Home -> Magazines -> Issues -> Articles in this issue -> View
Chip Parade (Part 5) | |
Article from Electronic Soundmaker & Computer Music, January 1984 |
More on CEM
This month Robert Penfold looks at new additions to the Curtis Electromusic IC range.
Last time we looked at some of the well established devices in the Curtis Electromusic Specialties range of music ICs. Recently the Curtis range of devices has been augmented by some interesting additions to the family, and in this article we will consider the three devices in question. In the past, Curtis integrated circuits have been much used in commercially produced electronic instruments, and it seems likely that these new components will be equally successful (the CEM3372 is already used in the new Prophet 600 polyphonic synthesizer for example).
This is a dual voltage controlled state variable filter having separate exponential control inputs for each filter. There is no provision for linear control of either filter incidentally, but in music applications linear control would not normally be needed anyway. The filters are identical and are 2nd order (12dB per octave) types. Fig 1 shows a basic circuit for one section of the CEM3350. The numbers in brackets are the equivalent pins for the other section of the component. An important point to bear in mind is that this device can be used in a variety of interesting filter types, and that this circuit is really only intended to help to familiarise you with the basic features of the device.
There are two inputs to each filter. With the 'fixed' input there is no variation in gain within the passband as the Q of the filter is varied, but with the 'variable' input the gain in the passband decreases as Q is increased. By applying a proportion of the input signal to each input any characteristic between these two extremes can be obtained.
Each filter has an internal variable transconductance amplifier which is used to control the level of positive feedback that in turn sets the Q value. The Q of each filter is therefore individually adjustable and voltage controlled. Like the filter frequency, the Q control inputs have a standard exponential scale. A wide Q range of typically 0.5 to 150 (50 minimum) can be achieved.
There are two outputs available (which will usually require external buffers) and these are the standard state variable bandpass and lowpass outputs. As for any normal state variable filter, a highpass action can be provided using a different configuration. In practice the two filters would normally be combined into a single 4th order (24dB per octave) filter, and a variety of filter types can then be produced. They could be used as a conventional 24dB per octave lowpass filter with resonance, but remember that the two filters do not have to operate at the same frequency, do not have to provide the same type of response, and do not have to have the same Q values. This enables interesting frequency responses to be easily produced, and, presumably, interesting sounds that are not obtainable using a conventional filter circuit.
The performance figures for the CEM3350 are quite impressive, with a wide sweep range of typically 15 octaves, low frequency and Q control feedthrough, a maximum audio output noise level of just 2.5uVRMS, and typical passband distortion of 1%. A positive supply voltage of between three and 18 volts is required, and the negative supply range is also between three and 18 volts. However, the supplies do not have to be balanced, and the total supply potential should be no more than 26 volts. Unlike the Curtis ICs discussed last month, no dropper resistor is needed in the negative supply lead.
This is a dual voltage controlled amplifier, and is in many ways similar to the CEM3350. To give a few examples, the low noise, low control voltage feedthrough, supply voltage range, and summing inputs are common to both devices.
As can be seen from the block diagram and basic circuit of Fig 2, the CEM3360 is a somewhat more simple device, and it contains little more than the two transconductance amplifiers which are used as the basis of the two VCAs. There is a logarithmic converter for each VCA, and these can be used in series with the control inputs to provide a 0-2 volts linear control characteristic. However an exponential control characteristic can be obtained by leaving the logarithmic converters unconnected. The control signal is then coupled to the control input of each VCA via a 16k resistor, and, additionally, each control input is coupled to the 1.8 volt reference output by way of a 2k4 resistor. A useful feature is that the control inputs are referenced to earth.
The outputs do not have internal buffers, and an external operational amplifier buffer stage would normally be used at each output. A wide control range can be attained, and a minimum range of some 100dB is guaranteed.
The CEM3372 is described by Curtis as a 'microprocessor controllable signal processor'. Any voltage controlled signal processing chip could actually be controlled by a microprocessor, and this would not be difficult with the CEM3350/60 just described. However, with the CEM3372 things are even easier.
First of all it should perhaps be explained that the CEM3372 is a two input voltage controlled mixer, a 4th order lowpass voltage controlled filter, and a high quality VCA, all contained in a standard 18 pin DIL package. Fig 3 shows the internal arrangement of the chip.
