Modifications for the Yaesu FT-901
Improving the YAESU FT-901/902 type transceiver
By LA8AK, Jan-Martin Noeding,
Voielien 39/B, N-4623 Kristiansand S.
Several printed circuit boards in either model are very much equal.
There has been some modification kits available to obtain the new
bands for FT-901, and the RF-unit has also been replaced to achieve
the better intermodulation performance as with FT-902.
The FT-902 has been in use here for 3 years, and some changes has been
found important. Too often it seems not possible to buy a new trans-
ceiver at the required specifications, but many features now found on
receivers add only marginally to the on-air performance and just too
often there are many new bugs, some of these are caused by incorrect
ways of operating a semiconductor device, some rigs have some boring
background noise which does not disappear, even for strong signals
(IC-730/740). Some of the bugs with FT-902 has been experienced with
the later YAESU models. 5-10 years old rigs are much easier to modify
and find room for the additions you prefer, compared to last models.
But there is of course a limit for when a piece of equipment is
worth modifying, exceptions are of course rigs with historical
attraction. The first owner has paid most part of the price to the
dealer. Therefore it seems quite interesting to buy a secondhand rig.
The following changes/improvements has been made for FT-902 (FT-901):
PB-1705A AF Unit. Audio Quality.
This is in my opinion awfull. Very bad bass-response. Looking at the
FT-7 audio circuit, will show the solution. C511 to the AF-power-
amplifier is decreased to 47nF (.047). Improvement is just inbelie-
vable! Measuring the audio RF to audio response now indicate that
there are still more room for improvement, but it sounds good .......
even with the built-in load-speaker, at least.
PB-1994 NB unit. Carrier Level Control.
This is a 5kW potmeter, with effective 'range' of about 1000 W, ie.
80% of the rotation is outside practical control range. This potmeter
is shunted with 1200 W, it is easiest to connect the resistor on
PB-1994 (NB-unit) pin 5, and ground, or under the PCB-socket.
PB-1720B RIT. (Receiver Incremental Tuning = Clarifier)
One must decide self which RIT-range should be. In my experience the
requirements depends on the operation.
.HF SSB . + 5kHz ... HF CW . + 1kHz ... 6M aurora. + 5kHz.
.6M CW . + 5kHz ... 6M SSB . + 5kHz ... 6M CW MS . + 5kHz.
.2M CW . + 2kHz ... 2M CW MS. + 5kHz ... 2M SSB . + 5kHz.
.2M Aurora . + 5kHz ... 70cm CW . + 5kHz ... 70cm SSB . + 5kHz.
.70cm Aurora . +10kHz ... 23cm CW . + 5kHz ... 23cm SSB . + 5kHz.
I discovered that sintran letter for ohm is converterted to W, so 5kW should
read 5k ohm
First I modifyed my rig for HF CW because it is very important to use
correct tuning in HF CW contests. The modification shown in fig.2
makes +1.6/-1.5kHz (original is +5.5/-4.5kHz). VR06 is shunted with
1500W resistor. Switch RIT on, set VR06 to center, tune RX to
7000.0kHz, switch off RIT and adjust VR01 to frequency reading 7000.0.
The modification as shown above was used for about one year, but it
was a desire to have some further changes. The circuit shown in Fig.3
switches automaticly between two RIT ranges, dependent on modes
1: AM/SSB/FM "wide" (+4.8kHz), 2: CW/FSK "narrow" (+ 1.4kHz)
This is solved by using a dual-OP-amp, LM358=MC3358P1 (TL082 will not
function). The circuit is connected on a printed circuit board. Some
time was needed to find a suitable location, but at last it was
mounted next to RIT-control, at the XTAL OSC. A little angle- bracket
was soldered to one end of PCB and fixed using the two available
screws which fixes the oscillator assembly. VR01 is adjusted such that
RX with and without RIT operate on the same freq. The transistor is a
P-FET, which is not too commonly available, I used the only available,
a J177 Switch-FET. For "narrow RIT" the FET reduces the gain, but
center-frequency remains equal.
PB-1994 NB-unit: RF-clipper (so-called RF Processor)
It was not possible to notice any operation of this circuit, reports
on the air did not indicate that this circuit did operate at all! I
made some measurement using RF mV-meter connected to PB-1994 pin 17.
Whistling into the mike produced 70mV from NB-unit, while "blowing"
into the mike produced 200mV rms, this indicate that RF clipper does
not work on most sort of speech, while perhaps very good for single
tone transmission....... The LIMITER (Q206 TA7060P) was suspected, a
pair of 1N4148 diodes were mounted in antiparalell across the output
from this IC, and the RF CLIPPER functioned !! Just as easy it was to
improve this. The same suspect circuit has been seen in some other
PB-1994 NB-unit: Processor Level Control. (NB unit Q206).
