ENIGMA 2000 Newsletter - Issue 31

November 2005
Articles, newsreports and Items of interest :enigma2000-owner@yahoogroups.com

Morse stations | Voice stations | Oddities | Polytones
Beginners corner | German branch | Numbers predictions
Confirmed end user of number transmissions
Verwaltung Aufklärung | E10 schedule research
HJH's watch | Secret service on the short wave | Killing of Colonel Callan
News Items | Web sites | Contribution deadlines
Index | E2K NL Home


Polytones

Polytone Daily Logs [If tally mark present in signal waveforms then entries underlined]

September 2005

XP [1+12 Russian Intelligence Multitone System]     XPA [MFSK-20 Russian Intelligence Multitone System]
   
1. 0600z: 9084kHz 2. 0620z: 10384kHz 3. 0640z: 11584kHz 1. 2000z 9101kHz 2. 2020z 6971kHz 3. 2040z 5758kHz
ID935   ID197  
   
    dk/gc       dk/gc      
02 Fri 935 1 00912 00257 197 000 00853 00001 00000 10140 [anonuk]
06 Tue 935 1 00912 00257 197 1 00906 00113  
09 Fri Unreadable – str carrier, weak audio [see notes] 197 1 00906 00113 [auto recording]
13 Tue 935 1 04736 0211   197 000 01263 00001 00000 10140 [auto recording]
16 Fri 935 1 00865 00183 [Good Sigs sl fade 0600z] 197 000 01609 00001 00000 10140 00000 00000  
20 Tue 935 000   197 1 00863 00161
23 Fri 935 1 05684 00083 [str sendings, all] 197 1 00863 00161 [see notes]
27 Tue 935 1 00891 00275   197 000 08323 00001 00000 10140 [see notes]
30 Fri 935 1 03918 00125 197 1 05321 00001 00000 10140 [see notes]

XP Notes

First sending of XP at 0600z 02 September 2005

XP spectrogram 2005-09-02 XP Tones
303 : Space   319 : End                        
335 : 0 351 : 1 367 : 2 383 : 3 399 : 4
415 : 5 431 : 6 447 : 7 463 : 8 479 : 9
495 : Start 511 : Repeat                  

For September XP used the same freqs of Sept 2004 and also kept the strange ID of 935. From the figures taken from the 100kHz figures one would assume the ID of 035. It is not known why this practice is repeated unless the zero is not recognised by the machinery for some reason.

The sending of 09/09 was appalling, strong carriers with little or no audio to be heard. Either faulty equipment or faulty operator.

On Tuesday 13th September all started well, carrier up at 0552z leading on to a transmission of the well known cyclic intro; all at 20dBs. 935 935 935 1 was seen in tonal form on the spectrogram, only to be followed by a 2.29s break in the sending which removed the 9 & 3 from the second cyclic ID: --5 935 935 1 etc.

XP spectrogram

The static heard as the carrier dropped can been seen as a patch of blue. [The reason the tone for the 9 is not as well represented as the other is because PLondon had his FFT filter set too tight].

The actual message for all sendings of 13/09 reads:

935 935 935 1 __5 935 935 1 935 935 935 1 935 935 935 1
935 935 935 1 935 935 935 1 935 935 935 1 935 935 935 1
935 935 935 1

*****************04736 0211 60375 43810 46550 30443 65729 20094 91918 78113
46229 59681 71422 28276 13725 81694 21671 14426 92678 71739 00006 14445 48259
02797 45188 05224 48872 76217 97500 31422 72849 61172 96047 10877 53630 89042
04855 68146 02078 95940 17878 70754 12392 23993 21284 58903 22254 26587 24191
44180 47580 43131 63705 14257 47045 24468 78621 49920 87121 06709 31127 44789
70477 85575 44146 69672 01447 75171 35571 63474 89756 68375 79082 53962 34401
20989 34600 80536 63146 96565 52074 88703 45798 58308 66810 80837 70358 17285
15040 00149 74817 11183 36840 54804 09674 96517 14214 20288 15606 51944 58962
51584 27280 46606 06673 67399 76274 97214 57596 36057 68400 23692 16018 19173
71308 22873 31609 61733 73481 68140 02933 26314 33602 79666 73982 40561 89945
70309 97966 44207 48842 40548 23410 67098 07122 73170 46776 64563 65359 71597
62884 82705 41990 87036 56613 03831 56279 89334 49745 85571 14105 57573 42542
11648 93745 65080 04476 81178 30120 34737 88781 92475 97630 22917 03018 90393
45547 50792 77754 08760 89906 34880 89608 26393 67619 23948 90441 71364 20302
71809 74038 44160 36780 78048 81979 18621 91606 19528 36246 00496 81616 47437
14723 65943 88782 87629 25839 81534 39220 35566 20064 91324 97857 04832 41640
57821 11162 27953 91246 74479 45782 49100 54740 ********

Note the second set of idents – the missing 93 can be easily seen.

