RF-Amp

RF Amplifier Features

From Charlie Morris' (ZL2CTM) Go QRP Portable SSB Rig
- Charlie references Solid State Design for the Radio Amateur (pp 19-20)
Single 2N3904 NPN transistor
- Ft = 300 MHz (Gain Bandwidth Product)
  - Theoretical gain
    - +20 dB at 30 MHz
    - +29.5 dB at 10 MHz
    - Reality is lower due to capacitance, etc.
Measured +22 dB gain @12V, +25dB gain @14V
Input connectors: SMA or BNC
+12V nominal power
49x49mm card
4x 4-40 mounting holes

From Charlie's notes with mods for my use
- Charlie Morris' (ZL2CTM) Go QRP Portable SSB Rig
- Charlie describes the design in detail in his video Simple SSB Rig: Part 6 - IF Amplifiers (Feb 2021)
- Based on the Class A RF Amplifier in Solid State Design for the Radio Amateur pp 19-20
2N3904 data sheet
Emitter Resistance - helpful paper

Gain bandwidth product divided by operating frequency
- Assume operating frequency of 9 MHz (IF frequency)
- = 300/9 = 33.3

Max HFE RF gain at CE current of 10 mA
- If Vce = 6V, this is 60 mW power dissipation
Assume Ve (voltage across emitter resistor) = 1/10 Vcc = 12V/10 = 1.2V
- R3 is Re (emitter resistor) = 1.2V/0.01A = 120 ohms
VCE = 0.7V (typical from data sheet)
V(emitter) at 10% of Vcc rule of thumb = 1.2V
V(base) = V(emitter) + VCE = 1.9V
Base current is collector current divided by Beta DC
- Biasing resistors = 10x current needed by base current
- 10 mA in C-E, beta DC less = 10 mA/173 = 58 uA
- 10x the current in the biasing resistors = 580 uA (calculated)
R2 is 1.9V at 580 uA = 3.29K use 3.3K
R1 sources current to R2 and transistor base
- Voltage = Vcc (12V) - 1.9V = 10.1V
- Current = 577 uA + 58 uA = 635 uA
- R1 = 10.1 / .635 mA = 15.9K, use 15K

Measured with no input
Vcc = 11.96V
Current draw = 12 mA
- Quick test for wiring and more or less correct parts
- Expected 11 mA - close enough
+BUFF = 11.84V
- 0.12V which is 12 mA through R4 10 ohms - expected
V emitter = 1.41V
- 12 mA through 120 ohm = 1.44V - close
V on input divider = 2.06V
- Vbase + 0.7V - close

Xc for 0.1uF cap from emitter to ground
- C=0.1uF
- F=10MHz
- 1/2*pi*F*C = 0.16 ohms
Parallel resistors R1, R2 paralleled with transistor input impedance
- R1=15K, R2=3.3K
- Transistor resistance = Beta AC (33.3) times re
  - re = 26 / Ie (10 mA in mA) = 26/10 = 2.6
  - SSDRA uses 25 as constant - close enough
  - Beta AC * re = 33.3*2.6 = 83.3 ohms - predominates
- All in parallel are 80.8 ohms

Output transformer (T2 on Charlie's - T1 on this board)
T2 - different than Charlie's design since my Crystal filters are all 50 ohms in/out
SSDRA suggest presenting 200 ohm load to the collector
- Can't find reference in SSDRA
- Reflecting back 50 ohms load to 200 ohm collector...
200:50 ohms
n = sqrt(200/50) = 2.0:1 turns ratio
10:5 turns
- 10 turns primary (on transistor collector)
  - 10 turns = 35 uH
- 5 turns secondary (towards output)
  - 5 turns = 8.75 uH
- 15 turns = 9.5 in

Goal: Measure RF-Amp performance using a NanoVNA running NanoSaver software on PC
S21 (gain) needs to be measured with a 40 dB attenuator on input to RF-Amp to avoid compression on the output
S11 (reflection) input impedance can't be measured with input 40 dB attenuator because S11 just ends up measuring the attenuator
- Output should be terminated to 50 ohms for S11 measurement
DC current = 12 mA

Put 40 dB attenuator on RF-Amp input, measure S21 at output
- NanoVNA provides 50 ohm load to RF-Amp to properly terminate output
Measure S21 with 9:11 input transformer
- S21 @ 100 KHz = -8 dB dB
- S21 @ 1.45 MHz = 35.4 dB (peak gain)
- S21 @ 9.1 MHz = 24.3 dB
- S21 @ 16 MHz = 20.1 dB
- S21 @ 30 MHz = 12.7 dB
Peak gain justifies use of 40 dB attenuator to protect NanoVNA

LTspice simulation was pretty similar to NanoVNA results
- -10 dB at 100 KHz
- +32 dB at peak
- Lower output at higher frequencies

Is there compression if the NanoVNA drives the input directly?
- Test by driving directly from NanoVNA set to CW = 9 MHz
- Measured output with scope - not clipped at 9 MHz
  - Approx. 1Vpp input = +22.1 dBm gain which matches the S21 with the attenuator on the input
  - Vpp = 12.4V with 50 Ohm load resistor
- Starts clipping at 7 Mhz and down
Therefore, can measure input impedance at 9 MHz
Other evidence of compression
- Compare S21 gain with no input attenuator, put external 40 dB RF Attenuators on output of RF-Amp to protect NanoVNA input
- S21 shows lower gain in lower frequencies so clipping/compression is happening
- Was: 35 dB at 1.4 MHz
- Is: 23.1 dB at 1.5 MHz
Due to compression can't accurately measure lower frequencies with attenuator at output
Compression below 7 MHz matches what was on scope

Above had 9:11 turns ratio
Change to 7:9 turns ratio
Slightly better gain at higher frequencies
Was: S21 @ 30 MHz = 12.7 dB
After: S21 @ 30 MHz = 15.3 dB
Small additional gain at 8 MHz
- Was: S21 @ 9.1 MHz = 24.3 dB
- After: S21 @ 9.1 MHz = 24.8 dB