I2C Master-Slave如何工作

它遵循它的状态机。它需要知道地址是否适合它。该地址将是启动后的第一个字节。



      以下为原文

    It follows its state machine.
It needs to know if the address is for it.
The address would be the first byte after the start.

正确的。该地址将是启动后的第一个字节。那么，如果在地址之后收到一个字节，状态机会自动知道接收到的下一个BTE不是一个地址吗？也许我不完全理解I2C。地址总是指设备吗？这将意味着所有后续字节可能是与设备相关联的数据或内部地址。这是正确的吗？



      以下为原文

    Right. The address would be the first byte after the start. So, then, if a byte is received after the address, does the state machine automatically know that the next bte received is NOT an address?
 
Maybe I just don't fully understand I2C. Does address always refer to the device? That would then mean that all following bytes might be data or internal addresses associated with the device. Is this correct?

更高级的协议（每个字节意味着什么）完全由从属决定。您需要读取您试图访问的设备的数据表。启动后的第一个字节总是包含芯片地址和RW位，但有时也包含内部地址位。下一个或两个字节包含内部设备地址-寄存器地址，或EEPROM或Flash或RAM的内存地址。您不能编写一个通用的主I2C接口。你必须为你所说的设备写接口。



      以下为原文

    The higher level protocol (what each byte means) is determined entirely by the slave. You need to read the datasheet for the device you are trying to access. The first byte after the START always contains the chip address and RW bit, but sometimes contains internal address bits as well. The next one or two bytes contain internal devices addresses - either register addresses, or memory addresses for EEPROM or Flash or RAM.
 
You can't write a generic Master I2C interface. You have to write the interface for the device you're talking too.

我明白了。我的问题不在于大师。它只关于奴隶。这两种设备都是相同的18F2510。所以，主发送开始，然后7位地址（让我们说0xB0）。由于RyNoTyw为0，这将告诉奴隶这是一个写操作。下一个字节既可以是寄存器地址，也可以是数据字节。我的问题是奴隶如何知道这个字节是什么。硬件是否理解第二个字节是内部地址或数据，还是我编码的固件的一部分？我假设，如果从接收到的字节不是从属地址，它将被解释为数据（关于I2C硬件）？如果是这样的话，那么只有奴隶知道第二个字节是内部地址/寄存器是固件告诉它。这是正确的吗？



      以下为原文

    I undertsnad that. The question I have is not about the master. It is about the slave only. Both devices are identical 18F2510.
 
So, master send start, then 7-bit address (let's say 0xb0). Since R_NOT_W is 0, this would tell the slave that this is a write operation. The next byte would either be a register address or a data byte. The question I have is about how the slave knows what this byte is. Does the hardware somehoe understand that the second byte received is an internal address or data or is that part of the firmware that I code?
 
Am I to assume that if the byte received by the slave is not the slave address it is to be interpreted as data (with respect to the I2C hardware)? If that is so, then the only way for the slave to know that the second byte received is an internal address/register is for the firmware to tell it so. Is this correct?

对不起，我在第一句中漏掉了“两个”。除了第一个字节上的地址匹配检测外，没有任何内置的字节序列。所以我的评论仍然是有效的- Slave决定序列。因为你在编程奴隶，所以你决定序列。这是你想要的任何东西。如果你需要一种寻址或命令来让奴隶知道该怎么做，那么你必须创建状态机（两边）来处理这个问题。



      以下为原文

    Sorry - I missed the "for both" in the first sentence. Except for address match detection on the first byte, there's nothing built in that determines any byte sequence. So my comment is still valid - Slave determines the sequence. Since you're programming the slave - you determine the sequence. It's whatever you want it to be. If you need some kind of addressing or command for the slave to know what to do, then you have to create the state machine (on both sides) to handle that.

好啊。我能做到。最后一个问题…我想：如果我必须处理字节序列（正如你所说的），我还必须以编程方式发送ACK，或者当接收字节时，从属设备会自动发送ACK吗？谢谢你的帮助。



      以下为原文

    Ok. I can do that. Last question... I think: If I am having to take care of the byte sequence (as you put it) will I also have to programatically send the ack or will the slave automatically send the ack when it receives a byte?
 
Thank you for the help.

将ACK发送到接收的字节，从属发送ACK给接收的ACK。



      以下为原文

    you send ACK to received bytes, slave sends ACK to the ones IT receives

将ACK发送到接收的字节，从属发送ACK给接收的ACK。



      以下为原文

    you send ACK to received bytes, slave sends ACK to the ones IT receives