Table of contents

A key requirement when designing the RFzero was to make spectral performance as good as possible while, using the cost effective Si5351A digital clock generator, balancing cost and performance. However, if spectral performance is extremely critical, e.g. for a very high power beacon, transmitter or VFO then the Next Generation Beacons platform is the right choice instead of the RFzero but also in a different price range. Therefore you should see the RFzero as a low cost and fair performance RF unit.

Characteristics	Frequency	Value (typical)	Notes
Frequency range	3907 Hz to 200 MHz	2289 Hz to 290 MHz	Above 200 MHz is outside the specification but the Si5351A will often work close to 300 MHz
Output power	137 kHz 400 kHz to 200 MHz 240 MHz 280 MHz	>8 dBm >13 dBm >12 dBm >11 dBm	Above 200 MHz is used above the Skyworks specifications 432 MHz using 3 x 144 MHz: > -5 dBm 1,3 GHz using 5 x 260 MHz: > -30 dBm No time limit unterminated on any frequency
Third order IMD (below PEP)	1 MHz 10 MHz 50 MHz 200 MHz	35 dB/- 35 dB/75 dB 20 dB/55 dB 25 dB/-	T1: transformer/combiner 1 kHz tone spacing For use of combiner please see the modifications page
Second harmonic	137 kHz 1 MHz 10 MHz 50 MHz 200 MHz	-60 dBc -65 dBc -57 dBc -42 dBc -37 dBc	The shown frequencies are the fundamental frequencies Please see the spectrum plots below for more details
Third harmonic	137 kHz 1 MHz 10 MHz 50 MHz 200 MHz	-5 dBc -15 dBc -15 dBc -10 dBc -20 dBc	The shown frequencies are the fundamental frequencies Please see the spectrum plots below for more details
In-band spurious	1 MHz 10 MHz 50 MHz 200 MHz	-84 dBc -75 dBc -70 dBc -68 dBc	+/-500 kHz from carrier Please see the spectrum plots below for more details
Phase noise	100 Hz 1 kHz 10 kHz 100 kHz 1 MHz	-103 dBc/Hz -128 dBc/Hz -136 dBc/Hz -138 dBc/Hz -144 dBc/Hz	Measured at 9 MHz
Ref. freq. suppression	27 MHz	>100 dB
Short term freq. stability*	27 MHz	<105 mHz/3,8 PPM <13 mHz/0,48 PPM	Naked crystal, <40 Hz on 10 GHz Foam covered crystal, <5 Hz on 10 GHz
Long term freq. stability GPS PPS	27 MHz	<30 ns RMS 99% <60 ns

*: The RFzero has been tested generating WSPR signals on 260 MHz and receiving the fifth harmonics on 1,3 GHz with 100% decoding performance. The RFzero has also been tested generating PI4 signals on 108 MHz that were fed into a x96 multiplier resulting in a 100% decoded signal on 10 GHz. The tests were carried out with a foam covered crystal.

All values have been measured with a standard RFzero into 50 Ω, i.e. no output filter and with a T1 transformer, and using 8 mA as output drive strength except where otherwise stated.

In general it applies that if the IMD of the input signal to an amplifier is -12 dB then the contribution to the output is 6%.

The table below shows the typical output power vs. the current in the output stages running in push-pull with a T1 transformer, i.e. default H/W.

Current [mA]	137 kHz	1 MHz	10 MHz	30 MHz	50 MHz	200 MHz
8	9,5 dBm	14,2 dBm	14,5 dBm	15,0 dBm	14,5 dBm	13,3 dBm
6	9,2 dBm	12,8 dBm	13,3 dBm	13,7 dBm	13,0 dBm	11,8 dBm
4	8,3 dBm	10,3 dBm	10,7 dBm	11,0 dBm	10,5 dBm	9,7 dBm
2	5,2 dBm	4,7 dBm	5,0 dBm	5,5 dBm	5,0 dBm	4,5 dBm

The output level can be controlled with the si5351a.rfLevel(RF_LEVEL_X) function, where RF_LEVEL_X is the output stage off, on or current value: 2 mA, 4 mA, 6 mA or 8 mA: RF_LEVEL_OFF, RF_LEVEL_ON, RF_LEVEL_2, RF_LEVEL_4, RF_LEVEL_6 or RF_LEVEL_8.

Why care about spectrum performance?

Many people use the very popular Si5351A uncritically in combination with RF. The Si5351A is a programmable clock generator for digital environments where the spurious are of much less concern or perhaps even irrelevant. Designing RF circuits take knowledge, time, care, resources, iterations and money.

Spectral performance, or signal purity, is expressed as

Phase noise
Spurious both out-band, close-in and most importantly in-band
Harmonics

Of these parameters only the harmonics are easy to do something about using a low pass filter. For some odd reason the harmonics are in many constructions the only thing that seem to matter – what about the spurious and the phase noise?

The performance is relevant both on-air but also inside a radio. When transmitting poor spectral performance means that other radio amateur stations will be disturbed more easily. But also non-radio amateur receivers, radio stations, TVs, electronic devices etc. will be affected. When receiving the phase noise, and spurious, both in own receiver and from transmitting stations, mean that unwanted mixing takes place in the receiver resulting in unwanted signals and poorer reception of the wanted signal(s).

You may say that a QRP signal should not live up to the same requirements that a 100 W transmitter must. Well, let’s try with some real numbers. The very popular IC-7300 has, according to the datasheet, an unwanted signal level suppression that is better than 60 dB at a power level of 100 W, i.e. 50 dBm. This means that the unwanted signals are max: 50 dBm – 60 dB = -10 dBm. If a QRP signal, e.g. from a Si5351A, Si570 or Raspberry Pi, is 100 mW/20 dBm and has a suppression of the unwanted signals better than 30 dB, which is a real number in some constructions, then the unwanted signals are max: 20 dBm – 30 dB = -10 dBm. So the unwanted signals are at the same level for the IC-7300 and the QRP signal. Thus when the QRP signal is amplified to e.g. 5 W, i.e. 37 dBm, the QRP signal becomes 17 dB worse than the IC-7300.

For more information please see e.g. Radio-Electronics.com that has a number of concise tutorials for RF.

Spectral performances

Below are a series of pictures showing both wide and narrow spectrum performances measured with a Rohde & Schwarz FSIQ 26, 20 Hz to 26 GHz spectrum analyzer with Low Phase Noise B4, FFT Bandwidth 1 Hz to 1 kHz B5, Vector signal analysis B7 and DSP & IQ extension 2 x 512 kB B70 options and with firmware options K11, K21, K72 and K74.

The measurements were carried out with a 200 MHz low pass filter mounted on the Z1-Z10 pads irrespective of the fundamental frequencies.