|
ICOM IC-R75
Modifications
Well, I have purchased the general coverage receiver, ICOM
IC-R75, for my 50-year-old hobby.
This is a great product. It features a lot of useful functions as well as
marvelous sensitivity and selectivity. Nevertheless, it carries somewhat
disappointing problems. These glitches are sometimes easy to fix.
This page is for the challenge of modifications to pull better performance
out of this gear.
First off, the list of wishes:
1. Audio sound is less than satisfactory.
I do not mean all radios must have Hi-fi sound, but IC-R75
creates uncomfortable feeling in long listening.
2. S-AM does not work properly.
Many users complain about this problem. Thanks to Rado who wrote
(this) forum
article, we are able to fix the problem.
3. AM band has RF attenuator inserted.
Quick patch can improve to some extent.
4. Synchronous detection is not for SSB.
I was disappointed about missing Exalted Carrier Single Sideband capability.
S-detection is only for AM. You cannot select upper or lower sideband to
eliminate unwanted signal, or distortion caused by phase shift. We cannot
do anything to modify.
5. Sleep timer makes me wake up.
This timer is a joke. Why does it beep five times? I want to sleep.
This "feature" cannot be removed unless I am provided with the firmware code.
6. Pre-amp for long wave band does not work.
Pre-Amp 1 and 2 does not work when you receive LW band, especially below
60kHz.
As I do not use this band, I will not implement the fix, but will suggest
possible cause in this page.
7. AGC and S-meter
S-meter does not work when you turn off the AGC. I traced the circuit and
found that the S-meter signal came through a buffer amp, which was fed from
AGC voltage. The AGC generator was controlled by a logic signal that came
from a shift register data fed from MPU. So, this is out of consideration.
8. SSB sound level become low when filter is set to 3.3kHz and AGC is on
Because of wider bandwidth, AGC input power grows making IF gain lower. However,
relative sideband spectrum level at input does not change. Net result is
lower audio level. You can confirm this by turning RF gain lower. Audio
level will not change by bandwidth SW if RF gain is clipped. This is a
basic design nature, so, I must accept it.
9. Strong signal causes heavy audio distortion
Heavy noise by audio saturation takes place when S9+20dB or stronger signal
is received. This is because level setting among AF stages is inadequate. This
is the most annoying problem.
10. Design flaw in AF filter circuit
The AF filter assumes zero signal source impedance, but it is actually very
large. See the analysis below. The same flaw exists at the FM hi-pass
filter, but I will not modify it as it is not significant.
11. Strong AM signal has distortion when AGC is turned off
When you turn off the AGC, strong AM signal gets distorted. This is because
an emitter follower placed between IF amp and AM detector is saturated.
12. Miscellaneous
Battery life, power pack voltage, and missing noise suppression capacitor on
rectifier are discussed.
Modifications:
1. Sound improvement
I found that there were two things to be done. One was the speaker
replacement. I am not in success to find a new one yet. The second thing
was to modify low-pass filter parameter as Kiwa suggests.
1.1 Speaker
The built-in speaker is a sort of toy (see the right-upper
photo). 4cm diaphragm of flat shape can never produce good sound. ICOM
designer must choose better one at the beginning. The stamp on magnet says
8 ohm 2 W, but the heat mass of the voice coil can never deal with this
power continuously, nor the travel stroke cannot be such large. You need to
replace anyway.
I surveyed in several parts shops in vain. The photo on the right is 4cm
diaphragm one and has good sound, but the dimension is about 1.5cm too deep.
Another one of 3cm speaker on
==> this page looks little better than the original, but low frequency
response is not sufficient. (I heard it at a shop.)
So, this modification is yet to be done on my machine.
Any external box speaker with larger than 8cm diaphragm produces comfortable
sound by IC-R75.
［Edited 3/10: I bought a speaker
attachment for iPod at a nearby shop. The speaker unit in it was a 4cm
diameter one, which fit to the R75 box. Picture on the right is the one just
pulled out of attachment. The magnet used on the speaker was rather small
than I hoped, but not so bad.
I measured the f0 and found it was 180Hz. The original speaker of R75 had f0
of 670Hz, which was very poor. We need 300-600Hz because it is an important
#1 formant frequency band for vowel. The sound became very clear after the
replacement and the filter circuit change written below.
1.2 AF filter
There is a 24dB/oct Butterworth low-pass filter placed after detection and
before power amp. The cut-off is set at 3kHz.
