My GI7b amplifier started
out following the YU1AW design as shown here on the LEFT. In the
building of the anode compartment I made some changes, mainly because
when looking at the YU1AW design, and information that can be found
using Google, it appears that certain parameters are not that stringent.
The size and diameter of the coils is one area, some versions of the
YU1AW design have one coil, some have two and all pertain to be the same
design. Subsequently, there is a reasonable amount of latitude to 'play
around' with, there's more on this below.
The real reason I opted for
this design of amplifier is because back in 1998 I acquired a couple of
N.O.S. GI7b tubes. And after sitting on my shelf ever since I thought it
was about time I did something with one of them.
The next item I picked
up for the project was the socket. I didn't go for one of the many hand
made sockets you can find available these days, not that there's
anything wrong with those! It's just I saw this socket on eBay and it
was cheap, so I bought it.
The blower
I've used, some people will recognise as the unit pulled from the GS35b
amplifier of mine, this little blower supplies enough air flow to meet
the requirements for the GI7b tube, and once I'd cleaned it up, it
worked fine.
The cabinet is the 19
inch rack mounted unit I've had sitting here for some time. This picture
shows the cabinet with the transformer I've used, plus the GI7b tube
and my original tune and load caps. Included here is the junk 'voltage
doubler' PSU board
that I'd also planned on using, but due to it's design, haven't.
I quickly realised that
although I'd created a grid compartment (the blue/silver box) as my
first step in this little project, this wasn't the way to go. So, pencil
and paper to the ready and I designed a complete grid and anode
compartment as one unit. That may sound grand, but it isn't really! All
I did was I'd drawn what was required. i.e. a big box sitting on a
little box, then I created drawings that would allow some one to cut my
some sheet metal out of aluminium so I could simply bolt the sheets
together. (see below)
This is
what I drew and then had cut in aluminium. I simply made the compartment to fit my 19"
rack unit, making sure that the tube, caps and coils would fit and that
was it... At the same time, I had holes cut in the front panel for the
C1 and C2 capacitors, plus the hole I required for the GI7b socket, and
for the hot air exhaust. I also had the holes cut for the RF in and RF
out.
Once the
panels arrived here, I quickly realised that you have to be perfectly
sure you've got your measurements right! On some panels I had forgotten
to allow for the 2mm wall thickness, not a huge problem, but it meant
that the panels although bolted together, there were some places
where they didn't quite site right.
Here's the
set of panels bolted together using aluminium angle. I also had the
holes cut for the C1 and C2 capacitors. I used 6mm nuts and bolts to
hold it all together with a set of captive nuts for the anode lid.
This is
another view of the anode compartment, this time with the lid fitted and
showing the hot air exhaust.
Here's the anode
compartment with the GI7b in place, C1 and C2 capacitors and L1 and L2 coils fitted.
Yes, TWO coils! I noticed that some copies of the YU1AW circuit had just
one coil, others had two coils. Through experimentation I found that I
needed the two coils.
There's a
1,000pF door knob on the top side that is where the HV comes in via a
porcelain feed through, the door knob capacitor if to decouple the DC. Then there's the
RF choke, details can be found on the web on how to wind this, but I
simply copied a choke I had made for yet another 50mhz amplifier. The DC
connects to the bottom of the choke and the top end to the GI7b tube
itself.
The green wire is a
resistor 'test set up' that will simulate the presence of the HV DC,
then, it's possible too use an analyser on the RF output to test the
resonance of the RF circuit, as I have an
MFJ-269 analyser that's what I used. This makes adjusting the coils and caps very simple to
do while winding your own coils.
Then, on the left of the
GI7b there's the DC blocking door knob and a copper strap to C1. C1
consists of a small vacuum variable in series with the 5pf door knob
giving me circa 5 to 15pF, on the left of C1 is C2, a 30 to 110 mF
vacuum capacitor. These two capacitors do not need to be vacuum caps,
I've used them as that's what I had here.
With my MFJ analyser connected to the
RF output socket, and with the resistor connected to simulate DC voltage
on the tube, I was able to 'play around' with the coils and C1, C2.
Eventually I ended up wit the match as shown on the right.
Here's a
copy of an amplifier by 9A6C that uses TWO GI7B tubes, in the end I used
the coil dimensions given in this circuit for my single tube amplifiers
anode RF circuit.
