Tag Archives: featured

Triangulation of a VHF signal with RTL-SDR !!!

Locating the source of a signal

Privacy and security is a big topic in today’s society. With the advances made in technology, the commercial and military techniques of the past is suddenly becoming available for everyone.
With almost everything being wireless today anything and anyone can be tracked.

With this in mind I decided to try and track a signal with easy to buy equiptment.
I started with a cheap TV tuner known as RTL-sdr and the the stock/cheap supplied antenna.
Scanning the VHF band I noticed a strong permanent encrypted signal. Judging from the Norwegian frequency plan this looks to belong to either “Norwegian Directorate for Civil Protection” or the local police that is known to operate in the same area provides an interesting subject for this exercise. The signal strength was higher than commercial FM radio stations in the area suggesting the transmitter would be very close by.

Equipment used:

  • RTL-SDR TV dongle (http://sdr.osmocom.org/trac/wiki/rtl-sdr)
  • GPS receiver (NMEA or GPSd compatible)
  • rtlsdr-scanner (http://eartoearoak.com/software/rtlsdr-scanner)

Using the software and data collecting from a 15 min ride in my car I was able to track the source of the digital signal broadcasting with reasonably accuracy. The signal originates from a modern residential area. For security reasons this test is done with a cheap antenna and a low sampling rate. Enough to show the approximate area, but will not show what building.

Heatmap overlayed on Google Earth

 

Source: theanatomize.blogspot.no

RTL-SDR: $16 Dual-channel Coherent Digital Receiver

I have been playing around with the cool RTL dongles (more on rtl-sdr dongles on superkuh’s web page or rtl-sdr.com) that you can buy on e-bay for about US$8 (including shipping). These are very capable 8-bit digital receivers that have up to 2.4 MHz bandwidth and can tune anywhere between 24 MHz and 1850 MHz
I recently came up with a trivial hack to build a receiver with multiple coherent channels using the RTL dongles. I do this basically by unsoldering the quartz clock on the slave units and cable the clock from the master RTL dongle to the input of the buffer amplifier (Xtal_in) in the slave units (I’ve attached some pictures).

I originally drove the master crystal with both dongles, which also worked. However, Ian Buckley pointed out to me that a more typical way of doing this is feeding the signal into Xtal_In (in the pictures below). So I tried that too, and it also worked. I’m still not sure what the optimal setup is, as there is no schema for the dongle, but both methods I’ve tried so far have worked in practice.

This is how you make a dual coherent channel digital receiver with $16.  The clock drive probably won’t be enough for many of these, but this can be fixed with a buffer or some other active splitter.

 

The oscillator is wired using a piece of 75 Ohm antenna coax that came with the dongle. It’s like they designed the dongle for  multi-channel coherent applications.

This has some implications for low cost geophysical instruments. It will be possible to use this receiver for the 150/400 MHz beacon satellite receiver, as this only requires that the receivers have clocks that are locked with each other. Interferometry and passive radar are other application examples. With more than two locked channels, applications such as imaging start to become possible.

I’ve made some relative phase noise measurements, and the systems don’t have detectable sample drift over two hours, and their relative phase is also pretty stable.

Spectrum at 1 Hz sample rate of the relative z_1/z_2 phase signal going into two receivers.

 

IQ plot of the z_1/z_2 relative phase signal over ~6000 seconds at 1 Hz sample rate.

And oh, by the way, I found this nice usb hub, which I’m going to use to hopefully get a 7 channel coherent rtl system.

Hub with the right usb port orientation for rtl dongles.

Stay tuned for more results. I already have some pretty nice passive radar results using the system, which I’ll be posting in a few days.

Update: Apparently three dongles will also run fine from one master clock. I know the clock isn’t split correctly, but adding any components would increase the total cost and the whole point of this exercise is to determine what the lower bound is for software defined radios.

Three channel coherent RTLSDR receiver.

Source: kaira.sgo.fi

Kalibrate-RTL: Calibrate #SDR (SDR Sharp) Linux/Windows TUTORIAL.

My dongles drift about 8ppm from cold start to warm up after around 40 minutes.
And converter will drift also.

Kalibrate, or kal, can scan for GSM base stations in a given frequency band and can use those GSM base stations to calculate the local oscillator frequency offset.

Download it from here:

Linnux source: https://github.com/steve-m/kalibrate-rtl
Windows : http://rtlsdr.org/files/kalibrate-win-release.zip

See the list of options below.

(I have the ‘Kalibrate’ files on my D: drive in the folder named ‘Kalibrate’)
(The commands below are what are in my shortcuts. If you want to run from a DOS window, just enter everything after the ‘C:\Windows\System32\cmd.exe /k’)

Use this command to find a GSM850 signal in your area.
C:\Windows\System32\cmd.exe /k “D:\Kalibrate\kal.exe” -g 42 -e 22 -s 850

Then, once you have identified a GSM signal in your area, run calibrate using the command below.
C:\Windows\System32\cmd.exe /k “D:\Kalibrate\kal.exe” -e 41 -c 234 -v

For the above command:
GSM Channel 234 (-c 234)
-e is roughly the error rate expected. ’41’ in this case. (not real critical) (-e 41)

That will give you the:
‘Frequency correction (ppm)

Where options are:
-s band to scan (GSM850, GSM900, EGSM, DCS, PCS)
-f frequency of nearby GSM base station
-c channel of nearby GSM base station
-b band indicator (GSM850, GSM900, EGSM, DCS, PCS)
-R side A (0) or B (1), defaults to B
-A antenna TX/RX (0) or RX2 (1), defaults to RX2
-g gain as % of range, defaults to 45%
-F FPGA master clock frequency, defaults to 52MHz
-v verbose
-D enable debug messages
-h help

Then, if you are using a converter, you need to set ‘Shift’ or ‘Offset’ for it.
I don’t use a converter so I can’t help with that part.

If you have ‘Snap to grid’ selected, SDRSharp will land on an exact multiple of your selected step size.

Rich

Edit: Here is an older run that I did.
The search found channel 136 to be the strongest on that antenna.
Kal_Search

Using channel 136, the ‘Frequency correction offset’ rounds up to 38ppm. (SDRSharp accepts whole numbers only)

Kal_Offset

Source : Extracted from forums.radioreference.com

“SOME” DIGITAL MODES USED IN HAM RADIO (and signal ID guide, Thanks hfradio!)

Digital modes are becoming more and more popular on the amateur bands. This is mainly due to the following reason: Affordable home PC’s with built in soundcards. This has brought forth a multitude of decoding software, some free, others not. There are new modes being invented all the time and keeping track of these is turning into a full time job! One of the main problems encountered by the newcomer to digital modes (or digimodes as they are known) is how to identify what they are seeing/hearing. Most of the decoding software uses a visual ‘waterfall’ display to facilitate easy tuning.

