A low-power experimental HF amplifier, using the famous (and old 🙂 2N3632 transistors, capable of delivering approximately 6W from 3 to 20MHz. A different design compared to the classical class AB approaches for this configuration.

ROWPA-01EXP

An experimental linear HF amplifier application of approximately 10W, covering 3 to 20MHz, was proposed by W7ZOI in one of the American magazines (most likely ARRL, unfortunately I couldn’t identify the reference). Following this schematic, I also built and experimented with the current project, implementing some minor modifications to achieve better linearity and higher gain.

1. Functions, details and explanations from schematic

A frequency-compensating attenuator (RFC1 / 68Ω / 150pF) is located at the input of the amplifier. The phase-splitting transformer T1 (2:1) provides the two sinusoidal signals for amplification to the bases of Q1/Q2, biased through RFC2 at approximately 0.7V via diode 1N4719, directly from 28Vdc. A primary drawback related to the approach of this scheme would be the high supply voltage (between 20….30Vdc/2A) but even so i obtained positive and encouraging results. The emitter “degenerative” resistors of 1.2Ω x 3 are necessary to balance the current consumed by each transistor. A preliminary measurement of the transistor gain factor would be useful. A Bal-Un (balanced-to-unbalanced) transformer ensures the summing of the two collector signals from Q1/Q2 into the output signal, at 50Ω impedance according to the transformation ratio provided in the original scheme. Below is the original design taken from the database of VA3IUL.

The 10W HF amplifier schematic proposed in the VA3IUL database

2. Calculations, implementation and practical details

The execution was carried out on an industrial printed circuit board, designed by me. I did not want to try this scheme on a test circuit this time (usually I do). The input-output connectors are 90-degree SMA female types. Anyway, if they will not solder well, the large-sized holes in the footprint can be used to solder the input-output coaxial cable. Capacitor C3 in the diagram in Fig.4 is silver-mica type as recommended by the author [1]. The original details of the execution of the ferrite transformers and RF chokes, extracted from the original schematic:

For the redesign, it was obviously necessary to redraw the electronic schematic, in the Eagle Autodesk software that we are working with:

Fig.2 – The HF amplifier schematic after YO6TJJ modifications and redrawn.

The table below provides detailed information on the new values and construction references for the aforementioned schematic:

Schematic reference	Value	Execution details
RFC1	15µH	It can be a prefabricated RF choke (recommended) or it can be constructed on a small ferrite rod with an impedance of approximately 10Ω @ 100KHz (1.2Ω @ 1KHz).
RFC2	3µH	8 turns CuEm Ø 0.4mm, FT37-61 FairRite
RFC3	3µH	7 turns CuEm Ø 0.6…0.7mm, FT50-61 FairRite
T1 (TR1)	n/a	Conventional wideband transformer, transformation ratio 2:1 PRI: 14 turns of 0.35mm CuEm wire SEC: 10 turns of 0.35mm CuEm wire, wound over PRI T1 assembly (TR1) composed of 2 X FT50-61 ferrite cores, stacked
T2 (TR2)	n/a	Conventional wideband transformer, transformation ratio 1.2:1:1, bifilar + single or trifilar, on 2 x FT50-61 cores 1-3: 18 turns of 0.5mm CuEm wire, single winding 2-4-5: 15 turns bifilar (~8 turns/inch), CuEm 0.5mm

The general power supply of this HF amplifier is achieved through an AK-500 circuit board connector, for direct current, with wire connections secured by screw fastening.

Fig.3 – The placement of HF amplifier components, following YO6TJJ modifications and redraw.

The implementation approach was with through-hole components, transistors T1 & T2 are of type 2N3632 / 2N3375, TO-60 package. They were mounted and cooled on two completely separate elements of a heatsink, for ease of testing and mechanical fixation. A small sample of 5 transistors was measured for this project, using the TC-1 multimeter (see ebay / Alieexpress for a reference….).

