|
TX Gain Menus
It’s great to see a manufacturer offer improved close-spaced
SSB transmit performance! The MK V Yaesu reverses other modern radio’s downward
spiral of transmitter SSB IM performance. Yaesu included a
class-A mode. Even
without class A, the
HV finals can be
very clean when
compared to other
radios if the rig is
kept out of ALC.
Unfortunately
Yaesu missed
correcting two
important flaws:
- The
noise blanker,
like in other
earlier Yaesu’s,
creates receiver
IM distortion even
when off
- The
transmitter has
virtually no
wave-shaping on
CW, causing
terrible keyclicks
Preliminary tests
show very strong
keyclicks +1kHz and
-1kHz.
An early FT1000MK V
tested here had 1mS rise and 2ms fall
times with sharp edges.
Later FT1000MK
V’s, are slightly
better on and look
like this:
The very sharp
falling edge is a
particular problem
in later FT1000MK
V’s.
What does a 1mS
fall mean?
A CW signal is
actually a 100% AM
modulated signal.
The fall (or rise)
time makes up 1/2
the time required
for a full rise and
fall cycle.
If the time
period of the fall
is 1.04mS, the
frequency of the
modulation frequency
is the reciprocal of
the rise (or fall)
time divided by 2.
1/.00105 = 952
952/2= 476 Hz.
The lowest order
sidebands possible
at ANY keying speed
are 476Hz away from
the carrier, one
above the carrier
and one below the
carrier. This means
the narrowest
possible bandwidth
with a perfectly
shaped raised sine
wave envelope would
be 952 Hz. Since the
FT1000 MK V does not
have a perfect
raised sine
waveform, the actual
sidebands contain
high levels of
harmonics. This
makes the bandwidth
even wider. It is
the really sharp
turn-over on the
edges that kills the
bandwidth.
Radios almost
always have a
different bandwidth
on make and break.
This is because
envelope shape is
different on rise
and fall.
A stock FT1000MK
V from Europe sounds
like this in the
USA!
Wav file link
Keyclick and Noise Blanker Mod for FT1000MP MK V
Since a portion of the keyclick mod requires accessing the same general area as the noise
blanker, it would be prudent to fix the receiver and transmitter at the same
time. This article offers a combined modification that patches both problems.
The NB Problem
The worse thing
about the receiver
in any of the FT1000
series radios (I’m
sure this carries
over to other Yaesu
models ) is the
noise blanker. The
noise blanker, even
when OFF, causes
deterioration in the
close-spaced IM
performance of the
FT1000MK V.
1.) All signals inside the wide roofing filter
(70 MHz) BW of the MK V
are applied to the
IF input of the IF
unit.
2.) They
immediately go
through a very good
balanced mixer
(Q2020 and Q2024 in
push-pull). This
mixer has very
little distortion
and very good strong
signal handling.
3.) 70MHz signals
are converted to
8.2MHz .
4.) The 8.2MHz
signals reach the gate of Q2009
through C2043
(darkened line on
schematic). This point precedes
all narrow 8MHz IF filtering–allowing a
rather wide swath of unwanted signals to reach the gate of Q2009
along with the
desired signal.
5.) Q2009 is left
operating even when the noise blanker is turned off and can have substantial
gain depending on bias voltages at TP2001. Bias voltages at TP2001 can be varied
by changing menu settings for NB gain, but
even the lowest NB
gain settings never fully turn off Q2009!
6.) Q2009 acts
like a mixer,
creating unwanted
mixing products of
desired and
undesired signals.
Accumulated level of
all signals reaching
the gate of Q2009
produce a large net
voltage at the drain
of Q2009. This
voltage (and
resulting net
current) causes
overload and
distortion by
driving Q2009 and
2010 into
distortion.
7.) The unwanted
distortion products
feed right back down
the same connection
into the 8.2MHz IF.
These new IM products appear as “phantom splatter” on SSB and
” phantom CW signals” on
CW. On CW, they
sound like random
blips and bloops
that aren’t real
Morse characters. We cannot actually hear the distortion on
the frequency of strong signals.
The IM products
simply appear as artificial interference when we
attempt to copy
weaker signals within
about 5-10kHz of two
or more moderately
strong signals.
Anything we do
to increase IF or RF
gain in front of
this point will
greatly increase
unwanted noise
blanker
intermodulation
products. This
includes
increasing 70MHz IF
system gain.
The NB Mod
The NB mod is a simple effective mod.
It improves close-spaced IM3 dynamic
range about 10dB on
average. In rare
cases I have seen as
much as 20dB change!
Moving just one foil
trace, a very simple
manufacturing change, would have
made the MK V receiver noticeably better
in close-spaced
performance. Fortunately this mod is fairly easy
for owners.
