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.

TCA9517: TCA9517 WITH ADUM1250 WITH DRV10983

Part Number: TCA9517
Other Parts Discussed in Thread: PCA9517, DRV10983, TCA9509, , ISO1540, TCA4311A, PCA9306, TCA9406, TXS0102, LSF0102

I2C MASTER SIDE Run At 5V, This is on borad A, 1scl and 1sda are the master,sclo1 and sdao1 are connect to board B.

I2C SLAVE SIDE, This is on borad B,sclo1 and sdao1 are connect to board A,

1scl and 1sda are the SIDEA 1 of ADUM1251, AND connect to PCA9517 SIDE B,

3scl and 3sda are connect to  DRV10983 Run at 3.3V。

I wonder if there is some problem in the schematic,

can someone help me to check this ?

  • Hey User,

    The issue is that ADUM1250 uses a VoL stated in their datasheet of around 600~900mV and is connected to our B side of of 9517 device (assuming its TI since we do not use PCA9517). The B side of our device does not recognize anything above ~400mV as a low due to the ViLc spec stated in our datasheet. You essentially have two static voltage offset devices connected to each other which will not work.

    [Note our device states it cannot be used with other buffers with static offsets if B side is connected to the other offset buffer at the static voltage offset side. Page 10 of the ADUM device hints to this as well.]

    The only way to fix this is to either respin the board or try to swap out the TCA9517 with a TCA9509 and remove and pull up resistors from the A side of the TCA9509. An alternative would be to look for an I2C isolator which uses a static voltage offset on the opposite side in reference to the ADUM1250.

    -Bobby

  • As the borad A can NOT be changed,

    I am considering how to change board B.

    Do you mean I shoud remove the ADUM1250 of board B,

    and connect the side2 of adum1250 ,the sclo1 and sdao1,

    directly to TCA9517,

    But which side shoud work ?

    ASIDE

    BSIDE

    @ swap out the TCA9517 with a TCA9509 and remove and pull up resistors from the A side of the TCA9509

    Do you mean this ?


    @look for an I2C isolator which uses a static voltage offset on the opposite side in reference to the ADUM1250.

    I don't understand this,

    Do you mean I should replace the ADUM1250 with another IC,

    Can you give me an example of this kind of IC ? 

  • user6011360 said:

    As the borad A can NOT be changed,

    I am considering how to change board B.I'm unsure from your schematic which board is A and B, did you label them side A and B for board A and board B?

    Do you mean I shoud remove the ADUM1250 of board B, My suggestion is to make it such that the I2C buffers do not have static voltage offset sides touching each other.

    and connect the side2 of adum1250 ,the sclo1 and sdao1,

    directly to TCA9517, B side of TCA9517 has the static voltage offset.

    But which side shoud work ? Side 1 of ADUM has the static voltage offset. So B to side 1 does not work.

    ASIDE

    BSIDE

    @ swap out the TCA9517 with a TCA9509 and remove and pull up resistors from the A side of the TCA9509

    Do you mean this ? TCA9509 looks like it may work with the ADUM device in this configuration but I am unsure where the '3SDA/3SCL' net is going towards. The A side of the TCA9509 should not connect to another device on a static voltage offset side.


    @look for an I2C isolator which uses a static voltage offset on the opposite side in reference to the ADUM1250. Side 1 of the ADUM has a static voltage offset. If you need an I2C isolator device then finding a pin to pin replacement for this device with the static voltage offset on side 2 would fix the problem.

    I don't understand this,

    Do you mean I should replace the ADUM1250 with another IC, Yes, if you can find one that flips the static voltage offset.

    Can you give me an example of this kind of IC ? I am not aware of one, LTC/ADUM don't look like they offer one for side 2 and same goes for our ISO1540. I don't see any I2C power isolators for NXP either.

  • @I'm unsure from your schematic which board is A and B, did you label them side A and B for board A and board B?

    Yes,I motioned it in the first level post.

