ENIGMA 2000 Newsletter - Issue 32

January 2006
Articles, newsreports and Items of interest :enigma2000-owner@yahoogroups.com

Morse stations | Voice stations | Oddities | Polytones
E2K master index | German branch | Numbers predictions
E03 & E03a prediction charts | E22 uncovered | An interesting piece
G06 schedules over a year | E11 schedules over a year
Numerals | If it had not been for 15 minutes (1/6)
News Items | Web sites | Contribution deadlines
Index | E2K NL Home

Polytones

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

Before we proceed to the logs PLondon has received information that the original XP was generated on the keyboard of a Commodore 64 pc. The tones were captured with the use of an acoustic modem and used, presumably from a recorder. It is also now known the original devices were referred to as FM – Zahlen Tonen and that the messages were destined for a KGB officer, in an embassy, with the cover title of ‘Military Attaché. Such covers are well known. [I recall the Russian trade mission in London being bugged by an Irish glazier, on behalf of the real ‘Harry Pearce’ and his boys. After the event the Daily Mirror used the story on the front page. This followed the expulsion of a large number of Russians who were stated as being KGB. Of course a few diplomats were slung out of Russia too].

PLondon even remembers being at Cardiff University in 1986/87 when the SB came to him and asked if a certain Russian Diplomat had visited him; apparently the Dip had broken the terms of his travel and should have gone no further than a stipulated distance from the Embassy. No doubt Cardiff was a lot further. See XPA 11/11. What’s the wwelsh for Russian then boyo?

November 2005

*XP changed to XPA on 29/11 but results will be shown in the XP section for continuity of results. Full XPA notes follow on.

PoSW also noted this change and was prompted to note,

“The Tuesday + Friday XPA Polytone starting at 2100 UTC has been heard in December on 5,424 + 4,968 + 4,473 kHz with strong signals, the middle transmission over-riding an "XJT" roarer on the same or very close frequency. The Tuesday + Friday schedule starting at 0700 UTC in the UK early morning has changed in December from the XP Polytone mode which has been used for years to the same XPA as heard at 2100. The morning schedule is at a rather inconvenient time for some of us but the first sending at 0700 is on 9,364 kHz, repeated 0720, 10,264 and 0740, 11,464kHz.”

[Tnx PoSW]

XP Notes

The above spectrograph was taken from the first sending for November 2005. The cyclic 818 818 818 1 [sent nine times] is clearly seen as is the 87 group message.

In concert with the annual retardation of time by 1 hour XP once again changed from 0600z to 0700z, meaning the sending occurs at the same time. The frequencies used are those used for November 2004. The audio was excellent for the first two sendings – the last being a little weak.

The sendings heard on 11^th November started well with the 0700z signal easily topping S9 with PLondon and allowing a crystal clear spectrogram for easy signal analysis.

The second sending was as good but strength measured at S3 only. Reception of the last sending at 0740z was very poor – little movement on the S meter and very dim traces on the spectrogram. A phone call to JoA resulted in PLondon concluding that his location was not favourable for the 13MHz band signals. JoA’s were in excess of S9 and measured with a peak of +5dBs, or slightly greater.

On Tuesday 29/11 PLondon was expecting XP and was surpised when his interception started with the 511/1273Hz toggling start of XPA. The spectrogram for the full sending , with abbreviated start, was as follows:

Be aware that within the above spectral offering a number of tones and their associated functions cannot be displayed as part of the message proper. For instance, XP went straight into its ID, XPA does not. Ten 511/1273Hz tone toggles to start are shown here [the norm is 60 such intro tones]. Then we can see ten cycles that lead into a peculiar set of tones that can only have machine administration purpose. This leads into the nine cycles of ID [in this case 818] and message number. Another ten like cycles of tones takes us into the admin tones prior to start pulses and message proper.

