Method and Apparatus for an Improved Multiple Channel Sensor Interface Circuit

This invention generally relates to sensor systems having multiple channels, wherein it is desired or necessary that each channel provide a uniform output for the same input. Since the channels are physically separated, the difficulties of matching the channels such that each channel has the same physical characteristics may become prohibitive. Further, when building a multiple channel sensor system on an integrated circuit, the necessity of matching channels which are physically separated may make fabrication impracticable as the process variations across an integrated circuit may become so great as to prevent proper operation of the device. The circuit technique introduced herein avoids this problem, making large well-matched sensor arrays possible.

Without limiting the scope of the invention, its background is described in connection with a multiple channel video sensor having multiple analog video inputs which are sampled in parallel and having a single time multiplexed analog output channel, the output reflecting the value of a single one of the input channels at any particular time, such as might be found in a facsimile machine, as an example.

Heretofore, in this field, the standard method of implementing such a sensor interface was to build a bank of integrator circuits followed by a bank of sample-hold amplifiers, followed by a multiplexer which determines which channel's signal is to be transmitted to the sensor interface output. This approach requires two amplifiers per channel, and it also imposes a requirement that each channel be identically matched to all the others. Further, any normal offset voltages which occur in the sample-hold circuits are transmitted to the output and result in further channel to channel non uniformity.

Accordingly, it is a goal of this invention to create improvements which overcome the requirement of channel matching, the sensitivity to offset voltages in the sample-hold circuits, and which is more efficient to design, test and manufacture.

Generally, and in one form of the invention, an improved multiple channel sensor interface circuit is described which comprises a plurality of input integrator circuits coupled in parallel; a switched capacitor multiplexer coupled to the input integrator circuits; and an output integrator stage coupled to the switched capacitor multiplexer.

An additional embodiment is described which is a multiple channel voltage sensor interface circuit, comprising a plurality of switched capacitor storage elements coupled to a plurality of inputs; a plurality of integrator amplifiers coupled to the switched capacitor storage elements; and timing circuitry coupled to the switched capacitor storage elements and to the integrator amplifiers operable to selectively enable sampling of the inputs.

In the drawings:

FIG. 1 is a block diagram of a prior art multiple channel sensor system;

FIG. 2 is a block diagram of an improved multiple channel current or charge sensor circuit embodying the invention;

FIG. 2A is an alternative embodiment of the invention depicted in FIG. 2;

FIG. 3 is a schematic diagram of a typical single channel circuit in the device of FIG. 2;

FIG. 4 is a schematic diagram of one of the last channels circuit in the device of FIG. 2;

FIG. 5 is a waveform diagram of the switch timing of the single channel shown in FIG. 3;

FIG. 6 is a block diagram of an integrated circuit implementing a multiple channel sensor interface using the improved multiple channel sensor circuit of FIG. 2;

FIG. 7 is a schematic diagram of an alternative embodiment implementing an improved multiple channel voltage sensor circuit.

Corresponding numerals and symbols in the different figures refer to corresponding parts unless otherwise indicated.

In FIG. 1, a prior art multiple channel sensor interface is detailed. Integrator circuits 1 are connected to sample the analog current or charge inputs. Sample-hold circuits 3 then sample the output of the individual integrators. Analog multiplexer 5 is used to sequentially transmit the sampled value of each individual channel to output buffer 7.

In FIG. 2, the improved multiple channel sensor interface of the invention is shown. This embodiment is suited to applications wherein the interface is to sense charge or current as inputs. Input integrator amplifiers 9 are coupled to each of the analog inputs. Switched capacitor multiplexer 13 captures the outputs of the integrator amplifiers. Integrator 15 is used to sequentially transfer the sampled charge stored by each channel. Sample-hold circuit 17 then provides a constant voltage output.

FIG. 2A depicts an alternative embodiment wherein the sample hold circuit 17 of FIG. 2 is replaced with the analog to digital converter 18 shown in FIG. 2A, for use in applications where a digital representation of the output is required.

In FIG. 3, a single channel n of the multiple channel sensor interface of the invention is shown. Amplifier 25 is coupled with capacitor C.sub.3 (n) and switch S5 to form the integrator input stage. Switches S4 and S3(n) are coupled with capacitor C.sub.2 (n) to form the switched capacitor multiplexer stage. Amplifier 27 is coupled with capacitor C.sub.1 and switch S2 to form the sampling integrator stage. Switch S1 is coupled with capacitor C.sub.0 and amplifier 29 to form the sample-hold output stage. The switches may be fabricated using any one of several alternatives known in the art, typically through the use of CMOS or bipolar transistors. Timing circuitry to open and close the switches at appropriate times is fabricated using any well known design technique, such as digital logic comprising a state machine with a PLA to decode the switch control signals.