The NB correction removes surface mount 220-ohm resistor (R2046)
from the source of Q2009, replacing it with a 220 ohm
leaded resistor connected between
Q2009’s source at C2027 and Q2016’s (2SC4047) collector and the junction of R2049 (also
a 220-ohm).
Keyclick Problem
Raised- sine rises and falls would provide the fastest possible CW
speeds for a given bandwidth. With properly filtered rise and falls, we would hear
little or no change or softness when listening on-frequency. Tuning off- frequency, clicks would
quickly vanish.
If a 2 or 3 millisecond rise and fall is used for operation at very high CW
speeds in a single-pole R/C filter (this radio uses a simple RC filter) a transmitter
is almost guaranteed to interfere with less strong signals within 1kHz or so.
Transmitters with fast rise and fall times should stay at least 1.5kHz away from
operators working weak signals, especially when the CW transmitter has 1930-era
transmitter CW shaping.
Rumors sine-shaped waveforms impact tone or
readability of signals are false.
Some people even
claim clicks rolling off
at some
“X-dB-per-octave
rate” beyond a few
hundred Hz are a
necessary part of
life. Such
statements are
misleading, likely
being based on the
incorrect assumption
the receiver has
very wide bandwidth
and the transmitter
is filtered through
a single stage
simple
resistor-capacitor
click filter.
If you want to hear the
sound of proper shaping, listen to this click-free signal recording as I tune
across the signal. Off- frequency (even a few hundred hertz), we hear no clicks at
all. On- frequency the CW
is “hard” sounding, allowing copy to 60-WPM or more. When the tone
disappears in the deeper receive filter’s skirts, clicks also disappear.
In contrast,
compare the MK V recording as I tune past the signal.
This signal is from Europe on 40 meters!
There is a day-and-night difference off frequency between the no-click and
loud-click signals. On-frequency both signals sound the same.
MK V clicks are caused by excessively fast
rise and fall, and very poor shape of the rise and fall.
Unfortunately when we patch poor CW transmitter designs, we can not
make perfect corrections. Without major modification we can not
modulate the MK V ( or most other transceivers) with properly filtered (which also
means perfectly shaped) rise and falls. This modification, like the MP and 1000D
click mods, is a patch…not a
perfect cure…but
it is about 20dB
better 1kHz away
than doing nothing
about the problem.
Patches Vs Cures
Because mods on existing radios are patches, the radio owner must make a
choice. If the user operates speeds faster than 45
or 50 WPM, the rise and fall required for legal close- spaced operation may be less than
ideal. This does not mean the ability to work weak signals at modest speeds (up
to 30 WPM) would be compromised even the slightest amount. It means
high- speed ops (speeds over 45-50 WPM) may find the CW slightly mushy when
adequate for close-frequency operation.
Operators with stock MP MK V’s should always try to operate at least 4kHz away from weak or
moderate signal- level stations. Part 97 rules prohibiting keyclick emissions
that interfere with adjacent frequency operations. The specific rule is
97.307(b) “Emissions outside the necessary bandwidth must not cause
splatter or keyclick interference to operations on adjacent
frequencies.”
The Click Mod
The actual click mod requires changing two stages. The first stage modified is on the IF
board. The IF mod slows the rise and fall of mixer transistors Q2033 and Q2038.
Note: This stage is easy to modify, and is located on the same board as the
noise blanker. This allows the noise blanker to be corrected at the same time.
By itself, modification of the easy-to-reach IF board is NOT effective for
substantially reducing clicks. A later stage on the RF board also has truncated
rise-and-fall times with a very poor R/C edge-shaping system. This later stage
continues to add clicks even after earlier stages are modified.
RF amplifier stage Q1001 has the fastest rise and fall in the RF section.
Q1001 must have proper gate bias shaping and
timing to reduce clicks to acceptable levels. Removing D1002 and altering
components around Q1002 slightly reduced clicking, but I concluded any effort
wasn’t worth the result with bias rise and fall more rounded at Q1002.
As designed, Yaesu uses a square wave very
rich in harmonics to drive a simple R/C filter. This poorly filtered square wave
amplitude modulates the RF and IF sections. The poor basic filtering design,
combined with non-linear amplitude response, requires great care in component selection.
It also means we never will achieve the optimum bandwidth for any give rise and
fall time and ultimate CW speed.
Making the Mod
This mod is a little more complex than the FT1000MP mod because the chassis
of the MK V is a little more complex and unfriendly. Like any service work,
having a clean open bench and a spot to separately store screws and other
hardware removed in every step in order will make the job smooth and easy.
If you do NOT want
to modify your MK V,
I can make the
modification for a
nominal fee. You can
e-mail me at my
callsign@mycallsign.com
Some may wish to remove and change
surface mount
components… but I prefer to wire the click-mod
to a single terminal strip. This will allow you to customize the mod, switch the
mod in and out, or correct any errors without dismantling the entire
radio.
