GR-78 opinions, anyone? (long narrative)

[Robert P. Okas]

Hi Mark & Gang,

    Thanks for all of the replies. I picked up my GR-78 early this year
at a swap meet, and it had very weak reception. When the manual arrived,
I traced the problem down to the mosfet in the mixer stage. Curious it
was there rather than the RF stage. The NTE replacement for the 40673 is
an NTE222 and the unit came to life when the failed part was replaced.

    Of course, the NiCd battery was dead. I'm temporarily using a
regulated lab supply set to 9.8V. This radio draws significant current.
With the volume control turned down and no station tuned in, the current
draw is 41 mA on my unit. With a strong station (full scale S-meter) the
current drops to 37 mA (volume down). In the standby position, the current
drops to 28 mA. Increasing the supply voltage to a little over 10V
doesn't noticeably improve the sound quality.

    Once the radio was functional, I was disappointed by the audio. I ran
a quick response check of the audio amp with an audio generator and
discovered that there is very little output below 100 Hz and a broad peak
at 2.5 KHz. The crossover distortion was readily apparent to my ears and
on the scope. The amplifier is a sort of a poor man's op amp, and the
feedback caps set the response. Rather than try to fix the amp in situ, I
cobbled up a reasonable facsimile on a solderless breadboard. I spent
considerable time trying to wrench the original circuit into something a
bit more pleasant. Remember that old saying, "You can't get there from
here?"

    I tried modeling and tweaking the circuit in Spice, intuitive
component substitutions on the breadboard, etc. The goals I had set of
wider, flatter response and lower distortion couldn't be met simultaneously
without a significant redesign of circuit. That's when the notion of
substituting the whole stinkin' amp with an LM386 became attractive.
That's when I put it on the shelf until now...

    I had assembled a little utility amp based on the LM386 and thought
I'd try it out on the GR-78. The audio wasn't that much better! I've got
a distortion analyzer (HP) and the LM386 amp weighs in at around 0.2%,
just like the data sheet specifies.

    Following the signal upstream, I observed obvious waveform distortions
in the IF stages. The 2nd IF amp, Q403, is heavily loaded by the AM
detector. Lifting the anode removes the 10K load resistor and the load on
Q403. The audio and IF waveforms got better. Changing the 10K load
resistor to something higher, like 470K, would be the end of the story,
except that the volume pot is 5K. I haven't measured the impedance of the
ANL signal guantlet, but there is substantial signal attenuation there.

    The drain of the first IF amp, Q401, also shows significant
distortion, a sort of asymmetric compression. I was using an HP signal
generator modulated by 400 Hz when I noticed this. Injecting a 455 KHz
signal into G1 through a .05uF cap, the waveshape at the drain looks
reasonable at signal amplitudes up to about S2 on the meter. The generator
output is about -86dBm. At levels higher than this, the positive half of
the IF waveform shows a flattening while the negative half looks OK. With
the AM detector lifted, the same waveform appears on the collector of
Q403. Increasing the bias on G2 of Q403 from 1.5V to 2.5V doesn't seem to
improve the waveshape.

    OK, enough for now. More later...

73,
Bob - W3CD

-----------------------------------------------------------
This list is a public service of the City of Tempe, Arizona
-----------------------------------------------------------

Subscription control - http://www.tempe.gov/lists/control.asp?list=HEATH
To post - HEATH at LISTSERV.TEMPE.GOV
Archives - http://interactive.tempe.gov/archives/HEATH.html