For the past two years, I've been woking on a duo-band 80m/160m amplifier
using three GU74Bs. I've debugged and bench-tested the control grid and
screen grid regulators, QSK circuitry, plate and grid-trip circuits, and
various interlocks (e.g., to kill the screen voltage if the HV fails). The
amp uses a pi-L circuit with Q=10 (bandswitched with vacuum relays), with
2500V on the anodes (although I can change taps on the plate xfomr to raise
it up to 3000V if desired).. The plate trip circuit is set at 2.1A. and I've
initially set the operating bias voltage at -58V and the screen voltage at
300V. The tubes are fused with three 900mA HV microwave oven fuses, each
rated at 5kV. The input network is a passive 50 ohm 6dB power divider rated
at 150W. For initial testing, I removed all but one of the fuses. Before
applying HV, I conditioned the tubes by running them with just filaments
(and cutoff bias) for 50 hous.
Yesterday, was the initial checkout of the amp with HV and RF drive. My
initial test was just to key up the amp with no drive and look at the
resting plate current. It was about 250mA (600W), which strikes me as fine
for CW but a bit low for SSB. Any comments on this point will be very
welcome. I used an IR thermometer to measure the airflow temperature and
found that two of the tubes were matched to within a few degrees F, but one
tube was clearly running hotter, by 15-20 degrees F. I've got extra tubes,
so I plan to swap them around to find three that are more closely matched. I
have no way to measure their transconductance, but matching the resting
current is probably good enough.
With drive applied, the amplifier tuned easily and in accord with the Pi-L
network calculations. Everything seemed very stable, although with the top
cover off the amp the dimmer for the track lighting in my station went
crazy. Power output rose to about 2500W with 60W drive (15W after the 6dB
pad), but with no indication of any grid current at all. I don't measure
screen current directly, but have a red LED that lights when the screen
current goes positive. Tuning, therefore, consisted of peaking the power
output and adjusting the loading until the screen LED barely lighted. The
plate current dip seemed very shallow to me, possibly an indication that the
screen voltage is a bit low.
The tests stopped when there was a boom and the primary circuit breaker
tripped. The amp had been running for about 60 sec at about 2200W output (I
have a 2500W Bird dummy load), with about 1.5A plate current. At that
point, I remembered I'd removed two of the HV fuses, leaving just one 900 mA
fuse in the HV lead. I'm not surprised the fuse blew, but don't know why
there as a boom and the circuit breaker tripped. I'm guessing I may have
flashed over one of the tubes, probably by not nudging up the plate current
slowly enough to give the news tubes a chance to burn in under load.
So that's the status of my project. I don't know why I'm not getting any
grid current at full power, but I'm guessing that's probably normal. With
the tubes' bias tilted more toward AB2 (i.e., with a lower bias) I'd
probably see some current. I've not got any experience with the GU74Bs,
however, so I don't really know what to expect.
So that's the status of my project. One of these days, hope to you work you
all with it on the LF bands.
73,
Jim W8ZR
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|