Now onto the next fault with this message:

XP spectrogram

The missing ‘93’ was not the only fault with this Russian Intelligence message. [Looks like being strapped b*ll*ck naked to an old chair, with a smelly rotten sack over your head at the rear of your hinterland dacha and beaten with a bit of lead loaded hosepipe Petrov, until you learn to do your job properly]!

Look at the group count, the second 5 character group in the text – sorry, I meant 4 character group, 0211.

It should read 00211 and perusal of the Spectrogram above shows why – a repeat tone was left out, obviously after the preceding zero of the group count.

XPA Notes

Finding the XPA Sept frqs was a task, thanks to AnonUK we received the initial 2020z freq, ident and null message details.

XPA spectrogram XPA Tones:
  Start Low : 511     Space   671   End Low   705
0 : 751   1 : 791 2 : 827   3 : 871   4 : 911
5 : 951 6 : 991 7 : 1031 8 : 1071 9:   1111
  Repeat : 1187   Start High   1273

That above shows the XPA tones discovered by AnonUK and then checked to prove 197 000 00853 00001 00000 10140.

Note the cyclic nature of the ident – repeated 3 times before message indicator. It would appear that ‘00853’ is some sort of message serial, whilst the 00001 00000 10140 remains the same on null messages. It is feasible that these are reset/ready commands that the receiving apparatus recognises

The sending on 6th September was a full message. The 2020z sending was in very close proximity to a a fast Morse sending by FDG but use of a tight filter adjusted to 620 to 1284Hz removed the interfering machine sent Morse [had a tone of 515Hz]

XPA spectrogram

For the 2020z sending the signal cleaned up well as can be seen above. The dk/gc 00906 00113 easily deduced from tonal values.

The group count was checked by PLondon who manually counted that sent as 64 + 53 – 2 = 113. [-2 to remove the groups giving dk/gc]

The sending on 13/09 was a null message, 197 000 01263 00001 00000 10140, apparently transmitted on every frequency as MoK’s intercept notes prove:

2000z 9101kHz v weak almost u/r
  19:53.00 carrier up
19:54.20 6 x 1k short tones
20:00.00 intro tones
20:02.00 Mssg, sig so weak ending not identified.
 
2020z 6971kHz BC QRM from 6973, slightly better sig than 20.00z
  20:12.00 carrier up
20:12.15 2 x 1k short tones
20:12.18+ long series of x2 tones, not heard this before!!
20:19.58 intro tones
20:21.58 Mssg
20:22.13 ends, carrier down
 
2040z 5758kHz strong S9+20 some fading.
  20:32.00 carrier up
20:32.23 4 x 1k short tones
20:39.58 intro tones
20:41.57 Mssg
20:41.12 ends, carrier down

PLondon used his auto recording technique to intercept due to a variety of reasons. He was unsure if the first sending was just a blank carrier – thanks to MoK it would appear that all freqs were treated to a null message instead of the first two only.

But here’s the end of the null transmissions recorded by PLondon:

XPA spectrogram

The sending on 16/09/05 was a little different. It was a null message and was sent on all 3 frequencies again. Usually the transmission ending appears to be : nnnnn 00001 00000 10140. It was thought the nnnnn was a serial number for that particular null transmission [the cyclic preamble is 197 197 197 000 repeated thrice] and the 00001 an indicating ‘statement’ for the receiving apparatus to show a null message triggered by the 00000. The 10140 group was thought to reset the apparatus for the next message prior to the end tones to place the receiving apparatus into the standby mode. However, this time we see the ending as 01609 00001 00000 10140 00000 00000 which is different to other null transmissions received so far.

XPA spectrogram

Compare this derived spectrogram with that of the previous 01263 ending shown previous.

The XPA sending of Tuesday 20/09 started badly for PLondon as it was obviated by a whopping great XJT. However the next two sendings produced splendid signals of 30 and 20dBs strengths respectively. [It is worth noting the decode keys for both XP and XPA; the last XP was 00865, the current XPA being 00863.

