Hi!

Something I've never seen anybody referring to is the striking
similarity between a Sony ICF2010 and a Grundig Satellit 700. Inside
that is. I personally own samples of both radios and also have the
schematics of both radios.

The Satellit 700 is more or less an upgraded ICF2010 (ICF2001D).
The synthesizer is using the same chips in a similar configuration
(Sony CX7961 and NEC uPC1037 chips two of both). The AM/FM-IF chip is
the same (Sanyo LA1205) as is the synchronous demodulator (Sony CX857).

The major differences of a SAT700 compared to ICF2010 are the inclusion
of RDS on FM and the use of switched varactor tuned filters on SW.
Otherwise the receivers look almost identical to an engineers eye.

It seems it would have been very unlikely for Grundig to be able to
source the Sony chips without a mutual agreement on the design ie.
Grundig must have bought the basic design from Sony! In the 60's and
70's the japanese very accused of ripping european and US designs.
Later on it seems that the roles did change a bit!

Coming back to changes on the original design Grundig did produce
a few bugs. The first one is quite often referred to as rumbling
synchronous AM. The main reason for this is that Grundig did mess
with the time base of the SAM-PLL. In SAT700 the LP-filter has a
56k resistor and a total of 11nF capasitance. In the ICF2010 the
same components are 10k and 11nF. Correcting the Grundig circuit
to comply with the Sony original does away with most of the rumble.
At it's easiest you just replace the 10nF of "CC817" with a 1nF capasitor.

The second bug is rather serious as well. The varactor tuned SW-
frontend filters are switched with PIN-diodes. In the Grundig
design the diodes of the filter in use have low forward bias.
The diodes of the remaining filters have their cathodes and anodes
at the same voltage ie. they have zero bias. For best intermodulation
performance one would reverse bias the non-conducting diodes. In my set
I reverse biased the diodes and the improvement in large signal handling
was remarkable.

What you need to do is to disconnect the resistors CR228,233,238,239,
245 and 246 from +AM+5V rail and connect them somewhere higher. In my
set I used the +30V tuning voltage generator and by a high impedance
voltage divisor produced +10V that I connected the resistors in to.
This resulted in a few volts of reverse bias for the diodes.
(Actually in the easiest implementation you remove the aforementioned
resistors and add new ones on vias on the PCB near the connector of the
cable going to the CPU-board...)

-Vesa-