The 3CX3000F7 power grid triode was chosen for my HF linear amplifier because its 4000 watt plate dissipation could easily handle the heavy duty demands of Amplitude Modulation, plus I only had to build one D.C. power supply. The rugged 3CX3000F7 has a grid dissipation of 225 watts vs the popular 8877 triode which has only 25 watts of grid dissipation. The 3CX3000F7 (rated to 30 MHZ) is exactly the same electrically as the 3CX3000A7 (rated to 110 MHZ) however the 3CX3000A7 requires a socket whereas the F7 is bolted to the chassis. I chose the 3CX3000F7 rather than the A7 because I didn't expect to change tubes very often and I wasn't going to run this RF linear amplifier above 30 MHZ. The cabinet and PI-L tank components with right angle drives were leftovers from a partially constructed Gates transmitter that I purchased at a hamfest. With the major mechanical parts in place, the rest of project was fairly easy to build. The amp design was for 80, 40 and 20 meters. In 2007 I added the 160 meter band, which is described, as a seperate section, near the end of this webpage. (Schematics are also included).
Created a homebrew heavy aluminum template for the 3CX3000F7 and mounted it in an old Gates cabinet. Installed the 3CX3000F7 by setting it's grid ring on the top of the chassis (template) and holding the grid ring down with two heavy homebrew aluminum half-circle mounting strips. I had an Eimac SK-1406 tube chimney for a physically larger 4CX3000 so I made up the space difference by inserting a ring of teflon between the tube and the chimney and "glued" it in with GE Silicon Kitchen and Bath caulk.
Added a homebrew plate choke, glitch resistor and homebrew parasitic choke (plate circuit schematic at bottom of this webpage with wire size and turns count). Full power tests on the 3CX3000F7 were conducted on the popular AM frequencies of: 3885 kHz, 7290 kHz and 14286 kHz with no signs of instability. (Check on my 160 meter modifications at the bottom of the webpage - no parasitic choke required).
Installed the 3CX3000F7 filament transformer, two 30 amp filament chokes in parallel, the RF drive coupling caps and installed eleven 6 amp diodes between the center tap of the filament transformer and ground to set the bias at -7.7 volts for a no drive idling plate current of 325 mills. The original 1000pf coupling caps were replaced by 6800pf (yellow ones) for a little better match to the cathode as noted by less reflected power. Since there is still some reflected power with the 6800pf capacitors it suggests that an even greater value would be better.
Built a full wave bridge, capacitive input, Peter Dahl transformer based 5000 Volt power supply into a standard 3 foot Bud cabinet (Schematic at bottom of webpage). A soft-start delay was built using a 30 second solid state cube (ABB SSAC Solid State Timer part number TS14130) in series with a 60 amp contactor that shorts out a large wire-wound start-up resistor. For diodes I paralled two sets of 14KV K2AW Silicon Alley rectifier diodes and mounted them on a 5" x 7" x 2" heat sink. The high voltage fuse between the two standoff insulators was made from a single four inch piece of #30 magnet wire. If there is a high voltage short in the 3CX3000F7 compartment, the fuse material explodes and disappears (this feature was tested). High voltage wire between the power supply cabinet and the amplifier has a 30KV rating and is encapsulated in shielding stripped from LMR-400 coax cable, which is grounded at both the amp cabinet and the power supply cabinet. The power supply has a rating of 5000 Volts at 1.7 amps CCS with a 133 MFD 8100 volt bank of computer grade caps inside a Bud cabinet. The power supply cabinet and the 3CX3000F7 RF Deck cabinet are connected by Scotch 25 Electrical Grounding Braid which also connects to the common point station ground.
Click HERE to see my other high voltage supply.
Outside view of both the 3CX3000f7 linear, the power supply and the remote viewing wattmeter. (The top section of the 3CX3000F7 amp cabinet has a switchable inductor and a vacuum variable for tuning out the 2nd harmonic; however, the 2nd harmonic is so far down that I keep it in the switched-out position). Even though I have an oscilloscope at the operating desk in the ham shack, I added an inexpensive oscilloscope and RF pickup loop in the garage to save me those back and forth trips while tuning up. The wattmeter reading camera has a 900 mHz transmitter. Video output from the remote 900 mHz receiver plugs into a video monitor in the ham shack.
POWER UP TESTS:
Tests were conducted at 3.885 mHz with the following setup (refer to block diagram below):
The AM driver transmitter was 100% modulated by a 1000 cycle audio tone. The input and output waveforms were sampled with RF pickup couplers and an oscilloscope. After the initial highest power test, with amplifier loading controls adjusted for linear operation, the 1 KW Navy dummy load (DA-75/U) was given a cooling off period, then the rest of the power output tests were completed.
Block diagram of test setup (print in landscape orientation).
It was noted that if the final tank circuit was retuned using only the Bird wattmeter, significantly more output power was obtained for a given input; however, the AM waveform on the scope was distorted. For true linear amplifier service under amplitude modulation conditions, with a plate voltage of 5000 and a grid bias of -7.7 volts, the following 3CX3000F7 data applies to my amplifier:
25 watts in----350 watts out
50 watts in----700 watts out
100 watts in---1400 watts out
200 watts in---2800 watts out
300 watts in---4200 watts out
400 watts in---5600 watts out
500 watts in---7000 watts out
Since the output power is fourteen times the input power, I set the AM driver transmitter for a tad over 26.5 watts of carrier (aprox 107 watts PEP) which yields a 375 watt carrier and in AM service it produces a clean 1500 watts PEP. Another nice thing about the extra 3CX3000F7 tube plate dissipation is that on those occasions when I work between 3600 and 3800, where the SWR is a little high, I don't have to retune, I just increase the drive and talk.
A
B
C
A: Schematic of the 3CX3000F7 plate circuit (print in landscape orientation).
B: Schematic of the 3CX3000F7 cathode circuit (print in landscape orientation).
C: Schematic of the 5000 volt 1.7 amp plate supply (print in landscape orientation).
The 3CX3000F7 amplifier photographed and described above has been operating flawlessly for three years on 80 and 40 meters. Documentation on the original Gates amp indicated that it was a 2 to 30 mHz design. Being interested in 160 meters, I measured the PI-L with an MFJ 259 and discovered that the PI-L would just make 1885, which is a popular AM frequency, so I decided to make mods to this amp. In 2007, I completed another 3CX3000F7 amp and took ideas from that amp and incorporated them into the mods. Click HERE to see the other 3CX3000F7 amplifier.
MOD #1, REMOVED PARASITIC CHOKE AND REPLACED PLATE CHOKE:
1
2
3
4
1: Test for PI-network range. (using clip leads, a carbon resistor was added for the test).
2: Removed the unnecessary parasitic choke and removed the old homebrew plate choke and replaced it with two Peter Dahl 90 uHy plate chokes in series. (180 uHy total now in the circuit).
3: Added a 2nd plate choke bypass capacitor. (fat red wire shorts B+ to ground if the cabinet door is accidently opened).
4: Schematic of the plate circuit mods.
MOD #2, REPLACED DUAL FILAMENT CHOKES AND REPLACED COUPLING CAPS:
1
2
3
1: Replaced the B&W filament chokes with a Peter Dahl 180 uHy 75 amp filament chode.
2: Replaced the 6800 pfd caps with .015 mfd @ 2.5KV mica caps.
3: Schematic of the cathode circuit mod. Notice the removal of the bias diodes which are not needed. With diodes the idling current is around 235 mills with no diodes the idling current is around 360 mills.
The mods are complete and the amp makes full power on 160 meters.