NOTES ON THE HW9

By Gary AB7MY




VFO DRIFT FIXES

I've found some more areas of the HW-9 that need addressing. Everybody
knows how badly the HW-9 drifts, so I have read everything that I could
find that was been printed about curing the drift. I now have it down to
at least tolerable levels, though some additional improvement is a
definite possibility once I get some more capacitors with other
temperature compensation values.

Note that most HW-9's drift upward in frequency, some as much as 3 Khz
over an hour or so. Changing a single cap in the VFO circuit can
counteract the drift, or at least it did on my rig. Receiver and
transmitter frequency drifts upward,while the actual VFO freq. is
downward.

So to begin with, here are some things to try:

Remove the VFO inductor shield for L118 and paint the inside of it with
some flat black paint. Clean the paint off the solder tabs and outside of
it with a cloth soaked in paint solvent so it can be re-installed. Allow
the painted shield to dry thoroughly while you do the following steps.

Remove L118 - the VFO inductor - and check it to see if it is epoxy
coated. Some HW-9's had this inductor coated with epoxy and they are
reported to have the worst drift. Mine seemed to only be coated with RF
shellac, as it has that characteristic smell that used to be common with
all the old tube radio equipment inductors and chokes.

Whether or not it has epoxy or shellac coating, you can still try the
following measures to see if you can stabilize it a bit. Use a hair
blow-dryer, heat it up to a good temp but watch that you don't get it so
hot that the carboard form starts to turn color. My hair dryer would not
go that high in temperature, but you should be careful with the inductor
since you probably can't get another one. Once you have it good and hot,
set it aside  for 30 minutes or so to cool. Repeat this process at least
four times.

Next, look at the components inside the VFO coil shield area, and move
any caps and/or diode D118 so that no parts touch the coil shield when it
is replaced. 

Check your junkbox to see if you have a 33pf or 36 pf N150  to N220 disc
cap. Replace C184 with this value. I used a 36pf N220 in mine and the VFO
freq. drift has decreased to only 400 hz (downward in freq.) in the first
thirty minutes, stabilizing within a few hz after 30 minutes. Try another
N-value if you don't get good results with this recommendation. 

I noticed that by taking a plastic drinking straw, and blowing gently on
*only* NPO caps C186 and C188 there was a big decrease in VFO frequency.
So although they are NPO's - they don't seem to be very temperature
stable.

They are also outside of the VFO inductor shield, so they are exposed to
a different environment than the caps within the VFO inductor shield.
With the HW-9's cover on this might not pose such a problem, but I would
like to try some other NPO's for these caps or maybe some that have
negative temperature coefficients. Just be aware that these caps are a
potential problem for temp. stabilty since they are exposed and have a
large influence on the VFO freq. 

Compensating C184 inside the VFO inductor shield is not a cure-all, but
it helped a lot with my rig. I would recommend not adding dial or meter
lamps to the rig, since that would present a large source of heat that
would adversely affect the temp. stability.

T/R CIRCUIT CHANGES

There was a mod by S.W. McLellan, ND3P, to the T/R switching circuit that
appeared in QEX for October 1990 and was later reprinted in the ARRL's
"QRP Power" book. It is fairly involved, but is said to increase the
sensitivity on 12m and 10m. Plus, it also cures a problem of transmitted
RF getting into the receiver circuits, particularly the receiver R12
power source. At RF output levels just a little above 1/2 watt to 1 watt,
D407 starts to rectify some of the transmitted RF and places a negative
potential on the positive R12 line that goes toward ground potential on
transmit and should stay close to 0.4 or 0.5 volts. The rectified RF
drives the R12 line more negative and upsets the base - emitter junction
of Q103 in the VFO RIT circuit. This is a bad thing.

If you listen to the transmitted signal of the HW-9 on another receiver
while you vary the rig's output with the CW level control, you will
detect a big change in the transmitted frequency. Not only that, the RIT
control that is supposed to be deactivated during transmit has a
considerable effect on the transmitted frequency! Because of this, your
TX offset that is supposed to be determined *only* by the BFO being
pulled 700 hz lower in frequency won't be correct at all. And this
problem varies from band-to-band with different power levels. What a pain
this is!

In addition, the TX return adjustment provided by R131 is badly upset by
the negative R12 voltage applied to Q103's base. If you put your digital
voltmeter on Q103's emitter, you will see what I'm talking about. The
rig's frequency wanders all over the place with CW level adjustments and
RIT position during transmit. Here's how to fix it. One way is simple,
the other is not.

You can elect to do the full T/R mod as suggested  by ND3P in QEX or the
reprint in QRP Power. This requires removing the T/R printed circuit
board and adding some genuine PIN diodes, a transistor, and changing a
few additional parts. It is pretty difficult to do, but the reward is
said to be improvement sensivity on 10m and 12m. I decided to pass on it
because of the complexity involved.

There is simpler fix I have come up with. Clip and lift one end of R132,
a 10K 1/4 watt resistor on the OSC printed circuit board. Try to leave a
little bit of lead between the resistor lead and its body, and the
clipped end at the PCB. Solder a 1N4148 or 1N914 silicon diode between
the clipped leads, with the cathode (banded) end facing toward Q103's
base lead. You're done! Now if you check the freq. while changing the
transmitted output level with the CW Level control, the transmitted
frequency is rock stable. Similarly, the RIT control during transmitt
will have *no* effect on the transmitter's frequency at all. Success!

Note that the TX Return adjustment with trimmer R131 is stable now, and
it does not change the voltage at the measurement end of R127 when you
set it with your DVM and vary the CW Level control. Put a 5.6K or 6.2K
resistor across R131 if you want it to be less critical to adjust.

