[HBR] HBR2K -- Chapter 14 -- Large Signal Performance, Part 6

[hbr@mailman.qth.net]

A couple weeks back, Jim, N2EY commented:
> I'm beginning to wonder if the performance of the HBR2K is actually much
> better than the test setup permits to be shown?

I looked at the design of the URM-25, found much good, and went on 
at length about it, but the comment stayed in my thoughts.   The 
other day, I thought of an issue I had not previously recognized, 
namely, second harmonic output from the generator.

Why does the second harmonic matter?   Well, receiver 3rd order 
IMD is measured by feeding it two signals 20 kcs apart.   The two 
signals are presumed 'pure'; the non-linearity of the receiver 
generates the second harmonic of each and mixes it with the other 
fundamental to give new signals 20 kcs higher than the higher input 
and lower than the lower one.   The ratio of these 20 kcs +/- signals 
to the input signals is a simple test of a receiver's ability to hear a 
'weak one' without being bothered by strong signals elsewhere on the 
band.

But what happens to the test if the signal generator isn't pure?  In 
particular, what if it supplies some 2nd harmonic energy?   Obviously 
the receiver performance 'looks' worse than it is.  I can't decide 
whether making measurements at various input levels would allow 
determining that the measurement was in error but I'm inclined to 
think not.   The distortion signals we're watching for (20 kcs above 
higher test signal and below lower) are actually generated in the 
receiver, whether the 2nd harmonic is generated there or comes from 
the sig gen.   Thus It seems to me that they would behave just like 
receiver distortion caused by overload from pure inputs and if so, Jim 
would be right ...

When measuring at the antenna, the receiver itself might provide 
useful attenuation of the sig gen's second harmonic.  But one tuned 
circuit isn't a lot -- maybe 20 db?  And most of the measurements I'm 
making now don't include a matched tuned circuit so there's no help 
at all.

I measured the 2nd harmonic output of one of the URM-25's.  Gee ... 
only 30-some db below the fundamental.   That seems to me to be  
really, really crummy if you want to measure dynamic ranges in the 
80-db area.   And the URM-25 design -- so good in other ways -- 
can't be helped in this department because everything is single 
ended and the design is optimized for wide frequency range rather 
than excellent linearity.   For example, the only tuned circuit is the 
oscillator tank itself.

So I spend the last week building a test oscillator.   Using the 
mechanicals and power supply from a GPM-15 LORAN test set I 
built up a push-pull Colpitts oscillator with a 6ES8 dual triode and 
followed it with a 6KE8 with the triode as a cathode follower and the 
pentode as a class 'A' linear amp.   The amp is cathode driven with 
negative feedback to the grid to set the gain and improve linearity.   

The oscillator covers 3-6 megacycles which includes all frequencies 
needed to test the HBR2K.   (Guess the source of the tuning cap 
and dial!)  The output stage is also tuned, to provide harmonic 
attenuation.   At present the 2nd harmonic is about 80 db below the 
fundamental, which should be good enough.

('About 80 db below ...' -- I have some leakage back through the line 
and some radiation so I cannot attenuate the fundamental more than 
about 80 db.   I know about some problems that could be causing 
the leakage, so in the next day or so I should have a better 
measurement.)

The project isn't quite done.  I have an amplifier output of about 5 
volts RMS but want 10 volts so a built-in 40 db attenuator can be 
used to get down to 100,000 uV at 50 ohms.   That 40 db of 
attenuation will give isolation from the other generator in addition to 
that from the hybrid combiner, helping to assure against non-linearity 
caused by the other signal; the URM-25 has only 20 db of internal 
isolation.

10 volts across 5000 ohms ought to be easy -- 20 mW, right?   But 
the stage has to operate well within the class 'A' range to minimize 
2nd harmonic generation.   So far it looks like the 6KE8 (basically a 
VHF TV oscillator/mixer tube) will only deliver about 5 volts unless 
the drive is raised enough to cause much more harmonic output.

I have some 6LQ8's on order; the pentode side of these is designed 
for video output service and has almost twice the gm and about half 
the plate resistance of the 6KE8.   I hope this tube will easily hit the 
10 volt mark If the 6KE8 won't make it.

Then I have to build a second generator.   And then the tests on the 
HBR2K can be repeated to see if the so-so IMFD showing is really 
the receiver or not.  Three weeks of preliminaries, 10 minutes of 
measurement.  Thanks Jim.   Really.  :)   

Walt Hutchens 
KJ4KV