A few years ago, I bought a HF amplifier. It wasn't working, but I decided to repair it rather than build one from scratch and hunt down the necessary parts. This amplifier was built according to the G2DAF circuit, which is also commonly referred to as a passive grid circuit.
Dočítate sa:
What is the advantage of a G2DAF amplifier?
The G2DAF amplifier generates the voltage for the g2 grid by rectifying part of the drive signal. This means that a Ug2 stabilizer is not required, and the tube’s quiescent current without drive is zero. This has a positive effect on efficiency, allows for the use of higher anode voltages, and results in slightly higher efficiency.
Schematic of the G2DAF amplifier
The disadvantage is that the amplifier requires a slightly higher excitation level, a suitable input circuit design, and monitoring of DC parameters during operation to ensure the signal is processed with minimal distortion.
Anode choke
The original choke was destroyed, either due to operation outside its rated load or within its resonance range. However, the core was only minimally damaged, so I removed the remaining wire, cleaned the core, and rewound it with new CuL wire. This required patience, as there are 300 turns divided into sections of 165, 65, 35, 20, and 15 turns.
Pi-cell of the amplifier
The high-voltage isolation capacitor was only sooted by smoke from the choke. It simply needed to be cleaned. This was followed by a test with the nanoVNA to verify the output pi-cell and the natural resonances of the anode choke.
Procedure for Checking the Output Pi-Cell and the Natural Resonances of the Anode Choke
Simplified diagram of the amplifier’s anode circuit measurement
- Disconnect the Ua power supply from the amplifier (extremely important!)
- Then, load the anodes of the tubes with the expected resistance of the anode circuit
- Power up the amplifier and switch it to transmit mode
- It is easier to measure the resonant frequencies of the anode choke with the amplifier switched to the 28 MHz band and both variable capacitors set to nearly minimum capacitance
- The resonant frequencies of the anode choke will be visible on the nanoVNA screen. These may be difficult to see if the number of measurement steps on the nanoVNA is too small. It is then advisable to measure across multiple ranges, such as 1 to 10 MHz, 10 to 20 MHz, and 20 to 35 MHz
- I detected resonance in the anode choke as high as around 34 MHz. However, if the resonance falls within an amateur radio band or very close to it, the amplifier cannot be tuned to that band; it may spark, or the anode choke may burn out. Such a choke is unsuitable and must be rewound.
- If the anode choke is in good condition, we proceed to check the π-circuit. We tune to a VSWR of 1.0 in each amateur radio band. If such a VSWR cannot be achieved, the π-circuit requires adjustment
- In my case, this required soldering the parallel capacitor back in (one terminal had been disconnected) and changing the tap on the 80-meter band. Without this, it was impossible to tune the pi-cell to this band; its resonance was higher. The other bands were correct.
- Finally, don’t forget to disconnect the resistor from the tube anodes and reconnect the Ua supply
Reception Problems
Anode stops; on the left, a coaxial cable connecting the amplifier’s input and output relays
The original owner reported problems with the relays during reception. However, he was reluctant to replace the relay in a very hard-to-reach location. Since I had also noticed occasional loss of reception, I set out to investigate with an ohmmeter. I was pleased to find that the problem was not caused by the relay, but by the short coaxial cable connecting the input and output relays. Replacing 25 centimeters of coaxial cable was much easier.
Testing the QY3-125 Tubes
I had no previous experience with QY3-125 tubes. Visually, they didn’t look very good, but I decided to give them a chance. First, I let them warm up for 24 hours while gently blowing air on them. Then came the critical test: applying the anode voltage. The tubes withstood it. Since this configuration doesn’t allow for a quiescent current, it was time for another test.
Putting the Amplifier into Operation
So I had to start exciting the amplifier. It was confirmed that the tuning using the nanoVNA was correct. The amplifier immediately delivered power across all bands. However, the excitation required to achieve the same output power differed significantly across each band. Measurements showed that the cause was a high input VSWR. At 28 MHz, the VSWR was as high as 2.5:1. A look at the load resistors explains why:
The original resistors in the amplifier’s input circuit
The designer used power resistors with high inductance. As an experiment, I replaced them with a resistor of the same value in a miniature, non-inductive version and remeasured using an antenna analyzer. With this resistor, the PSV dropped below 1.5 and reached the ideal value on several bands.