The standard way of using the chip is to have two VCOs, one feeding each input of the device. The signal level of each oscillator could be controlled just by using potentiometers to provide the control voltages, but apart from this simple manual control system, it is also possible to individually modulate the two inputs, or modulate them out of phase, and a number of interesting effects could be readily produced.
The next stage is a voltage controlled filter, which is a fairly conventional 24dB per octave (Butterworth) lowpass type having variable resonance. The passband gain remains constant so that there is no fall in volume at high Q settings, and the Q can be advanced to the point where the circuit goes into oscillation. A VCA is used in the positive feedback path so that the filter's Q can be voltage controlled. The frequency control input has the usual exponential control characteristic, and an extremely wide range of some 14 octaves can be covered.
The final stage is a high quality VCA which has a very low level of feedthrough and does not require any trimming. It also has a wide control range of typically 120dB (100dB minimum), and a low noise level. This stage is used to provide the envelope shaping of the signal. Distortion is typically only about 1%, which is perfectly adequate for most electronic applications.
What makes the CEM3372 an attractive proposition for a microprocessor controlled application is the 0 to 5 volt control range of the VCAs, and the very high input resistance at their control inputs. The 0 to 5 volt control range is convenient since it is easy to design a digital to analogue converter which will match this. A high quality circuit of this type is relatively expensive, and in practical systems it is common to use a single digital to analogue converter plus a multiplexer to enable it to sequentially drive several devices. The multiplexer is simply an electronic switch, such as the CMOS 4051 which was discussed in an earlier 'Chip Parade' article.
There is an obvious problem with this system in that the converter can only drive one input at a time, and some way of holding the signal level at each input during the absence of drive from the converter must be found. Conventionally a simple sample and hold circuit would be used ahead of each input. A circuit of this type is basically no more than a capacitor connected across the input of a very high input impedance buffer stage. When the converter drives the input it forces a certain charge voltage on the capacitor. When the converter is disconnected from the input the capacitor holds this charge, and maintains the input at its previous potential so that the converter is free to drive another input. The charge on the capacitor will not be held forever, and the charge will eventually subside due to a minute discharge current flowing into the buffer, or due to leakage in the capacitor itself. However, the charge can be held accurately for at least a few seconds, and in most practical systems this is long enough as the converter modifies the control voltage quite frequently. There is no difficulty in designing the system so that each capacitor has its charge refreshed even if the charge potential is not altered.
Due to the high input resistance at the control input of each VCA within the CEM3372 there is no need to have a sample and hold circuit between the multiplexer and each control input. All that is required is the capacitor to maintain the input voltage between refresh or control cycles. This keeps the number of interface components to an absolute minimum, and with the ever increasing use of microprocessor controllers in electronics, makes it seem highly likely that the CEM3372 will be used in many future designs.
The CEM3350/60/72 devices and data sheets are available from Digisound Ltd, (Contact Details).
Read the next part in this series:
Chip Parade (Part 6)
(ES Feb 84)
All parts in this series:
Part 1 | Part 2 | Part 3 | Part 4 | Part 5 (Viewing) | Part 6 | Part 7 | Part 8 | Part 9 | Part 10 | Part 11
Workbench - Signal Processors — the saga continues |
Amdek DMK-200 Delay Machine Kit |
Lab Notes: Potpourri & The Apple Connection |
Lab Notes - Computer Drums |
The RackPack |
Workbench |
Fader-Panner Project |
ElectroMix 842 (Part 1) |
An Emulator for £10 |
Workbench - CHECKA |
BeeBMIDI (Part 1) |
Browse by Topic:
Topic:
Series:
Part 1 | Part 2 | Part 3 | Part 4 | Part 5 (Viewing) | Part 6 | Part 7 | Part 8 | Part 9 | Part 10 | Part 11
Feature by Robert Penfold
mu:zines is the result of thousands of hours of effort, and will require many thousands more going forward to reach our goals of getting all this content online.
If you value this resource, you can support this project - it really helps!
New issues that have been donated or scanned for us this month.
All donations and support are gratefully appreciated - thank you.
Do you have any of these magazine issues?
If so, and you can donate, lend or scan them to help complete our archive, please get in touch via the Contribute page - thanks!