• TI Thinks Resolved

ISO1541: ISO1541

Part Number: ISO1541

Hi Sir,

I'm Len from GMT.

We have a application of ISO1541. We use it to isolate I2C interface. we connect a level shift TCA9800 to side 1, and connect a sensor to side 2. we found the pulse of the waveform is changed between the two side obviously. The waveform as below, the blue one is side 2, white one is side 1. pls help to check it. 

  • Hi Len,

    Thank you for your post on E2E.
    I see you have shared two waveform and the blue waveform in the second second doens't look very good, is this what you are referring to in the post? Can you please confirm if this SDA2?

    From the waveform, I see you are applying input to SDA2 and monitoring output at SDA1. This means that the input waveform at SDA2 doesn't look good, do you know why the input waveform is clean for first capture and not for the second capture? Can you please explain the difference between the test setup for first image and second image?

    Please confirm if my understand of your question is correct, if not please elaborate on the issue you are observing. Thanks.


    Regards,
    Koteshwar Rao
  • In reply to Koteshwar Rao:

    Hi Sir,

    Yes. the blue waveform is different with the white waveform. both of the blue waveform are side 2. Resend SDA waveform.

    We measured two times for each signal. The white signal is measured side 1 at first time. we save it in oscilloscope. and then we measured side 2, the blue signal. because the SDA signal of side 2 is changed, the communication failed.

    BR,

    Len

  • In reply to user4931969:

    Hi Len,

    I am not clear if I understand your issue correctly, please help me with answers for the questions below individually.

    1. Can you please tell me where is input signal applied? SDA1 or SDA2?
    2. What is the difference in test setup in the first image and second image?
    3. From the image that you have shared, I see that the LOW voltage for SDA1 is about 0.7V and it makes me feel as the output. Please confirm this.


    Regards,
    Koteshwar Rao
  • In reply to Koteshwar Rao:

    hi Koteshwar,

        Do you understand what I2C interface means? It's a protocol of serial interface include I2C_CLK and I2C_SDA. one image is clock signal. the other is data signal. each image has two waveform. they are belonging to the two side of ISO1541.

        Do you familiar with the ISO1541? The ISO1541 is a unidirectional parts. Side1 is input. Side 2 is output. My question is that the pulse width of output signal is changed. I want to know why it happened.

    BR, 

  • In reply to user4931969:

    Hi Len,

    Thank you very much for clarifying on the first two waveform that you have shared, the first image showing the clock applied at SCL1 & monitored at SCL2 while the second image has data applied at SDA1 & monitored at SDA2.

    Please do note that I2C is a multi-master, multi-slave interface where the data is bidirectional and transferred both ways while the clock can be unidirectional for single master system (ISO1541) and bidirectional for multi-master system (ISO1540). If you have a single master system but implementing clock stretching, we recommend you use ISO1540 to support clock stretching. Please refer the below diagrams of ISO1540 & ISO1541.

    I am sorry for asking too many questions, the intent was to understand your query correctly and not make any incorrect assumptions. Now I understand your question better. I still have one more question on the SDA1 signal applied because it violates datasheet spec. I see the LOW voltage at SDA1 from the waveform is about 0.7V whereas ISO154x accepts a maximum of 0.5V for it to be considered a valid LOW, do you know why is this so?

    Could you please share the schematic showing ISO1541 and the devices it is interfaced to on either side?

    Regards,

    Koteshwar Rao

  • In reply to Koteshwar Rao:

    Hi Koteshwar:

    The connection is simple as below. it's a simple system of I2C. I don't know why the output of TCA9800 is over spec. I want to know why the waveform of ISO1541's input and output is different. why it happened? 

  • In reply to user4931969:

    Hi Len,

    Thank you for sharing more details regarding your application, including the block diagram.

    The reason I am asking why LOW state voltage on SDA1 is 0.7V is because the value of SDA1 LOW state voltage is dependent on the direction of data transfer. Based on the direction of data transfer, the LOW state voltage can be either between 0.65V to 0.8V or less than 0.5V.

    To understand more on why the LOW state voltage is different for either direction of data transfer, please read through section 8.4 on page 19 of ISO1541 datasheet. After you go through section 8.4 in datasheet, you will understand that a diode has been added on SDA1 output section (B) so that input section (C) of SDA1 pin differentiates between LOW output state from output section (B, let's call it VOL1) and LOW input state from an external input (let's call it VIL1). If we do not do this and since the channels are connected back to back as shown in the block diagram, the device will lock into a LOW state irrespective of the external input.

    I am sharing below a sample waveform that shows SDA1 and SDA2 voltage levels (voltage scale - 1V/div). The blue waveform is SDA1 and green waveform is SDA2. Every time SDA1 is acting as input, VIL1 is close to zero. When it is acting as output, VOL1 is close to 0.8V.

    Going by the above explanation, the reason for SDA1 LOW state voltage to be about 0.7V is that this voltage is produced by SDA1 (VOL1) and is not an external input. In your case, I believe you may have applied input at SDA2 and output has appeared at SDA1 accidentally. I recommend if you could re-verify your device connections and re-test to confirm device behavior. If you could share us the schematic, I can also re-verify at the schematic level.

    I have another reason to state that input is applied at SDA2 and output is being monitored at SDA1. In the waveform that you have shared, I see that SDA1 waveform follows SDA2 after about 2µs propagation delay which otherwise should have been the opposite. The large propagation is propagation delay is probably due to the large external pull-up resistor values.

    As next steps to address the issue, I would recommend you testing the device once again making sure you are referring to the sides correctly. If you could share me the schematic, I could also review it and let you know if there are any issues.

    Looking forward for your feedback, thanks.

    Regards,

    Koteshwar Rao

  • In reply to Koteshwar Rao:

    Hi Len,

    Please do let us know if you were able to re-test the device ISO1541 in your application to see the direction of data? Like I said earlier, helping us with schematic or PCB layout will help us understand your application better.

    If you would like us to reach out to you on the email you have used in your profile, I can also do that. Do let me know, thanks.


    Regards,
    Koteshwar Rao
  • In reply to Koteshwar Rao:

    Hi Len,

    I am not sure if the issue is addressed, if it is not addressed please share the information requested in my earlier post. If you could address the issue at your end already, please do let me know so that I can close this thread.

    In case if you want me to reach out to you over the email that you have used for your profile creation, please do let me know so that I can do so. Thanks.


    Regards,
    Koteshwar Rao
  • In reply to Koteshwar Rao:

    Hi Len,

    I am not sure if you were able to address the issue. I believe I have provided the relevant information regarding device ISO154x and example waveform that helps you look into fixing the input waveform shape. I am closing this post, please do respond if you have any update on this discussion.

    Thanks for your contribution on E2E Forum.


    Regards,
    Koteshwar Rao