With that in mind I went on the bands and captured images of the most common digital modes in use at the moment. Below you will see images of each mode together with some brief notes on the mode. The images show the most common variant(s) of the mode, although some have many different ‘flavours’! I will add to this list as and when I hear/identify a new mode that is being used on a regular basis (last popular ‘new’ one is Olivia which wasn’t around when I did this page on my original site)

Click on the name of the mode (where the name is underlined) to hear an mp3 of how the mode sounds on air (these are to give you an idea of how that mode sounds, not for analysis purposes) I have included some sound files of mode variants – more to come as I find them).

PSK31

PSK, or Phase Shift Keying has become the most popular of the newer digital modes. There is a wealth of information on the web regarding BPSK (Binary PSK) and QPSK (Quadrature PSK)

Because PSK31 has a bandwidth of only 31Hz, many signals can fit into the same bandwidth that would be occupied by an SSB signal (2.4kHz approx.). It is quite common to see 15 or more signals on a 2.5kHz waterfall display.

A ‘clean’ BPSK31 signal. This is how your signal should look!

Here are a couple  of  BPSK31 signals that are badly distorted. This is probably due to overdriving. Reducing the input to the soundcard or reducing the output level would improve the quality of this signal. Note that although some way from the adjacent signal on the left, the distorted signal is sufficiently wide to cause interference to the other signal.

Here we see a station that has an unstable signal and is drifting badly. A stable and ‘clean’ transmitter is vital when using narrow modes such as PSK31 and it’s variants so as not to cause QRM to nearby stations

PSK63

PSK63 is gaining popularity, with many programs now supporting this mode. The pro’s for this mode are the fact that data is sent and received at twice the rate of normal PSK31, so is great for chatting and contest exchanges. The con’s to this mode are the increased bandwidth required over PSK31, the increase in power required to maintain the same level of copy as PSK31 and that not all software decoders support PSK63.  PSK63 can be identified quite easily as it looks like a ’fat’ psk31 signal!

Other variations of PSK31 are PSK16 (half bandwidth/speed of PSK31); PSK125 (4 times bandwidth/speed) and other experimental variations such as PSK10 (to be found in MultiPSK) and even PSK250. The other  common variant of BPSK31 is QPSK31,  (the ‘Q’ stands for ‘Quadrature’, rather than the usual B which is ‘Binary’ Phase Shift Keying), which is sideband dependent (i.e. both transmitter and receiver must be using the same sideband) but is not in common use despite it’s superior decoding ability during poor conditions.

Here is an waterfall shot of QPSK63 (the wider of the signals. If you compare it to the BPSK63 signal above, and also on the left of the picture  you can see there appears to be more information contained within the same signal, this is the easy way to tell QPSK from BPSK.

This picture shows the bandwidth difference between PSK31 and PSK63, PSK63 being the wider signal.

Here we can see the difference between a PSK31 signal and a PSK125 signal. The PSK125, although much faster, takes 4 times the bandwidth and requires 4 times the power for the same s/n ratio as PSK31. It is a great mode when conditions are good and signals are strong, especially on the higher bands where there is more space.

 

SSTV (Slow Scan TV)

Slow Scan TV has been popular for many years, although the vast majority these days is computer generated. The most common modes are Martin and Scottie. Robot still has a following. Most SSTV programs handle these modes and others too. The received pictures are built up line by line over the course of nearly a minute so you need to be patient! Quality can be very good, even over long distance paths. Here are two pictures received by me — the topmost one is from Hawaii (KH6AT) and the bottom one is from Sweden (SM7UZB).

sstvkh6atsm7uzb

 

RTTY (Radio Teletype)

The ‘original’ data mode. RTTY (pronounced ‘Ritty’) has been around for many, many years and is still just as popular. Years ago the only way to get on RTTY was to use a mechanical terminal unit such as the Creed 7 series, which were big, noisy and messy. These days, virtually all RTTY is done by the computer/soundcard combination. Amateurs (hams) use 45 baud (the speed) with 170Hz shift (between mark and space). Commercial stations use 50 or 100 baud with shifts of 425 or even 850Hz. Most software caters for differing speeds and shifts. Unlike most digital modes, RTTY is transmitted on LSB.

rtty_120x75

 

MFSK

MFSK is similar to the commercial Piccolo system. MFSK is very good under poor propagation conditions. The usual variant of MFSK is MFSK 16, but other types such as MFSK 8 are in development, along with other similar modes to MFSK (such as Domino). MFSK is sideband dependant, so you must have your receiver set to the correct sideband in order to decode it properly. Also tuning is quite critical, although AFC helps somewhat. The top image is of an MFSK16 signal and the lower image is of an MFSK32 signal (which as you can see is nearly 500Hz wide, twice as wide as an MFSK16 signal).

mfsk2_106x109MFSK32

 

MT63

MT63 is very robust and offers 100% copy when other modes fail. The tradeoffs however are bandwidth and speed. MT63 is quite slow and occupies anything from 500Hz to a full 2kHz (which is still less than a single voice channel). Because of the wide bandwidth, MT63 is usually confined to 14MHz and above, where there is sufficient space to accommodate it.

mt63 long2_308x85

 

HELLSCHREIBER (HELL)

Hellschreiber (or Hell as it is commonly known) is a bit different from most other data modes. When receiving a Hell signal, your eyes do the filtering! The decoded text is displayed on a ‘ticker tape’ display (as shown in the picture). Hell has a very distinctive ‘grating’ sound and is a narrow band mode. The Hell signal is on the left of the picture (with the green flag above it), with an MFSK signal on the right—note the bandwidth required for the MFSK signal compared to the Hell signal. Even weak signals can be decoded as your eye/brain combination can ‘fill in the blanks’ where the signal fades.

