|
Harmonic Number |
Level |
| 2 | -45 |
| 3 | -43 |
| 4 | -52 |
| 5 | -62 |
The transmitter is based on a novel double-ended free-running voltage controlled oscillator (VCO) operating at half the output frequency. The idea behind the VCO operating at half the output frequency is that it should be more resistant to having the centre frequency pulled by variations in loading further down the transmit chain. The active devices in the VCO are a pair of BF494. The mono audio input circuit is totally passive. No pre-emphasis is provided. The audio input is a PCB mounting phono connector. A small 5V voltage regulator provides a stabilised DC reference for the Varicap diode in the VCO.
The output device is a trusty old 2N4427 operating in class C. This picks off the 2nd harmonic of the VCO and amplifies it up to 1W. There are four adjustable capacitors. The first sets the centre frequency of the VCO. The second is on the input match to the 2N4427, the third and fourth form the capacitive legs of a pi matching circuit for the output match of the 2N4427. The only harmonic filtering will be by this pi circuit. A simple push-on aluminium heatsink is fitted to the output device. The RF output connector is a SO-259 (UHF) socket, mounted directly onto the PCB. A simple rectifier/amplifier circuit drives an on-board LED which allows the circuit to be tuned without additional test-equipment, by adjusting the trimmers for maximum LED brightness. The circuit is reverse polarity protected by a series diode in the supply. This diode drops 0.78V with a 16V supply.
The circuit took one hour to build, and worked first time.
The unit was checked at the extremes of the frequency band, with a 16V supply. With the coils as first fitted, the maximum frequency that could be obtained was 106MHz. Pulling apart slightly the inner turns of the VCO tank inductor L1, allowed the frequency to be increased to 108MHz. This was as stated in the supplied instructions. The other three capacitive trimmers required adjusting to maximise the output power when the frequency was changed, this was easy to do. It was found that the maximum brightness of the on-board LED corresponded closely to the maximum power as indicated by a power meter. I recommend using a proper insulated trimming tool to adjust the variable capacitors, in preference to a small screwdriver. The table below shows it is possible to get virtually 1W of output power across the FM broadcast frequency band.
| Frequency (MHz) |
Supply Current (mA) | Output Power (mW) |
| 88 | 203 | 965 |
| 98 | 194 | 976 |
| 108 | 172 | 915 |
Even at full power, the temperature of the output transistor stayed at an acceptable level.
With a 12V supply, the frequency was initially set to 104.00MHz, and the trimmers adjusted for maximum LED brightness. After one and a half hours the frequency had drifted to 104.06MHz, a shift of 60KHz. Another hour later, the frequency remained stable at 104.06MHz.
The supply voltage was increased to 16V, and the frequency set to 98.046MHz (approximately mid-band). The table below shows the supply current, output power and frequency shift as the supply voltage was reduced from 16V.
| Supply Voltage (Volts) |
Supply Current (mA) | Output Power (mW) | Frequency Shift (KHz) |
| 16 | 194 | 976 | 0 |
| 15 | 183 | 887 | -70 |
| 14 | 172 | 791 | -144 |
| 13 | 159 | 689 | -220 |
| 12 | 144 | 587 | -297 |
| 11 | 128 | 483 | -380 |
| 10 | 109 | 370 | -470 |
| 9 | 83 | 228 | -547 |
| 8 | 49 | 67 | -540 |
| 7 | 19 | 1 | -627 |
The table shows:
At 16V supply, centre frequency 98.00MHz, 0.97W output power, the harmonics were measured on a spectrum analyser.
|
Plot
of Spectral Purity, 1W out, 500MHz span |
Plot
of Spectral Purity, 1W out, 200MHz span |
The only significant frequency component other than the harmonics was at half the output frequency. This was measured at -58dBc. Apart from this, the output was clean down to the -70dBc measurement noise floor. With this units 1W maximum output power, these levels of harmonics are unlikely to cause any problems.
With a 16V supply, the centre frequency of the unit into a good load was 97.979MHz. Changing the load to one with a 5.4dB return loss (3.3 VSWR) resulted in frequency shift of between +40KHz and -84KHz depending on the phase angle of the mismatch. This demonstrates probably the major drawback of a VFO based transmitter. Frequency variations due to supply voltage can be minimised by using a stabilised power supply. Variations due to temperature can be minimised by allowing the unit to warm up before going on air and keeping the ambient temperature stable. But for a given design, there is no way to prevent loading variations altering the output frequency. To minimise this effect, the aerial (antennae) should be kept well away from anything else, especially anything else moving, and shouldn't be allowed to flap about in the wind. Adding an additional buffer stage between the VCO and the output stage would help. PLL based transmitters do not suffer from this problem, as their output frequency is referenced to a stable crystal oscillator.
Due to the small output power capability of the unit, coupled with the low supply voltage, bad VSWRs are unlikely to damage this unit.
A dummy load was supplied with the unit. This was made of a UHF plug with a standard resistor soldered inside it. The measured return loss of this dummy load was 22dB, which is fine for this application. The power rating of the resistor was probably only 0.6W, so I don't recommend using this dummy load for extended periods of time at full output power.
Audio response was not measured, due to the simplicity of the audio part of the circuit. By ear it sounded fine, with no hums or whistles.
A high quality unit, good value for money and clear assembly instructions. Good specification considering the simplicity of the unit. I recommend using the unit with a stabilised power supply, so as to guard against frequency variations caused by supply voltage fluctuations, and to prevent any supply ripple and interference from being modulated onto the audio output. I also recommend performing the final frequency adjustment after the unit has warmed up (say twenty minutes) and after the real aerial has been connected, due to the frequency pulling effect detailed above. Contact Veronica FM, info@veronica.co.uk