    B side of TCA9517 has the static voltage offset.

    Side 1 of ADUM has the static voltage offset. So B to side 1 does not work.


    So I shoud have a test to connect A to side 1 ?

    But what about the Supply restrictions VCCA <= VCCB,

    I want a level shift from 5v to 3.3v.

     

    TCA9509 looks like it may work with the ADUM device in this configuration but I am unsure where the '3SDA/3SCL' net is going towards. The A side of the TCA9509 should not connect to another device on a static voltage offset side.

    The 3SDA/3SCL net go directly to DRV10983. I have ordered four  TCA9509 chip AND will have a real test later.

    I want to ask, can I connect more than one device to the A side of the TCA9509? with or without an I2C isolators ?

    If the EN pin has an interal pullup in the chip as same as TCA9517 ?

     

    @I am not aware of one, LTC/ADUM don't look like they offer one for side 2 and same goes for our ISO1540. I don't see any I2C power isolators for NXP either.

    So we should not talk about a thing that have never existed.

     

    I found this on LTC site,I want to ask if this can be used with ADUM1251 on the left side, If you have time to have a look.


     

    Does TI have an similar device with a Rise time accelerators (RTAs) as LTC calles it。

    will RTAs help in this kind of configuration ?

  • My order of TCA9509 arrived,

    And I have a real test yestoday ended up with a failure.

    I think TCA9509 maybe used with something else AT 3.3V,

    BUT NOT DRV10983,

    DRV10983 I2C input low is 0.6V,

    TCA9509 I2C input low is 0.3V.

    This may be the problem.

    TCA9509

    DRV10983

    Can you help to explain more details of DRV10983 I2C port ?

    how low can this chip drvie the SCL/SDA to ?

    In TCA9509 datasheet,

    Does this mean TCA9509 also need pull-up resistance ?

  • "So I shoud have a test to connect A to side 1 ?

    But what about the Supply restrictions VCCA <= VCCB,

    I want a level shift from 5v to 3.3v."

    The TCA9509 does have voltage level restrictions (I overlooked this) so B side is meant to be larger than A side by atleast 1 V. A side of TCA9509 will not work due to the ViLc requirement.

    "I want to ask, can I connect more than one device to the A side of the TCA9509? with or without an I2C isolators ?"

    The only limitation on A side of this device is it wants to see a low voltage lower than the ViLc in order to pass a low from A side to B side.

    "If the EN pin has an interal pullup in the chip as same as TCA9517 ?"

    No, this device does not have an internal pull up on the enable pin.

    "I found this on LTC site,I want to ask if this can be used with ADUM1251 on the left side, If you have time to have a look."

    The LTC devices which use RTAs typically utilize dynamic offsets instead of static voltage offsets. This has its advantages and disadvantages. If you connect this to side 1 of the ADUM, it will not work because the LTC device wants to see a low below 400mV and will pass lows with a shifted VoL. I don't think this will work on side 1 of ADUM.

    "Does TI have an similar device with a Rise time accelerators (RTAs) as LTC calles it。

    will RTAs help in this kind of configuration ?"

    TI has the TCA4311A however I am skeptical this device will work with side 1 of ADUM.

    "Can you help to explain more details of DRV10983 I2C port ?"

    The 0.6V ViL is the INPUT low threshold which our device will be below. I do not see any VoL specs (the one we care about) in the datasheet for the DRV device.

    "Does this mean TCA9509 also need pull-up resistance ?"

    B side needs a pull up resistor but A side does not. It means if you use a pull up resistor on A side, you will make it harder for the device to recognize a low on A side because the pull up resistor supplies additional current which will shift up the Ron/VoL of any device on A side.

    Since you said you cannot change Board A but can adjust Board B, can you just flip the ADUM device on board B? Make it so side 1 connects to side 2 Board A? If you keep the ADUM device on Board B unchanged, you will never be able to make this work because side 1 from board A and side 1 from Board B are connected to each other (this is against the rules set by the ADUM device). After you flip the ADUM device on board B, you can then connect TCA9517 side A to side 2 of the ADUM device.