You’ll need the following parts:
(1) one foot each of two small insulated wires, #20-#26 one ( preferably) green and one
blue to
make connections
(1) four-lug (with ground) terminal strip
(3) .1uF 50 volt disc capacitors (C1-C3)
(1) 22k 1/8w fixed resistor (R2)
(1) 680k 1/8w fixed resistor (R1)
(1) 220-ohm 1/8w fixed resistor
In addition you need a well-lit bench, along with some hand-tools such as
soldering pencil and solder, screwdrivers, and cutters and strippers.
Populate the terminal strip as follows:
IF Board Mods
Clicks
1.) Remove top and bottom covers to gain access to internal circuitry.
2.) Remove the screws holding the IF board in place, and the minimal amount of
plugs to allow flipping the IF board over.
You should be able
to flip the board
over by removing
only two ribbon
cables and one
shielded cable. Draw a roadmap of all plugs
and cables that
must be disconnected! This will help you remember where everything goes.
(click here to download expanded IF board if
needed)
Click IF Board
Attach one end of the blue wire to the ungrounded end of C2148.
The the other
end hangs loose
for now.
Noise Blanker
1.) Locate and remove R2046
220 ohm surface
mount.
2.) Form the
leads properly and
solder a 220-ohm
1/8-watt leaded resistor to the Q2009/C2027 source and capacitor
connection point
3.) Connect the other lead of this resistor to the junction of Q2016’s
collector and R2048.
4.) Reinstall the IF board with the flying lead exiting the closest edge of
the PC board. The
blue wire should
just hang out the
closest edge. Take care to avoid pinching any wires.
RF Board Mods
(improves clicks
only)
1.)
Remove all hardware necessary to access the RF board.
This includes the PA
section and fan,
both of which can
flip over the case
edge if a few wires
are unplugged..
2.) Remove the RF board.
You will have to
remove a small
shield panel and
unplug some wiring.
Be sure to remove
ALL necessary
screws, including
black screws near
the DIN jack on the
back panel of the
radio!
(Click here to download an expanded view of RF
board if needed)
3.) Attach the green wire to the junction of C1004 and R1003/R1004 at the gate
of Q1001.
4.) Route this wire up
through any opening
near the middle of
the radio to an area
near the IF board.
5.)
Reinstall the RF board and all other hardware taking care to not pinch any
wires, and to reconnect all unplugged wires in proper locations.
6.) Select a clear area on the IF board and mount the terminal strip under
a convenient clear
mounting screw area.
(Note: I have simply
heat shrunk the
resistor capacitor
network without
using a terminal
strip.)
7.)
Connect the green wire (from RF board) to C3, and the blue wire from IF
board to the junction of C1 and R1. R1 is a 680k resistor, and R2 is a 22k
resistor
(Example
of heat shrunk
construction below)
Reassemble, test the radio, and reinstall the covers.
The FT-1000 MK V
has hidden transmit
gain menus. They are
accessed by pushing
and holding FAST and
LOCK while turning
the POWER switch on.
Both of my MK V’s
and every MK V
serviced here has
had the TX IF gain
set too high. This
causes first
character clicks on
CW and spits and
splatter on SSB.
Here is how to
correct the IF gain
to prevent ALC
clipping on leading
edges of CW and
voice:
Press and
hold FAST and LOCK
before and during
initial POWER on.
Press FAST and ENT
at the same time.
You are now in the
MENU’s and the
display should say
“0-1 GrPI-cH”.
Turn the VRF/MEM CH
counter-clockwise to
9-2. The display
should say “t iF –
GA in” This is the
transmit IF gain
menu.
Turn the SUB
VFO knob clockwise
one position to
” t iF – 018″. This
is the 1.8MHz
transmit IF gain.
Press the ALC/COMP
meter selector until
the bar graph says
“ALC”. Set RF
PWR knob to full
power.
With the
radio on CW and a 50
ohm dummy load
connected, close the
key and adjust the
MAIN VFO-A knob
until the ALC
display is about
75-85% of full scale
on the illuminated
bar marked “ALC”.
Press the next band
button (3.5), make
sure the radio is
still on CW,
and turn the SUB
VFO-B knob clockwise
one band to “t iF –
035”.
Again adjust MAIN
VFO-A until ALC is
at 75-85% of full
scale.
Repeat this
process through all
bands.
Most radios
I have tested
require a setting of
2 to 4 on TX IF
gain, with 3 being
the most common
setting.
This
change will reduce
SSB bandwidth and
distortion. It will
also reduce
keyclicks and
annoying thumps on
the leading edge of
each Morse
character.
Copyright W8JI 2004.
No portion of this
page can be copied
or reused without
permission.
|