Like Tuesday the repeated message heard on 23/09 started badly due to XJT [at PLondon’s QTH in any case]. Having started just a few seconds earlier the usual short low and long high tones were heard – thanks to the Carrier Insertion Oscillator being used to correctly match the tones – but when the change over was expected the tones continued. Eventually the tones changed to the admin tones and ident and the message proper started, its ending coming at 2006z

The second sending at 2020z was ok but was in near proximity to a CW signal which has suddenly appeared. It’s Morse read:
NG VVV VVV VVV de FDG AR’, the French Military at its best!.

The best sending was the last at 2040z allowing a splendidly clear spectrogram to be seen.

The first 27/09 transmission at 2000z was virtually inaudible due to the ongoing presence of XJT. PLondon was able to usefully resolve the signal to the correct tones by providing a better bandwidth response by insertion of the USB. The null message reads:
197 000 08323 00001 00000 10140.

Whilst the 2020z sending read a massive 48dBs on PLondon’s RX320 the signal from 5758kHz at 2040z [expected to be as good] was a measly 20dBs at times as the carrier varied, but nonetheless, the best signal seen in the observed 75kHz bandwidth below:

XPA spectral analysis

At 5 seconds into the sending the carrier suddenly vanished as the illustration below indicates:

XPA spectrogram

This break lasted 4.7 seconds although an attempt to raise the carrier can be seen to last just under two seconds, 22 seconds into the failed sending. The illustration above is of the entire sending, start to finish 5758kHz 2040z 28/09 [Also reported by JoA]

On Friday 30/09 the received sending was as expected, a null message. PLondon watched the tones displayed on his spectrogram and noticed the format was not as it should be for the cyclic message sequence. Instead of displaying
space 1 9 7 space 1 9 7 space, 1 9 7 space 0 repeat 0 space
it showed:
space 1 9 7 space 1 9 7 space, 1 9 7 space 1 space.

Totally wrong. PLondon then measured the frequency to prove the point, as illustrated below. The freq displayed was 800Hz, within tolerance for 1.

XPA spectrogram

The ‘blue shaded character’ is the tone selected for measurement and is 800Hz [1].

October 2005

1. 0600z: 9184kHz 2. 0620z: 10788kHz 3. 0640z: 12208kHz   1. 2000z 6842kHz 2. 2020z 5924kHz 3. 2040z 5178kHz
ID172   ID891  
   
XP [1+12 Russian Intelligence Multitone System] XPA [MFSK-20 Russian Intelligence Multitone System]
         
dk/gc dk/gc
   
04 Tue 172 000 [see notes] 891 1 00217 00097 [see notes]
07 Fri 172 1 00814 00265 [see notes] 891 1 00217 00097 [see notes]
11 Tue 172 1 00838 00229 [see notes] 891 1 00698 00149 [see notes]
14 Fri 172 2 00399 00039 & 00838 00229 [see notes] 891 1 00698 00149 [see notes]
18 Tue 172 1 00692 00225 [see notes] 891 1 02345 00001 00000 10140 [see notes]
21 Fri 172 1 02047 00149 [see notes] 891 1 03425 00001 00000 10140 [see notes]
25 Tue 172 1 01025 00105 [see notes] 891 000 09673 00001 00000 10140 [see notes]
28 Fri 172 000 [Good sendings] 891 000 [see notes]

XP Notes

The first sending was a strong 172 172 172 000. Also heard and confirmed by JoA who measured the frequency as 9184kHz. He remarked the line up was as last year. 10788kHz was the second freq and found to be weak with PLondon. Prior to the sending a Morse station hand sent [badly] on 10788kHz : VVV de TDN NR

The spectrogram below shows the excellent spectrogram gained by PLondon followed by the text [that in spectro in bold]:

XP spectrogram

172 172 172 000 172 172 172 000 172 172 172 000 172 172 172 000
172 172 172 000 172 172 172 000 172 172 172 000 172 172 172 000

R>>>>>>>>>>>>>>>>>>>

The 07/10 0600z sending started well with 00814/00265, strong sigs that tipped the meter at 40dBs, the second sending was variable between S3 to S9 with an average of S7 rising to 20dBs. However the 0640z sending had difficulty maintaining S3 with a signal loss at 0644z. PLondon blames his location on this rather than any propagational aberration. JoA heard this sending without signal loss but at a generally reduced strength of S6 towards the end of the transmission that had touched 20dBs initially.

All the AM scheduled sendings of 11/10 were of good strength with PLondon at his London SW7 works QTH. The resultant 229 group message was received well. As mentioned for 0620z 04/10 an ICW transmission was heard and likely on a nearby frequency, around 1500Hz away, rather than that used by XP.

