Solution of ISD chip content duplication circuit under single chip microcomputer control

1 Introduction

The ISD series voice chips are high-quality, record-and-play voice chips launched by the American ISD company (acquired by Taiwan Winbond at the end of 1998). With the company's unique analog speech and multilevel storage technology, the ISD chip directly stores the voice signal in the form of an analog signal in a non-volatile multi-level storage array (a kind of E2PROM), making the storage and playback of the sound different from the general electronic synthesized voice, so that the recording and playback process does not require the traditional A/D and D/A conversion, and thus can better achieve the effect of reproducing the original sound. At the same time, the ISD chip has a simple peripheral control circuit, the control signal is compatible with the TTL level, and it is easy to use with a microprocessor. It has a power-off resistance function, the information storage is not easy to volatilize, supports multiple recording and playback operations, can be used alone, and multiple chips are cascaded, with good flexibility of use. Therefore, ISD chips can be widely used in consumer, communication and industrial control, intelligent instrumentation and other products such as voice alarms, information prompts, and voice messages.

However, since the ISD chip is a combination of analog and digital circuits, the level information stored in its internal E2PROM cannot be read through a programmer like a traditional digital memory, so batch copying is difficult. At the same time, the ISD chip provides an address input line, supports direct address operation, can store voice information in segments and accurately control the recording and playback of voice, which makes the consistency problem of batch copying also prominent. Although these problems can be solved through dedicated ISD device development equipment, the cost is high for a small number of users. Therefore, this article takes the ISD2532 series as an example to provide a simple solution for content copying of ISD series voice chips.

The second part of this paper briefly analyzes the principle of ISD2532 voice chip, and the third part describes in detail the implementation of content replication of ISD2532 voice chip.

2 Principle of ISD2532 Voice Chip[1]

ISD2532 is one of the ISD series of single-chip voice recording and playback integrated circuits. It has a recording time of 32s and a sampling frequency of 8KHz. It can provide sound quality better than that of a telephone. Its internal structure is shown in Figure 1.

When ISD2532 is working in recording mode, the small signal collected by the microphone enters from MIC, is output from ANA OUT (analog output) after preamplification, and is sent to ANA IN (analog input) after passing through the DC isolation capacitor. This input configuration method allows users to design a preamplifier with better performance to replace the existing functional components inside the chip. In addition, ANA IN can be used as a line input port (Line In) to directly input the electroacoustic signal. The signal entering ANA IN will be adjusted to the optimal level of the dynamic range of the storage circuit through the amplifier, automatic gain control (AGC) and low-pass filter, and will eventually be written into the E2PROM through the analog transceiver. In the playback mode of ISD2532, the recorded analog voltage is read out from the analog storage array in sequence under the control of the sampling clock, and the original waveform is restored after the sampling frequency component is removed by the smoothing filter on the output channel, and sent to the multiplexer to be mixed with other external signals, and then sent to the power amplifier, and output from the SP+ and SP- terminals, which can directly drive the speaker. [page]

The typical recording and playback circuit of ISD2532 is shown in Figure 2. The peripheral circuits include: microphone, speaker, switch, a few resistors, capacitors, and a common 5V power supply. Its basic control process is very simple. First, adjust PD to a low level to power on the chip and prepare for work, and set the P/R level (high level for playback, low level for recording); then enter the address to set the starting address of the recording and playback operation; finally, make CE a low level, that is, start the playback operation at its falling edge, and start the recording operation at a low level. In general, the playback operation ends automatically; when PD or CE becomes a high level, the recording operation stops immediately.

The ISD2532 chip can work in two mutually exclusive modes: address and mode, depending on the input status of the highest two bits of the address pin. When the highest two bits are both high, the remaining address pins serve as the operation mode selection terminals, and the chip can record and play in a variety of preset modes, so as to achieve the most functions with the least peripheral devices. Otherwise, the input of all address pins is interpreted as an address bit and used as the starting address of the current recording and playback operation. It can be seen that the 9-bit address in the ISD2532 chip can only provide a maximum of 384 addresses. Fortunately, the E2PROM storage space in the ISD device is not addressed by bytes, but by lines as the basic unit. The 256K E2PROM memory inside the ISD2532 is evenly planned as 320 rows, and each address starting from 0 points to one of the rows. The recording and playback time of the ISD2532 is 32s, so its address resolution is 100ms.

ISD devices can perform multi-segment recording and playback operations. Each segment is called an information segment, which occupies one or more rows of storage space, storing audio data and an end mark (EOM). Both recording and playback functions start from the set starting address. When recording is finished, the chip automatically inserts an EOM mark at the end position of the segment; and when playing, the EOM mark is encountered, the playback of this segment is automatically stopped.