This strange circuit did not have any practical function. I used a P-
channel FET, the only one available was 2N5462, while it may be
possible to use the J177. Some resistors were changed too. 6-10dB
variation was achieved. The rotation is now opposite way, so the
center connection at the potmeter must be connected to the other side.
VR203 must be at maximum, but this was later removed and a strap was
inserted on the PCB to get maximum output.
The values for some resistors were later optimized and more than 10dB variation
was possible for 2N5462, but for J177 almost 30dB. By the way, I believe that
FT-101ZE uses similar speech processor circuit, so it may need the diodes.
Later reports on the air 'are possibly the best RF processor ever heard on the
bands, it is tested on 80m/40m, 6m, 2m, 70cm, 23cm, and some rather critical
persons who too often critisize have said this is good and RF processor may be
left on during local QSO's!
CARRIER UNIT PB-1706A. FSK-mode.
The dc input level is limited to +1.5V, which is an impractical
level, so I have inserted a 4k7 resistor in series with base (Q701).
It is now possible to use normal TTL-level input, noise treshold is
PB-1707B FM-unit.(PB-1706A *not cw-ms): Carrier level adjustments.
I believe that it is important to achieve as equal level as possible
for the different operations/modes, ie.(1) Normal SSB, (2) SSB with
RF-processor, (3) AM, (4) CW, and (5) FM. Difficulties arise because
the different units loads the common point, which is NB-unit pin 13.
Diode-switching is therefore used. A diode is inserted at the FM-
input to this point, see 'Overall Circuit for IF signals between the
different boards'. Another diode must be inserted in the other cable
from CARRIER Unit, possibly on this unit, and a similar modification
like the one for FM-unit must be made there. Fig. --- shows the
modification for FM board. In the physical location for TC901 is
mounted a trimpotmeter, a capacitor goes to the output, and resistors
are installed to bias the diode connected at the input to the NB-Unit.
It is now easily possible to adjust the levels from FM-unit, and CW/AM
signals from CARRIER unit) independently without interactions to
occur. When the diode is inserted between CARRIER unit and NB-unit the
capacitor in series with cable must be removed (this has a fixed
value, although it is drawn like a trimmcap). Some interesting details
with the FM-unit is that it is easy to find suitable connection points
for G3RUH 9600 baud modem, but it may be difficult to find some proper
free pins on the connector to wire these to.
It should be pointed out the difference between FM operation on VHF
and HF. On VHF/UHF 25kHz channel spacing is used, while on 10M 10kHz
channel spacing is employd, this transceiver is built for FM operation
with transverter. For 10M deviation must be reduced, and it is an
advantage to use a narrower FM RX filter. On VHF/UHF deviation is set
to +5kHz, while for 10m it should be adjusted to maximum +2.5kHz
(at voice peaks). This is important otherwise you will cause severe
interference with the adjacent channels on 10M. Using simple tools,
you can just reduce to audio to the varactor diode to about the half
of the setting as adjusted for VHF. You may whistle 'hard' into the
mike and measure at the input to the varactor diode with a sensitive
audio level meter, and adjust the reading to about half of the first
FILTER Unit PB-1995: NOISE-BLANKER-improvement.
I made an ignition noise simulator, this was earlier used to improve
the FT-7 noise-blanker, so it was interesting to compare FT-7 with
FT-902. The noise-blanker seems not to be too bad, but it seems to be
room for improvement. Without NB, S-meter reading was about S9, with
it was S-3, and after modification, the noise dissapeared almost
totally, the background noise could be heared without antenna.
Change: Add 0.1uF to the capacitor C328 (PB-1995 FILTER Unit),
bead-tantalum may be used.
After some years in use, the push-buttons will not function properly,
the problem is that some small springs, easily seen from above, will
not move too well in the tracks, some thin oil improves this.
PB-1715 PA-unit (2x 6146B): PA-valves protection.
To protect the power-supply against anode-to-screengrid voltage flash-
over, a small 1N4148 type diode is connected in series with screen-
grid supply to the 6146B's. This is not shown on the circuit diagram.
This may be good for power supply, but it may be fatal for the valves
in case of some secondary emission due to hard operation. Some com-
promise is therefore needed. The diode must be paralelled with a
10KW 0.5W resistor. To protect the screen-grid circuit against flash-
over I have connected a pair of SIEMENS varistors, SIOV 14K150, they
draw 1mA current @ 250V DC.
PB-1715 PA-unit, PB-1708 RECT-A-unit: BIAS-circuit improvements.