The message, as received (minus title) read:

• XPA 12162kHz 0720z 29/11/05 ID818 dk/gc 03076 00085

818 818 818 1 818 818 818 1 818 818 818 1 818 818 818 1
818 818 818 1 818 818 818 1 818 818 818 1 818 818 818 1
818 818 818 1

03076 00085 74136 52233 36564 30397 23889 46172 17913 35939
55349 24684 94607 94345 88824 56983 21796 97803 84259 86741
43725 38096 68845 55815 98810 07695 31959 87143 65360 79719
21034 00245 29452 30992 18530 43624 53272 32357 95330 48703
62940 35228 20454 28942 31371 19287 73159 74251 65916 51379
38655 74141 07853 05968 12770 13147 19901 26266 30427 90217
62620 95245 20024 59975 28951 28833 12178 65813 06130 73800
05214 51968 75985 05625 96716 80443 92458 19648 12101 94522
19621 43890 37525 60494 68142 06646 77560

XPA Notes

Not having copied XPA over the clock change we really did not know where the sending was going to occur. Thinking it would follow the same rules over frequency PLondon sent a suggested series of freqs where polytones were known to occur to certain members. Luckily he decided to include the 5MHz range [even if it was in the wrong place]!!!

The first sending at 2100z was terrible and PLondon stumbled onto it in desperation, the freq not correctly measured was found to be 5890kHz. Severe BC QRM hampered the signal. [Likely to be RAI International or Vatican Radio]. The 2120z on 5268kHz, confirmed by Gert and measured by JoA was better and allowed quick analysis to show 825 1 00317 00125. The groups had been counted and were 64+ 63 = 127 - 2 produces 125, the number of groups. The sendings lasted 3 minutes.

The final sending was 4571kHz and a reasonably good signal between S7 & 9 at both PLondon and JoA’s shacks. Despite the strong carriers and acceptable audio PLondon found it useful to apply an FFT filter post reception.The spectrogram was that shown above. It was a difficult task to actually intercept this sending without frequency cribs to help; the XP freq crib and ‘rules’ being of minimal guidance. Thanks to all those who listened for this most interesting of stations.

The second transmission of November 04/11 was poor, the bracketed dk/gc is not known to be accurate due to the poor signals received by PLondon using DSP and FFT filter techniques to recover what he did. PLondon was not alone in his thoughts on this most poor sending, as RNGB wrote,

“the broadcast on 5885 was a bit of a bugger last night wasn’t it? About 50db over 9 in [censored] at 2100. Amongst all the noise I could hear a carrier on 5891 but not one on 5879 before start of XPA. When XPA did start I could hear it clearly amongst all the noise on 5879 but it was of no use for analysis. Strangely, like you, I could also hear it between 5890 and 5891 although it didn’t quite sound in tune. Weird ! Is it possible that the strength of 5885 tx is overloading my receiver and playing tricks on it? When I tuned around 5268 I couldn’t hear XPA near it anywhere except when I was bang on 5268. Same with 4572kHz."

PLondon suggested far from being an overload the problem was caused by XPA being in between the skirts of two heterodyning stations, the third tone being a modified XPA tone of incorrect frequency.

The sendings on the 08/11 were better. Even the first sending was heard and PLondon actually extracted some spectrographic information by inserting an analogue and DSP filter on the input to his spectrogram.

The spectrogram above was taken from the sending at 2140z but is also representative of that from the 2120z sending also.

*On 11/11 the sendings were as usual – BC QRM at 2100z and a mixed bag for the other two. Thinking of the name of the XP system FM Zahlen Tonen PLondon thought he would set up a receiver for FM detection. This was carried out on the 2^nd and 3^rd sendings, the results comparable as spectrograms, shown below. Note the detail on the waveform of that to the corresponding spectral view.

The orange colour denotes the passband set by PLondon’s filtering, the yellow being outside that selected.

On 15/11 the first signal remained shrouded by the BC transmission. Mndbs claimed a move in frequency around the nominal 5892 – he showed 5891 [RNGB reckons it IS 5891 – Plondon thinks not], but both JoA and PLondon found the signal where it was last time. The FM experiment was not successful with the first sending as the signal was too weak. PLondon could not hear the signal on 5891 on any mode and JoA can confirm this as he was on the other end of the phone. The FM demod experiment was successful with the other two sendings; most successful with the third which produced the hefty 40dBs signal that used to be the hallmark of this station.