The grid
compartment is a T match circuit. In simple terms, the DRIVE RF
comes in and connects to a 1,000 pF disc capacitor, then through a coil,
and then a small 4 to 20 pF capacitor that's to deck, and then another
coil to the grid connection of the tubes socket.
The heater
connections are fed their supply via two 'feed through' capacitors then
through a bifilar wound choke before connecting to the socket. The bias
supply is connected to this heater supply on the outside of the grid
compartment. Cooling air flow comes from the left via the square hole,
air flows around the grid compartment and up through the socket to cool
the larger anode fins before leaving the amplifiers out of the top
cover.
With my MFJ
analyser connected to the input and with the tube fitted into the
socket, it was possible to test for a match. With adjustment of the
coils by pulling them apart, or squeezing them together. This, along
with adjustment of the capacitor got the match shown here on the
right.
The heart of
any good amplifier is the control it uses, and why try to re-invent the
wheel!
The control board seen
here on the 'top right' of the picture is a GM3SEK triode board. This
board has been modified for use with the torroid transformer also seen
in the picture. The modification is something that Ian, GM3SEK has come
up with and is the first time it has been implemented in this way, as
such, it's still in the 'beta test' stage.
The transformer supplies
heater voltage plus two 700 AC windings, and with these connected in
series to a HV board from WD7S (a full wave bridge rated at 4.5KV)
gives me a stable 2KV of DC. The board has provision for a HV 'glitch'
resistor, but I have used these connections for a high rupture, high
voltage fuse. The fused HV then leaves the board (red wire, top middle)
and connects to a 100w 50ohm glitch resistor that is fitted on the rear
of the brown main board.
240v control and step
start relays are located on the bottom of the cabinet on the right of
the transformer. These relays are controlled by the GM3SEK triode board.
The manual for this board gives details on how to formulate the step
start system.
The red/orange wires
shown on the bottom right hand side of the picture as the heater supply.
I found it easier to connect these to the RF deck via a 'quick release'
chock block connector.
Here's a picture of the
RF deck fitted within the 19" cabinet. At this point in time the
amplifier is almost ready to try.
This next
picture on the right shows the cabinet with the front panel waiting to
be bolted in place. The front panel is already fitted with grid and
anode current meters, C1 and C2 control knobs and the four LEDs
required
by the GM3SEK control board. There is also a 240v power switch and
'stand-by' switch. All can be seen on the next picture. I fitted this
panel with a multi ribbon umbilical cable to allow for its removal
with out any problems. If I need to re-test the amplifier at
some point in the future like this then I have a temporary set up just
four LEDs and meters ready to connect with a longer umbilical.
It's all a bit of a
squeeze!
With the blower fitted to the RF deck, and the TIP147 device
that comes as part of the GM3SEK triode board kit fitted to a suitable cooler (black fins on the right of the blower)
there isn't much room left!
In this
picture on the right, you can see that I fitted a pair of TX/RX relays
(under the left hand side of the anode compartment).
I've used two CX520 relays. These are also controlled by
the GM3SEK control board.
Grid and Anode meters
are shown here on the left. The Anode is clearly marked as 0-350mA as
being normal, with anything above 350mA being abnormal. The Grid meter
is marked out as 0-80mA as normal and anything above 80mA as abnormal.
The control board will trip the amplifier at 90mA.
The GI7b amplifier in its
finished state. As with all of my amplifiers, it has been given a name
and this one is Piglet.
This has nothing to do
with the Disney movies, i decided on Piglet as I consider the amp to be the
daughter of my GS35b amplifier that acquired the name 'The Pig', the
reasons for that can be found elsewhere on my web site.
So, the question most
people will want to know is "how does it work?"
Well, from the outset I
was looking for an amplifier that would deliver 300w.
As it is, I can get 350w
out with both the grid and anode reading within normal parameters for
about 40w of drive. 300w is achieved with 30w of drive. If I'm careful,
with some very fine tuning it will produce 400w, but would then
often trip on grid current.
I dare say that with a
few more volts on the anode, over 400w would be achievable. But that was
never the intention.
The amplifier was first
used in the RSGB 50mhz trophy contest in June or 2010, and it performed
admirably throughout, never missing a beat or complaining about
anything.