Monday, April 2, 2012

Light Communications Idea

Once again, the entertainment industry inspired me to this one. In stage illumination, the most recent development is the use of multi-color LED spots and washers. Those devices contain either three or four differently colored groups of ultra-bright LEDs. Controlling of the spots or washlight is usually done a serial protocol called DMX, by so called DMX-controllers or DMX control software.

In stage lighting language one controls different settings of "fixtures" (the lighting devices) and stores this control settings in "scenes". The scenes than can be called either manually or automatically as a sequence called "chase". The frequency in which the scenes of a chase are being called usually can be set by a sliding fader.

So, what's the trick about all this and where is the link to amateur radio?
Very simple, in long range light communication or cloud scatter experiment, usually QRSS is used. Now the link should be obvious... the fixture(s) are, very obviously, the light source(s), while the DMX-controller serves a beacon keyer.

A simple series of unmodulated dots (A1A) can be programmed with the following 2 scenes:
  1. red on all fixtures to 100%
  2. red on all fixtures to 0%
The next step would be to program of a chase of scene 1 and scene 2.

Unmodulated signals may be hard to discriminate. However, with the strobe function, the entertainment industry offers a solution to this problem. The strobe will create sidebands in the known fashion.
So, for a modulated signal (A2A) the following scenes can be used:
  1. red on all fixtures to 100% with a fast strobe
  2. red on all fixtures to 0%
Again, the chase would simply repeat scenes 1 and 2.

In order to know what I am writing about, I actually bought some material at a local pro-audio store:
Reasons for the decision on just those devices:
The washlight can be controlled by either 3 or 7 DMX channels. 3 channel resemble the control of the red, green and blue LED groups. 7 channels include said RGB-controls and some more stuff, which can be found on the respective webpage (#4=hue, #5=strobe, #6=color cycles, #7=luminance).
The lighting controller employs 8 faders to control 16 channels before switching to another "fixture" (i.e. bank) is required. This 8 channels fader control comes handy to control 7 channels of the SlimPar 38 or (and that's another trick) 3 channels of 2 SlimPar washers. In the latter case, two devices are controlled by a single fixture channel.
Just for the interested: the trick is the address of the washer or spot. The address of the first device (officially called fixture, but this can be confusing here, hence, let's call the individual washers or spots "devices" for now) will be "1".  If the device is using 3 channels, the address of the second device could be "4". In this case, provided the above mentioned controller is used, the first 3 faders would control the first device's R, G and B groups and faders 4, 5 and 6 would control the second device's R, G and B groups. The advantage, the two devices are now dealt with as a single fixture.
Advice: With a 16 channel controller (as the one I am using) one could potentially control 5 3-channel devices, however, the assignment of the fader will be rather confusing. Therefore, I recommend controlling 4 3-channel devices only. For sake of convenience, I would assign the second device to address (channel) 5, the third to address 9 and the fourth to channel 13.

Back to QRSS. Even the cheapest of DMX-controllers with the cheapest of LED-spots would make a real nice light beacon setup. OK, I went for something more sophisticated... since I see a secondary use in my light beacon setup... just in case I want to through a party, I now have a club-worthy lighting setup.

Concluding, there may be "red only" devices. However, stage worthy multi-color devices would even allow for multiplexing, depending on the receiver filters. The ones I use through out 1500lx @ 1m each, all LEDs engaged (at a power consumption of about 20W). Since I bought 4 (for good measures) that would be 6000lx @ 1m in white or about 2000lx using just one color.
Now I need to work out some receiver concept.

Friday, March 2, 2012


There is a new mode on the market which was named OPERA. The interesting bit about it, it uses A1A modulation. This makes it suitable for LF and MF operations in The Netherlands, reason enough for me to have a further look.

It seems, the mode is used on HF and VHF too. There are frequencies mentioned for all the bands.
Since one has to start somewhere, 30m is the band of choice.

The OEPRA center frequency is mentioned to be 10.1365MHz. This frequency should immediatly ring a bell, at least if the dear reader is familiar with my blog and stuff I posted earlier about.

There is a standart crystal (and canned oscillator), which comes along quite handy for a subharmonic receiver: 5.0688MHz. This as an oscillator would result in a direct conversion LO of 10.1376MHz. Consequently, such a receiver would create an audio signal of 1.1kHz on the lower side band. Since the modulation is A1A, this does not matter at all. It is conceivable to add a crystal notch filter for the upper side band.

As to the transmitter, the solution is similarly simple. Just an oscillator using a 5.0688MHz crystal with a tiny downwards pull of 550Hz. That should be easily doable. Should I intend to build such a transmitter, I will use a VXO as to just cover the frequency range 10.1363-10.1369MHz, which, translated into oscillator range reads 5.068150-5.068450MHz, reflecting a VXO range of 300Hz.
Frequency doubling could be done either actively or passively by just two diodes.
The trusty old 74HC240 could serve as a driver or power amplifier.