That heard at 2100z was clearer and PLondon took the opportunity to make measurements of frequency:

High 5896.5 - 5887.9 > 8.6/2 > 4.3 + 5887.9 = 5892.2kHz

On tuning for a better signal PLondon found the best reproduction to be seen on spectrogram at 5891.6kHz. The signal was too weak to allow full FM analysis but that was taken on the 2140z sending, producing a splendid spectrogram.

The Friday 2100z 28/11 transmission was of better strength with an apparently reduced BC activity. The audio received was displayed to spectrogram whilst the others were also of better quality. The frequency for best signal at 2100z was a measured 5891.5kHz

Morning XPA schedule discovered by RNGB

RNGB writes,

"During my frantic efforts searching for S06 ID 196 at 0700z 18/11 (which I didn’t find) I stumbled across XPA on 6901 at 0720.

It was S7 and I taped it. I found what was presumably the 3rd sending at 0740 on 7460 with a whacking signal of S9 + 20db and it decodes as
257 257 257 1 00687 00137 24473

I checked the previous recording and they are one of the same. The ID doesn’t make much sense to me though. Unless they made a mistake? Does happen doesn’t it?

In reply to RNGB's discovery PLondon replied, "This is excellent and will bear out the theory that many more XPA sendings [& XP] exist that we are hearing.

I am interested in the ID -- 257, as you suggest should be more like 'n94'.

There are mistakes - I am not aware that we have encountered the ID mistake so far with XPA but the message indicator has been corrupt for sometime [see NL 30/31] where a 'space 1 space' routine has been shown instead of 'space 0 repeat 0 space'.

In addition to that there also seems to be an inability of the operator to tune the transmitter correctly or match his tonal input.

[The wrong ID has happened more than once with XP though].

With all that in mind I would say the chances of a mistake is very high indeed!

This is another freq to watch - seems I'll need to set up more auto stuff!"

So, with this last comment is there anyone out there who would care to start monitoring XPA and hunt for new freqs with the view to building a deeper insight into the schedules? Be aware that despite others writings on these transmissions they are number stations and sent behalf of the SVR to ‘Military Attaché’s in Embassies worldwide.

Not heard Mon 21/11, Tues 22/11, Wed 23/11, Thurs 24/11,

For the 25/11 sending [missed by both JoA and PLondon] RNGB wrote,

"XPA this morning was a bit unusual. The delivery speed was much slower than normal, I would estimate half speed.

The transmitter sent on EXACTLY 0720 and 0740z (as read on my Rugby radio clock) I never found the 0700z transmission. The message was extremely easy to read and was 257 257 257 1 02936 00096 78180; the last group was 04251."

RNGB followed that with a spectrogram he had taken showing the section giving the dk and gc.

				Long space
0	2	9	3	6
Short space
0	0	0	9	6
Short space
7	8	1	8	0

[Tnx RNGB]

RNGB has continued to monitor, the total gathered being:

ID 257

The 29/11 sending was captured by PLondon’s second Autosystem but the signals were too weak to be of any conceivable use.

December 2005

Morning XPA

[read about ‘gc+ 1’ discovered by RNGB]

Not surprisingly the transmission on the first December Friday was XPA.

[The trace shown above has been altered to show the first few groups and last two groups of message – change at long space*]

This answered the question as to whether that on 29/11 was a mistake or one off. Both PLondon and JoA had set up for this well in advance and were ready when the XPA toggling intro started. Like that of 29/11 the ID matched that expected with the proposed freqs for December.

The first and last groups were 52641 and 56331, the transmission ending with ten 720Hz tones repeated ten times.

That sent on 06/12 0700z et al was also of good strong quality. A count of groups revealed 64+24>88-2 = 86 [gc+1] see below.

The Tuesday 09/12 0700z sending started well at around S9 with PLondon and a little better with JoA. After 1m02s the two tone input dropped ad was replaced by carrier for 34.02s. The transmission did restart and sent its two tone intro for the remaining 23.18s. This does beg the question as to how important is it to receive the total of 60 tones that are sent. However the message was received without other incident with the exception of some slight fades.

The message sent on all three sendings was the longest sent to date, since the apparent binning of XP.