    -Bobby

  • @The TCA9509 does have voltage level restrictions (I overlooked this) so B side is meant to be larger than A side by atleast 1 V. A side of TCA9509 will not work due to the ViLc requirement.

    So TCA9509 will NOT work at all.

    @If you connect this to side 1 of the ADUM, it will not work because the LTC device wants to see a low below 400mV and will pass lows with a shifted VoL. I don't think this will work on side 1 of ADUM.

    "the LTC device wants to see a low below 400mV and will pass lows with a shifted VoL."

    where did you see this ?

    the upper picture of datasheet of ADUM1250 says that should be about 500~700mv.

    @The 0.6V ViL is the INPUT low threshold which our device will be below. I do not see any VoL specs (the one we care about) in the datasheet for the DRV device.

    Can you ask the DRV team about this,or have a simple test ?

    @B side needs a pull up resistor but A side does not. It means if you use a pull up resistor on A side, you will make it harder for the device to recognize a low on A side because the pull up resistor supplies additional current which will shift up the Ron/VoL of any device on A side.

    But ,what doest this section mean?

    @can you just flip the ADUM device on board B? Make it so side 1 connects to side 2 Board A? After you flip the ADUM device on board B, you can then connect TCA9517 side A to side 2 of the ADUM device.

    Yes,I will have a try.

    And I want to ask, what if I just remove the ADUM device on board B ?

     @ If you keep the ADUM device on Board B unchanged, you will never be able to make this work because side 1 from board A and side 1 from Board B are connected to each other (this is against the rules set by the ADUM device)

    side 2 of board A and side 2 of board B  are connected to each other, this is the recommanded usage taken from AN913 of LTC,

    And this works well with other chips that run at 5V.

    you can have a look at this file at this URL 

  • user6011360 said:

    @The TCA9509 does have voltage level restrictions (I overlooked this) so B side is meant to be larger than A side by atleast 1 V. A side of TCA9509 will not work due to the ViLc requirement.

    So TCA9509 will NOT work at all.

    I do not believe so, or atleast it will not meet datasheet specs if you use it outside of it's intended

    @If you connect this to side 1 of the ADUM, it will not work because the LTC device wants to see a low below 400mV and will pass lows with a shifted VoL. I don't think this will work on side 1 of ADUM.

    "the LTC device wants to see a low below 400mV and will pass lows with a shifted VoL."

    where did you see this ?

    It is hidden in the text of the LTC4300A device (their generic RTA device)


    the upper picture of datasheet of ADUM1250 says that should be about 500~700mv.

    ADUM side one has 600~900mV of offset which you can see in the same picture you posted right below what you pointed out. 600mV~900mV output lows typically will not work with most devices which use RTAs (Rise time accelerators) because the RTAs have a minimum input low voltage they must see otherwise their logic could get disrupted.

    @The 0.6V ViL is the INPUT low threshold which our device will be below. I do not see any VoL specs (the one we care about) in the datasheet for the DRV device.

    Can you ask the DRV team about this,or have a simple test ?

    I can notify them about this post.

    @B side needs a pull up resistor but A side does not. It means if you use a pull up resistor on A side, you will make it harder for the device to recognize a low on A side because the pull up resistor supplies additional current which will shift up the Ron/VoL of any device on A side.

    But ,what doest this section mean?

    It means you MUST ensure whatever device is on A side has to be able to pull the voltage lower than ViLc of the device. If you add a pull up resistor, you make this harder to do. Any device that is I2C compliant should be able to do this easily.

    @can you just flip the ADUM device on board B? Make it so side 1 connects to side 2 Board A? After you flip the ADUM device on board B, you can then connect TCA9517 side A to side 2 of the ADUM device.

    Yes,I will have a try.

    And I want to ask, what if I just remove the ADUM device on board B ?