3k-4kHz is important frequency to listen clear sound of consonant, such as "sh"and
"ts". IF filters of CW and SSB are satisfactory for selectivity even if
this audio filter is modified to higher frequency. On the other hand, this
3kHz filter disturbs when you receive AM signals which comes through
6kHz/15kHz IF filter for AM listeners. So, I modify this to have higher
cut-off frequency, say 4.4-4.5kHz.
Four ceramic capacitors are to be replaced as this picture tells. Do not
use ferroelectric ceramic capacitors (BaTiO3) as they have non-linear D-E
curve, that in turn cause audio distortion when used in filter circuits. They
are OK for decoupling purpose as far as no voltage change is expected across
the terminal. Use film capacitors instead. Refer to
my webpage for further
information. (The page is in Japanese, but you can still see distorted
waveform.)
If you feel difficulty to replace capacitors, replacing the whole filter
circuit is an alternative. The signal you take as input can be stolen from
the lower side of the leftmost resistor or lower side of the second resistor
from the left in the chart. Output should go to the cable (not connector
land) from pin 1 (closer pin to Li Batt) of J1221.
If you plan to install a separate filter circuit, consider Chebyshev filter
that is written in my ==> other
page. See the last part of the page.
1.3 Speaker mounting
The front panel had very narrow slit of sound channel. I made it two-three
times wider with router grinder. You can see it on the speaker photo.
2. S-AM demodulator
The fix is two folds. One is to extend the AGC period in order to adopt to
fading signals that is typical in short wave. The second one is to have dual
AGC speed for tuning and after tuning. You can find detail description in
Rado's article in Yahoo forum, which is linked below.
The AGC time can be extended by adding 100-470uF to the pin 4 of MC13022A. The
28pin package IC is located between Li battery and front panel.
I added 220uF to the original 10uF.
The dual speed can be achieved by grounding pin 23 through
22-47uF. This pin is originally grounded directly, which is for fast tune
mode during station change. Although the internal circuit pulls this
terminal down by itself when in tuning action, I added external circuit for
faster action. 2SK679 discharges 22uF when PST1 signal comes in. PST1 is a
write strobe signal to PLL frequency setting register. As the pulse width
of 1.5us is not long enough for the discharge, I have added pulse extender
by 0.001u and 47k.
The most difficult work is to lift off the pin 23. The PC board effectively
spread heat of solder iron, so work very carefully. Do not damage the pin.
Pin 23 is named "Blend". Detail function and signal level of the pin is
described in data sheet of MC13122 that is a sister product of MC13022A.
Picture shows my rework of filter frequency change and two fixes around the
13022A. Components are glued so that they do not vibrate.
[Edited 2/10/2007
Above circuit is changed again to
this diagram.
Rado 2.0 uses 100u instead of 22u in the chart, and 100k-150k instead of
330k. ]
3. MW RF gain
Apply solder on the shorting pads near antenna relays.
This patch will improve AM gain by 1dB or so.
Note that strong AM station near your home may cause intermodulation.
As a matter of fact, above patch changes load impedance of the filter,
making the characteristic slightly deteriorated. So, correct modification
is as follows:
- solder the shorting pads
- remove two 270 ohm resistors
- add 820 uH choke coil at one of the 270 ohm resistors location
- remove 10 ohm and replace with 100 ohm
I made this change, which improves the gain by 3dB.
I measured the self resonant frequency of the 820uH. It was 2.5MHz, high
enough. The series resonance with decoupling of 0.1uF is 17kHz, low enough.
The ceramic decoupling capacitor has V-C and Temp-C characteristics. Those
two characteristics reduce capacitance in a real environment but the
frequency will not reach to 30kHz. So, 820uH is an adequate value. (On the
other hand, filter input has 330uH choke coil. The reactance is 62 ohm at
30kHz, which is too small.)
6. Pre-amp fix
Change several coupling/bypass capacitors of 0.1uF to 1uF. Addition in
parallel is acceptable. The reactance of 0.1uF at 30 kHz is 58 ohm that is
not acceptable at low impedance node. I believe that the coupling/bypass
capacitor is one of the causes of gain loss.
Capacitors indicated by red circles are subjective ones.
Transmission line transformers at input and output are also suspicious for
sufficient bandwidth..
However, I do not install this change because LW is not my hobby, and because
they are hard to access in shield case.
