MSI / 55 - an old terminal for ordering goods by a branch in a central store

The device shown on the KDPV was intended to automatically send orders from the branch to the central store. To do this, having previously entered the articles of goods ordered into it, call the number of the central store and send the data according to the principle of an acoustically mated modem. The speed at which the terminal sends data is estimated to be 300 baud. It is powered by four mercury-zinc elements (then it was possible), the voltage of such an element is 1.35 V, and the entire battery is 5.4 V, so everything worked from a 5 V power supply. Three modes can be selected using the switch: CALC - a normal calculator, OPER - you can enter numbers and other characters, and SEND - sending, but at first no sound was squeezed out. It is clear that you can somehow save the articles and then send them, but how? If you can find outthis program , or even somehow adapt the terminal to digital types of amateur communications.

The device is on the reverse side, the dynamic head and the battery compartment are visible:



The most important thing is how to squeeze the sound out of the terminal - the author learned from a person who once had the same terminal. You need to enter the initialization code, and then you can enter the articles. We put the switch in the OPER position, the letter P appears. Enter 0406091001 (the author does not explain what it is, probably the username) and press ENT. The letter H appears. Enter 001290 (and this is probably the password) and press ENT again. The number 0 appears. You can enter part numbers.

You need to start the article with the letter H or P (here the author was mistaken, there is no letter P on the keyboard, there is F), then the numbers follow. After pressing the ENT key, a line of the type 0004 0451 appears, where with each subsequent article the first number increases and the second decreases, which implies that this is the number of occupied and free cells, respectively. You can scroll through the entered articles with the arrow buttons, but the author does not know how to delete them (which means that the CLR key did not help). How to indicate the quantity for each article, is not said.

After entering the articles, you must then turn the switch to the SEND position and press the SND / = key. The message SEND BUSY will be displayed, and transmission will begin:



For 4.4 s, a tone with a frequency of 1200 Hz sounds. Then for another 6 s - 1000 Hz. The next 2.8 s go to the transmission of the modulated signal, and after them another 3 s - again to transmit a tone of 1000 Hz.

If you look closely at the spectrum, in fact, instead of 1000 Hz, you get 980, and instead of 1200 - 1180. The author wrote down the WAV file, installed the program mentioned above (“man” for it here ) and ran like this:

minimodem -r -f msi55_bell103_3.wav -M 980 -S 1180 300

Happened:

### CARRIER 300 @ 1000.0 Hz ###
�H00��90+�H00��90+�H00��90+�H��3�56��+�Ʊ�3�56��+��9��+�ƴ56+�H963�5���+�
### NOCARRIER ndata=74 confidence=2.026 ampl=0.147 bps=294.55 (1.8% slow) ###

This is similar to Bell 103 modulation . Although there are generally 1070 and 1270 Hz.

But didn’t the frequencies “sail away” at the terminal? The author edited the WAV file so that the speed increased by 1.8%. It turned out almost exactly 1000 and 1200. A new launch of the program:

minimodem -r -f msi55_bell103_4.wav -M 1000 -S 1200 300 -R 8000 -8 --startbits 1 --stopbits 1

And she answered:

### CARRIER 300 @ 1000.0 Hz ###
�H00��90+�H00��90+�H00��90+�H��3�56��+�Ʊ�3�56��+��9��+�ƴ56+�H963�5���+�
### NOCARRIER ndata=74 confidence=2.090 ampl=0.148 bps=299.50 (0.2% slow) ###

In both cases, the result carries a semantic load, despite errors. The article H12345678 is “pulled” out of the signal as H��3�56��- the numbers that we could make out are in place. The PSU may have poor filtering, which is why a 50 Hz background is superimposed on the signal. The program reports a low value of confidence (confidence = 2.090), which indicates a distorted signal. But now it’s at least clear how the terminal sent data to the central store computer when it still existed.