Radioclub de BevelandenPI4ZLD

Nederlands English

The 10GHz beacon PI7GOE
on 10368.075 MHz
now testing in Middelburg (JO11TL) at 35m asl!

Latest news:
On september 10 at 17:00 UTC, the beacon came back on-air, after the power supply was reset.
The beacon went off-air on september 7 after a heavy thunderstorm. Probably the over-voltage protection was activated by the thunderstorm.

Contents: Introduction Principle of operation Planning Realisation Pictures Reception reports

The 10GHZ beacon PI7GOE has been mentioned in various beacon lists for many years. Yet, the beacon PI7GOE is still not QRV. But this will change soon:
Some SHF-enthousiasts within our Radio Club have set up a plan to become active on 10GHZ with narrowband SSB / CW. We'll start with building and operating the beacon PI7GOE at 10368.075MHz. This beacon is eventually going to be located at the Churchtower in the town Kapelle (locator JO11XL), co-located with the 2m repeater PI3GOE at 145.725/145.125MHz.
Further, as an initiative of the Radio Club, a joint transverter building-project will be started with the intention to stimulate activity on 10GHZ in our province Zeeland. More information about this 10GHz transverter project will be given during the monthly Veron A33 meeting on october 30 1998, in the Radio Club (see main page).

Principle of operation
The principle of operation of the 3cm beacon PI7GOE is (for the time being):

Local oscillator: G4DDK004 2-2.6GHZ LO source: Crystal oscillator on 108.00078 Mhz, output: 10mW at 2592.01875 MHz
This module is in use at many beacons in Europe. The design has been described in the RSGB Microwave Handbook, volume 2 & 3 and in DUBUS Technik IV.
The Xtal is being held at a constant temperature of 40° Centigrade by means of a Murata crystal heater.

2.5 to 10GHz multiplier G3WDG001: In this module the 2.592MHZ LO signal is multiplied by 4 to the desired transmitting frequency of 10368.075MHz. Output power is approx. 100-150mW.
This module is also in use at many other 10GHz beacons. The design has been published in the RSGB Microwave Handbook, volume 3 and in DUBUS Technik IV.

1Watt 10GHz Power Amplifier G3WDG007: Initially we don't have plans to build this module and use it for PI7GOE as we're still looking for a sponsor for this module or at least for the expensive MGF2430A power GaAs FET.
If we later discover that the signal strength of the beacon is insufficient, we always can add this module, if we have the financial funds.

Slotted waveguide antenna: The signal of the x4 multiplier module is fed to waveguide (WR90) via a SMA-Waveguide transition directly at the multiplier output connector. The piece of waveguide is acting as a omni-directional antenna.
The dimensions of the antenna are calculated with the Waveguide Slot Antenna Calculator by Stephen Bell KB7TRZ, a program written in Mathcad, that I found at the Tech-tips page of Bryan W6BY's homepage.
The slotted waveguide antenna will have a length of approx. 30cm and both broad waveguid walls will have 18 slots (for a total of 36 slots) of size 14.5mm by 1.87mm at a distance of 18.7mm. The gain of this antenna is estimated at 12dBd.
Question: Now is this a 18-slot waveguide antenna or a 36-slot waveguide antenna? If you know the answer, please send Email to Eddy PE9GHZ, ex.PE1CIG
You can view a picture of a slotted waveguide.
Download Excel spreadsheet with all sizes of the slotted waveguide antenna in WR90.

Beacon CW keyer: For identification purposes of the beacon, OM. Willem PA7WLH (ex PE1FYO) has volunteered to develop a CW-keyer. Altough many designs do exist, this keyer is built in a 87C51 microcontroller, the program (firmware) for the keyer is also written by Willem. More information and the schematics of the keyer can be found at his website
With this design it is possible to build a versatile CW keyer with a very little amount of components: With 8 Jumpers, the CW-speed, polarity of the CW-keying signal, choice of several messages and other functions can be adjusted. The CW keyer also provides a 600HZ audio output, for use as a repeater identification.
The keyer is modulating the LO-signal in FSK, with a 850Hz frequency shift at the beacon transmitting frequency 10368.075MHz. This means a shift of only 9HZ at the crystal oscillator frequency!
During implementation the following feature was added: The keyer has a control signal to alternate the output power of the beacon-PA between normal and low-power, so the listener can hear the difference with 10-12dB lower transmitting power.