With RNGB’s gc+1 theory in mind PLondon again calculated the actual number of groups sent; it was indeed gc+1

64 + 64 + 64 + 64 + 64 + 4 – 2 > 322 [gc + 1]

The sendings of 13/12 and 16/12 passed without any aberration to the signal, 16/12 being very good. Again a check on the groups produced an extra, hidden, group.

64 + 64 + 64 + 60 - 2 > 250 (gc+1)

The morning sendings on 20/12 started with strong signals at PLondon’s work QTH. The dk/gc were quickly derived and the manual count once again proved RNGB’s ‘hidden group’ theory [74836 being the group in question].

To check:

64+64-2 > 126 (gc+1)

Whilst the first and third sendings went unblemished there was a very short break in the second sending during the sixth M/S procedure of the intro. The 6^th Space freq of 510Hz was sent , but only 14ms of the expected 1279Hz Mark was seen. The recovery was quick and although the seventh space was affected the mark appeared complete along with the subsequent m/s on the intro. See below:

Evening XPA

PLondon was expecting problems finding this schedule but armed with his frequency cribs was able to locate the first sending with little problem. AnonUK also helped and confirmed the first frequency via telephone. Reception was good at PLondon’s QTH – a respectable S7 and all the better because of the non-existent BC stations from November’s schedule. The second sending was likewise a decent signal at a variable S9 [vS7^9] and the third moved into the +10dBs region of the S meter on PLondon’s ailing IC-R71E.

On PLondon’s RX320 the spectrogram [last two sendings only] revealed how strong the signals were; seen here in comparison:

The tonal spectrogram of the 2100z sending was of good quality too:

The various tones available for easy measurement. The last time a polytone signal [XP] was intercepted by PLondon this low at 2140z was in November 96 and 97; the frequency being 4031kHz

Whilst the null sending on Friday 02/12 was apparently received with little problem PLondon noted the immediate signal strength of the 06/12 2100z in excess of 20dBs on his Icom. The display on his autosystem also went to full scale. That desireable signal was short lived as the carrier cut out a number of times. Before the intro strength dropped to S5 and then aimlessly wandered between S7 to 9 during the message sending.

2120z had an S7 carrier but the audio was weak and a number of tones were heard prior to the full sending.

2140z came up as S8 and rose slightly to S9 which it maintained throughout the full sending. Tones were also heard pin the run up to the message -- so some adjustments being made Igor?

Note the message group count is 00177. PLondon looked as RNGB had suggested re the extra group and found 90384 joined, without a space to the seven tone separating character.

A check was made on the groups as 64+64+25>180 – 2 = 178 This upholds RNGB’s discovery of gc+1

The evening sending of Tuesday 20/12/05 was excellently strong and PLondon gained the details listed above from the first sending at 2100z. The 2120z was likewise and heard by ML who was in contact with PLondon for the last two sendings.

Unlike XP, XPA sends on all three slots irrespective of message content. At 2140z all that remained was a carrier. PLondon inserted a carrier to prove, the resultant heard by ML on the other end of the phone. At 2142:50s twenty three seconds worth of the 510/1279Hz m/s tones were sent [12 cycles only] but ended at 2143:25s [Times verified by DHC77]. The tones that were received were very weak and watery.

The last known evening schedule transmissions received by PLondon were of poor quality, either too much Vodka – or not enough – plus badly maintained kit at the sending end or it was propagational problems. Nonetheless, an inventive repeat of the previous Tuesdays [and PLondon incorrectly stated the HG as 52137 when it should have been 52037 – 252 and whitewash detail]!

Recently discovered second Morning Schedule

[RNGB]

As mentioned at the beginning of this issue’s Polytones RNGB discovered another schedule that resides on 6901 and 7460kHz and sends at 0720 and 0740z. If there is an 0700z it has yet to be discovered although RNGB has been searching.

RNGB was expecting a change of frequency and perhaps, like PLondon, was a little ‘stressed’ with the possibility of finding the replacement frequencies. Luckily RNGB did not have to search further and discovered they had remained as November.

