mostly the same but one is modulated by positive currents from the driver, the other negative
currents. Since the driver is common to both of them and its own output is either positive or
negative, only one of the transconductance stages is active at any one time.
I have tried to draw the schematic so you can see the two complementary halves of the core
easily. Each one appears identical but for the type of transistors used. Where one uses a NPN,
the other will use a PNP, and vice versa. Note too, that the X input is applied to different sides
of the transconductance stages resulting in one of them being of opposite phase to the other.
The left hand side deals with Y modulation signals that are positive in value. The right hand
side deals with Y modulations that are negative in value.
Each transconductance amplifier is made from five transistors. Four are two sets of matched
pairs. A matched pair is exactly that; two devices that share with almost identical
characteristics. This cannot usually be obtained from simply buying two transistors of the same
type, since in manufacturing there are small discrepancies that make every device slightly
unique. In the original 4014 module, ARP would have hand selected their transistors by trying
hundreds out and grouping them together into large bins with contained those devices with
similar characteristics. Each pair would then be made by selecting two devices from the same
bin. This would take too much time when you only want to build one or two Oakley Ring
Modulators, so I have used pre-matched quad transistor arrays. The THAT300 is four
matched NPN transistors, and the THAT320 is four matched PNP transistors. The electronic
matching is near perfect and they are also made onto the same silicon die and in the same
housing thus ensuring a good thermal matching too.
The voltage at the top of each transconductance stage is fixed by diodes to be 1.2V for the left
hand side one and -1.2V for the right hand side. C4 and C17 provide some additional
decoupling to reduce any noise or audio signal bleedthrough at these points. C4 and C17 were
not in the original design, but I have added them here to reduce any chance of extraneous
noise pick up.
The output of both transconductance stages go to the inverting input of the output op-amp,
U2, pins 1, 2 and 3. This is wired as an inverting transimpedance amplifier, which is a 'current
in, voltage out' device and essentially the opposite of a transconductance amplifier. A
transconductance amplifier followed by a transimpedance amplifier make an amplifier, ie.
'voltage in, voltage out'.
To make sure the overall gain of the ring modulator is positive for Y being positive, we use
the inverting outputs from each transconductance amplifier. These are found on the collector
of the left hand transistor of the input pair, ie. U14, pin 14 and U1, pin 7.
The gain of the transconductance amplifiers is controlled by the driver circuit. The output of
the driver circuit is converted to a current by R18 which in turn affects the current through the
input pairs and hence the amplitude of the current output. However, the response of gain
versus input current is not linear when the output of U2 is less than 0.6V for the left hand
amplifier and above 0.6V for the right hand one. This is because to control the gain of the
transconductance amplifiers the voltage at the bottom of R18 must be enough to overcome the
inherent base-emitter voltage and turn on either of the input transistor pairs. This would give
rise to a dead zone of operation if it were not for the deliberate non linearity introduced in the
driver circuit.
8
