[HBR] One more quick question - HR10

[Walt Hutchens]

Peter wondered:

> Since the HR-10 frequency doubles the LO on the upper bands (one reason it
> sucked so much)  would it possible to preserve the dial calibration when
> using plug in coils?

As I recall the HR-10 doubles on 15-10 but the HBRs do it on 20-15-10 so
there is some difference in the pattern.

When you think about how to tune a certain range -- say 3.5-4.0 Mcs or
21.0-21.150 Mcs (plus or minus the IF), the main thing that matters is the
capacitors involved: Specifically the fixed caps and the variable -- plus of
course the shape of the plates on the variable.   But using the variable
that came with the dial assembly you shouldn't have an issue with plate
shape.   

Roughly speaking, the coil is just 'whatever it takes to get the right
tuning range with the capacitors you have.'   So tuning 22.681-22.831 Mcs
(oscillator tuning range with 1681 kcs IF) is the same (requires the same
capacitors) as half that -- 11.3405-11.4155 Mcs: the COILS give you the
different ranges.  

However 'roughly' isn't the whole story.   The distributed capacitance in
the coil and elsewhere in the circuit IS NOT equivalent to a different fixed
cap across the coil (hence correctable by changing the fixed cap in your
design).   It is 'distributed' -- here and there inside the coil and around
the circuit.   Furthermore, things like transit time (electrons in the tube
...) affect the frequency and do so in different ways at different
frequencies.   

Additionally, the HR-10 used a triode oscillator while my HBR-ization of
that uses a pentode ECO: That introduces yet another frequency-dependent
variation.   (Using the exact same oscillator circuit might have saved some
bathwater -- slightly better dial tracking -- but tossed the baby -- better
oscillator isolation, hence, stability.   At least, that is my thinking!)

So an oscillator for half the frequency isn't the same fixed and variable
caps with 4x the inductance as the simple theory would say.   And that means
that it won't have precisely the same tuning curve, either.

HOWEVER ... the differences aren't large.   By tweaking both the coil and
fixed capacitance away from what the simple theory says the errors can be
perfectly adjusted out at the end points and the narrower the band, the less
error there'll be at mid band.   And of course 80M is the widest band,
percentagewise.   All the higher bands are quite a bit narrower:  10M seems
wide -- 2 Mcs -- but it's at 8x the frequency so it's only half the width of
80M.   

With the coils I've made (80 and 20 on the Long SLOW set and 80 for the
HR-10 to HBR) I have had NO tracking trouble caused by the factors above.

For the Long SLOW project I tried both half frequency and full frequency 20M
oscillator coils: I had no trouble getting both to track the existing dial
calibration on the Eddystone dial that came with the tuning cap I am using.

The HR-10 dial isn't quite the equal of the Eddystone unit: The scale is
shorter and string drive isn't equal to anti-backlash gear drive.  I think
errors of 2 kcs or less are probably what you can achieve on the lower
bands; somewhat more on the high ones.  (These numbers would be double those
for an Eddystone unit.)  In addition the HR-10 has a CAL control that might
tune 50kcs on 80:  These sets are plenty good for general ham use (and way,
WAY better than  you could do with a National MCN/ACN type dial) but they
simply aren't frequency meters.

(I think it would be worth adding a 6:1 vernier drive to the CAL control:
Not only would that make this control easier to set precisely but it would
tend to hold it in place.   I will look at this change.)

Bottom line:  I don't think tracking a new plug-in-coil oscillator to the
dial calibration on an HR-10 will be a problem.  The errors are likely to be
in the 'noise,' that is, not easy to separate from the other small errors
coming from using a dial, drive, and circuit of this type.

Walt
KJ4KV