Since the mode is keyed, both RX and TX can be easily tied together in a QSK fashion.
As a bonus, if the mode should become out of fashion once, the rig can easily converted into a QRSS receiver by changing the crystal notch filter to a normal crystal filter and pull the TX oscillator up.

In ITU region 2, this transceiver could be used for Feld Hell. Seen that the Hell frequency is rather close to the local oscillator frequency, it may be desireable to change the RX oscillator to 5.0680MHz. CMOS oscillators are available for this frequency.

Wednesday, February 29, 2012

472kHz I/Q-SDR kit

Nothing is for free, up to now, there is no kit available for 472kHz.
However, there is one for 136kHz, which can be easily modified to match the new band.
Check out box73's longwave I/Q-SDR kit.

You will see that a 15MHz signal is divided by 25. This results in a 600kHz signal which is further divded by 4 in order to create the phase shift. All in all, this ends up in a center frequency of 150kHz.

We can use the same oscillator and divide the signal by 8. This results in 1.875MHz, which will be further divided by 4 providing a center frequency of 468.75kHz.
With a sampling rate of mere 24kbps, or +/- 12kHz bandwidth, the entire band (472 to 479kHz) will be covered.

The digital part is rather simple to modify. A suitable ripple counter could be the 74HC93.
The frontend is even more simple... just pick a 455kHz i.f.-filter/transformer and replace C1, C2 and L1.

Should the old experimental range, somewhere above 500kHz, be a desired range, the additional modification would simply replacing the 15MHz canned oscillator with a 16MHz one.

I will see if I can persuade the OM at to provide such a kit.

Tuesday, February 28, 2012

MF/HF Aerial

And along came an idea....
You may have heard about the CobbWeb aerial. Essentially, this is a cluster of dipoles for the bands 20m, 17m 15m, 12m and 10m. The cluster is fed via a coax choke.
Maybe there is a way to squeeze more out of this aerial. The amount of wire in the dipole array creates a decent capacity, I figure.
It may be worth a try to build such an aerial, feed it with RG-6. And, for MF purposes, use the feedline's shield (and core) as vertical and the dipole array as capacitive load. The rf choke could further help to increase the load on the (very) short Marconi for 600m.

This would be somewhat like the antenna disclosed in the U.S. Patent 3,569,970, (see Figs.7a,7b) but using the CobWebb in place of the stretched dipoles.

Wednesday, February 22, 2012

472kHz Phasing Transmitter

We have seen that a 1.8432MHz oscillator will provide us with a 460.8kHz I/Q-SDR LO.
This is very much in a comfortable range for of the new amateur radio MF band, i.e. 11.2kHz to the lower band edge and 18.2kHz to the higher band edge.
Now, how to generate the modulator signal? Phasing style, the easiest would be to build an oscillator for the 44.8 to 72.8kHz and use two Flip-Flops to generated the 90 degrees phase shift.

  1. Such an oscillator could be a rather simple function generator. Other solutions could be based on micro-controllers such as PICs, PICAXE, AT-Tiny, etc. With such a controller, it would also be possible to program features like memory channels, frequency display, beacon-keyer...
  2. Another approach would be to build a crystal oscillator, using cheap industrial xtals, and divide it down. Some ideas could be crystals from the XMHz range divided by N (by means of a binary counter) before feeding the Flip-Flops:
    • 3.000MHz / 64 = 46.88kHz resulting in 472.5kHz
    • 3.072MHz / 64 = 48.0kHz resulting in 472.8kHz
    • 3.2768MHz / 64 = 51.2kHz finally resulting in 473.6kHz
    • 3.579545MHz / 64 = 55.93kHz resulting in 474.78kHz
    • 3.6864MHz / 64 = 57.6kHz resulting in 475.2kHz
    • 3.93216MHz / 64 = 61.44kHz resulting in 476.16kHz
    • 4.000MHz / 64 = 62.5kHz resulting in 476.4kHz
    • 4.096MHz / 64 = 64.0kHz resulting in 4768kHz
    • 4.194394MhZ / 64 = 65.54kHz resulting in 477.2kHz
    • 4.433619MHz / 64 = 69.28kHz resulting in 478.1kHz
  3. In the light of the above, ham-radio crystal such as (in MHz) 3.530, 3.535, 5.540, 3.550, 3.555, 3.560, 3.575611, 3.880, 3.885 can fill in gaps. Those crystals are found at "expanded spectrum systems".
  4. With some luck, one finds tons and tons of surplus crystals in the range of 2.8672MHz to 4.6592MHz. As I recall, there where channelised commercial transceivers (e.g. military, maritime etc.) making use of crystals in that range.
  5. Similar to the crystal approach, one could consider to use 3.58MHz, 4.0MHz, 4.19Mhz, 4.50MHz and 4.91MHz ceramic resonators for a VFO. The 6.00MHz, 6.50MHz and 8.00MHz resonators would require one additional division.
  6. The deluxe version of it all would be a DDS for the range 2.8672MHz to 4.6592MHz. I wonder is there is any kit in which the LO offset can be easily programmed to (f/256)+460800Hz. Maybe a project with the DDS60 board.
When modulating the phase shifted AF signals, one has to consider that those are essentially square waves. In order to reduce harmonics, it would be required to do some severe low pass filtering at about 19kHz before injecting the signals into the I/Q mixers.

