Amp features:
- QSK using vacuum relays
- Hot-switching protection for the T/R relays
- Remote control and metering
- LCD display of settings and fault status
- Fault protection; grid current, anode current, reflected power, high voltage too low, HV step-start fail, and low air pressure
- Manual tuning; display shows typical tune/load settings for each band
- Anode-cathode phase detector to aid in tuning
- Automatic power up/power down heat/cool sequence
- High voltage supply step-start
- Heater soft-start
- High isolation (optical and magnetic) of metering and control circuits from high voltage circuit.
- QSK and relay hot-switch protection
- A transmission is initiated by a signal at the KEY_IN line. Relay hot-switch protection is accomplished by sensing RF at the input of the amp and preventing the RX-TX or TX-RX transition if RF is present. The processor looks at the state of the output relay and won't allow the tubes to bias ON unless the relay is closed. When both relays are in the TX mode and the tubes are biased ON, the KEY_OUT line keys the exciter. This method makes it impossible to hot-switch either the input or output relay. Even though the Kilovac HC-2 vacuum relays used to switch the input and output of the amp are fast enough for QSK operation, it was simple to add a speed-up circuit and software to help them a bit. The relay coils are 26V, but when going from RX to TX, the processor first hits them with 48V for 2mS and then backs off to 24V. The overall switching time of the amp is about 3mS.
- Remote control functions
- All amp operation is done through the remote control head and 50 foot cable. The only control on the amp or power supply boxes is the main circuit breaker. The tune and load capacitors are adjusted via a single optical encoder and selector switch. It works OK, but it would have been nicer to have separate knobs so I could do "two handed tuning".
- The LCD display shows:
- Amp state (OFF, WAIT, STANDBY, OPERATE, XMIT, COOL)
- Band
- Anode voltage
- Anode current and bargraph
- Forward power and bargraph
- Reflected power and bargraph
- Grid current and bargraph
- Tune and load cap settings and targets
- Fault name
- Soft-start
- The basic HVPS step-start technique is pretty familiar: When going from STANDBY to OPERATE, an SSR puts a resistor in series with the HV transformer primary for about 4 seconds. The processor then checks to see if the HV is rising properly (no shorts, etc). If all is OK, then a large contactor shorts out the SSR and resistor applying full voltage to the primary. Heater inrush current to the tubes is limited simply by using a small current limited switching power supply module to power the heaters. I like it better than a transformer because it's small, light, cheap, adjustable, and regulated.
- Tuning phase detector
- I used a phase detector on the YC156 amp and liked it enough to include it on this amp as well. It works by sampling the voltage at the anodes and cathodes of the tubes and providing and indication of how close the phase shift is to zero degrees. The anode and cathode voltages of a cathode driven amp will be in phase when the anode is properly terminated in a resistive load. In use, the red and green TUNING LEDs give a quick indication of whether the amp is tuned above or below resonance. This is especially handy when the tune cap is a multi-turn vacuum unit. A gain detector could also be included to detect the correct setting of the load cap. Together, these gain and phase detectors could form the basis of an auto-tune amp. Maybe on my next amp with a bigger processor and more memory.
- Sequence of operating states;
- OFF
- WAIT - 3 minute wait for heaters; blower on, HV off
- STANDBY - tubes warmed up, HV off; can select band with BAND button
- OPERATE - HV on, ready for XMIT
- XMIT - transmitting
- COOL - 2 minute cooldown with heaters off
- OFF
- Stability
- HF and VHF instability were never a problem with this amp even during development. The coaxial grid construction and grid ring of the 3CPX800 tubes allow very good RF grounding of the grid, and excellent stability. The YC156 is nice because the grid/mounting flange almost forces you to properly ground the grid. Instead of the usual finger stock for the grid connection, I used some springy metal mesh EMI gasket material to fill the small gap between the bottom of the grid ring and the chassis. The clamp bands on the Eimac chimneys insure that the gasket doesn't shift and the tubes stay firmly seated in their sockets. See the socket photo for more details of the grounding. Despite the simplicity of this technique, good performance has been confirmed through isolation measurements and actual use.
- RF chokes
- RF chokes can be troublesome is amps that cover all the bands 160-10m due to the inevitable resonances in this range. I don't really trust the idea of "parking" the resonances between ham bands and hoping they stay there during time, temperature, cover on/off, later modifications, etc, so I decided to use two chokes and a vacuum relay to switch them depending on band. I use both chokes in series on 160 and 80m, and short the larger one out on the higher bands. This works well, but I found out that even with separate chokes oriented at right angles, there still can be significant coupling between the chokes themselves and other components in the PA cavity. Another point to remember is that just the pF or so added by the vacuum relay and the wires going to it will change the parallel resonant frequencies of the chokes, possibly by many MHz. These points were the cause of several fireworks shows featuring 1.5 inch arcs that I didn't even get to enjoy because the cover was on the amp at the time, as it always should be whenever HV is present!
- Efficiency
- When I first built the amp I noticed that the efficiency on the higher bands was poor compared to what it was on 160m and 80m. I found that by adding a 120pF cap to ground at the cathode of each tube, not at the input PI network which is remote from the tubes via a length of coax, the efficiency on the bands above 80m increased by 5%-10%, a substantial increase. I believe this is due to lowering the cathode drive impedance at harmonics of the drive frequency, which helps efficiency because the tube input and output are in series in a cathode driven amp, and the current through the tube isn't sinusoidal(contains harmonics) in a class AB amp.
- Fault actions (possible cause)
- Anode current fault (too much tube current, arc, HV short):
- Amp goes to STANDBY (exciter switched to output, KEY_IN connects to KEY_OUT, exciter keeps going)
- Press UP to reset and return to STANDBY
- Grid current fault (load fault, overdrive):
- Amp goes to OPERATE (exciter switched to output, KEY_IN connects to KEY_OUT, exciter keeps going)
- Press UP to reset and return to OPERATE
- Reflected power fault (load fault):
- Amp goes to OPERATE (exciter switched to output, KEY_IN connects to KEY_OUT, exciter keeps going)
- Press UP to reset and return to OPERATE
- Air pressure fault (blower intake blocked, cover removed):
- Amp goes to OFF
- Press UP to reset and return to OFF
- Anode voltage fault (heavy load on HV supply, low line voltage, bad rectifier):
- Amp goes to STANDBY (exciter switched to output, KEY_IN connects to KEY_OUT, exciter keeps going)
- Press UP to reset and return to STANDBY
- HV stepstart failed fault (blown stepstart resistor, HV short):
- Amp goes to STANDBY
- Press UP to reset and return to STANDBY
- Press UP to attempt OPERATE again
|