In operation, switch S5 is closed in all channels to allow the inputs to be integrated onto capacitor C.sub.3 (n) in each channel simultaneously. After the integration period, switches S4 are closed to enable the capacitive multiplexer element in each channel to sample the integrator outputs in parallel, and then store the sampled charge onto the capacitors C.sub.2 (n) in each channel. At the end of the input integrator sampling phase, switches S4 are opened to stop the input sampling. Switches S3.sub.(n) are then closed sequentially in each channel to enable output integrator capacitor C.sub.1 to receive the sensed data from the currently selected channel. Switch S2 is closed between each transfer to ready the output integrator stage for receiving the sensor data from the next channel. Finally, switch S1 is closed while switch S2 is opened and the current channel switch S3(n) is closed to enable the sensor data to be captured at the output of the sample-hold output stage. The output of the sample-hold buffer will be a continuous voltage output.

To demonstrate the absence of a channel matching requirement in the improved multiple channel sensor interface circuit of the invention, the reader is referred to the example channel depicted in FIG. 3. The voltage V.sub.1 produced at the output of the input integrator stage is given by: ##EQU1## for a charge input, or ##EQU2## where T is the integration time for a current input.

When the channels are sequentially scanned to produce the output voltage at the output of the output integrator, the charge transferred from capacitor C.sub.2(n) to capacitor C.sub.1 is:

where V.sub.1 is available above. Thus, by substitution ##EQU3## for a charge input, or ##EQU4## for a current input. Since the voltage V.sub.2 produced at switch S1 is simply ##EQU5## for each input channel, it becomes clear that the output produced by a given channel n is directly proportional to the current or charge input and on the ratio of C.sub.2 to C.sub.3, and that the output has no relationship to the absolute physical characteristics of a particular channel. To produce highly uniform sensor interfaces all that is required is that the ratio of the two capacitors C.sub.2 and C.sub.3 be kept constant between channels. Since these two capacitors are placed in close physical proximity, the control of this ratio is easy to achieve. This is a large advantage over the circuits of the prior art, wherein capacitance matching was required between channels which could be several hundred mils apart.

The above discussion assumes the input integrator stage 25 exhibits no offset voltage. In practice, the input integrator stage 25 will add an offset voltage to the voltage output into the switched capacitor multiplexer stage. This offset will vary from channel to channel, and to assure uniform output for a uniform input the offset voltage must be canceled through offset cancellation or auto-zeroing techniques. By using switches S5 and S6, the offset cancellation is accomplished. Between sampling periods the capacitor C.sub.3 (n) has placed on it a voltage of - V.sub.os by the closing of switches S6. When the integrator is in input mode, switch S5 is closed and the capacitor C.sub.3 (n) receives the voltage Vi plus the offset voltage of amplifier 25. Thus, the output of the input integrator stage at V.sub.1 is given by: ##EQU6## which shows that the use of switches S6 eliminates the offset in each channel and maintains the channel-channel uniformity required.

The use of different ground voltages 31 and 33 in FIG. 3 allows the use of the switched capacitor multiplexer circuit to level-shift the input integrator sample to a different bias level at the output. Of course, this is entirely optional and for applications where the bias level requirements are the same at the input and the output, these two ground voltages may be the same or tied together.

Care must be taken in the implementation to prevent loss of the stored charge on capacitor C.sub.3(n) in the first cycle. To understand this, consider that reference voltage 31 is at a potential of 1 volt, reference voltage 33 is at zero volts, and capacitor C.sub.3 is charged to 3 volts. When switch S4 is dosed to being sampling the integrated value onto capacitor C.sub.2, the voltage at node V.sub.1 will instantaneously go to 0 volts, as the capacitor has to be charged to the voltage on C.sub.3. This results in the switch S6 having a -2 volt potential drop across it. In an integrated circuit, switch S6 may be implemented, for example, with a NMOS transistor. If the voltage at the source or drain is less than that at the gate by more than the V.sub.t of the transistor, current will begin to flow away from node V.sub.1 towards the input terminal and erroneously discharge C.sub.3. This is prevented in the current embodiment by precharging capacitor C.sub.2 before the first cycle begins, so that C.sub.2 in each channel has a voltage of the voltage at reference 31 minus the voltage at reference 33 placed on it, which will be the voltage on it after each cycle. By insuring that the instantaneous voltage at node V.sub.1 is such that the voltage across switch S6 is never greater than the V.sub.t of switch S6, this error is prevented for all cycles. This initialization step can easily be handled by the timing circuitry driving the switching sequences.