To stabilize grid supply against drift-off due to gassy valves, a
diode is connected in paralell with R05, a 1N4007 is preferred because
it is a poor RF rectifier. To avoid that the grid voltage should
dissappear because of bad contact in BIAS POTMETER VR01 a 100KW
resistor is connected from negative side to center. Often this occur
if you have left the potmeter in same position for years, and when
one needs to adjust the idle-current, the trouble sometimes happen.
PB-1708 RECT-A, PB-1717 RECT-C: IMPROVING THE CW-KEYING (Normal
Orignally the CW-keying is bad, lots of key-clicks. For HF CW ca.
4-6ms delay is optimum. A single capacitor cannot solve this problem,
because of different charge and discharge current. A suitable
compromise was made with simple capacitor/diode/resistor combinations
see figure ----
PB-1705 AF-Unit, PB-1846 VOX/MARK: AUDIO LINE OUTPUT.
For packetradio/RTTY/CW-meteorscatter it is desired to use an in-
dependent AUDIO LINE OUTPUT. Surprisingly, there was an unused 1/4-
LM324 on the AF board. There was marked but unused component-designa-
tions around this, one point was used only to connect two components,
they were removed from this common point and connected in the air.
Then the components as shown were mounted, the un-numbered components
are mounted below PCB or which is practical. It was an unused cable
from AF 'MJ6' to VOX 'MJ7', and further to J17 SIDETONE out on rear
side. I could not see any practical use for this TONE output, so I
decided to use this for AUDIO LINE OUTPUT. The 4k7 resistor is changed
to 1KW or removed, IN VOX UNIT R52 is loosened in one end.
Other changes (PB-1728A and other components):
MIKE contact is changed to the KENWOOD circuit for 4pin connectors,
such that it is possible to use same mikes for different rigs. CW-jack
is rewired. One contact is KEY, while the other is PTT in/out. The
ELBUG was removed, I like my own HB9ABO Keyer/Memory keyer better. A
separate CW-high-speed board was installed (keyfilter and keyed IF
amplifier). AM position is wired for CW-Wide (2.4kHz BW) for VHF/UHF
CW (normal CW/- Aurora/Meteorscatter) where larger bandwidth is
preferred. Perhaps someone else would prefer to have a switch to
select between CW-Narrow/CW-Wide, the AMGC switch may be used for this
PB-1715 PA: TRANSVERTER RF OUTPUT.
RF output level for transverters is far too high, this is reduced by
connecting a capacitor in series with C02=10pf. I used 3P3 with 1000
ohms in series (the resistor is used to avoid parasitics). About 20mW
is available on 10M.
PB-1717A RECT-C: +6V Regulator for RIT.
OZ1HDA (with FT-901) and I (with FT-902-DM) have been troubled by VFO
instantaneous instability, the +6V IC-regulator (Q01-TA7089) is sus-
pected. I changed the +6v regulator with a 78L06 regulator and the
problems seem to have disappeared.
Forward conductance for Q03=J310 must be adjusted to 20mmho (20mS) for
50W mixer termination.The drain-current must be ca 20 mA. Correct
value for R13 ca. 18 W, according to G4DGU, Chris (MUTEK).
Pse note that this manuscript is written in sintran and converted to
DOS, therefor some letters will be wrong, here W may mean ohm!
! Table 1 !
! COMPARING CIRCUIT BOARDS USED IN THE TWO MODELS !
! Board . FT-901 . FT-902 Note !
! RF unit ! PB-1702 ! PB-2154 Different !
! Filter ! PB-1716 * ! PB-1995 * Equal ? !
! IF ! PB-1704C ! PB-1704C !
! FM ! PB-1707B ! PB-1707B !
! CARRIER ! PB-1706A ! PB-1706A !
! AF(RX) ! PB-1705A ! PB-1705A !
! NOISE BLANKER ! PB-1703C * ! PB-1994 * Equivalent !
! Keyer ! PB-1728A ! PB-1728A !
! VOX/MARK ! PB-1846 ! PB-1846 !
! VFO ! PB-1440B ! PB-1440B !
! XO ! PB-1711A ! PB-2165 New bands !
! VCO ! PB-1710A ! PB-2166 New bands !
! PLL ! PB-1709A ! PB-1709A !
! Counter ! PB-1729A/PB-1730 ! PB-2086 New !
! Memory ! PB-1787 ! PB-1787D !
! Capacitor ! PB-1713 ! PB-1713 !
! RECT-B ! PB-1712 ! PB-1712 !
! RECT-A ! PB-1708 ! PB-1708
! RECT-C ! PB-1717A ! PB-1717A !
! Driver(12BY7) ! PB-1714 ! PB-1714 !
! PA (2x 6146B) ! PB-1715 ! PB-1715 !
