High speed CW meteor scatter (or HSMS) has been popular in Europe for more than 20 years. The traditional station setup was a high speed memory keyer and a modified tape recorder (MTR). In the last couple of years the more serious HSMS'ers in Europe have switched to the German-made Digital Tape Recorder (DTR). Recently HSMS was reintroduced in the USA. The 9A4GL Computer program MSDSP became quickly the most popular receiving and transmitting device for the American stations. The current situation is that most European stations use a dedicated device (DTR or MTR) for slowing down HSCW, while the American stations use a computer / soundboard and MSDSP software. The average speed in Europe for most skeds is 2000 lpm, while in the US most stations use 4000-6000 lpm. This difference in equipment and speed has raised some interesting questions.
-What is the difference in S/N ratio between the various types of equipment used for slowing down HSCW?
-How do the DTR and MTR compare during an actual meteor scatter sked to the computer-based program MSDSP?
The aim of this paper was to try to give some meaningful answers to these questions.
Equipment used in the various tests.
Modified tape recorder. (MTR)
The MTR used in these tests was a Radio Shack portable tape recorder (CTR 69, model # 14-1154). The speed was varied by a single switch connected to the internal motor speed control, making a 15x reduction possible. This switch changed the voltage supply to the motor. During receive the record and fast forward button were depressed simultaneously, giving about two revolutions /second tape speed. No up-converter or other audio devices were used on playback. More information on the MTR can be found on the N1BUG Web site (http://www.mint.net/~n1bug/tech/hsms-cir.html) and in the documentation of MS-SOFT (http://www.sci.fi/~oh5iy/index.htm#Latest%20software%20&%20data).
Digital Tape Recorder. (DTR)
The DTR is a German-designed portable digital recording instrument dedicated to HSMS. It was developed by Dithmar Daude (DF7KF) and is only available in Europe. The DTR is housed in a small aluminum box (170 x 100 x 45 mm). Inside the DTR are two printed circuit boards. On the front panel is a digital display and 7 pushbuttons. The first four buttons act much like a tape recorder (REW, PLAY, FF, REC). The last three buttons are for marking the ping, setting the speed reduction and adjusting the tone from the upconverter. On the left side are an audio input and output plug, the output volume pot, and a switch that allows simultaneous listening to the input on one channel and the slowed down recorded information on the other channel. (This can be very handy as one can record and playback at the same time with the DTR.) On the right side of the unit is a RS-232 plug and three switches. The RS-232 plug has 8 outputs that can be connected to a computer running MSSOFT for transmitting. The other outputs are for the transceiver PTT control, an external mark switch (eg. a foot switch), and the keyed (500 Hz or 2300 Hz) audio output for injection into the transceivers microphone. The three switches allow for selection of automatic recording when the keying stops, automatic PTT control, and the selection of the audio injection tone (500 Hz or 2300 Hz). The DTR is developed to work together very closely with the MSSOFT keyer software. It contains an RS-232 interface and an audio injection generator. Used together with MSSOFT, it provides a similar sophisticated receive and transmit MS unit much like MSDSP. The technical information and the complete owners manual can be found at the following address http://qsl.net/dk3xt/dtr.htm
Most information on MSDSP and its technical specifications can be found in several papers at the HSMS web site (http://www.nitehawk.com/rasmit/ws1_15.html).
For these tests MSDSP V. 0.51 was used at a sampling rate of 22050 Hz. (Version 0.70 is now used by most operators. This program was written by Tihomar Heidelberg, 9A4GL, and is available on the Web as shareware)
Other equipment used:
To record the same ping on the DTR, MTR and MSDSP, an interface box was used that split the signal equally among the units. All tests were analyzed with headphones, and care was taken to provide similar conditions when comparing each unit. For each recording unit the optimal conditions were set for analyzing the slowed down ping. At least five separate pings were used for each speed level. Tests were performed at 2000, 4000, 6000, 8000 and 10000 lpm. All pings selected were short, <1 second (underdense pings) and of various strengths.
To compare with MSDSP the ability of the MTR and DTR to handle the varied speed levels, tests were done analyzing the same ping and comparing the number of characters copied with the MTR, DTR and MSDSP. For each speed a minimum of 5 pings were analyzed. All pings were shorter than 1 second and were of varying strength. The greatest number of tests (>20) were performed at 4000 lpm comparing the DTR vs. MSDSP.
