> menu uitklappen <
> apparatuur foto's
> Racal Cougar
> meetapparatuur (info)
> Rigol DSA815-TG
> surplus apparatuur
> telefonie surplus
Pixie CW transceiver module
The Pixie CW (continious wave/telegraphy) QRPP (low output power) transceiver is a well known classic kit among hamradio operators. It's a small and simple design for a low power HF crystal operated telelegraphy transmitter. This kit can be bought in eBay for less than $10,00. The transmitter will work on 9...14 VDC. In standby mode it'll draw 10 mA @ 9 VDC. The RF output power is 800 mW (29,0 dBm) @ 9 VDC and 1,2 W (30,7 dBm) @ 12 VDC supply voltage. The "beating heart" of the transmitter is a 7,023 MHz crystal. Since there's a diode and variable resistor, the oscillation frequency of the received frequency can be set between 7,023...7,026 MHz. (TX frequency is fixed.) Since it works on 9 VDC, it'll work on a 9 VDC monoblock battery. When built into a small metal (Altoids?!) case, it's ideal for portable operation. The case protects the circuit from mechanical inpact also it's a good RF shielding to prevent influences from nearby signals.
There's is 3-pole LPF (low pass filter) filter onboard to reduce possible harmonics. The filter consists of two 470 pF capacitors and one 1 uH inductor. I calculated the response of the filter with (free) Elsie software. it is assumed that the filter in- and output is 50 Ohms. The loss on (f) 7,023 MHz is with 0,1 dB is negelectable. The signal reduction on 2*f = 8,9 dB, 3*f = 20,8 dB and at 4*f = 28,0 dB.
According to the Dutch "Gebruikersbepalingen amateurfrequentiegebruik" number 212-3.3 of the 20th of augustus 2012 part 3.1 it's not allowed to radiate spurious signals of -36 dBm (0,25 nW) or -60 dBc between 50...1.000 MHz wherever the highest number is applicable. The highest harmonic (at 14,046 MHz) is -8,9 dBc (decibell relative to the carrier). If the crystal quality is very good, the signal strength of the hartmonics could be small and therefore meet the legal limits. The plan is to verify the frequency spectrum of the Pixie transmitter. Even if the signal stength is within the legal limits, it would me wise to add/replace the 3-pole filter with a 5- or even a 7-pole filter. Examples of the suggested fitlers are shown below. These filters are designed with "real world" components. The 7-pole filter will met the legal limits easily with 68 dB signal reduction on 14,046 MHz.
20161101 - My first idea was to improve the low pass filter design, but after a while it seemed a better idea to replace/change the filter design to a band pass filter (BPF). By "adding" one component (and replacing the positions of the 7-poile filte rdesign) a band pass filter is created. After changing some parameters in Elsie I got a banad pass filter for 7,023 MHz. The filter width is 2 MHz. The supression of 2*f (20 m band) is 60 dB and 82dB supresison at 3*f (15 m band). There are some markers shown at other ham bands. Normally only spurious radiation will occur at harmonics and therefore is 60 dB reduction of signal very nice. The band pass filter wil also function as a preselection filter. Strong signals from other ham bands will be reduced by this design. De transmission loss (S21) of the band pass filter is 0,41 dB and 0,30 dB for the 7-pole low pass filter. The relative loss is rather small and if there's enough space my choice would be the band pass filter shown below.
I overthought the filter designes above. There are several criteria for selecting a suitable filter. Space, desired suppression at a specific frequency and maybe the desire to improve filtering whitout changing the design too much. For the last option I designed a filter shown below. The suppression of spurioous radiation is less than the "complexer one" shown above, but due to component reduction in the design below the "new" components can be placed on the existing board. The two original capacitors C5 and C6 should be removed and be replaced by two sets of a capacitor and inductor in parallel. The result is a rather simple band pass filter with 28 dB more signal reduction at f*2. This coule be a nice compromise between size/complexity and performance.