XP spectrogram

The tone of the ICW sending was 1994Hz and read:
VVVV VV VVVVVVV K H4 GP HV GP HV GP DE WKH
possibly hand sent and feasibly of military/naval origins.

The sending of the AM schedule of 14/10 was not particularly good, PLondon and JoA received the first sending well but the other two 0620/0640z transmissions were weak. This was also noticed by mndbs who showed the first sending for 0600z as 10dBs [that isn’t big for XP transmissions by any means Mike] and commented on the variable strength signals that followed.

The transmission received by PLondon at his work QTH on 18/10 were not very good, the best being heard at 0640z. JoA also experienced similar signal variance.

The 20/10 sendings were stronger, the repeats allowing the message to be decoded in full.

The sendings of 25/10/05 were strong although the last sending was affected with some QSB

172 172 172 1 172 172 172 1 172 172 172 1 172 172 172 1
172 172 172 1 172 172 172 1 172 172 172 1 172 172 172 1
172 172 172 1

R>>>>>>>>>>>>>>>>>>> _________01025 00105 _33764 61061 98671
 24318 41878 71222 47460 73944 49417 83681 71466 24648 60896
 44886 39074 16271 60939 72183 10223 44070 49731 02919 71216
 47563 08292 88177 89620 77956 22823 18818 80264 23863 30306
 22259 56835 45421 26319 02915 62844 06449 76199 95925 65509
 97839 83373 67447 92906 61503 75905 62889 14962 58199 68203
 72735 24467 72370 14239 20503 16577 48424 29165 43989 47562
 29996 54107 97552 67668 26425 38251 92199 52767 63012 92992
 23912 05740 45236 42942 25663 31576 99797 47235 81755 02363
 52295 87063 11624 56499 22843 70151 09654 79771 78014 45940
 22291 48686 34119 09906 46129 33474 64702 29810 60110 52606
 27203 80896 <<<<<<<<<<<<<<<<<<<<

Perusal of the above sending shows the tally mark appeared between the groups containing the dk and gc and the actual message body.

An investigation of the spectrogram below indicates the tally mark was intended, as can seen below:

XP spectrogram

Measurement of the highlighted space tone indicated a longer tone, c265ms, instead of 130ms duration. The tally mark was a common sight on the now defunct Schedule 2 and 3 messages, appearing at group 50. [This use of the 50 gp tally mark was discovered by RNGB].

XPA Notes

A decrease was expected in frequency for the October month but the decrease was seen to be 2.25MHz. In conversation with JoA PLondon, as he extrapolated the ID of 891.

XPA spectrogram

PLondon stated that whilst he would expect the 2020z frequency to be within the 5.9MHz range he thought the 2040z freq to be a little too low if it was in the 4.1MHz area. Indeed a quick scan at 2015z produced the 2020z freq of 5924kHz some 32dBs and almost the strongest signal in the 35kHz scan on his RX320.

XPA spectral analysis

From the onset JoA kindly measured the freqs for this station whilst PLondon derived the ID, message indicator and dk/gc ready for the following two sendings [891 1 00217 00097].

After the second sending the 2040z freq was tackled and found to be 5178kHz; surprisingly the 5m range freqs were not troubled by BC stations, although the 2020z sending was within the skirts of a BC station [R.Prague]. A check on last months 2000z freq, 9101kHz found XJT to be silent. Mere coincidence?

However on Tuesday 11th October things has drastically changed; 2000z was beset with a variable signal strength [S5 -7] and heavy QRM. The 2020z was no better for the huge amount of BC QRM, thought to be from R.Prague and the 2040z was beset with a 930Hz het tone that rather disrupted analysis. A surprise was to be had though. Out of sheer curiosity PLondon had set up his recorder with his Satellit 2100. Despite being used on its telescopic antenna the recorded signal was at least useable for analysis. This sending was repeated on Friday 14th but heard with difficulty by all reporting monitors, 2000z being very weak whilst the 2020z being affected by the general noise and QRM associated with the BC band. The best sending was the 2040z offering but a constant 918Hz tone caused further problems, obliterating the 911Hz tone for ‘4’. A little FFT filtering allowed PLondon to analyse the dk/gc.

A null message was sent on 18/10. However, much like that of 30/09 the header was faulty, instead of 891 000 etc we saw 891 1 into the null message routine.