As shown in FIG. 3, there is only one sample-hold amplifier in the circuit of the invention, so that any offset voltage which occurs in the sample-hold circuit affects all of the channels equally thus providing further improvement over the prior art circuit in channel-channel uniformity.

Also as shown above, the voltage V.sub.2 at the output of the output integrator 27 is inversely proportional to the value of the capacitor C.sub.1. Since this same capacitor is used to sample each of the channels, and the output voltage of each sample will depend on C.sub.1 in exactly the same way, capacitor C.sub.1 can be adjusted to uniformly adjust the gain across all the channels. This adjustment could be made electronically, by changing the interconnect information on a mask level for an integrated circuit which includes the system, of possibly through the use of nonvolatile programming either at the time of manufacture or in the application itself, or through any other means.

FIG. 4 details one of the channels which is at the end of the array and therefore sampled at the end of the sampling sequence. The basic structure of FIG. 3 is implemented, however there are now two switched capacitor banks, the first comprised of switches S3A(n), S4A, and C.sub.2A ; and the second comprised of switches S2B(n), S2B, and C.sub.2B.

In applications where the input data stream is continuous, it is necessary to add the additional capacitor circuitry to the channels which are sampled last in the sequence. The data being integrated onto integrator amplifier 25 is alternatively sampled onto capacitor C.sub.2A(N) and capacitor C.sub.2B(n). This enables the sequential sampling circuitry to transfer the sampled charge to the output integrator Stage formed by amplifier 27 and capacitor C.sub.1 for one bank while the alternate bank receives the newer integrated sample data. Without this provision, the channels sampled last in sequence would lose data in those applications where the sample rate at the input exceeds the transfer rate at the output, or in continuous data stream applications. Switches S3A and S4A are alternated with S3B and S4B in these channels by the switch timing circuitry, which is not shown.

FIG. 5 depicts the timing of the switches in the sample and output stages of operation of the embodiment shown in FIGS. 2 and 3. At the beginning of each cycle, switches S6 are closed and S5 are opened to enable the autozeroing feature to charge capacitor C.sub.3 in each channel with the appropriate offset voltage. During the input integration stage, switches S5 are closed and S6 are open, to allow the input integrator circuit of each channel to be sampled onto the capacitor C.sub.3 (n) in each channel in parallel. Switches S4 are then closed to sample the integrated data onto capacitors C.sub.2(n). For those channels which are high ordered channels, such as the one depicted in FIG. 4, switch S4A will be used alternately with S4b. This is indicated in FIG. 5 by the S4A and S4B signals. At the end of the input sampling phase, the switches S4 are opened. Switch S3(n) in each channel is operated sequentially to allow the selected channel to be sampled by the output integrator stage, and switch S2 is closed between channel sampling to ready the output integrator to receive the next channel's sensor input. Again, in high ordered channels switch S3A will be used alternately with S3B. Note that as shown in FIG. 5 for the channel (n-1), while switch S3A is operated to discharge capacitor C2A, switch S4B can be operated concurrently to capture the next data sample on capacitor C2B. Switch S1 is alternated with switch S2 to allow the output of the currently selected channel to be sampled by the output sample-hold stage. At the end of the sampling of the last channel, the input integration stage begins again with the closing of switches S5. The input integration phase, indicated in those time periods when S5 is closed, overlaps the sampling of the individual channels which occurs in those time periods when switches S3 are operated. Put another way, capacitors C2 may be discharged, transferring the current input sample to the output, while capacitors C.sub.3 are charged, capturing the next set of input currents. This technique maximizes the data throughput rate of the circuit.

FIG. 6 is a block diagram of an integrated circuit including the multiple channel sensor circuit of FIG. 3, and including the timing and control circuitry to produce the switching timing as shown in FIG. 5. Input integrators 35 are coupled to the sensor input pins, and feed into switched capacitor multiplexer 37. The output of multiplexer 37 is coupled to the output integrator circuit 39 and finally to the output sample-hold circuit 41. Timing circuit 43 provides switch controls to each of the switches in each of the above blocks. Details of timing circuit 43 are not shown, the timing circuitry can be implemented using straightforward sequential state machine design, logic trees, or programmable logic components using techniques well known to those experienced in the art.

