Keyclick and Noise Blanker Mod for FT1000MP MK V by
W8JI
Download Keyclick and NB Mod Information by PA1JR
Key Click Mod Report by IC8POF
Preliminary tests show very strong keyclicks +1kHz and
-1kHz. The LSB clicks peak 500Hz below the TX frequency,
and the USB clicks (in the LSB mode) are very obnoxious
because they are at twice the rate as LSB clicks (strong
on both make and break).
The FT1000MK V shows a 1mS rise and a 2ms fall with
sharp edges and a very poor rise and fall slope.
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. Unfortunately close-spaced receiver performance
has been neglected, and this affects all modes. The
transmitter also has nasty CW clicks. Stock MK V's
appear to be as bad as or worse than older 1000 series
radios in two important areas:
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.
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 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.
Rumors sine-shaped waveforms impact tone or readability
of signals are false, as are claims clicks rolling of 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.
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 cure.
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.
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.
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.
The NB Problem
Signals inside the wide roofing filter BW of the MK V
reach the gate of Q2009 through C2043. This point
precedes the narrow 8MHz IF filtering, allowing a rather
wide swath of unwanted signals to reach the gate of
Q2009. 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 never fully turn off Q2009.
The 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 in the non-linear
characteristics of Q2009 and 2010 (my manual has very
poor printing, but I believe it says 2010).
New signals created by this distortion and the resulting
mixing products are fed back through C2043 to the IF
strip. The IM products appear as "phantom splatter" on
SSB and " phantom signals" on CW. We can not actually
hear the distortion on frequency of strong signals.
Instead the IM products appear in the form of artificial
interference when we attempt to copy weaker signals
within 10kHz of a mixture of signals containing a few
moderately strong signals.
The NB Patch
The NB mod is a simple effective mod, and improves
close-spaced IM3 dynamic range around 10dB. A simple
factory change of moving one foil trace would have made
the MK V receiver noticeably better, but fortunately
this mod is fairly easy for owners.
The NB correction removes the surface mounted 220-ohm
resistor (R2046) from the source of Q2009, replacing it
with a 220 ohm resistor connected between Q2009's source
at C2027 and Q2016's (2SC4047) collector and the
junction of R2049 (220-ohm also).
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 someone sends step-by-step
text, I'll put it on this page.)
Some may wish to remove and change parts, 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 1/4w fixed resistor (R2)
(1) 680k 1/4w fixed resistor (R1)
(1) 220-ohm 1/4w 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:
Remove top and bottom covers to gain access to internal
circuitry
Remove the screws holding the IF board in place, and the
minimal amount of plugs to allow flipping the IF board
over. Draw a roadmap of all plugs that must be
disconnected. This will help you remember where
everything goes!
(click here to
download expanded IF board if needed)
Attach one end of the blue wire to the ungrounded end of
C2148, let the other end float
Locate and remove R2046.
Add the 220-ohm resistor to the Q2009/C2027 source and
capacitor connection point
Connect the other lead of this resistor to the junction
of Q2016's collector and R2048.
Reinstall the IF board with the flying lead exiting the
closest edge of the PC board.
(Take care to avoid pinching any wires.)
Remove all hardware necessary to access the RF board
Remove the RF board
(Click here to download
an expanded view of RF board if needed)
Attach the green wire to the junction of C1004 and
R1003/R1004 at the gate of Q1001
Route this wire up to the area of the IF board
Reinstall the RF board and all other hardware taking
care to not pinch any wires, and to reconnect all
unplugged wires in proper locations
Select a clear area on the IF board and mount the
terminal strip under the screws
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
Test the radio and reinstall the covers |
