Welcome to my favorite part
of my web site. The photo above is the famous
Eimac 4CX1500B/8660 ceramic tetroide. This is the heart of my HF amplifier project.
After two years of planning, looking for parts, construction, and testing, it finally became
a reality. A legal limit "Big Gun" amplifier. Who would of thought I could do it? Well,
Ill tell you who...K1JI, John Isidoro, that's who! John pushed me into making an attempt
to build such a unit. He had build a 4CX1000A back in the 70's and has a vast knowledge
of electronics. So with guidance in the technical department, and my mechanical
know how...it lives!
This amplifier uses a passive
grid drive configuration. This means the grid is swamped
using a 50 ohm resistor. This keeps things stable and requires no neutralization. I built my
own 50 ohm input resistor out of 27-150 ohm resistors. There are 3 sets of 9 resistors. Nine
resistors in parallel, and 3 sets of those are in series, for a total of 50 ohms. Drive requirement
is typically 35 to 45 watts in to achieve 1500 watts out at the maximum
rated plate current of 900 ma.
Typical operating values @ 14.000 Mhz
Efficiency = 62.5% meter error +/- 3%
Ep:2880v x Ip:570ma = 1026 watt out with 18 watts of drive
Ep:2800v x Ip:680ma = 1190 watts out with 25 watts of drive
Ep:2720v x Ip:790ma = 1349 watts out with 35 watts of drive
Ep:2720v x Ip:850ma = 1445 watts out with 40 watts of drive
The 4CX1500B is rated at 1100 watts output,
continuous duty FM//CW/AM . In SSB
service, this tube will last virtually forever. This tube has low IMD qualities and is a great tube for SSB service. To me, this much more important than having more power output and a higher IMD figure. The difference between 1100 watts and 1500 watts is about 1.5db. If you can see that on your S meter, you have one hell of a radio!
power supplies are all home brew except for the 24vdc unit that
relays and indicator lamps. The high voltage supply is basic. It uses a 4400vac/1amp
transformer that is fully adjustable via a 20amp/240vac variac on the transformer primary.
It uses a full wave bridge rectifier and 10-400uf/450vdc capacitors. The B+ is supplied
to the tube via Belden "wire core" spark plug wire. The bias and screen supplies are
regulated using 2 power transistors in each and kept slightly loaded. Both of these supplies
are fully adjustable. Setting the plate idle current correctly demands that the bias supply be
adjustable. The screen supply is loaded to about 60ma to keep it as stable
as possible. The filament supply uses a 6vac/15.5amp transformer that is adjustable from
a small variac to tailor the voltage to the filaments. Eimac calls for 6vac on the nose. This
will prolong the life of the tube if set correctly.
The relays used in this unit
are basic reed type relays. They all operate on 24vdc.
The antenna relay (TX) is a modified reed relay that is rated for 20 amps. I removed the wire
that feed the contacts and replaced it with RG-58 coax braid. This beefs up the wire
size and keeps the impedance down as much as possible. The bias and TX relay are
delayed. When the amp in keyed up, the TX relay closes, 80msec later, the bias relay
closes and turns the tube on. When the amp is un-keyed, the bias relay opens and turns
the tube off, and 80msec. later, the TX relay opens and returns to its normal position.
This does two things, it prevents the amp from transmitting into a "no connection" condition
while the TX relay is in motion (point bounce). It also saves the contacts on the TX relay from pitting. This is done with a simple RC circuit before the relays and is confirmed on a scope. So why
didn't I use PIN diodes? Big bucks that's why!!! And the fact Im trying to keep things as simple as
I can. It would have been nice to make this amp run QSK, and I suppose it probably could. But
the cost of PIN diodes or fast vacuum relays is just to much for my needs.
Looking at the rear of the intermediate supply
you can see several connection
points. First a jones style plug carries the screen, bias+, filaments, B-, and a 24vdc feed that
controls the TX relay. The cable used is 10 conductor from a police light bar! (I wonder where
that came from?) There are two 10ga. wires in this cable that are perfect for feeding the filaments. The rest are 16ga. just right...