The number of characters copied with the MTR at 2000 lpm was not as good as for MSDSP (see table). On average, 3-5 more dits or dahs were copied with MSDSP. This accounted generally for a letter or part of a number. The code on the MTR was very clear and easy readable. When the signal levels dropped, as at the end of a ping, the ability of the MTR to pull out the weak code was significantly less.
Comparisons of the DTR versus MSDSP (see table) shows that both DTR and MSDSP performed equally well up to 6000 lpm. Above 6000 lpm, MSDSP did copy 3-5 more dits or dahs than the DTR. This difference was noticeable at the end of a ping when signal levels dropped off. At 10000 lpm, MSDSP was still very useful while the DTR had significant problems. The slowed-down code on the DTR sounded extremely distorted and the Morse code became very difficult to read.
Comparison of MTR and DTR versus MSDSP
Discussion of results:
MTR vs. MSDSP
The MTR (modified tape recorder) is useful up to about 2000 lpm. Its S/N ratio is slightly worse than MSDSP, giving MSDSP a clear edge even at this relatively low speed. The lack of an audio upconverter in the MTR might have contributed to this, as MSDSP was able to copy one or two more characters at the beginning and the end of a ping when signal levels were the weakest. No test were done with the MTR at higher speeds as the MTR could not slow down the recording to a manageable code speed.
MSDSP was a clear winner with better audio quality and better ease of use. The MTR is cumbersome and lacks the ability to mark a ping and find it back easily. The counter on the MTR was not as precise as the marking used for MSDSP. It was difficult to analyze more than one ping with the MTR in a one minute period. However, for simplicity (if one does not own a computer [386 or higher]), the MTR cannot be beaten. The MTR used cost less than $30.00 and was converted in about 30 minutes for HSCW use. A quicker way to become active on HSMS cannot be found. It is surprising that so very few American stations have not used this method as it is very effective and its poorer S/N ratio than MSDSP is most likely not significant for the outcome of an actual sked. Hams that want to become active on HSMS but do not own a computer that can run MSDSP should consider this very simple and proven alternative. The MTR is also a great unit for portable use. Its real limitations are the more cumbersome handling and its maximum speed. The MTR was not compared to the DTR, but similar S/N results should be expected.
DTR vs. MSDSP
The results show that the S/N ratio of the DTR and MSDSP are equal up to about 6000 lpm. At 10000 lpm the code on the DTR became difficult to read, and MSDSP handled this speed better. It was rather surprising that the DTR still managed at this speed, as the sampling rate of the DTR (6 kHz) is significantly lower than for MSDSP (20 kHz or higher). The DTR was not tested at higher speeds as the code became too distorted and unreadable at speeds above 10000 lpm. Of greatest interest to the American amateurs is the performance of the DTR at the common US speed levels (2000-6000 lpm).
The DTR has 10 fixed speed reduction ratios. At 4000 lpm the x30 reduction is most useful and brings the code down to about 26 wpm. The 50x reduction at 4000 lpm made the code very slow, < 12 wpm, and sometimes difficult to read. At 6000 lpm the 30x or 50x reduction can be used and the 50x reduction gives acceptable code. With MSDSP, one can set the speed reduction at any level, making it somewhat easier to set the decoding speed to one's own code ability. If your maximum code ability is less than 20 wpm, the DTR will give you problems at speeds above 4000 lpm.
The speed-reduced audio sounded different between the two units. On the DTR one can choose between three levels of upconverting, and in these tests the speed-correlated up converting (PL02) and 8x upconverting (PL03) were most useful and brought the tone pitch to about 200-300 Hz. On both settings the code was perfectly readable with a slightly greater amount of ringing than when analyzed on MSDSP. The upconverter on MSDSP has more settings, giving the operator more flexibility to set its own preference.