! Miscellaneous ! PB-1718A/PB-1720 ! PB-1718A/PB-1720B !
Almost equal boards but new numbers are marked with an asterisk *)
Note the availability for modification kit to have the 'new' bands
Some IF measurements to be used for fault-tracing.
All readings are in mV-RMS(eff.Werte) +/- 3dB
* Adjusted level, see text for modifications
! P.C.B. ! PB-1994 (PB-1703C) NB-UNIT ! PB-1706A CARRIER-UNIT !
! MODE (TX) ! Pin 8 ! Pin 13 ! Pin 17 ! Pin 4 ! Pin 2 !
!FSK CW AM ! - ! 55 ! 70* ! 65 ! 75 !
!FM ! - ! 65 ! 70* ! - ! - !
!SSB(Normal) ! 60 ! - ! 70* ! 160 ! - !
!SSB(Proc.) ! 60 ! - ! 20-80 ! 160 ! - !
! RX ! - . - . - ! 30 ! - !
MC3356 S-meter for FT-901/902
MC3356 improves S-meter for HF transceiver (FT-902-DM)
HB9MIN describes in DUBUS nr 1/92 the use of MC3356 to improve
the S-meter for IC202E 2m ssb transceiver. For this transceiver he
achieved 50dB dynamic range. Since there is few if any good VHF
transceivers I decided to use my HF transceiver with transverters for the
different VHF/UHF band on cw/ssb modes. I have tested this circuit in my
FT-902 and achieved 70dB S-meter range. IF is ca. 9MHz.
The most suitable room for the IC and associated components is at the room
for additional AM filter. It was fairly easy to mount the components. I
used a 0.5mm tinned iron plate and soldered the groundings to this plate.
The components were premounted and tested and this was installed at last.
It is important that no stage ahead of this board has AGC control, co I
removed AGC to the RF amplifier and connected it directly to RF gain
control. It is now possible to reduce RF gain, and RF gain control may be
calibrated as such, range at least 30dB.
An amplifier is used ahead of the MC3356 to match the IF level and
impedance of the connection point of receiver. The circuit is fed from
IF which has passed the selected filter on the board (CW or SSB).
It was found important to adjust the output signal from this amplifier, and
with a BF905 quite little gain was needed, it is however important that
current in the transistor is not too low, then maximum deflection is not
possible! If gain is too high the S-meter will read with only background
noise. So a compromise must be found, transistor gain must be adjusted such
that S-meter just starts to deflect.
With no signal input at receiver the dc voltage from MC3356 detector is
0.6V. I believe HB9MIN did a fault in order to removed the false
deflection, and it is very important to remove this 0.6V the correct way!
The use of a series diode is not good at all!!! Remember that voltage drop
across at diode at 10uA is about 350mV, increasing logarithmic to about
0.55-0.6V at 1mA, so at bottom end of S-meter deflection two diodes are
necessary to obtain zero reading, while the two diodes will have about a
volt drop at maximum deflection. Better is to use an emitter follower,
because emitter-base voltage drop is fairly constant compared to a simple
I used the circuit for about a year now, the only problem is to find a
proper S-meter time constant, it must be fast 'attack' and slow 'decay',
just like AGC performance. An improved circuit is shown in VHF/UHF dx book
(G3SEK), called peak detector, the circuit uses an LM358 (half LM324), this
is made at a time constant of 0.5-1second. I have used this additional
circuit with several power meters tested last month, and they have rather
AGC is left untouched for the IF. It runs it own life, while S-meter is
derived from MC3356 - with additional transistor (level correction) or
LM358 (level correction and peak detector).
The circuit has 70dB range over the noise. (93-11-07)
Circuit is described in Amatýr Radio (NRRL) nr 12/92 page 5.
circuit diagram is available if you send SASE to address below or as a part
of FT-902 modifications
convert FT-901 front-end
I saw a message from you titled "FT-902: Improving
RX-MIXER.....". I own both a FT-901 and a FT-902. Some years ago
I changed RF unit of the FT-901 (PB-1702) by a FT-902 RF unit
(PB-2154). I can tell you it works fine! The only thing you
should be aware of is that the three connectors on top of the
module have changed position as follows:
position |FT-901 FT-902
top |Red Yellow
middle |Yellow Green } colour mark of
bottom |Green Red / the coaxial cable
If you make the right connections the PB-2154 works right away
in an FT-901. No further modifications necessary.
This counter shows the number of hits since the
9th February 2000
Go Back To The Yaesu Mods Page
Go Back To The Main Modifications Page
Go Back To The CB / HAM Radio Main Page
Go Back To The Main Home Page
Copyright © The Defpom 1997-2008