Here is a waterfall of a Hell signal, together with a decode (showing how it appears on screen)

feld hell 2feld hell

 

PACKET

HF mailboxes etc. use packet to forward messages to users. The usual data rate on HF is 300 baud, with 1200 and 9600 baud being common place at VHF and UHF. The picture shows a mailbox/BBS in Turkey negotiating with a BBS in the UK. The short burst at the bottom of the picture is header and callsign information whereas the longer burst is the actual data. Several of these packet BBS/mailboxes can be heard chirping around 14.1MHz.

packet2_131x75

 

PACTOR

HF mailboxes etc. also use PACTOR  to forward messages to users. PACTOR has had a lot of bad press recently, mainly due to the actions of a few inconsiderate operators who are apparently causing interference deliberately to existing users of the digital sub bands. I cannot comment on this as I have not experienced it personally. The picture shows the PACTOR signal trying to establish contact. Once established the transmission of data can begin. Because PACTOR uses error correction, it can take quite a time to send a message particularly over a less than perfect path—but the transmitting station will keep trying until the message is received perfectly. The picture is of a PACTOR 1 signal, however there are PACTOR 2 and 3 variants, but these require hardware encoders/decoders.

pactor_140x86

 

THROB

Throb is one of the newer digital modes and although it can be heard, it is nowhere near as popular as other modes such as PSK31 or RTTY. As with the other modes, there are various variations of Throb, 1 throb/second; 2 throbs/second and 4 throbs/second. 1 throb is the slowest and 4 is the quickest. Throb is actually quite a slow mode and is therefore probably quite resilient to the effects of fading etc. although is does take quite a time to complete a contact!

throb1-1_76x111throb2 - 16_78x108throb4_99x102

Left to Right: Throb 1; Throb 2; Throb 4 (click each type to hear the different sounds).

 

OLIVIA

Olivia is a fairly new digital mode and it seems to be extremely resistant to fading and QRM. I can get full copy on stations that are barely audible (even ones that fade down to almost zero seem to still print well). As with other modes, Olivia has different variants each having a different bandwidth (from 500Hz to 2kHz) and different number of tones. Olivia can be very slow (in the order of 2-3 characters per second) but a slow contact is better than none at all! In the below pictures, the 8/250 indicates 8 tones over a 250Hz bandwidth and 32/1000 is 32 tones over a 1kHz bandwidth. To avoid interference to other stations is it usual to start an Olivia transmission on a full kHz (i.e. 14.108.0 rather than 14.108.3 for instance).

olivia

Here are some waterfall shots of some other Olivia modes:

olivia 8-250olivia 32-1000

 

CONTESTIA

Contestia is another very new mode to be found on the ham bands. It is not, as yet, very popular and so far I have heard only one station transmitting this mode. Again I have included a sound file and a waterfall capture so that you may see what it looks and sounds like. This image is of a Contestia 4-250 signal from RW3AS on 20m.

 

contestia

JT6M

JT6M is a specialised mode found in the WSJT software suite  (from Professor Joe Taylor, K1JT) designed for weak signal working (such as EME—Moonbounce and Meteor Scatter). JT6M is the favoured mode for MS and Sporadic E and can be heard on 6m around 50230.  I have done some monitoring recently using JT6M and have seen full decodes from stations that were not audible to me by ear, which I think is quite impressive!

jt6m

HAM DRM

Similar in principle to the broadcast DRM signals heard on the SW broadcast bands. DRM is a very experimental mode at the moment, with the main exponents being found on 80m around 3733kHz. I have not had much success with this mode as yet, despite having good signal levels. The signals need to be very clean and strong in order to decode. Pictures can be sent using DRM, but time will tell as to how/if this mode grows in popularity. Below is a waterfall ident from DD9ZO, sadly this station was not strong enough to decode. This mode does not seem to have taken off in the way that others have, mainly I think due to the fact that it requires a very strong and noise/fade free signal in order to decode. This is much like broadcast DRM. Unfortunately, the disadvantage with this mode and most other digital modes is that the signal is either fully readable or not copiable at all. With analogue signals (and some digital modes) you can usually fill in the blanks when you get fading or noise on the signal, also they are copiable when they are barely audible. Some newer modes do work very well at signal levels that are at or below the threshold of human hearing (WSPR for instance, which is mode that has only been around for a few months but unfortunately it is a one way mode, that is it is a beacon mode rather than a ‘conversation’ mode).

drm2

 

DIGITAL / HD SSTV

If you tune on 14.233 you may well hear a strange signal that sounds very similar to the HAM DRM signals mentioned above. This will be one of the new Digital SSTV modes. Like all DRM modes, Digital SSTV produces excellent, noise/distortion free pictures which can be in high definition. However for this to occur, the received signal needs to be very strong and relatively free from noise etc. If the program loses any part of the signal, due to a noise spike or a brief fade, the whole picture is lost. This is the big disadvantage with this mode on HF, it really is all or nothing. The software I am using to decode this mode is called  ‘EASYPAL’ and is available from http://www.qslnet.de/member/hb9tlk/ . If you don’t want to transmit a picture, you can send short text messages in the ‘waterfall’. I haven’t used this program much so don’t know all the ins and outs but it’s good to be able to decode this new mode. Below is a waterfall grab of a digital SSTV signal (as seen in the MixW waterfall)

hd sstv-4

Example of waterfall text received in EasyPal

hd sstv-5

Digital / HD SSTV picture received on 14.233MHz in August 2010

100813142255-Clip1

Here are some more pictures captured using EasyPal, on both 20m and 80m. I am finding that there is more activity now than when I originally wrote this piece. The pictures are of stunning quality and, as they are digital, there are no traces of noise, QSB or any of the other problems that affect analogue SSTV pictures. The downside is that with these you either get the entire picture, or nothing at all – plus the transmission time seems rather lengthy. You do need a strong/clean signal in order to use digital SSTV, but it is well worth it. There are some received pictures I cannot show on this site as they are rather ‘risqué’ and show women in various states of undress – pleasing to the male eye, perhaps, but not really suitable for ham radio.

These were received on 80m (again, note the quality of the decode, even on the noisy 80m band at night:

111219145846-Last_TX_Picture1

111219223939-9l1ovc1

And here are some from 20m:

111219155219-Clip1

111219154924-Clip1

DOMINO

Here is another of the new modes which can be decoded by many different software packages. Domino is another mode that uses MFSK (Multi-Frequency Shift Keying). MFSK sends data using many different tones, sent one at a time. As with ‘normal MFSK, it has excellent performance but was developed specifically to cope with the noisy conditions of the lower HF bands. For more information and technical details, visit:

http://www.qsl.net/zl1bpu/DOMINO/Index.htm.

As with the other MFSK modes (such as MSFK16, Throb, Olivia etc.) Domino is used with different parameters, the best mode variant to use is dependent on band conditions.

domino1 - 8_109x121domino1 - 16_169x106domino2 - 8_111x102domino2 - 16_175x108domino3_169x95DominoEx

Above Left to right are:

DominoF 1-8; 1-16; 2-8; 2-16, Domino 3 and lastly DominoEX. However DominoEX (click to hear sound file) has superceded DominoF. Like DominoF, DominoEX has a multitude of variants to suit various bands/conditions..