By the way, I added 0.33u chip capacitor in parallel to the 0.1u at the
right most position in the chart. This must be a chip type if you will add,
otherwise the lead inductance will have parallel resonance with the 0.1u
cap.
9. Distortion with strong signal
Strong signal reception causes very heavy distortion. RF
attenuator saves this problem, but little modification in AF circuit will
easily correct the situation. I traced the signal and revealed the cause
instantly. Signal path is as follows:
AM/SSB/FM detector, Analog SW, Low pass filter, DSP, AF Amp, Electronic
volume control, Mute SW, Attenuator, Power Amp
The AF Amp is the culprit. The output swings up to 7V while Vcc is 8V,
causing upper side saturation. So, I made the gain reduction by 4.4dB at
this OP amp. Also, I gave 3.6dB attenuation at the switch for AM and S-AM
signals. That compensate imbalance between SSB and AM.
No heavy noise is observed after this modification.
10. AF filter design flaw
The AF filter is intended to have 3kHz Butterworth low-pass characteristic.
However, it can be achieved only when the signal source has zero impedance.
They are actually very large as you see on the chart. Assuming 100 ohm of
analog switch resistance, they are:
FM 2.2k
SSB 28.0k
AM 23.5k
I calculated the frequency curve that is affected by this higher impedance.
This narrower frequency response is another culprit of poor sound quality.
The best solution is to insert a buffer amp prior to the filter input (just
in front of the first 39kohm). Or implement the frequency change described
in 1.2 as a compromise, although the objective frequency response is
deteriorated. For further analysis data, see
==> another page. I
implemented the case-4 of the page.
11. Strong AM signal is destorted when AGC is turned
off
The emitter-follower buffer placed between IF-amp and AM detector cannot
deal with signals more than 7Vp-p. It is saturated by a strong signal
caused by no AGC.
I applied a modification as shown on the left. It improves the dynamic
range by 1.5 times, or about 4dB. Also, I adjusted the 220kohm so that the
emitter voltage bacame 6V. This biasing method is problematic because each
transistor has different HFE.
12. Miscellaneous
Somebody reported that the internal battery lasts only for a couple of
months. I measured the load current. Drain current is 4uA when no power is
applied at the connector. 220mAH can last about five years, so there should
be no problem. No change is needed in my case.
Add battery holder for CR-123A that has 1.3AH capacity, if you have the
problem of short life with CR2032.
Another strange design is the power supply.
Power pack voltage is 19.8V before power-on and 17.7V after power-on. The
ripple is 0.5Vp-p.
17.7V is off-spec, but it turned out to be harmless. The LDO
PQ30RV31allows 35V input and 20W power dissipation at 25degC. I do not see
excessive heat sink temperature near the regulator. Another regulator on
the line is AN78L05M, which also allows 35V input.
I will keep it at that.
Adding snubber capacitors to rectifiers may be necessary for lower noise
level, although there is a noise suppressor across AC input to the bridge.
Refer to my webpage.
However, I feel this is not needed in my case, as the noise level is very
low without modification.
I have measured frequency stability with original Xtal and
CR-282. With original Xtal, the 60MHz reference frequency drifts about
-0.092ppm/degC between 20 and 40 degrees C. The temperature is measured at
the shield case around the Xtal. The curve is not linear, and becomes less
sensitive at higher temperature.
The high stability CR-282 has much better figure, about one seventh of the
original one. The graph on right side shows both of them relative to
power-on time.
The initial transient of CR-282 suggests that it has heating mechanism
within the case.
If you want to see relationship between LO2 drift and receiver frequency
drift, visit ==> this page.
Opening the case cover, you will find large vacant space underneath that can
house additional functions if you prefer.
All in all, I like this receiver very much. It is worth $800 (w/DSP and
3.3k Filter) cost.
Links:
IC-R75 Commands
http://www.plicht.de/ekki/civ/civ-p0a.html
Remote control program, HAM RADIO DELUXE
http://hrd.ham-radio.ch/
Kiwa
http://www.kiwa.com/R75.html
Forums
http://www.eham.net/reviews/detail/427
http://groups.yahoo.com/group/icomr75/messages/
Rado's article in Yahoo forum
http://groups.yahoo.com/group/icomr75/message/9 for ver. 1
and follow up articles of ver. 1.1, 2.0, and 2.1.
MC13022A
As to the pin23 information, please refer to
MC13122 spec
ENJOY
|
