The Cly Institute has been given the opportunity to resurrect a VHF amplifier based on the February 1971 design in QST magazine by W1SL. Over the years this design has become known as the "Watt Weasel". It employs two 4CX250B tetrode tubes for an achievable output of 1 kw when driven hard with high plate voltage. Since the first 500 watts is what counts, the goal of this restoration is to achieve at least that output linearly. Our benefactor K3WHC has used this amp 30 years back routinely putting out 800 watts.
The original components consist of the PA deck and a Screen / Grid power supply. The original B+ supply has been lost to time and smoke. Therefore, the first big missing piece is to design a new B+ supply with at least 2000 vdc capability. Since our experience at the Institute is with Heathkit SB series HF amplifiers our first thoughts are to employ the technology that we know and trust. An SB-200 runs roughly the same plate voltage and current, therefore a Harbach Electronics PM-200 kit will work nicely here for the B+ supply. It is a voltage doubler circuit that contains all the parts we need to rectify and produce the B+ voltage. The power transformer chosen was an
Antek AS-8T800 toroidal-wound transformer. This should give us 800 vac at the high voltage secondary, (with a tertiary winding available for 12 volts or 6 volts if needed.) The transformer can be air-cooled to work into its 1.2 safety factor on voice peaks. Supplying just about half an amp of current will be attainable, and more than we need for our minimum goal.
The PM-200 Harbach supply uses the standard 0-200uA meter movement in the Heathkit. If we want to monitor both current and voltage simultaneously, all we need is two meters labeled as such. A quick operation on Inkscape after scanning in the original scale, allows us to trace over it and create the appropriate scale which can be printed out 1:1 scale and put on the back of the original scale with contact adhesive.
Easily removed with four screws so we can apply new artwork
The Harbach boards and toroidal transformer work nicely in the 3U rack enclosure box. Cutting out the front panel is the fun part which required a 2" diameter hole and lots of elbow grease.
The power supply layout allows for the low voltage windings to be used for a case circulation fan through the back of the rack enclosure. 5/8" standoffs were used to mount all the boards. As you can imagine the center of gravity WAY to the left side. It produces a healthy 2.1 kilovolts with no load; right on design target.
Here are some photos of the expertly crafted PA deck.
Modern Technology dictates that a pair of VDR's be installed at each screen to cathode.
Before
Shaft coupler repaired and rotation stops added
After
The next task is to insure that the Screen / Grid power supply is operational. This supply as it originally came to us, has a full wave bridge and unregulated Screen voltage which balances out around 310 vdc. In doing some research the Screen voltage around this value may be ok for Class C operation, but not for Class AB1. Some others have insisted that good regulation is needed on the screen supply, whereas this supply has none. The voltage is metered, but the current is not presently hooked up. This Screen current is a good tuning method and should employ the hole in the cabinet for one with a centered needle. A better regulated supply needs to be employed to hold the Screen supply stable with load changes and reduction of secondary emission. That means that the original Screen supply will need to be completely re-worked.
The Grid supply as it arrived to us, is apparently half-wave rectified. It is unregulated and produces about -130 vdc. (Other articles indicate that this voltage should more more around -55 vdc for Class AB operation. There is an adjustment pot for the grid supply, but it only changes a few volts. This seems unsuitable.
To aid in neutralizing the tubes, it will be essential to install a pair of grid current meters to individually monitor grid current. Presently there is no instrumentation on this line. A spare conductor run to the 8-pin Jones-Cinch connector will do well to provide individual grid current measurements. I found some nice edge-reading mA meters for the individual Grid Currents. They will work in the remaining panel real estate.
The final and perhaps the most important issue is the control and interlocking of all the components.
The
VHF / UHF DX Book is an excellent publication to refer to on the requirements of feeding the 4CX tube family. The control requirements are like this in order of priority. The availability of the first item on the list permits the second item to start as a permissive. The Grid is negative, so it doesn't matter. Danger exists when the Screen supply is on before the B+. If the Screen starts acting like a Plate, then there will be quick destruction of the tube. If the Grid or B+ fail, the Screen supply must shut down.
1) Air Flow
- Engage Fan and wait for flow switch.
2) Filament Voltage
- (2 minute warmup time)
3) Grid supply
4) B+ supply
- interlocked with relay to Grid supply
5) Screen supply
- interlocked with relay to B+ supply and Grid supply
The more research that was done on regulation of the screen supply lead to a the need for a better screen and grid supply, which would also incorporate a sequence / protection system.
Tetrode Boards by GM3SEK offered a ready-made solution with everything that was needed according to my research. The manual is extensive, and requires the boards to be stuffed with the appropriate parts based on the tubes used and design intended. Ian produces a set of printed circuit boards which offer exactly what is needed for tetrode operation and protection. This was the right thing to do. The screen regulator is a shunt type, which can absorb as well as deliver current to the screen. In an unfortunate case of upset and potential flashover, the tube is protected by absorbing energy through the regulator. Zeners are too noisy and cannot absorb the energy. Neons can absorb in the reverse direction but are not precise enough, and cannot guarantee a tight voltage regulation required to eliminate IMD on the tubes caused by variations in screen voltage.
Rough wiring before the final tye-wrap job.
The implication of these boards are that additional voltage is required to allow the regulator to work. The transformers that came with the old supply would not deliver the voltage or current needed. A new transformer was required. Ideally a transformer with a 400 volt winding and two 120 volt windings. An industrial control power transformer is ideal which is why I selected a Hammond 75VA
PH75MQMJ. One 120v winding will be used as a primary. The 480v winding is used for the screen supply voltage. To take the stress off the screen / grid regulation circuits, I put an auto-transformer in front of this transformer to dial in the exact voltage required to the Grid / Screen supply.
Additional markings are needed on the tuning knobs. This will allow a repeat of initial tuning by setting per the dial. These dial markings were made in Inkscape and then printed out and adhered to the plate behind the knob.
I may not leave the fan where it is mounted in the photos. It is loud!
The TX and RX relays are both Tohtsu coaxial units. The TX relay has to handle the full power, so its got N connectors. I chose an RX relay as well instead of running separate lines from my XV144 Elecraft transverter. This allows me to run barefoot with the transverter. The limitation of the transverter is that is has no ALC input. So in practice watching the drive level will be critical.
Initial power-up test into a dummy load were good. The amp was able to idle without internal oscillation or unbalance. The neutralization was not disturbed in the PA deck, so it must have been effectively neutralized in the past. The long wait of the warm-up timer before the HV enable relay kicked in was an anxious moment. All the magic smoke stayed in.
The first on air test was done December 18th, 2019 with WA3USG and K3TEF. Other than a little mic hum the signal reports came back clean. The amplifier is operating well within its design limits at 600w out, with no spurious emissions.
The initial relative knob settings are:
Grid Tune = 115
Grid Load = 40
Plate Tune = 134
Plate Load = 13