THOR

Thor is a new mode and is very closely related to DominoEx. It is an extremely robust mode and is well suited to HF weak signal conditions. A single carrier of constant amplitude is stepped between 18 tone frequencies in a constant phase manner. This means that no unwanted sidebands are produced, and it does not require the same kind of linearity requirements as some modes (PSK in particular). The tones change according to an offset algorithm which ensures that no sequential tones are the same or adjacent in frequency, considerably enhancing the inter-symbol interference resistance to multi-path and Doppler effects. Thor, like other similar modes has a variety of speeds and tones to choose from, dependent on band conditions and signal levels. The modes are Thor 4, 5, 8, 11, 16 and 22. Speeds vary from the equivalent of 14wpm right up to 78wpm for Thor 22. and bandwidths vary from 173Hz up to 524Hz.

.

THOR 16 better 2

This is a waterfall grab of a Thor 16 signal.

JT65

JT65 was developed originally as part of the WSJT weak signal modes software package by Joe K1JT. JT65 can also be decoded by other packages, such as MultiPSK. The screen grab below is taken from MultiPSK. JT65 has found a use on HF and can be found around 14.076MHz and 21.076MHz amongst others. Signals that are virtually inaudible can give perfect copy so its  performance is excellent on the noisy HF bands. The transfer rate is slow, as are most modes that excel in low signal decoding. I am now monitoring JT65 most days and am amazed how well signals come through, day after day!

jt65a in multipsk 2jt65 2jt65 1

jt65hf1.07

This is a screenshot of another JT65 program. This is JT65-HF by W4CQZ, a free decoder that is specially written for those of us that like to play around with JT65 on the HF bands. I have been using this for quite a time now, and I am very impressed by it. I like the interface, it is easy to use and easy to set up. It will decode multiple signals at the same time and stores the results in a CSV (comma separated values text) file for later analysis. Another useful feature is that, like DM780, it sends spots to the PSK Reporter website and they are displayed on a map. You can interrogate the system to display all spots seen by any particular callsign etc.  As you can see from the above screenshot JT65 works very well on HF and I have heard signals from all over the world from stations using fairly simple equipment and low power. Also the program reports to the RB (reverse beacon) network, which is found at: www.jt65.w6cqz.org/receptions.php . The above screenshot shows stations in Brazil, UK, Netherlands and Asiatic Russia, all on a very quiet 10m band.

 

JT65 grab

This screenshot shows the spots my receiving station has sent to the PSK reporter network over a 24 hr period, using JT65a. As you can see, all four corners of the world have been heard! There are some stations active from Africa, mainly in South Africa (ZS) and it is not uncommon to hear them – however none were active then this screenshot was taken.

I have decided that in 2011 I will be devoting a fair amount of time to JT65a (each year I choose a particular mode or band that takes preference – in the past I have done CW, 160m, PSK31, WSPR, 6m and now is the turn of JT65). There has been a marked increase in JT65a activity on the HF bands, mostly due to the efforts of Joe and his excellent JT65-HF software. I am intending to promote activity on this weak signal mode wherever possible. When using this mode, it is advisable to keep the transmitter power to 50w or less or no more than half the rated output of your transceiver. Most of the time 20-30w is more than adequate and quite often contacts over amazing distances can be achieved using just a couple of watts. It is interesting to note that even when a band appears to be closed, the chances are that there may well be a JT65 path open. If you have restricted antennas, power or both, this could be the mode for you. Signals that are 24dB below the noise level can be decoded with relative ease.

Joe is constantly developing the software and releases new versions regularly. Each release sees a further improvement in functionality, but without sacrificing ease of use.

To download the latest version of the software (v1.093), click HERE.

WSPR

This introduction is taken from Joe Taylor (K1JT)’s WSPR 2.0 online user guide. WSPR (pronounced “whisper”) stands for “Weak Signal Propagation Reporter.” The WSPR software is designed for probing potential radio propagation paths using low-power beacon-like transmissions. WSPR signals convey a callsign, Maidenhead grid locator, and power level using a compressed data format with strong forward error correction and narrow-band 4-FSK modulation. The protocol is effective at signal-to-noise ratios as low as –28 dB in a 2500 Hz bandwidth. Receiving stations with internet access may automatically upload reception reports to a central database.

For more detailed information on WSPR and to download the programs, see the WSPR website or the WSPRnet website where the database etc. are hosted.

wspr waterfallwspr screen grab3

 

Here are the USB dial frequencies for WSPR:

Band Frequency (MHz)
LF 0.502.4
160m 1.836.6
80m 3.592.6
60m 5.287.2
40m 7.038.6
30m 10.138.7
20m 14.095.6
17m 18.104.6
15m 21.094.6
12m 24.924.6
10m 28.214.6
6m 50.293.0
2m 144.488.0

 

Below is a summary of stations heard on WSPR over course of a month (27 November to 27 December 09).

30m has been the band I have spent most time monitoring as can be seen by the fact I have decoded over 270 stations on that  band alone. What I have found strange is the lack of African stations – not a single one has been decoded to date. I guess it hasn’t caught on over there yet. Hopefully as the mode develops and interest increases some stations will take it up as it would be interesting to see how the propagation paths to Africa open and close during the course of a day or month. Asia is quite sparse too here. I know there are stations on from Japan etc. but I am not hearing them here very often. Same with Australia. No path to VK for me yet. It is nice to see some North American’s on 80m, proving that there has been a path most nights and it should be workable with modest powers and antennas. If I can hear a 1 watt American station on 80m WSPR at a reasonable level,  then I should be able to hear ones running 100w of CW/SSB or 30w of PSK with little problem.

By Continent
BAND TOTAL STNS Eu NAm SAm As
17m 8 1 7 0 0
20m 28 17 11 0 0
30m 273 223 48 1 1
40m 30 29 0 1 0
80m 177 169 8 0 0
160m 63 62 0 0 1

Over the past few weeks I have been transmitting on WSPR (and a few other modes) with 5 watts or so into my inverted Vee and have been very impressed with the results. I still feel that I would be better with proper dipoles for each of the main bands I am interested in and I will hopefully work on that during the summer.

UPDATE:

I have now been using the OCF dipole/Windom arrangement for a few weeks on various bands/modes and can report that it works well although it does not have quite the same impact on 30 and 40m as the G5RV. The reason is because tthe OCF is designed for use on 20m and above so operation below 14MHz is a compromise. That said, it still radiates a fairly good signal on 30m as borne out by the WSPR reports I have collected.