The (adjusted) time schedule of this project is:

september 4-18 1998: Functional design of the Beacon CW keyer.
september 1998: Building the CW keyer by Willem PA7WLH (ex PE1FYO).
september-december 1998: Building the G4DDK004 2.5GHz LO module, by Eddy PE9GHZ, ex.PE1CIG.
october 1998: Machining the slots in the waveguide, by Arjen Verhaar (who also made our 13cm 2C39-cavity PA).
december 1998 - january 1999: Building the G3WDG001 2,5 to 10GHz multiplier module, by Eddy PE9GHZ, ex.PE1CIG.
april 1999: Testing the beacon. Operational test on temporary location (Middelburg JO11TL).
june 1999: Setting up the beacon at it's destined QTH: The Churchtower in Kapelle JO11XL.

(Problems encountered during) realisation
October 1998: The oscillator signal of the 2.5GHz LO module is not stable enough, it sounds jabberish/howling when listened to the 12th harmonic on 1296.0093MHz, with a 23cm SSB receiver.
Listen to a recording of the signal in WAV format.
After discussing this problem with Sam G4DDK, several parts (such as the crystal) have been checked.
Eventually an unstable carbon resistor R4 (470ohm) across the coil L1 appeared to cause some of the instability. After replacing this resistor with a metalfilm one, the signal sounds more stable (but isn't perfect yet).
November 1 1998: Replaced the 2 BFY90 transistors (TR1 and TR2) in the oscillator and buffer circuit bij new ones. Sam G4DDK advised to mount the metal-case almost direct on the PCB with leads as short as possible, and the cases soldered to the groundplane of the PCB. Unfortunately this still not results in improvement of the signal stability.
November 4 1998: Tried another crystal (106.5MHz) in the oscillator circuit. The 12th harmonic (on 1278MHz) sounds much more stable. So my conclusion is that the original 108.00078MHz crystal must be causing the instability. I have sent Email to the crystal supplier with a complaint about the bad crystal. We're awaiting his reaction...

December 11 1998: An envelope containing a new 108.00078MHz crystal was in my mailbox today. Unfortunately I had no time to test it right away - Today is my son Julian's first birthday!
December 13 1998: I've mounted the new crystal in the 2.5GHz LO. After quickly aligning the operating frequency two things were noticed: The signal (at 1296.009MHz) is perfectly stable, no jitter at all. Additionally, the LO module now delivers 6dB more output power (now almost 20mW) than with the unstable crystal. Clearly an improvement!
December 17 1998: The LO module has been powered on for several days now. The frequency drift has been nearly zero (just a few hundred hertz measured at 1296.009 MHz!). This was measured with the tin enclosure closed, but with no thermal isolation around it.
Next week during the holidays i hope to build the G3WDG001 2.5 to 10GHz multiplier module.
(Note: Yours truly got the flew just before the holidays so this plan could not be realised).