The G2DAF amplifier’s input circuit
So, what should I use to replace the unsuitable resistors? Searching online, I found a power resistor labeled RPT100 with the appropriate resistance. These resistors are labeled as non-inductive, which is exactly what I needed. However, it turned out that this resistor is not actually non-inductive. Fortunately, I got my hands on several 100-ohm non-inductive Alcatel 39-4696 resistors, so I assembled the appropriate resistance from a set of three. While this method isn’t ideal, it’s sufficient for HF. The input PSV has thus improved significantly. The amplifier now delivers 500 W of power on most bands.
Enclosure for the Amplifier and Aesthetic Improvements
Since the amplifier has to be on display, its visual appearance matters as well. One of the previous owners drilled two rows of holes in the top panel. However, they were very unsightly, with varying shapes and misalignments. After some thought, I decided not to replace the panel. In Corel Draw, I designed the lines and sizes of the holes so that, once drilled, they would conceal these imperfections as much as possible. I printed the resulting template onto a self-adhesive label. This then served as a guide for drilling the holes.
The original front panel of the G2DAF amplifier
The amplifier also got a new coat of paint. I sanded off the original paint. I then sprayed it with a base coat followed by black. The top layer is additionally coated with clear spray varnish.
The second question concerned the front panel. Its dimensions are larger than A4, which is the maximum size I can print on a self-adhesive label. I toyed with the idea of a custom-made aluminum front panel through https://www.quickpanel.sk. However, the price is not insignificant. So I designed a new front panel in Corel Draw, drawing inspiration from the designs of other amplifiers. The front panel design was printed on two self-adhesive labels. After applying them, the edges were carefully trimmed, and the panel was sprayed very lightly three times with clear varnish. This method ensures that the panel is resistant to both abrasion and dirt. You must spray very lightly, because the label tends to warp once it absorbs the varnish.
Work on the Amplifier
The tuning capacitors and the switch in the pi-cell were fitted with new instrument knobs. I had some TESLA WF-series instrument knobs with a diameter of 40 millimeters in my stash, which came in very handy. These old knobs are very well-made and surpass most of the instrument knobs available on the market today.
Spacers with a diameter of 25 millimeters were placed between the knob and the front panel. They were cut from the remnants of an old foam sleeping pad. Pressing them between the knob and the front panel creates friction that provides just the right amount of resistance for comfortable tuning. At the same time, they fill the gap between the knob and the front panel.
G2DAF amplifier schematic: detail of the output capacitor scale for the pi-cell
Originally, the amplifier had two indicator bulbs in red and green sockets on the front panel. One indicator lamp was missing, and the other would light up intermittently. I soldered 5-millimeter-diameter LEDs into the sockets and added a limiting resistor so that the LED current would be approximately 15 milliamps.
The G2DAF Amplifier in Practice
The G2DAF amplifier after modifications
Conventional amplifiers are characterized by a strong emphasis on stabilizing the Ug1 and Ug2 voltages. The quality of this stabilization significantly influences the amplifier’s spectral purity and linearity. I was therefore curious to see how an amplifier whose Ug2 varies depending on the drive signal would behave in practice. I conducted a comparison using an SDR receiver with the amplifier turned off and then on. Perhaps surprisingly, no intermodulation products were visible on the spectrum after turning on the amplifier—only a significantly stronger signal. In the future, I plan to conduct a two-tone SSB test as well.
If you happen to get your hands on a G2DAF amplifier, I recommend reading “G2DAF Linear Power Amplifier—History, Theory, and Practice” https://www.radiozurnal.sk/radioamater/download/cisla/2000-6.pdf and https://www.radiozurnal.sk/radioamater/download/cisla/2001-1.pdf
The material contains a wealth of information and advice on how to build such an amplifier and what its characteristics are. However, when working with vacuum tube amplifiers, it is important to remember that you are dealing with life-threatening voltages! If you want to be sure of your safety, it is better to opt for LDMOS amplifiers, which operate at safer voltages.