So far this year I have been heard in 56 entities, which are:

Prefix Country/Entity
4X Israel
9A Bosnia
9H Malta
CT Portugal
CX Uruguay
DL Germany
EA Spain
EA6 Ballearic Isl
EA8 Canary Isl
EI Eire
ES Estonia
EX Kyrgyzstan
F France
FR Reunion Isl
G England
GI Northern Ireland
GM Scotland
GW Wales
HB9 Switzerland
I Italy
JA Japan
KH2 Guam
KL7 Alaska
KP4 Puerto Rico
LA Norway
LU Argentina
OE Austria
OH Finland
OK Czech Republic
ON Belgium
OX Greenland
OY Faroe Islands
OZ Denmark
PA Netherlands
PY Brazil
S5 Slovenia
SM Sweden
SP Poland
SV Greece
T6 Afghanistan
TF Iceland
TK Corsica
UA0 Asiatic Russia
UA3 European Russia
UN Kazakhstan
UR Ukraine
VE Canada
VK Australia
VU India
W USA
YO Romania
YU Serbia/Montenegro
YV Venezuela
Z2 Zimbabwe
ZL New Zealand
ZS South Africa

As you can a see there are a spread of countries from all continents, which is gratifying as it means that I am at least radiating some sort of signal in every direction.

Here is a summary, by band, of where my WSPR signals have been heard (The term ‘ODX’ refers to ‘longest distance’, so in this table ODX means the furthest that I my signals have been heard):

 

Band ALL 10m 17m 20m 30m 40m
Hrd by number of stations 630 46 26 366 336 49
ODX (km) 18934 2078 12028 18934 18934 14549
Avg Distance (km) 2534 793 3057 2420 2538 1188
Hrd in DXCC 52 11 15 46 40 15
Lowest SNR reported -33dB
Highest SNR reported +11dB
WW QRA locator Squares: 264

So from this I can see that on 10m there have only been reports from single hop sporadic E openings and the average distance on that band is quite short at just under 800km. This short distance is due to the MUF being high and the skip distance decreasing as a result of the higher angle reflections. I have not spent much time on 10m because the high MUF sent me packing up to 6m where the band would usually be open. There have been multi-hop Es on 10m, but I have been on 6m when these have been about! On the 20 and 30m bands there is not much to choose between them and the average distance is about that of a single F-layer hop. The 40m average is about half the distance so my antenna must have quite a high angle of take off to reflect at this short distance. Experiments like this are interesting in assessing antenna effectiveness. I can see that even though my antenna is 3/4 of a wavelength above ground, it is physically short (a quarter wave in length, but the feed arrangement does not make it an efficient DX antenna on this band). However it does seem to be suited to local/semi local (i.e. within the UK and Europe) working – with the odd longer distance QSO possible when conditions are favourable. Looking at the signal to noise ratios (SNR) there is a huge difference between the strongest and weakest (44dB – that means that the best SNR is over 20,000 times stronger than the worst SNR!). -33dB is extremely weak and must be right at the very limit of the decoding capability of the software. To quantify this and put it into figures that are more easily understood, say I was running 1 watt output and my signal was received with an SNR of -33dB, in order for me to improve that report by 10dB, I would, in theory,  need to increase my output by 10 times (which is 10dB), therefore I would need to run 10w. To increase my received SNR by a further 10dB another 10 times power increase is needed – taking my output power to 100w (10x10w). A further 10dB increase (giving a 30dB overall increase) would take the power output to 1kW (1000w) or a 1000 times our starting power. Reversing the situation, let’s see how that kind of change would affect the s-meter of a receiver. S-meters are not usually that accurate and calibration can be all over the place. there is a standard though and that is 6dB per s-point, so if we say the worst SNR gave a reading of s1, the best SNR would indicate just over s7. If we were talking about pure signal strength this would be the case, but with Signal to Noise ratios this would not be true as we are talking about dB above the noise level and dB below the noise level. The human ear can decode CW signals down to about -15dB or possibly -18dB SNR but certainly not much, if at all, lower than that (and believe me, that is a weak signal and takes all your concentration to hear it!). That is still a good 15-18dB above the weakest SNR that WSPR has decoded my signals at!

ROS

ROS is a fairly new mode that is in it’s first year or two of use. ROS uses multiple tones over a 2kHz or 500Hz bandwidth, (the frequencies for each mode/bandwidth are hard coded in the software which is causing some annoyance amongst some users. ROS has three main speeds, 16 baud, 8 baud and 4 baud. There are some special modes, such as 7bd/100Hz for 136 and 502kHz (and 80m for some reason), plus an ‘EME’ mode for use on 2m and some other bands, for weak signal work as it has, in theory at least, the capability to decode signals that have a Signal to Noise Ratio (SNR) of -35dB, which is even lower than  WSPR. There are, however, questions as to the legality of using the mode on HF in North America, as spread spectrum is not allowed below 222MHz and the authorities are still undecided if ROS is SS or not.

This will all be hammered out as the mode grows. It will either become widely used or, as is sometimes the case, just disappear through lack of interest. Only time will tell on that. One thing that has come to light is that ROS has just been accepted into the ADIF standard (Amateur Data Interchange Format), which is the common ‘language’ that is used for exporting and importing log data. Also eQSL.cc (the electronic QSL exchange centre) has updated it’s system to accept ROS QSO’s.

 

Click on pic below to hear the sound of 16 tone, 500Hz ROS

ROS 500Hz-4ROS 16

The above pics are taken from the ROS program, which uses a monochrome waterfall. As most software uses a color waterfall, I have included a screenshot of a ROS signal (in this case the same as above, 16 tones / 2000Hz) as viewed on the HRD waterfall.

ROS 16-2

To download the software and users guide: http://rosmodem.wordpress.com – it is worth checking there at least weekly as new versions are being released frequently in response to users requests for features etc. (the latest version is v 7.0.8). Here are the latest frequencies, as of May 2012. Note: these are considerably different to the previous version that appeared on this page.

QRG BAND
136 kHz 2190m
500 kHz 600m
1840 kHz 160m
3583 kHz 80m
3585 kHz 80m
3587 kHz 80m
3589 kHz 80m
5367 kHz 60m
7040 kHz 40m
7044 kHz 40m
7046 kHz 40m
7048 kHz 40m
10132 kHz 30m
10134 kHz 30m
14101 kHz 20m
14103 kHz 20m
14116 kHz 20m
14118 kHz 20m
18107 kHz 17m
18111 kHz 17m
21110 kHz 15m
21115 kHz 15m
24916 kHz 12m
24926 kHz 12m
28185 kHz 10m
28295 kHz 10m
50245 kHz 6m
144160 kHz 2m
144980 kHz 2m
432097 kHz 70cm

There is a buzz of excitement as Joe Taylor, K1JT, has now released the first version of WSJT 9 – which can be downloaded from: http://www.physics.princeton.edu/pulsar/K1JT/wsjt.html There are some new modes in this version although I have not had time to investigate them as work is very busy at the moment and will be for another few weeks. There was a brief beta test version of WSJT, version 8, that investigated a host of new modes and this research and it’s findings have gone into producing the public release of version 9. I will provide more information on these modes as it becomes available (screenshots and mp3’s etc). It’s a busy time in digimodes, there always seems to be a new mode or refinement of an existing mode being released. ROS is still finding it’s feet and is under constant development – it still has its critics, but that is true of most modes but it does seem to be getting quite popular, although I doubt if it will ever be as widely used as, say, PSK31.