In operation, this integrated circuit receives analog inputs, and concurrently samples the inputs onto all of the input integrator circuit capacitors C.sub.3. These integrated values are then sampled onto the switched capacitor multiplexer capacitors C.sub.2 for each channel. Then the sampling period is ended and the circuitry sequentially transmits the sensed value for each channel onto the output integrator amplifier capacitor C.sub.1 and out through the sample-hold output buffer using capacitor C.sub.0. The timing circuit provides the control lines to the switches to implement the sequence timing as shown in FIG. 3.

An alternative embodiment for applications where the input is a voltage level is shown in FIG. 7. Switches S26 and S27 and capacitor C.sub.22 are coupled to create a first input to the first sensor channel. Switches S28, S29 and capacitor C.sub.24 are coupled to create a second input to the first channel. Voltage V20 is the output of the selected one of the two inputs. Amplifier 51 is arranged with capacitor C.sub.20 and switches S20 and S22 to create an integrator with offset voltage cancellation which takes as its input the voltage V20. Switch S24 allows the output of the first channel to be selectively output as V0. Second channel 55, comprised of switches S35, S37, S39, S41, S31 and S33 with capacitors C.sub.28, C.sub.30 and C.sub.26 and amplifier 53 is identical to the first channel 50. The output will be a signal array representing the demultiplexed, sampled data.

In this embodiment, a time multiplexed input signal is fed into the switched capacitor multiplexer formed by switches S26, S27 and capacitor C.sub.22, and the output of the switched capacitor multiplexer is then fed into integrators, for example the one formed by the amplifier 51 and the capacitor C.sub.20. Switches S20 are used to place an offset cancellation voltage on capacitor C.sub.20 when the capacitor is not being charged with sensed data.

The arrangement of two capacitors in parallel in a single input channel, for example capacitors C.sub.22 and C.sub.24 in FIG. 7, allows the sampling of time multiplexed input data. This single input channel may receive sensed data into capacitor C.sub.24 while capacitor C.sub.22 is being discharged into capacitor C.sub.20. Additional multiplexing may be achieved by adding additional switched capacitor circuitry in parallel.

In operation, the charge stored onto capacitor C.sub.22 is given by:

The charge stored on capacitor C.sub.22 is then transferred to capacitor C.sub.20 so that the charge across C.sub.20 is given by:

Finally, the output voltage due to the sampling of one of the channels is given by: ##EQU7##

As before, the output is proportional to the ratio of capacitors C.sub.22 and C.sub.20. Since this ratio is easily tightly controlled, a uniform output for uniform input in any given channel is realized without the need for channel matching. This embodiment is useful for LCD applications where voltage levels are sensed, instead of charge or current sensing as used above.

Note that the integrator circuit in FIG. 7 will also contribute an offset voltage across capacitor C.sub.20. As before, switches S20 are used to place an offset cancellation voltage on capacitor C.sub.20 between sampling periods. The voltage across capacitor C.sub.20 where there is a non zero offset voltage at the input of amplifier is given by: ##EQU8## The charge Q.sub.os on the capacitor C.sub.20 due to the offset voltage supplied by amplifier 51 is:

Substituting the offset charge into the V.sub.o equation gives: ##EQU9## so that the offset voltage is canceled when C.sub.22 is equal to C.sub.20. Although this requires matching among the capacitors in a single channel, no matching is required across channels, and again the capacitors C.sub.20 and C.sub.22 will be fabricated in close physical proximity, so that variations in their values should affect both capacitors and the overall output of the channel should be unaffected. If needed the output of the circuit in FIG. 7 may be followed by a sample and hold circuit as in FIG. 3, or by an ADC or other circuitry as required.

Possible system embodiments of the improved multiple channel sensor interface include facsimile machines, handheld electronic scanners, bar code readers, cameras, imagers, and other optical systems, and any type of multiple channel current or charge input interface.

A few preferred embodiments have been described in detail hereinabove. It is to be understood that the scope of the invention also comprehends embodiments different from those described, yet within the scope of the claims.

Internal and external connections can be ohmic, capacitive, direct or indirect, via intervening circuits or otherwise. Implementation is contemplated in discrete components or fully integrated circuits in silicon, gallium arsenide, or other electronic materials families, as well as in optical-based or other technology-based forms and embodiments.

While this invention has been described with reference to illustrative embodiments, this description is not intended to be construed in a limiting sense. Various modifications and combinations of the illustrative embodiments, as well as other embodiments of the invention, will be apparent to persons skilled in the art upon reference to the description. It is therefore intended that the appended claims encompass any such modifications or embodiments.