• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 15 subscribers
  • Views 236 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

ISO6721-Q1: About galvanic isolation of debugger

andy
Prodigy 80 points
Part Number: ISO6721-Q1
Other Parts Discussed in Thread: MSP-FET, ISOUSB211

Tool/software:

Hi. There is something related to galvanic isolation which I am confused about. It is not actually related to ISO6721-Q1, it involves microcontrollers and SMPS as well. I am also not sure about how to ask it but I will try my best.

There is a PC and there is a custom development board.

There is nothing special about the PC. It is just a regular computer.

We have our development board. It has built-in SMPS to power the board from AC line. The SMPS is non-isolated.

The pc and the devboard are connected to each other with a debugger such as MSP-FET. The pc and the devboard can be connected with a USB-to-UART line but it is optional.

What happens is due to lack of galvanic isolation, a short occurs and the PC's motherboard gets fried and the PC dies. Why does this happen? Where does the short occur? What precautions are needed to ensure safety? Can you point to me application notes or something to read about this to better understand it? I had found an application note from Analog Devices years ago that explained this but I can't find it anymore.

Some additional questions.

Where should the "galvanic isolation block" be inserted in the picture below? Between PC and Debugger, between Debugger and DevBoard or does it depend? If it depends, what does it depend on? Would we still need another "galvanic isolation block" for USB<=>UART connection?

Again for the connection diagram in the same image below, seperate from the previous question. Why does connecting one of the PC or the devboard to AC line through an isolation transformer solve the issue?

Does it make any difference if the AC line is coming from electric grid or from UPS? If so, what is the difference?

Would changing the SMPS in the devboard from a non-isoalted one to isolated one be enough to solve the problem?

In the application note, which I can't find anymore. It was saying something like, it is not possible to know which wire of the AC line is the 0V reference for the internal circuits of the PC and the devboard. And it is possible that a short didnt occur but may occur if you flipped the AC-line plug of the devboard. I am not sure if this was what it said or if I understood wrong but I am leaving this here just in case, if helps anything.

Thanks in advance.

  • 0
    TI__Genius 12485 points

    Hi Andy,

    Thanks for reaching out.

    andy said:
    What happens is due to lack of galvanic isolation, a short occurs and the PC's motherboard gets fried and the PC dies. Why does this happen?

    This can happen due to several reason. Firstly, I believe its important to discuss the function of isolation. Not only is an isolator used for breaking the ground loop (separating two circuits with different ground potentials), but it can also be used hold signal integrity of the system and protect against transients from one circuit to the other circuit.

    Considering this, if there is some type of transient that occurs on the PC motherboard, debugger, or devboard, an isolator can be used to keep each side from affecting one another, but this does not mean it will protect against the short you are describing and a transient on one circuit can possibly damage the corresponding side of the isolator itself.

    From what you are describing, it seems that the short is occurring on the devboard itself. This may be something you will have to investigate and see where this short/transient is occurring on the devboard. It's important to fix the short/transient issue before we can investigate using isolation. Can you please describe this more in detail?

    Here are some helpful links regarding the above points:

    andy said:
    Does it make any difference if the AC line is coming from electric grid or from UPS? If so, what is the difference?

    Yes, there is a difference. An AC line from the electric grid is typically not as regulated as a UPS system. Considering this, UPS systems may have less variation and will be more regulated than an AC line.

    andy said:
    Would changing the SMPS in the devboard from a non-isoalted one to isolated one be enough to solve the problem?

    Yes, this may make a difference and possibly be a good idea because it can protect against variation from the AC power line, but I'm not sure if this will fix your issue. Once you share some more information on how the short/transient is occurring, I can make a better suggestion on where isolation is most suitable in your system.

    Regards,
    Aaditya Vittal

  • 0 in reply to Aaditya V.
    Prodigy 80 points

    Hi Aaditya,

    It is hard to describe it in more detail but I will try to give more context.

    A colleage was debugging a device that was self-powered through a non-isolated smps. Motherboard of his computer died. After that, company got isolation transformers for all our PCs. We plugged the isolation transformer to UPS-supplied AC-line and plugged the PCs to the isolation transformer. I had asked the senior engineer to understand why the fault occured, I serached on internet and found some application notes but I still didn't understand what the problem was. This was 4 or 5 years before. So I don't remember much and it was not my project, so I don't have much detail.

    Fast forward to today. I am also in a similar position now. I am working on a self-powered device, I am connecting to it with debugger and logic analyzer. I am not responsible for the power converter part of the device but I can ask for changes. It has been powered by debugger so far, but soon I need to power it through AC. I want to ensure safety and avoid killing things. Therefore I have been reading about isolation. There are USB isolators, there are JTAG isolators, there are isolated power converter topologies. But I am still confused about how to decide what I need to isolate.

    I may not be able to use JTAG isolation because of timing constraints.

    I may go with a USB isolator but the only reason I would go with this one is because I think something needs to be isolated. But is this actually the thing to isolate? I am still not certain.

    I still didn't understand why the solution to the problem was "powering the PCs through isolation transformers" 5 years ago. I can't call if this is the solution for my case today or not.

    Your message cleared things a little, I guess. Just to check, if I understand correctly, I have a question. So, let's assume the development board has two different GNDs, for example. Multimeter reports open-loop between GND1 and GND2. In this situation I shouldn't use one logic analyser to probe signals that are referencing different GNDs (GND1 and GND2). If I do, I would connect two seperate GNDs together through the logic analyser. If the voltage difference between two GNDs is high, a large current may flow and kill something, possibly logic analyser in this case. Is my way of thinking correct for this demonstrative theoretical situation?

  • 0 in reply to andy
    TI__Genius 12485 points

    Hi Andy,

    Please give me an additional day to respond.

    Regards,
    Aaditya V

  • 0 in reply to Aaditya V.
    Prodigy 80 points

    OK. Aaditya. I ham looking forward for your reply.

  • +1 in reply to andy
    TI__Genius 12485 points

    Hi Andy, 

    Sorry for the delay. 

    Thank you very much for the background on the issue you are experiencing. As I mentioned before, understanding where the issue originates from would be greatly helpful in making a more informed recommendation.

    From what you have described regarding past experience with including a isolation method, there may be several options on how to solve the issue. Since the device you have connected to the PC is independently powered vis a non-isolated power supply, it is possible that the power supply having a certain instability/variance may be causing transients to occur, therefore frying the PCs motherboard. I think it is worth investigating more in terms of providing isolation to the power supply, which can help stabilize and possibly prevent high transients. 

    andy said:
    I may go with a USB isolator but the only reason I would go with this one is because I think something needs to be isolated. But is this actually the thing to isolate? I am still not certain.

    This may be a solution to further add protection to the USB signal lines. The ISOUSB211 device is our USB isolator. Adding a USB isolator may protect the signal interference, if any, due to the transients. However, if the transients are from the power supply or some other component on the custom board, this may damage the isolator itself, but still protect the PC. (This may be useful: Protect your Designs with Low-Emissions USB 2.0 Compliant 480 Mbps Isolators)

    Overall, The best place to start is to first figure out what exactly is causing the issue and the above mentioned points hopefully guide you in the right direction. 

    Regards,
    Aaditya Vittal