This photo shows the head unit of the 47Ghz DB6NT transverter with, from lower left to right the 11.7 to 23.5GHz LO doubler and the x2 diode multiplier and mixer. The 3-stage LO doubler uses NE32584 GaAs fets in an single amplifier configuration followed by a pcb filter, a X2 stage and a final amplifier to produce ~ 50-60mW at 23.5Ghz. The upper pcb is the IF and switching. It was recognised in the first attempt that it was quite tricky to optimise the 23GHz multiplier strip due to the integrated approach. Thus this second version gets round this by having a removable 23Ghz strip which can be optimised into a power meter in its own right. Using this method 60mW has been seen at 23Ghz enhancing the mixing properties of the following mixer diode. A small screw is also visible which allows matching into the diode for best performance. High stability and good phase noise are achieved by using QT crystals in the LO from by Eisch-Kafka in Germany.
The LO shown below comprises a multiplier chain which provides a signal at 11Ghz mounted in a back to back pair of die-cast boxes. One side contains the crystal oscillator and RF multipliers, the other side shown here, a PLL using the RDDS system, which is locked to a double ovened OCXO 10Mhz source available on eBay.
This shot on the right shows the transverter mounted on the waveguide switch and cassegrain plastic 30cms antenna with the 4mW FM source on the Tx port. The LO crystal driver multiplier chain for this 4mW source can be either keyed or audio in the form of FM applied.
The 4mW GaAs FET multiplier, an example is shown here, comprises GaAs fet multiplier stages using the NE32584 devices, driven by an 11.772Ghz source capable of being FM modulated or keyed for CW.
MVDS Unit Conversion
Some time ago a Philips ex Microwave Video Distribution System (MVDS) unit was obtained with a view to modifying it from 42Ghz to 47Ghz. These are interesting units comprising a full multiple conversion receive unit down to an IF in the 1GHz region with a free running DSO local oscillator and a dielectric lens antenna. Receive performance should be good as a 42GHz MMIC precedes the single diode mixer. However the spec. sheet for the MMIC seems to indicate that the MMIC response above 42GHz may roll-off rather rapidly. So some modifications were made to the unit to check it out for use on 47GHz. For 47GHz replacing the DSO with a fixed xtal multiplier chain at 9GHz was tried and modifying the following multiplier stages and filters. However the sensitivity of the unit was disappointing with performance no better than a DB6NT bare 47GHz mixer.
And if you thought this is advanced for its time ...look at this link.
Listen here to a typical FM QSO on the band
Why not try a link budget Excel spreadsheet here. This calculates s/n for a given path length knowing the RH and ambient temperature either for an SSB or FM QSO for 24, 47, 76 and 120Ghz and gives some idea if the path will work or not. The Excel workbook comprises 4 sheets. Pick the one for the band your working on and enter the figures where they are shown in blue.
Waveguide coupling into the unit just following the MMIC was also tried ahead of a DB6NT mixer but although the noise level due to the MMIC increased there was not a corresponding increase in sensitivity. MMIC gain at 47Ghz was estimated to be ~6dB.
The photo on the right shows the unit with an injection of the LO at 23.47Ghz to produce an IF at 145Mhz. In this state it was on a par performance-wise with a bare DB6NT mixer and 30cms dish although it should have be considerably better.
In a further effort to make it work well on 47GHz the two pcbs containing the MMIC's were cut up and paired as shown on the left to try and produce a gain block, but despite the combination seeming to have gain and being quite ready to become unstable, its seems the gain is not at 47Ghz!
The verdict seems to be that these MMIC's do not operate up to 47GHz although quite happy at 43GHz. The data sheet for the device does not indicate performance at 47GHz and it must be concluded that this is the reason......unless someone knows different................
A beacon designed and built by G8ACE is now running under my own callsign for control purposes, at the South Coast Beacon site on the south coast. The picture shows the small rectangular box housing the beacon hardware on the mast.
Visit the web site of G8ACE for the full story and the latest news on the 47GHz beacon.
Listen to the beacon here over a 42km line of site path in the New Forest.
47GHz South Coast Beacon
Operation at Batcombe down , Dorset
Using the LMX2541 synthesiser chip on 47GHz
Some recent experiments have been made on the bench using the Texas Instruments LMX2541 synthesiser chip. If you purchase the correct version of the device it can provide 3.7GHz signals locked to an external reference which are comparable in performance to a xtal oscillator driving a multiplier chain even up to the 47 & 76GHz. As an LO for a receiver or a stand alone source for an FM transmission these devices seem to be very good with respect to stability and phase noise as long as the reference is good.
I put together a beacon source with JT4G modulation on it as per the G4JNT examples and the resulting signal can be heard here.
The picture at the right shows the lash up used together with the NMEA GPS feed needed to synchronise the timing of the JT4G signal.