4GL, whose programming abilities and hard work started this modern age of HSCW operation; and to those HSCW operators, both in North America and Europe, who have helped with testing, ideas, and suggestions for the best ways to operate over here.

06/2000 -