January 2 1999: Because of financial advantage, I got the opportunity to invest in a piece of measuring equipment that is much desired in our hobby: I've purchased a (2nd hand) spectrumanalyser, suitable up to 22GHz. Now the signal of the 2.5GHz LO-module and later on the multiplier module can be analysed really good. (Pictures of this will follow).
January 8 1999: Due to an unknown cause (probably a powerpeak on the supply), it appears that the 2.5GHz LO-module does not deliver any output power anymore. After several evenings of measuring and successively replacing the transistors and some trapezoid-C's whose capacity had become very temperature-dependant, the output signal is now back to normal level, even more than that: It now delivers around +13dBm (clean).
January 22 1999: During the past week I've spent several evenings building the G3WDG001 2.5GHz to 10GHz multiplier/PA module . The very first measuring results are positive: Output reaches +16dBm at 10.368 GHz and the spectrum is reasonably clean.
I suspect this can still be improved: In the final stage I did not use a MFG1302 GaAsFET (as per design), but I applied a ATF-26884, which according to specifications can deliver +18dBm at 12GHz and should be able to deliver at least as much at 10GHz, maybe even 20dBm. The (teflon) PCB has been designed with etched in- and output "Lumped element capacitors", providing proper matching for the MFG1302 but naturally not for the applied ATF-26884 device. In the forthcoming time I'll be trying to improve matching to the ATF-26884 so see if output power can be improved a few dB's.
January 29 1999: After some experimenting with small pieces of copperfoil used as tuning capacitors on the microstriplines near the last amplifier stage, a power level of +20dBM was measured for a while. Unfortunately the ATF26884-device was destroyed immediately by a foolish soldering action, while the module was operating... Since it was the only one available, I had to temporarily put a MGF1302 in place and will order some new ATF's later on. With the MGF1302 as final stage the output power is only +15dBm (32mWatt).
March 26 1999: The ATF-26884's are in back-order at my supplier. So I continued with the mechanical work on the beacon: Put the modules in a watertight, insulated case. Put a thin plastic tube around the slotted-waveguide antenna to make it waterproof. (white waterpipe of a watercloset, after an idea in Electron issue April 1999).
March 31 1999: The beacon now has been outside on top of my shed, in the rain - No problems with that. tonight I've mounted the beacon in the top of the antenna tower (12m ASL) at my home-QTH (JO11TL). If everythingl keeps working as it should, the beacon will be moved to an other (higher) temporary QTH after april 12th. Maybe the ATF-2668 devices will be delivered in the mean time.
April 2 1999: The first reception reports of the tests from my home-QTH were received: The beacon has been heard in Goes, Katwijk en Breda.
April 4 1999: Adjusted the transmitting frequency of the beacon some 10kHz, since reports mentioned the beacon was transmitting on 10368.055 - 10368.059. It should now transmit around .070 .
April 7 1999: Today the ordered ATF-26884 fets dropped in the mailbox. The evening was dry and not windy, so I removed the beacon from my tower at once. Later in the evening I replaced the fet in the final stage with a ATF-26884 device. After some tweaking with small pieces of foil along the microstriplines, the output power from the G3WDG multiplier/PA module reached 19-20dBm (80-100mW), 4dB more than with the original MGF1302 device. Then I built the normal/low-power switchover circuit into the module, so the CW-keyer module can alternate the TX-power. Power reduction is set at 12dB, which means the beacon transmits 80mW in high power and 6mW in low power mode. The transmitting frequency also was adjusted some kHz upward.
April 8 1999: The beacon will be placed in the antennatower (12m asl) at my home-QTH (JO11TL) again, for a week of testing. After one or two weeks the beacon will be moved to an other testlocation (20m asl) in JO11TK.
April 9 1999: Willem PA7WLH brought along the final version of the Beacon CW-keyer and handed it over to PE9GHZ, ex.PE1CIG. Some added features of the new keyer: Controlling signal for normal/low-power that becomes active every 5 cycles of the keying loop, with CW-text indicating low power.
Jumper-programmable are: Call-ID CW-speed, Infotext CW-speed, Selection of Normal QTH / Test QTH (with adjusted Infotext).
The new CW-keyer will be built into the beacon during next week. After that, the beacon will be measured by Rob PE1KHX at his QRL, where he has access to a well-calibrated Frequency Counter and Power Meter up to 18GHz.
April 29 1999: The beacon is moved to the testing-QTH Middelburg (JO11TL) at 35m asl. Special thanks to Paul PE1RJV for making his qth on the top floor of the tower-flat available for testing the beacon.

Picture gallery
The 2.5GHz oscillator module and the 2.5 to 10GHz multiplier module
A peek in the G3WDG001 multiplier module:
The copper screen-side, with voltageregulator and cavity. Down-right: the SMA/Waveguide transition.
A peek in the G3WDG001 multiplier and G4DDK004 oscillator modules:
Center: the tracksite, below: the Oscillator module G4DDK004. At the extreme upper-right: SMA/Waveguide transition.
Slotted waveguide antenna for 10.368 GHz
The 18-slot waveguide antenna, made by Arjen Verhaar:
More pictures of the slotted waveguide antenna:
These 8 photo's were taken with a Creative Webcam and therefore the sharpness is not as good as it could be.
Spectrum analysis of the signals from the modules:
The frequency spectrum of the G4DDK004 2.556GHz LO module:
Frequency spectrum of the G3WDG001 2.5 to 10GHz multiplier module
More pictures will follow later..

Reception reports
Your reception reports of the beacon PI7GOE are welcome!
PSE QSL PI7GOE via PI4ZLD, Region 33 (via the Dutch Qsl Bureau or direct to PI4ZLD).
You may also contact us by submitting the comment form below.

Reception reports received so far during test at location JO11TL, Power=30mW, antenna at 12m asl:

 date     time    call  locator  QRG  RST
 19990402 20.00Z PI4ZLD  JO11WM .075  519 - 529
 19990402 21.00Z PA0WWM  JO22FE .055  519 - 549
 19990403 09.00Z PA3DYS  JO21JO .059  53RS
 19990405 10.20Z PA0WWM  JO22FE .065  52RS QTF 220°
Reception reports received so far during test at location JO11TL, Power=80mW, antenna at 12m asl:
 1990410  16.45Z PA0WWM  JO22FE .067  52RS QTF 220° - 195°
 date     time    call  locator  QRG  RST


This page was viewed times since march 30 1999.
Back to Radio Club de Bevelanden Homepage
Links to other Amateur Radio sites

If you have any comments or suggestions: Please submit a comment form

Eddy van Loo , PE9GHZ, ex.PE1CIG
Last modified:

Graphics by The GIMP Validated with
CSE 3310 HTML Validator