The 2020z sending was blighted by a nearby BC station although the signal was useable. However the 2040z transmission has a fault on it. A 918Hz tone as previously reported runs in the sending [911Hz is fig 4]. PLondon used Spectrogram 12 [tnx Mike] to display the entire sending. The tone rose with the open he carrier at 2032z and disappeared at the same time as the message ends, as can be seen below:

XPA spectrogram

The tone can be seen as a solid line to the centre of the trace. Note the defined edges from the insertion of an active filter, the edges defined by the selected cut-offs. [Note this tone is 918Hz and not 780Hz as claimed elsewhere].

With this tone firmly marked for further attention the 21/10/05 sendings were a mixed bag. 2000z was poor by anyone’s standard with fades and general noise, not helped by a watery audio. 2020z was expected to be poor due to the high power BC station that splashes its sidebands over the freq, But, not so! The 2020z sending was the most useful and allowed PLondon to lift the serial number from the trace on his spectrogram.

5178kHz 2040z 21/10/05 was pretty dire. As soon as the carrier rose there were not one unwanted tone – but three. The strongest tone was measured at 918Hz, later it shifted to 923Hz [but read about the message being sent later]. Of good audible magnitude was a 773Hz tone that happened now and then when its presence was deemed necessary to really ruin things. As this was occurring another weak tone was seen and measured as a 620Hz tone.

XPA spectrogram

At 2040z the station started but it was in real trouble, the start tones 511/1259Hz were off frequency requiring an adjustment to made to return them to there nominal values. This action meant that the receiver was tuned to 5180kHz, but the sending was terrible indeed. No tones were really heard amongst the products of a faulty or badly adjusted transmitter – or maybe both.

PLondon was fortunate to operate the receiver section of the massive transceiver in the East German Embassy, but the complexity of the device! If this lot are using something similar one can easily imagine how difficult the isolation of a fault must be.The het tone remained on freq right up to PLondon going QRT at 2145z 21/10.

Further spectral analysis illustrates the magnitude of the interfering tones:

XPA interferences diagram

Perusal of the display above the wanted signal can be seen between the two Freq Markers at 500 and 1280Hz; it is between these limits the message tones appear. [The sharpness of PLondon’s filter can also be seen at 400Hz – the upper limit increased to 1500Hz to permit complete view if the situation]. On this occasion the entire message was detuned, the nominal value for the Start Low tone of 511Hz seen to be offset to 489Hz with a corresponding difference reducing the Start High tone [1273Hz] to 1251Hz. It was necessary to retune the receiver to 5180.4kHz to place the tones visually in the ‘right’ place on the spectrograph.

On Saturday 22/10/05 [time not noted but around 1200/1300z] PLondon was monitoring 5178kHz and a carrier came up. The characteristics were the same; it appeared that maintenance was being undertaken on the transmitter because the same unwanted tones were seen, although in varying strength and presence and, at one stage, frequency.

The null sendings on 25/10 were constructed correctly, the 5 part tone arrangement seen in the header, producing the 000 indicator. However the troublesome tones were yet to be resolved or removed on the 2040z offering, the actual sending being detected this time:

XPA spectrogram

Note the 920Hz tone through the centre of the trace, and that below, at around 775Hz.

The last sending of October, before the clocks move to GMT, was horrendous on all frequencies. Weak at 2000z the presence was only just audible. There was no significant heterodyning with the sidebands of the BC stations resident on either 5920 or 5926kHz. For all that could be seen and heard the sending was not there. At 2032z the carrier rose along with its attendant 918Hz sprog and its sub harmonics. The carrier was present but on the start of sending the effect of the very strong 918Hz sprog and its offspring was to heterodyne with the very weak wanted tones and offsetting them by around 800Hz. Luckily PLondon was able to see the familiar pattern of the header, which he now knows to be 891 000; the rest remains a mystery!

Let’s hope that when the clocks go back the Radio Officer in charge of this polytone shambles employs someone who can at least read and write!!

XPA spectrogram

From the above spectrogram produced from the 2040z sending the strong tone can be seen dominating the screen [and its finish as the carrier drops at the end of the sending].

In the lower darker red horizontal band the tones which make up the cyclic 891 891 891 000 can just be seen, unfortunately the rest of the sequence was too weak and remained beyond analysis.


Morse stations | Voice stations | Oddities | Polytones
Beginners corner | German branch | Numbers predictions
Confirmed end user of number transmissions
Verwaltung Aufklärung | E10 schedule research
HJH's watch | Secret service on the short wave | Killing of Colonel Callan
News Items | Web sites | Contribution deadlines
Index | E2K NL Home

---