As to receiving, the 11.2kHz to 18.2kHz is in the comfort zone of any 48kbps sampling sound card.

There you have it, my presently preferred solution for the new 600m amateur radio band.

Sunday, February 19, 2012

600m Signal Source

As we know by now, 472-479kHz it will be. In an earlier post I revealed some "cheap" frequencies which would mix into the new band.

Some further options using industrial crystals:

DigiKey sells a 4.754687MHz crystal... count to (divide by) 5 and further divide by two results in 475kHz. A super-VXO could be an option here.

The above mentioned count to 5 solution applies to the following crystals, found at the same source: 9.494531MHz, 9.509375MHz and 9.545MHz. Other crystals would allow for an out of band I/Q-SDR LO: 9.600MHz, 9.625MHz and 9.7941MHz. Here, the chain would be count to 5, divide by 4.

Further: 18.9375MHz, 19.0625MHz  and 19.069928MHz and for I/Q-SDR: 18.869MHz, 18.8696MHz, 19.200MHz, 19.280MHz and 19.440MHz. Consequenctly, the chain would be count to 5, divide by 8.

Plus: 28.59375MHz, 28.5938MHz, 28.636MHz, 28.6363MHz and 28.63636MHz (count to 3, count to 5, divide by 4).

Taking it even higher: 38.000MHz and 38.00053MHz (count to 5, divide by 16). I/Q-SDR: 38.400MHz and 38.880MHz.

Also found at DigiKey: a 7.680MHz crystal. Divide by 16 results in 480kHz. Again, a super-VXO and some (severe) down pull should generate a signal in the band. This crystal provides easy access to I/Q-SDR: LO spot om 480kHz, even a mere 24kbps sample rate would cover the whole band. The same applies to the 15.360MHz found at the same store.

All the above mentioned crystals are of industrial kind. One option would be order one for the favorit solution, the other option would be to carefully watch out for those frequencies before dumping old computers & Co.

Tuesday, January 17, 2012

600m SDR RX (TX)

As we know, presently there are a couple of frequencies of the 600m band open to amateur radio operators.
Most of authorities allow transmission somewhere above 500kHz. In The Netherlands the permitted range is 501-505kHz. In the future, depending on the decision of the WRC12, this will possibly change to 472 to 480kHz. The U.S.of A. proposed the following ranges 461-469 and 471-478 kHz.

Lets look at the (inexpensive) option the box73 SDR. The 80m version of this receiver uses a 14.000MHz oscillator. Operation on the 600m band can be achieved by changing the front-end filter and the SDR-LO.
Considering 48kbps sampling, the LO-frequencies would be the following
  • QRG: 470kHz - LO: 1.843MHz
  • QRG: 500kHz - LO: 2.000MHz
This will result is RX-ranges of:
  • 460.8 -/+ 24 = 436.8 .. 484.8
  • 500.0 -/+ 24 = 476.0 .. 524.0
Concluding, a decent 600m receiver can be built with LOs using regular canned oscillators.

600m with CB-XTALs follow up

Hi there... I am back!
Presently, amateurs in The Netherlands are really happy, not only have we got the permission to transmit on 4m, we also may transmit on 600m again. For the time being between 501 and 505kHz.
For this range (well, to the upper QRG of 504kHz) this article could lead to a cheap signal source.

However, in about a month's time, we will know is we need to redesign our exciters, i.e. to the range 472 to 480kHz. So, lets have another look at the cheap crystal solution.
The new range is not that easy really. One solution would be the mixing of two standard xtal frequencies:
  • 4915.2 - 4433.6 = 481.6 somewhat high, could be pulled into the range
A spot on solution would be a 10m QRP crystal (28.060MHz) on its fundamental mixed with a 9.8304MHz standard crystal:
  • 9830.4 - 28060/3 = 9830.4 - 9353.3 = 477.1
Another spot on option: a 40m QRP crystal, mixed with a 6.5536MHz standard xtal.
  • 7030 - 6553.6 = 476.4
Starting from lower frequencies, crystals from
  • 14745.6 - 14270 = 475.6
  • 7025 - 6553.6 = 471.4 (pull!)
  • 3550 - 3072 = 478
  • 4000 - 3530 = 470 (pull!) (*)
  • 3555 - 3072 = 483 (pull!)
Subractive mixing has the advantage that some thermal effects may cancel. Assume to us several standard crystals in a super-VXO.
Mind you, the higher the frequencies, the easier to pull, i.e. create a nice tuning range. However, lower frequencies will give better stability.

(*) Note, there is a 4.00MHz ceramic resonator, which will allows for a nice VFO. Use a Pierce oscillator to obtain a QRG above the ceramic resonator's series resonance.