Transistor no.	hFE	Vbe [V] @ hFE	Comment
1	20	0.58	–
2	43	0.59	chosen
3	27	0.63	–
4	43	0.62	chosen
5	22	0.65	–

Transistors 2 & 4 were chosen for this design. The total idle current for the amplifier is set to approximately 250mAdc per transistor, precisely 518mAdc in total.

the voltage drop measured across the three resistors R3||R4||R5 is 110mV at a 30Vdc supply
the voltage drop measured across the three resistors R6||R7||R8 is 90mV at a 30Vdc supply

R8 was removed to increase the voltage drop across the R6/R7 group from 90mV to 113mV. The new current values for the two transistor “arms” / emitters are:

Rtot. (R3||R4||R5)= 0.4Ω and I=U/R= 0.11V / 0.4Ω = 0.275Adc for transistor T1
Rtot. (R6||R7)= 0.7Ω and I=U/R= 0.113V / 0.7Ω = 0.161Adc for transistor T2

The original transformer T2 from Fig.2 was rewound because the output parameters were completely unsatisfactory. Thus:

Fig.4 -The initial measurements with the first version of the output transformer for the HF amplifier.

I’ve chosen a new configuration with a center-tapped (CT) winding on the secondary winding. I tested several configurations and transformer ratios for transformer T2. The input impedance into transformer T2 is approximately 93Ω (which is approximately the output impedance of the final transistor collectors).

Zp (imped. primary) = 93Ω
Zs (imped. seconday) = 50Ω
N=sqrt(Zp/Zs) resulting the transformation ration of aprox. 1.36.
N=1.36
np (no. turns primary) = 15sp.
ns (no. turns secondary) = 15/1.36= aprox. 11 turns.

The resulting configuration would be 1.4:1.4:1. However, too many turns in the secondary winding would significantly degrade performance at high frequencies.

3. Adjustment, testing and results

A new configuration, and the best one from my experiments, was tested with a 1.2:1.2:1 ratio. With 18 monofilar turns in the primary winding, it’s easily achievable to obtain approximately 10W from 160 to 20 meters, with 1 to 1.5W at the input.

Band [m]	Frequency [MHz]	Pin [W]*	Pin [dBm]	Pout [W]	Pout [dBm]	Icc [A]	Gain	Vcc [V]
160	1.91	1.1	30.4	11.5	40.61	0.545	10.21	30
80	3.69	1.1	30.4	9.94	39.9	0.57	9.5	30
40	7.09	1.1	30.4	11.52	40.6	0.617	10.2	30
30	10.116	1.3	31.1	8.24	39.1	0.662	8	30
20	14.285	1.2	30.8	8.16	39.1	0.654	8.3	30
17	18.13	1.2	30.8	5.12	37	0.593	6.2	30
15	21.285	1.3	31.1	3.53	35.5	0.565	4.4	30
12	24.95	1.4	31.5	3.38	35.3	0.506	3.8	30
10	28.365	1.3	31.1	3	34.5	0.394	3.4	30

Assembly does not require special adjustments. However, it is necessary to ensure the “matching/balancing” of transistors T1 & T2 by measuring the emitter resistors and adding/removing them to achieve currents as close as possible on the respective transistors. Special attention must still be paid to the two transformers TR1 & TR2 and their correct winding connections.

Above is a plot showing the output level versus bandwidth.

4. Conclusions and future perspectives

Clearly, the amplifier requires a low-pass filter at the output, or a block of switchable filters. It is an experimental amplifier that can be successfully used as a QRP amplifier in a small transmitter or transceiver if the high supply voltage is not an obstacle.

For those interested, there is the possibility of producing PCBs for certain projects, including this HF amplifier, upon request and in quantities of at least 5 pieces per project. You can write to us at the contact email address, and we will provide you with more details.

73 de Andrei – YO6TJJ (Sibiu, Ian. 2024)

5. Bibliography

[1] website VA3IUL / YO3DAC, PA Amplifier Schematics: https://www.qsl.net/va3iul/Homebrew_RF_Circuit_Design_Ideas/HF_15W_Linear_PA_W7ZOI.gif

3-30MHz HF Amplifier 3-10W | ROWPA-01EXP