WHAT IS RSID?

RSID is a method of identifying certain digimode signals. RSID stands for “Reed Soloman IDentifier”. The way it works is that the station who is transmitting can, in certain software, enable RSID. When enabled, RSID sends out a numeric code in the form of an MFSK signal which is identified by the receiving software (if it is capable of RSID) and sends alerts to the user and advises of the mode used. This is particularly useful with the lesser used/recognised modes, however it can be rather tedious to have an alert every time someone starts transmitting in, say PSK31 or RTTY as these are extremely common modes and you can find yourself inundated with alerts. In some software RSID can be configured to be active for certain modes only. In DM780, a box appears on the desktop alerting the user to a transmission and gives the frequency and mode (such as Olivia 23/1000 on 1500Hz) and you can tune to that transmission by clicking on the alert box which will change the mode and frequency of your radio (providing you have computer/software control set up).

VIDEO ID

Video ID is different to RSID in that it can be seen by anyone using software with a waterfall display. Again, video ID is enabled in the transmitting software and it is not supported by every software. If you are watching a waterfall trace, you may, at the end of the transmission, see some text or even a crude picture in the actual waterfall – this is video ID and confuses many people who have not seen it before.

 

waterfall pic sstv 2
psk125 with 73 pic

SOME OTHER MODES

Below are some less common modes, together with sound files. This list is certainly not exhaustive (not by a long, long way!) but it might just give you an idea of what that strange noise was that you heard when tuning around on SW. New data modes are emerging all the time and it is difficult to keep up with them. I usually follow the development of new modes that I think might interest me – some of these modes become popular, some much less so.

Unfortunately I do not have waterfall traces for these signals yet, but I will endeavor to track some of these signals down and capture them (that’s why they call it ‘hunting’ 🙂

For technical details of the modes below, and others, visit: http://f1ult.free.fr/DIGIMODES/MULTIPSK/ (where some of the following information has been verified/sourced.)

SITOR: SITOR  is a commercial teletype mode it means SImplex Teletype Over Radio) used for sending text messages between stations, SITOR may be run in interactive (ARQ – Automatic  Repeat reQuest) mode, which is known as SITOR-A (also called AMTOR). When SITOR is run in broadcast (FEC – Forward Error Correction) mode, it is known as SITOR-B (or NAVTEX). The mode uses special error correction techniques. There are many ARQ and TOR modes to be found on the HF bands, SWEDISH ARQ, G-TOR and CLOVER are just some. If you start investigating, you will be surprised by just how many commercial digital modes are in use.

CHIP: This mode was created by IZ8BLY back in 2005. Chip is a PSK mode which uses “Spread Spectrum” modulation and comes in 2 variants, Chip64 and Chip128. Chip has a bandwidth of almost 600Hz, but is an extremely robust mode and has a good throughput speed.

ALE: (Automatic Link Establishment), is now finding more use in amateur circles, thanks to the efforts of the writers of some of the multimode decoders, such as MultiPSK. ALE, when running correctly, can initiate and establish connections between two stations without human intervention (hence the ‘Automatic’ part.)

PAX, PAX2: PAX is another MFSK mode which is derived from Olivia. PAX2 is the same as PAX but runs at a higher baud rate (62.5bd for PAX and 125bd for PAX2), therefore PAX2 is twice as fast as PAX but requires a better signal to noise ratio. The following frequencies are used for PAX/2: 3.590, 7.042, 10.148 and  14.075MHz .

STANAG: Used by the military, comes in many different variations. Pretty much all Stanag traffic is encrypted. STANAG 4285 is the NATO standard for HF communication. It consists of several sub modes (75-2400 bps) and two different interleaving options (short and long). The receiver should be in USB mode and provide flat frequency response from 600 Hz to 3000 Hz. Another Stanag mode is 5066, which includes mode identification in it’s data and can therefore be detected correctly by multimode auto detectors that are built into some software.

HFDL: HFDL stands for High-Frequency Data Link. It is used on the HF bands and is a comprehensive, global, air-ground, communications system. (It also uses VHF and satellite). ACARS (Aircraft Communications Addressing and Reporting System) is part of  this system and with the appropriate software, aircraft routes and progress can be seen on a pc screen. HFDL is a single-tone, phase-shift keyed, text-based, error-checking mode, with a base band audio carrier frequency of 1440 Hz. It is tuned in USB, and the 1440 Hz center is critical for decoding. (Thanks to Monitoring Times blog for the info on this)

NAVTEX: NAVTEX is a world-wide system which transmits navigational and meteorological warnings, and urgent information through coastal stations. Transmissions are on 490kHz (National NAVTEX, broadcasts in local languages) or 518kHz (International NAVTEX, in English), in FEC/SITOR-B mode with special coding. The receiver should be in USB on 489 or 517kHz. Transmissions are at fixed times, the schedules for which can be found on the internet (http://www.dxinfocentre.com/navtex.htm is one source).

navtex 517-2

NAVTEX transmission received on 518kHz on 9th October 09, the text shows the co-ordinates of two special buoys.

SYNOP: Synop is used by ships, and shore based weather stations to report prevailing conditions and forecasts. It is RTTY that has specially coded messages which allow the display of ships and weather buoys/stations and shows weather forecasts and conditions, when used in conjunction with the appropriate software. There is a SYNOP station on 10.101MHz in Germany, it runs at 50baud, with 425Hz shift and is reversed/inverted (when received on USB).

COQUELET: Screenshots and audio are being prepared. This mode is not used a great deal on HF but can sometimes be heard on 18.181MHz where there is an Algerian station (possibly a diplomatic service) that uses Coquelet and broadcasts some non encrypted text in French.

