I bought a used HW-9 that works well except for dial accuracy. I'd
appreciate comments from other HW-9 owners with regard to accuracy of
theirs.

So far, I've found the 80 Meter heterodyne oscillator crystal is about 5
KHz off frequency. All the others are within about 1/2 KHz, which is
not significant compared to the width of the lines on the VFO dial.

Except for the 80 meter crystal problem, it appears that frequency
errors are almost entirely due to VFO tracking error. VFO frequency is
correct at 0 and 250, but as much as 10 KHz off in the middle of the
range.

One of the VFO capacitor rotor plates is bent away from an outer stator
plate. That might be the result of an earlier owner attempting to
correct the problem, or it might have come that way from Heath. Can
anyone tell me if the stock VFO capacitor plates were originally
straight (flat) or bent?

The VFO Shield is missing. If I can find some copper, I'll try to make
one, but doubt that will correct the problem.

Fred
K4DII

Hi Fred:

Many receivers capacitors have the outside plates separated into sections,
like a pie. This is for the purpose of calibrating the receiver to the
dial. But carefully moving the plates, excellent calibration can be
achieved.

It requires a bit of logic and patience. Find out the frequency range of
the vfo (not the output frequency). The simples approach is to put a
frequency counter on the output. Then, note the calibrated frequency to the
actual frequency. Let is say at the 100 kHz range, the dial reads 100, but
the actual frequency is 103. That means the vfo is 3 kHz high, which means
not sufficient capacitance at that point. I would draw a correction graph
for the entire range.

You know that at 100, you need more capacitance, bend the plate in a bit at
that point where the rotor is just meshing with the stator. A little bit at
a time.

Note that not all capacitor plates are designed to be bent like that.
Generally, the high precision variable capacitors are not designed to be
adjusted by moving the plates.

It is 1/2 science and 1/2 art.

Once you get the hang of it, you can use a small file and file a little
notch off at any point to make fine corrections. Warning, do not do this
without thought first, or you will make things irreversibly worse and we
will see it on eBay.

73, Colin K7FM

In article ,
"COLIN LAMB" wrote:

Warning, do not do this
without thought first, or you will make things irreversibly worse and we
will see it on eBay.

Colin-

"I cut it off twice and it's still too short!"

Thanks for your thoughts. It is difficult to see whether the bent plate
has sections or not. The possibility of damage is the main reason I
haven't already tried straightening it to see if there was any
improvement.

This VFO works backwards. Zero on the dial is the high frequency end,
and 250 is the low frequency end. I've gone through the alignment
procedure so everything is right at the dial edges. Unfortunately the
WARC bands occupy the central portion of the dial, so it would be easy
to transmit outside the low end of each band.

There is an article in an ARRL QRP handbook, that mentions the HW-9
having poor dial accuracy. I'm at the point where I'll probably go
ahead and carefully straighten the capacitor plate to see if it makes an
improvement. If not, I'll then take the approach you suggested to
re-bend the plate.

73,
Fred
K4DII

On Fri, 1 May 2009, Fred McKenzie wrote:


There is an article in an ARRL QRP handbook, that mentions the HW-9
having poor dial accuracy. I'm at the point where I'll probably go
ahead and carefully straighten the capacitor plate to see if it makes an
improvement. If not, I'll then take the approach you suggested to
re-bend the plate.

Remember, it was a relatively low cost rig. The money went into a fuller
design than the HW-7 with its direct conversion receiver (that nobody
seemed to like, since there were so many attempts at improving the
detector section). It was no SB-101 or HW-101, that from the start
was expected to have linear frequency readout. After a certain time,
people got used to that, but it wasn't the norm before that. A lot of
rigs didn't have particularly good dials anyway.

Michael VE2BVW

In article ,
Fred McKenzie wrote:

I'll probably go
ahead and carefully straighten the capacitor plate to see if it makes an
improvement.

I now believe the bent plate was an effort to correct another problem.
Talk about a can of worms!

With the plate straightened, it was not possible to calibrate the dial
ends within the range of adjustment of the capacitor's trimmer.

Then I noticed that the capacitor plates had not been lined up with two
"guide holes" on its frame. I did that, and things got worse.

Now I find that frequency peaks at around 50 KHz on the dial.
Apparently it is going out of oscillation, or into some other mode of
oscillation, when it nears the end of rotation.

The capacitor plates are NOT shorting. The inductor's slug is in the
correct (lower) half of the coil. The next step is to verify that the
right capacitors are in the feedback voltage divider.

Another thing I may try is to replace the VFO transistor. The original
is an MPF-105. I have a couple of spares, but perhaps a higher gain
device would be better.

Fred
K4DII

Hi Fred:

It sounds like you have become one of us, and you have one of those rigs
that things get worse and worse, instead of better and better, the more you
work on the problem.

One common alignment technique is to align somewhat inside of each dial end.
That will often give you a better overall alignment. Heath may not have
specified that technique, in an attempt to keep you in the band. But, you
should not trust the dial calibration when you get near band edge (and what
used to be a pink ticket).

Oscillators go out of oscillation for a reason. It can be that the
frequency limit of the transistor (or other component) has been reached.
That usually does not apply to a solid state circuit that has been working
fine for decades, although it can happen with tubes. It can be a change in
feedback. If it occurs just after you made some mechanical changes in the
capacitor (or any other component), take a hard look at that component. In
the case of the variable capacitor, the capacitor depends upon one or more
wipers making good contact. If they do not make a good contact, the
capacitor will not be doing its job. Oscillations can either change
frequency or stop.

The good news is that when you are able to correct the problem, there is a
great deal of satisfaction. Sometimes, you can do it better than the
original design team (which included the bean counters, of course).

73, Colin K7FM

In article ,
"COLIN LAMB" wrote:

One common alignment technique is to align somewhat inside of each dial end.
That will often give you a better overall alignment.

Colin-

I hadn't considered exactly that. I did try aligning at the 50 and 200
KHz points, but the trimmer capacitor got too loose. I might try
alignment just outside the dial ends, perhaps aligning zero for 10 KHz
and 250 for 240 KHz.

My problem of the oscillator stopping may have been a case of
overdriving the counter. Trying a different counter with an attenuator,
seems to have cleared up that problem. It is a good thing I tried that
before changing any parts.

Now, with the capacitor plate straightened and VFO aligned at zero and
250, the middle of the dial is off about 9 KHz. This may be a KHz
better than before. Maybe the capacitor plate was bent in the wrong
direction? I think I'll just leave well enough alone!

73, Fred K4DII