However, not only did RNGB intercept the two freqs, he also made a very interesting discovery. This is told in an email exchange between RNGB and PLondon:

"As I knew you would all be catching 'XP' (or so I thought this morning) I looked for XPA at 0700 in 4 & 5 MHz, but nothing found. XPA came up at 0720 on 6901 (last months freq) with a weak signal. I monitored 7460 and 0740 XPA came up. Now I had another radio parked on 11464 to see if there as a message this morning on XP, and imagine my surprise when I heard identical signals coming from both radios. However, when the messages came the 7460 was slow speed and 11464 high speed.

The slow speed decoded to 257 257 257 2 00197 00143 91993 last group 25255 then 00000 00000 00428 00159 60600 last group 77710.

A repeat of last Tuesday's message [29/11].

This recording was very good, and after the 00000 00000 there were 163groups where I would expect 161

Did you count the number of groups in high speed XPA after the end of first block (ie.62626 etc). Was there a group before and after the first space and last space (giving number of groups nspaces+1) ?? Were there 99 groups or 100 ?"

PLondon has a small problem with numbers - due to reasons he is not going to discuss he suffers number blindness [and damnable belligerence too].

This difficulty of quantifying such an amount of numeric information prompted him to write to RNGB, "I am not being deliberately thick but I am not totally sure of exactly what you are saying. I think you are suggesting that in your 6 & 7MHz sendings there is an additional group giving the group count [other than the gc] and inserted before and after the first 6262626 group?

See last weeks entire XPA message [as shown above]. In picking it I ignored the 64 group separator but noted the tones used which gave 6262626, but, I have just found 93415 which I thought was part of the 6262626 sequence. That gives the n+1 unless my counting is crook."

RNGB replied

"Yes, I’ve looked at the recording in the files section of IC and there are 35 spaces after the 6262626 separator. There is a 5 fig group immediately next to the separator (which I had previously assumed was part of the separator) and there is one 5 fig group after the last space and before the stop signal. Giving 36 groups plus 64 in the first block = 100 !!

"I never noticed this before as all messages have been singular. But when 2 messages are sent it becomes more obvious as you end up with 2 extra groups which is much easier to spot.

I suspect the extra group/s are at the end of the messages, like when there is a null message; ie 825 000 00374 00001 00000 10140; the 10140 probably the extra group.

When 2 messages are sent – look for 2 extra groups right at the end before the stop signal."

Then RNGB asked PLondon if it was making any sense [!] and although PLondon did find it somewhat confusing [furree years ago I wanted to be a fizisist and now I am one] he was able to truthfully reply, "By God I think I've got it!! [thanks to 'My Fair lady']. I see what you are saying. The 10140 is strange - what's it doing there? I wonder if it is a machine function?"

RNGB’s discovery is certainly very interesting and has led to the discovery of the additional group.

December 05 6901 & 7460kHz 0720 & 0740z

PLondon shows the extra group using the 0720z sending of 06/12:

Highlighted is group 64 – consisting of 6 characters, space + 5 tones + space – followed by the seven separating character.

The separating character is seen thus [/]:

This equates to: 6262626.

Note that an extra group continues from this ‘separator’ character without a space. Both RNGB and PLondon originally understood those characters to be a continuation of the ‘separator’ function.

However, counting that group [28127] gives a total of 24 groups. Some simple maths produces 64+24>88-2 = 86.

64 is the number of groups prior to the separator - gp 64 [53253] is shown highlighted before the ‘separator’ but with two spaces.

Then comes the separator 6262626 followed, immediately by the previously unseen group, in this case 28127. If you care to count the spaces back from the ten repeated end tones you will count 24 groups. The product of adding the 64 and 24 groups is 88. However, in keeping with XP and XPH the groups containing the dk and gc [in this case 00916 00085] are not in any group count so 88-2 produces 86, one more than the gc of 85. That produces RNGB’s gc+1

This hidden Group will obviously be checked for on subsequent sendings.

[Tnx RNGB].

Finally another possible XPA sighting found by WUN member and sent to PLondon via Mndbs for ident:

Tnx to all reporting members!

Morse stations | Voice stations | Oddities | Polytones
E2K master index | German branch | Numbers predictions
E03 & E03a prediction charts | E22 uncovered | An interesting piece
G06 schedules over a year | E11 schedules over a year
Numerals | If it had not been for 15 minutes (1/6)
News Items | Web sites | Contribution deadlines
Index | E2K NL Home