 

coqulet 8 on 18181-3

 

coqulet 8 on 18181-4

PSK AUTO-LOGGING

Using DM780 (part of the excellent and free Ham Radio Deluxe radio control, logbook and decoder package, by Simon Brown HB9DRV) and PSK Reporter (By Phil Gladstone) I am able to set my radio up on any one of the designated PSK31 frequencies and let it monitor for as long as I wish. Each time a callsign is decoded, it is logged, along with its WWlocator. These locators are shown on the PSK reporter page as pins in a scalable map (the map can be displayed as a map, a satellite image or a hybrid of both). Occasionally a callsign is recorded that is not correct (might be a fragmented call or a call that has been corrupted due to noise or QRM). These are surprisingly rare considering the amount of traffic that the program decodes and shows just how good the programming is. It is fascinating to see the map at the end of a monitoring session, and seeing what the chosen band has had to offer that day. I have been doing this for some months now and am noticing interesting propagation trends that tie in with established theory. Below are some screenshots from the DM780 program taken at various times and dates that  illustrate the diversity of stations on the air. Some callsigns appear most days, others of course are only heard once.

Click on each map for a bigger and more detailed picture, then use your browser ‘back’ button to return here.

1_192x1122_192x1123_194x118

L-R:  PSK31 stns on 20m (23 Aug 09); PSK31 stns on 20m (25 Sep 09); DXCC count on 20m PSK31 (2 weeks monitoring)

4_192x1125_194x1186 oct 09 20m_341x129

USA on 20m PSK31 (25 Sep 09); DXCC count on 20m PSK31 (1 month monitoring); Screenshot of 20m taken on 6 Oct 2009 showing stations heard on all 6 continents (for ‘All Continents’ awards, Antarctica is not counted) on 20m PSK31. Although I have shown stations from all continents, they were not the first heard from the particular continent. I have chosen the stations listed below because they have the shortest time frame from first to last.

Elapsed time between hearing the first continent and the last was a mere 53 minutes! just think how quick it could have been done if I was using good antennas!

Continent Time Callsign
Africa 1620 EA8CCA
Europe 1624 IW5BAX
Asia 1631 RU9SL
N.America 1637 W5NA
Oceania 1639 YB4IR
S.America 1713 PT7DX

 

It is not unusual for me to hear over 90 countries on PSK31 during the course of a week (last week was 96 for instance), certainly that has been the case over the spring/summer, with 50-60 countries per day being about average. – and that is at the sunspot MINIMA! I would think that in a few years time I should be hearing over 100 countries in a week without a problem. Bearing in mind I am using a very low antenna, it just goes to show that even when you think the bands are quiet and there is no propagation – most of the time you would be wrong! 20m has been open from early morning to late night, and probably 24 hours but I have not monitored in the small hours of the early morning – yet!

DIGITAL SUB-BAND ACTIVITY

Below are some shots taken from the ‘Spectrum Laboratory’ program (freeware). Each shot covers about 1 hour of elapsed time and shows the amount of activity on the given sub band during that time. Click on each thumbnail for a full size image, together with explanatory text.

tn_cap8tn_cap28tn_cap37tn_cap62tn_cap64tn_cap90tn_cap106

tn_cap119tn_cap126tn_cap128tn_cap129

PLEASE NOTE:

**If you have problems playing back the MP3 files  please read this, it might help:-  I found that Windows media player would not play one of my mp3’s but Quicktime would. Having done some research I see it is a common problem with WMP.  Use the free Apple Quicktime player or similar – also consider installing a codec pack (again many can be downloaded and installed for free) such as the AC3 codec pack and that will (or should allow you to play any, or most,  mp3’s you come across as well as other formats)**

To read a more updated info you can visit the Source of this article here : hfradio.org.uk

LNA4ALL and LNA4HF presents: Homemade LNA

I made this LNA at home using minimal materials and money (MMIC chip and SMA connectors were bought on eBay for few dollars). It uses era-3sm+ MMIC which has 17­-23 dB gain and NF of 2.6-2.8 dB. I am fairly sure it should not be called low noise amplifier because of that noise figure but it’s OK for me, I couldn’t really find a better MMIC. I was inspired to make this LNA by lna4all and lna4hf so the board is somewhat similar. If you are not into making PCBs and want to buy one you should definitely choose lna4all or lna4hf. http://lna4all.blogspot.com/ http://lna4hf.blogspot.com/

http://www.changpuak.ch/electronics/mar_era_bias.php I used this site to calculate values. Lower limit is defined by coupling capacitors and is 4 MHz because I was choosing from the components I have. 4 MHz is enough for me because I am also currently making an upconverter for rtlsdr and it will have the same MMIC on board but with lower limit of few kHz.

Schematics.

Board layout. I assume that RF design gurus are bleeding from their eyes right now but that is what I came up with, I have only designed few boards in my life, let alone RF boards. (If you can suggest any improvements feel free to comment).

Ready for printing.

Printed with laser printer on glossy paper. The board is 2.5×2.5cm.

Transferred top side. I masked all imperfections with CD marker.

Bottom side.

I use sodium persulfate to etch.

I find it better than ferric chloride because it stays transparent and does not leave permanent stains on everything it touches.

Etched board.

Bottom.

Drilled holes.

Vias are ugly now but the paint will hide it.

Painted with glass paint and baked.

Cleared all pads using scalpel.

Finished board. Soldering was done by hand with regular iron so it is not ideal but I am satisfied.

FM station without LNA and with 0 dB gain in SDR#.

Same signal with LNA.

Simple Xtal Oven: “Reduce the PPM of the crystal and your RTL-SDR with this c00l mod”.

Simple xtal oven for accurate clocks
A one-transistor xtal oven gives stable xtal temperature for very accurate clocks.

What does it do?

This is a very simple and easy to make temperature controller and heater to be attached to a xtal (crystal). The xtal is normally used to clock a microcontroller (PIC or ATMEL etc).

Normally xtals provide an accurate clock to 50 or 100 PPM (parts per million) making them useful for real-time clocks, like in your wristwatch. However their frequency output changes with temperature.

This circuit keeps the xtal at a constant temperature – commonly called a “xtal oven” also called an “Oven Controlled Xtal Oscillator” or “OCXO”. Hence the xtal error is reduced to 1 – 10 PPM and the clock will keep almost perfect time.

Thsi circuit can be built with very common cheap parts and means you don’t have to find or buy an expensive xtal oven.

Coupled with my 1-sec PIC timer algorithm HERE you can use ANY value of xtal to build a very high accuracy clock.

How does it work?

Three small resistors are glued to the body of the xtal. These act as “heater elements” and get warm when current is passed through them and heat the entire xtal body. This current is controlled by a little darlington transistor. The temperature is sensed by a cheap NTC thermistor, which controls the current; increasing current if too cold, decreasing current if too hot. Simple negative feedback.