3 Implementation of ISD2532 chip content replication

Unlike digital memory copying, the ISD chip stores continuously changing sound analog quantities, so the content copying standard for the ISD chip can be relaxed to only require that the sound information played back according to the stored information sounds the same. In view of this, while controlling the source chip to play, the destination chip can record to achieve the purpose of content copying. See Figure 3. Under the control of the single-chip microcomputer, the same starting address is sent to the source chip and the destination chip, and the recording and playback operations of the two chips are controlled at the same time to achieve the copying goal.

Since the ISD chip can specify the starting address for recording and playback, the content copying must ensure that the source chip and the newly copied chip store the same sound information at the same starting address. In other words, the copying process must ensure that the information segments stored in the two chips are aligned in address. Therefore, obtaining the starting address of all information segments in the source chip is crucial to completing the content copying task.

3.1 Obtaining source chip address information

The address pins of the ISD chip can only be used as inputs, so the starting address of its internal information segment cannot be directly read without professional development equipment [2]. However, in general, in order to make full use of the storage space, the starting address of the first information segment in the chip is 0, and the next address after the end of the previous information segment is the starting address of the current segment. Therefore, for the ISD2532 chip, let the starting address of the i-th information segment be Add[i] and occupy Len[i] rows of storage space, then:

Thus, to obtain Add[i], Len[i] must be known. Since the address resolution of ISD2532 is known to be 100ms, the time length of the i-th information segment should be between (Len[i]-1) x 100ms and Len[i] x 100ms. When the source chip is playing, the end of a certain information segment can be known by capturing the EOM signal it outputs [2]. Therefore, by timing the playing time of each information segment, the size of the storage space occupied by the information segment can be determined, and thus the starting address of each information segment can be obtained.

[page]

In the case where the information segments are not stored continuously in the chip, we may consider the blank line between the i-th information segment and the i+1-th information segment as a component of the i+1-th information segment, which is equivalent to recording the information that cannot be heard in the front part of the i+1-th segment, and does not affect the hearing. In addition, since the input address of the ISD chip remains unchanged during the playback process, if the system circuit where the source chip is located is easy to test, the starting address of all information segments can be directly obtained by measurement.

3.2 Key points of content replication circuit design

The ISD chip content copy circuit is shown in Figure 4. The power supply, automatic gain control, XCLK and other related peripheral circuits are omitted in the figure. For this part, please refer to Figure 2. The address signal, control signal and status signal are all digital signals, which can be directly connected to the I/O port of the microcontroller (such as the MSC51 series), but considering that the driving ability of some microcontroller pins is limited, pull-up resistors should generally be added. For example, P1.0 and P1.1 in the P1 port of 89C2051 are comparator terminals and have no internal pull-up resistors, while the other terminals of the P1 port have internal pull-ups. Therefore, when the P1 port is used to control the address line of the ISD chip, pull-up resistors should be connected to both ends of P1.0 and P1.1, otherwise the address control will be wrong.

In the replication circuit of Figure 4, the source chip works in the playback state. The sound waveform output from SP+ can be sent to the ANA IN input terminal of the destination chip through simple adjustment of the voltage divider resistor R and the DC blocking capacitor C, and is stored in the E2PROM of the destination chip during the recording process.

In the design of single-chip computer program, the starting addresses of all information segments in the source chip are usually placed in a table, so that the starting address can be obtained through the table lookup instruction, which is convenient for playing and recording each information segment. At the same time, attention should be paid to the delay processing in the program. Usually, before giving the CE low level to start the recording and playback operation, a delay of 1~10ms should be inserted to ensure the stability of PD, P/R and address signals; when querying the effective EOM output state of the source chip, considering that the EOM low level width of the ISD2532 chip is 12.5ms, a delay of about 12ms should also be added here; finally, before entering another cycle to process the copy of the next information segment, some delays should be added accordingly to ensure that the chip can work stably.

4 Conclusion

Based on the introduction of the principle of ISD2532 voice chip, this paper comprehensively provides the method of obtaining the starting address of multiple information segments of the source chip for the problem of content duplication of ISD2532 chip, and also provides a solution of ISD chip content duplication circuit under the control of a simple single-chip microcomputer. Under the condition of adjusting relevant parameters, this solution can be applied to the duplication and copying problem of ISD series chips with similar processing principles.

Since the ISD chip has the function of live recording, the ISD chip may record incorrectly when the power supply is unstable, especially during power on and power off. At the same time, when allowing users to customize recording, there may be problems with recording errors. In response to the above problems, the content copy circuit proposed in this article can also be embedded as a module in the actual system as a solution to restore the original recording.