Revox G 36, Service Manual

Page: 18 / 52

The generator signal level should then be reduced by 20 db and the output level noted.
After changing the frequency to 10 kc/s, bias should be adjusted to an output identical
to that at 1 kc/s.
3.1.6 Measurement of overall frequency response
Before any attempt is made to verify the overall response, it is advisable to see that all
tests described in the preceding paragraphs have been properly observed and that the
settings arrived at have been left undisturbed.
A 1 kc/s signal should be recorded via the radio input socket at reference level, viz:
the same output voltage should appear on the cathode follower as from the reference
level of the calibration tape. The level of the signal should then be attenuated by 20
db and the “overall” response curve taken. It is essential for this test that the signal
level is some 25 db below peak record level (3 % dist) otherwise the record pre-
emphasis may lead to tape saturation at high frequencies thus producing false test
results.
Where the frequency response deviates from the specified tolerance of + 2 / - 3 db,
a slight re-adjustment of the bias current may be necessary. Any variation in bias has
a pronounced effect on the frequency response.
If a fault is suspected to lie in the record section of the recorder, the record head
should first be changed. If this does not prove to be the cause of the trouble the
response curves of the amplifier should be checked with the bias oscillator valve
removed.
To measure the response characteristic of the record pre-emphasis use a test circuit as
shown in fig. 12.
The frequency response obtained should coincide with the graph shown on the back
of the respective circuit diagram. Considerable deviations would indicate valve
aging, or faulty components in the frequency conscious networks of the record
amplifier.