Beacuse the time constant is slow, determined by the thermal mass (mechanical time constant) the circuit stabilises to a regulated temperature at about 35’C (about body temp, slightly above room temp) and remains at that temperature at all times, unless exposed to a temperature extreme that it cannot cope with. This is designed for indoor use and with the parts shown has plenty of range to cope with typical indoor temperatures.

Clever Roman Black-style minimum-parts shortcuts; A single transistor seems a poor choice for a temperature controller as it’s performance changes with temperature! However by including the transistor itself in the thermal mass, the transistor is now kept at a constant regulated temperature like the other parts!

To look into this further; the NTC thermistor (and the other resistors) act to reduce the heater current as temp increases. This is the desired effect. However the transistor tries to increase the current with temp increase, the opposite to the desired effect! I was worried that this may be a problem until I had done enough testing with the actual device, but there was no problem! It acts like a balance; the thermistor (NTC of I) tries to do the “right thing” and the transistor (PTC of I) tries to do the “wrong thing” and because the thermistor has a MUCH greater effect on operation the whole circuit acts globally as (NTC of I) and it works.

How to make it!

PARTS LIST;

 

  • the xtal
  • TO-92 (small) NPN darlington transistor (any type, or 2 normal NPN’s wired as darlington)
  • cheap NTC thermistor (I used DickSmith 100k type, =55k @ 40’C)
  • 220k resistor
  • 0.01 to 0.1uF small capacitor (value not critical!)
  • 3x 39 ohm resistors
  • superglue
  • araldite (5 minute epoxy)
  • (for adjustment) 150k resistor, 220k trimpot, multimeterINSTRUCTIONS;1. Glue the 3 heater resistors and thermistor to the xtal body with TINY spots of superglue. Glue the thermistor in good contact with the xtal body, and touching the 2 heater resistors on that side of the xtal. Cure the superglue for a few minutes under a warm desk lamp.

    2. Glue the transistor (or 2 if you make your own darlington) to the xtal body.

    3. Using pointy pliers gently bend the legs to the right position and trim to length for neat construction. Put a small spot of solder on the joins and check with magnifying glass. Above you can see the green thermistor on the right (between the 2 resistors) and the black transistor and 1 resistor on the left.

    4. Trim the capacitor leads and solder it across the thermistor.

    5. Hook up 2 short power wires; the +5v end of the TOP heater resistor (this is the +5v connection) and the 0v (gnd) is connected to the emitter of the transistor (see schematic).

    6. Connect the 100k and 220k trimpot in series (chain) and connect them from +5v to the base of the transistor. See schematic.

    7. Turn the trimpot to centre and connect REGULATED +5v power through a 50 mA ammeter (multimeter on 200 mA range will do).

    8. (See above photo – testing and adjusting) Adjust trimpot to about 15mA total current draw. Allow temperature to stabilise (may take 30 seconds) and adjust the trimpot again if needed. The trimpot sets the temperature of the xtal, you can measure it if you have a thermometer (I use an infrared optical thermometer), otherwise 15mA raises the xtal about 8’C above room temperature so you can work from that.

    9. Now disconnect power and disconnect the trimpot and measure it, then replace with a fixed resistor to reduce size and make it neater and more reliable.

    10. Test it still works with just the fixed resistor, I chose 220k which regulated at about 35’C. Test that it draws less current when warmed by a desk lamp, and draws more current when cooled by a fan. If it all tests ok, shorten all the leads and make it more compact and neat ready to cover with epoxy.

    11. Put a thick layer of araldite (5 minute epoxy) all over the whole thing but make sure your 2 (or 3) xtal leads and the 2 power leads come out dry so you can solder to them later! The araldite forms a decent thermal conductor and makes the entire circuit into one thermal mass.

    12. Cure the araldite under a desk lamp (see above) at “warm sun” temp (50’C) for half an hour or so. Then test it again and hope you didn’t mess it up! You can clearly see above the green thermistor surrounded by 2 heating resistors.

    13. Cover the entire xtal oven with generous layer of styrofoam insulation. This reduces its power consumption. Seal any air gaps with more epoxy or silicone etc.

    14. Cover with aluminium foil or other RF shield, wrap a ground wire around that and glue in place (optional).

    Note! I haven’t covered this one with styrofoam yet until I build the clock because I want to check size and clearances etc. I did wrap it in a large loose wad of tissue paper (for testing) and current consumption dropped a few mA as expected. Overall, the xtal oven does seem to work really well! Heating or cooling it performs exactly as expected; the heater current responds appropriately and my infrared thermometer says it remains at 35’C at all times. Success!

    Uses for the xtal oven

    I built mine for a precision real-time clock for my loungeroom. This circuit will also be handy for anyone building my Binary Clock Kit or building their own clock using my 1-sec PIC timer algorithm.

    If you are making your own clock, you will need to calibrate your software to match the new exact xtal frequency. I suggest using a GPS, these display (and serial output) a time code that is locked to the atomic clocks on the GPS satelites. Adjust it once a week and if it gains/loses less than a second a week that will be less than a minute a year error. With a little effort to calibrate it, your xtal oven clock should be accurate to a few seconds a year.

    You could also retrofit this circuit to an existing clock, or a test instrument like a frequency meter. Or certain amateur radio (HAM) equipment that needs a stable frequency. All it requires is a +5v regulated supply and 0 – 40 mA, (usually less than 20mA).

    Adapting my design for SERIOUS USE

    If you need more serious temperature control, especially for outdoor use with large temperature variations, the circuit needs to be improved. I suggest using more current through the heater resistors (reduce their resistance), obviously to cope with lower (MIN) temperatures you need more heater power. Good thermal insulation will help both in performance and to reduce total current needs.

    To cope with higher (MAX) temps you need to run the xtal at a higher temp than any temp the device will experience. If it is to run in a car or robot in the hot sun this might be a considerable temperature. Again this may involve increasing the power to the heater resistors.

    Finally you might want to switch to a higher precision temperature controller, ie use a 8-pin comparator chip instead of a single transistor! Either way you can expect to do a lot of temperature testing. It might be easier just to buy a commercial OCXO providing it can handle your temp range needs.

    So why not just BUY a xtal oven??

    Where’s the fun in that?? 🙂 And anyway, I built this in about an hour from $1 worth of parts. A commercial OCXO costs a LOT more than that with postage and would have taken a week or more to get here.

    My circuit is also more energy efficient than commercial OCXO products and only needs a few mA as it was tailored for my needs (ie low power indoor use within a limited temp range).

    And this way I got to pick my own xtal frequency too.

    – end –

You can also find more interesting stuff about High accuracy PIC timing systems HERE at www.romanblack.com

Source:  www.romanblack.com