The Curtis CEM3340 VCO chip, introduced in the 1980s, has been used in many synths and synth modules. The Alfa AS3340 a is more or less drop in replacement; AS3340A is an improved version. Coolaudio has the V3340 version. The latter two come in DIP and surface mount packages.

The CEM3345 and AS3345 are less often used variants with two additional pins.

Regarding the AS3340A, the datasheet says:

AS3340A/AS3345A are improved versions of AS3340/AS3345. AS3340A/AS3345A and have [sic] much more output frequency stabilty against VCC and VEE…. AS3340A/AS3345A benefits from improved symmetry of rising and falling edge of triangle signal (approx.. 50%).

The AS3340-HYB is a surface mount AS3345 on a DIP-16 size breakout board with additional components to minimize pulse width pitch shift in a module that is drop in compatible with the AS3340.

All of my experience, if you don’t count frying the CEM3340 the first and only time I tried to use one with incorrect power connections, is with the Alfa AS3340 and AS3340A chips.

Pulse width pitch shift

When wired as shown in the datasheet, these chips have the defect that a change in the pulse width control voltage will cause a slight change in the VCO frequency, sending it out of tune. A frequency ratio change from 90% to 10% pulse width of about 5% is typical.

A number of remedies have been proposed and tried:

• As mentioned above, the AS3340-HYB provides stable frequency in a pin compatible package. It is, however, more expensive than the AS3340A, and less readily available.

• Add 100 nF capacitor from pin 4 to ground. In my tests this did not affect the frequency shift. It adversely affected the pulse width range.

• Sam Battle (LMNC) has tested adding a 20M resistor between the PW CV and the pitch CV summing node. stmllr found they needed a 6.6M resistor to get this to work, and I found in my setup 4.7M largely eliminated the frequency shift. This approach relies on a cancelation between the pitch shifts due to PWM and a change in control voltage, and that cancelation apparently varies from chip to chip or perhaps in different environmental conditions.

• Alfa recommends this fix . Note the recommended matched transistor pair is available only as a surface mount package.

• Negative voltage for the 3340 chips can be up to -15 V applied through a series resistor to pin 3, which connects to an internal Zener diode, or it can be a regulated -5 V applied to that pin, in which case the Zener is inactive. The datasheet circuit diagram shows the former but the Alfa datasheet recommends the latter “to minimize self-heating and improve-thermo-stability”. In my testing, using -5 V also minimized the PWM pitch shift. But if that is the case, why would Alfa recommend a different solution? I don’t know, but if there are circumstances where it doesn’t work, I haven’t found them.

Pulse wave ringing

With the datasheet wiring, one sees ringing on the falling edge of the pulse wave at low frequencies:

The remedy for this is a resistor between pins 4 and 5. In my tests I found a 1 M resistor eliminated ringing down to about 12 Hz.

Pulse wave pulldown

The CEM3340 and V3340 datasheets specify a 10k pulldown resistor on pin 4, the pulse wave output. The AS3340 datasheet specifies 51k. In my tests with AS3340A I saw that larger values gave longer fall times on the falling edges of the pulse wave (and longer rising edges, but the falling edges are slower). This has little to no audible effect. Comments in the datasheet suggest there is some benefit in keeping the pulldown current low, so 51k with AS chips seems reasonable.

Buffering and level shifting

Series resistors between the 3340 pins and output jacks are advisable. The triangle wave output must be buffered.

The different waveforms have different amplitudes. All three are positive-going from 0 V. The triangle wave amplitude is VCC/3 (+4 V for a +12 V supply), the ramp is 2VCC/3 (+8 V), and the pulse wave is around VCC-1.5 V (+10.5 V). Op amp gain stages can be used to shift and scale these waveforms. In the Analog Output Hero VCO all three outputs are scaled and offset to a ±5 V range.