The addition of the diode in Q103's base lead prevents the R12 voltage
that goes increasingly negative during  higher RF output levels from
upsetting Q103 and varying the voltage on RIT diode D118. The rig sure is
stable now, at any power level and on all bands. No doubt this fix makes
the rig sound better on the air, and when you change the RIT control
during a QSO, your transmitted frequency will stay put. 

BFO TX OFFSET ADJUSTMENTS

With the transmitted frequency now stable, regardless of power level or
RIT position, you might want to check the TX offset frequency to see if
it is 700 hz as it should be. Using a freq. counter or separate receiver,
check the freq. at TP104 on the OSC board. Set it to 8.831400 Mhz with
L135 through the hole in the metal shield. 

With the CW Level turned down to prevent the PA transistors from
overheating, check to see if the freq. changes to 8.830700 Mhz during
transmit. Or, if you are using another transceiver to check the
frequency, you'll need to allow a small amount of transmitter output in
order to hear the HW-9's signal while using a dummy load. See if the
transceiver you use for calibration has a 700hz TX offset and adjust it
if necessary.  My TS-570D tracks its CW sidetone with the TX offset
freq., so it is easy to match the HW-9's received audio frequency with
the TS-570D's sidetone and then listen to the HW-9's signal on the
TS-570D for the same audio note. No re-tuning of either rig's tuning knob
should be needed when the BFO (TX) offset is correct.

If the TX offset is not close to 700 hz, you'll need to change capacitor
C205 under the BFO shield. I used IC socket pins and selected a 75 pf
silver mica cap from seven or eight I had on hand until I got a 717 hz TX
offset. You could also use a combination of a smaller fixed cap of 47pf
or so, and a tiny trimmer of 6-30 pf  if you want to set it to exactly
700 hz. 

MORE ABOUT THE T/R CIRCUIT

While I was investigating the problem of poor RF isolation in the T/R
switching and the effect it has on the R12 control voltage, I noted some
things worth mentioning. Diode D407 is under a great deal of stress. On
my rig, the output on 80m can be as high as 8 or 9 watts if you aren't
careful with the CW level control position. This much RF voltage can
possibly destroy the 1N4149 diode used at D407, since it is essentially
across the PA's at T403 right before the output filter network. If D407
fails (and it did fail once in my rig), it shunts all of the PA output
and the PA transistors are in danger of destruction. If this happens, you
will hear the power supply groan a loud buzz, and the output on the
relative power meter will go very low.  If this happens, stop
transmitting immediately and change D407!

I put some IC socket pins on the PCB where D407 goes, and tried some
diodes to see which ones allowed the best RF isolation and best receiver
sensitivity without modifying the bias ciruitry as Heath designed it. A
number of 1N4148's and 1N4149's (the same as Heath's pn. 56-56) were
tried. Some got pretty warm when the output level exceeded 5 watts.
Normally, you wouldn't exceed 5 watts in operation - but you might during
tune-up or on 40m and 80m where a lot of output is available. I suggest
you check your HW-9 by touching D407 after (not during, or you'll get a
small RF burn!) a few seconds of 5 watts or greater output. If D407 is
warm to the touch, you have a problem. You are also potentially losing
some RF on 10m and 12m that is needed. The risk to D407 is greatest on
80m, where the RF output is maximum.

I tried several diodes in place of D407, including 1N4007's and even some
1N5767 PIN diodes. The 1N4007's got very hot and leaked too much RF into
the receiver and the R12 control circuit. The 1N5767 PIN diode failed
quickly when the output went past 4 or 5 watts, since it is not biased
adequately with the given circuit. I finally found a 1N4149 or two that
remained cool, even at RF levels of up to 9 watts on 80m. No other
changes were made.

Rather that redesign the whole circuit or do the ND3P mod, I settled on a
hand-selected 1N4149 diode for D407. It leaks only a little RF during
transmit into the R12 control circuit, and the diode added to Q103's base
described earlier corrects that problem. The receiver sensitivity is
good, but if your HW-9 lacks sensitivity - look at possible low BFO
injection into the product detector. Low BFO output at the TX /OUT point
on the schematic will also limit the transmitted output level, as will
low VFO output. Check these areas, and the previous HW-9 mods I wrote
about concerning the transmitter stages.

You might want to put a voltmeter on your HW-9's R12 level and verify it
does not go dangerously negative during full transmitter output. That
could potentially harm some other components, although a reverse-biased
diode across R12 to ground would hold it to -0.6 volts - just in case.
That might occur at high SWR levels or if you happen to accidentally
transmit into an open load. Just keep in mind that excessive RF passed 
through D407 represents a loss of desired transmitter output, and
possible problems with other areas of the transceiver.

MOD THE MUTE CIRCUIT

You should definitely do the mod to the mute lead of U301 that ND3P
suggests, as this quiets another source of audio pops and prevents the
S-meter from incorrectly displaying AGC voltage during transmit. It is
simple, and easy to perform. See either the ARRL's book "QRP Power", or
QEX for October 1990.

SUMMARY

With the added VFO stability and lack of frequency shift under varying RF
output and RIT operation, I know you will enjoy your HW-9 much more. Much
credit also goes to the contributors of the HW-8 handbook, and other
articles found in QST, QEX, etc. Those sources are a good place to look
if you want additional information on improving your HW-9.

72,

AB7MY
Gary Surrency
Chandler, AZ  (Near Phoenix), QRP-L #571, AZ ScQRPions, ARRL VE


I am grateful to Gary for giving me permission to include these notes.

Click here to return to first page