MSDSP uses a visual display to mark the recorded information and save it for later analysis. The DTR has no visual display, and a ping is marked with a single button. Thirty separate marks can be placed. Finding a marked ping is very easy on the DTR, and one can analyze 5 separate pings in a one minute period and still have time left. MSDSP's visual display makes marking a ping easier, as the ping can be marked either by the spacebar or with the mouse and saved to one of the buffers. MSDSP allows the operator extra time up to the full 1 minute period to mark and save the ping. Marking a ping is more critical with the DTR. The weaker and shorter the ping, the more important it is to mark the beginning of the ping. Any delay of marking the ping makes it more difficult to find it later. MSDSP gives the operator more time to save the ping, and on replay one can visually examine the audio around the ping and mark the ping for play back. This is not possible with the DTR. You only have one chance during the recording to mark the beginning of a ping with the DTR. (The extra work involved with the mouse makes MSDSP slightly slower for analyzing pings during the transmit period, although an experienced operator can probably analyze just as many pings in a one minute period as is possible with the DTR). The DTR has no ability to save the received information. The maximum recording time is 155 seconds. The DTR either stops recording or can function in a closed recording loop. Thus all previously recorded information is eventually lost during a sked. With MSDSP one can save a reflection to a buffer and the saved recording will not be lost (as long as the program does not crash!) One can also save the ping in a WAV file. This is a clear advantage of MSDSP. The ease of use with the visual display and the less critical marking of a ping gives MSDSP a slight advantage. (Although when using the DTR this was not seen as a handicap, and it did not effect the actual outcome of any sked.) An advantage of the DTR is its ability to play back a ping while continuing to record. The DTR has a switch that makes possible the simultaneous listening to the slowed down recorded ping and also the receiver input. While decoding a ping, one can still mark new pings. This enables the DTR to analyze the recorded information during the receive period. MSDSP does not have such a feature.
The size of the DTR unit makes it ideal for portable use. It runs on 12 V and can easily be operated in the dark. For the DTR one will need a separate high speed memory keyer. The OH5IY program, MSSOFT, running on a simple computer is the ideal set up. The DTR actually is designed to work closely with the MS keyer in MSSOFT. The computer can be connected directly to the DTR's RS-232 interface. The DTR contains the RS-232 interface and the 2300 Hz (or 500 Hz) audio generator. The keyed audio tone, generated with the DTR, can be injected into the transceiver microphone input just like when using MSDSP. Unlike MSDSP, the keyed audio tone does not require any additional filtering. The unit allows selection of automatic recording when the keyer stops. It also allows automatic control of the PTT if VOX control is not available on the transceiver. If one decides to use MSSOFT for keying, one will need a separate computer. MSSOFT will run on practically any computer under DOS. The combination DTR-MSSOFT will give a similar excellent MS keyer as found in MSDSP. MSDSP will run on a 386 or higher computer, but will need a genuine Creative Labs Sound Blaster stereo card. However, the latest version can run on some other cards, also (especially computers using the ESS Technology chip set). To use MSDSP, one should build a very simple interface box to connect the computer audio ports to the transceiver (Schematics of the audio interface box can be found on the W6/PA0ZN Web site, http://www.nitehawk.com/rasmit/ws1_15.html). It is rare to get MSDSP running immediately. It sometimes requires adjustments of the DOS settings. It is not unconditionally stable and does crash unexpectedly on occasions, thought the latest version (V0.70) seldom crashes. The DTR is much easier to set up, does not require any interface box or building skills, and is very stable.
At this moment the DTR is only available in Europe. The author picked up his unit in Holland. There was no waiting time and the service was excellent. The cost of the unit was about $350-400, depending on the exchange rate. MSDSP is easily available from the Internet as fully-operational shareware (but read the "Manual" or "Tips" before trying to set it up). For MSDSP, one will need a computer and a 100%-compatible SoundBlaster sound card, and one should build a simple interface box. The DTR is a very complete unit and can run with a simple memory keyer. If one chooses to run MSSOFT as a MS keyer, an old 286 computer will do fine. No interface box is required as the RS-232 interface and audio injection generator are all build into the DTR.Cost wise to get on the air with either method will be similar, unless one owns already a (386 or higher) computer.
The MTR, DTR and MSDSP all are very capable in recording and slowing down high speed CW. The DTR and MSDSP are both highly sophisticated and are currently the state of the art in HSMS. Both have their advantages and disadvantages. Cost-wise, the MTR is an excellent way to start with HSMS. If one owns a computer, MSDSP is the simplest and least expensive way to become active on HSMS. However, if one plans to buy a computer exclusively for HSMS, then the DTR could be a great alternative. The author decided to purchase the DTR for portable MS-work. The DTR as a portable unit easily beats the cost of purchasing a portable notebook computer to run MSDSP. The DTR is definitely suitable for the US market and hopefully the unit, possibly with a few changes, will become available here soon. For the future, a dedicated recording device (like the DTR) controlled by a program like MSDSP would produce an unbeatable HSMS unit.
I would like to thank Shelby Ennis (W8WN) and Steve Harrison (KO0U) for their support and help in testing the DTR and review of this paper.
73's Maarten Broess, W1FIG (ex PE1FIG, KD1DZ)
Rhode Island, USA
1998 October 20