    It depends why you are using the ADUM device to begin with. My understanding is that the ADUM device is for power isolation and also helps with any potential GND shifts due to high current loads in your system. If you have common GND between the two boards and run the device off the same Vcc rails, then there is no point to using an I2C isolator. You you are concerned with GND shifts or require power isolation then you need the ADUM device.

     @ If you keep the ADUM device on Board B unchanged, you will never be able to make this work because side 1 from board A and side 1 from Board B are connected to each other (this is against the rules set by the ADUM device)

    side 2 of board A and side 2 of board B  are connected to each other, this is the recommanded usage taken from AN913 of LTC,

    And this works well with other chips that run at 5V.

    you can have a look at this file at this URL 

    This is correct. Side 2 of each ADUM is connected to each other and none of the Side 1s are connected together. In your previous schematic, it looks like you have side 1 <--> side 1 which is my concern.

  • @It is hidden in the text of the LTC4300A device (their generic RTA device)

    I think I have never talked about LTC4300A,

    I am talking about LTC4313 and LTC4315

    They say they can deal with devices that have a high Vol.

  • Sorry for the delay in response,

    I took a look at the device and it looks like the ViL can accept side1's VoL but it uses a dynamic offset voltage so if side1's voltage is 900mV worst case then the LTC4315's VoL may be as high as 1080mV for the unsaturated case. If your Vcc is 3.3V then ViL is 0.99V (according to I2C standard) then you would have a VoL larger than the ViL spec'd in I2C. I'm also not sure what would happen to the LTC device if you went above it's ViL spec since the output voltage in this case could be larger than the allowed ViL of the device (not ideal).

    If the VoL generated by the device is okay with you, then you may be able to use this device (I would ask one of the LTC engineers about this though).

    -Bobby

  • @Can you ask the DRV team about this,or have a simple test ?

    I can notify them about this post.

    Have you get some responses ?

    @If the VoL generated by the device is okay with you, then you may be able to use this device (I would ask one of the LTC engineers about this though).

    Have you get some responses ?

  • Hey User,

    "I can notify them about this post.

    Have you get some responses ?"

    I did tag them in that previous email but I think what happened is when you asked a following question after my post and I replied, the group responsibility (an internal thing) was moved back to me because I replied. I'll retag them  again so they can be notified. (You could also make another post in the motor drive section to ask about the VoL of the device).

    "@If the VoL generated by the device is okay with you, then you may be able to use this device (I would ask one of the LTC engineers about this though).

    Have you get some responses ?"

    You will have to reach out to them about this, I'm not so sure they will tell me because we are a direct competitor to them.

    -Bobby

  • Bobby,

    Please summarize the latest and any pending questions.

    Regards,

    -Adam

  • https://e2e.ti.com/support/motor-drivers/f/38/t/832238

    They recommanded PCA9306,

    And I find TCA9406,

    I want to ask which is better for DRV10983 ?

  • The PCA9306 and TCA9406 use a pass FET architecture so it will not redrive signals (remember ADUM side 1 generates a static voltage VoL which we want to be lower than ViL of devices on side 1).

    PCA9306 and TCA9406 are strictly for level shifting purposes and not trying to make a large low signal lower.

    The customer is looking for the VoL of DRV10983 (VoL for SDA @IoL =3mA)

    -Bobby

  • BOARD B can be changed,

    If PCA9306 or TCA9406 Can connect ADUM1250 SIDE 2 AND DRV10983 and works fine,

    I can remove the ADUM1250 on BOARD B,

    I will be very happy.

    can you give me an DIRECT answer ?

  • I²C translators with active buffers (i.e., both the ADuM125x and the TCA9517) must use a static voltage offset (or similar mechanism) on at least one side, and those pins must not be connected to another device that uses the same mechanism. Side 1 of the ADuM and side B of the TCA have this problem, and the board has connected them.

    Exchanging side A and side B on the TCA9517 should make it work.

    The TCA9406/‌TXS0102 is not recommended when you have a cable with nontrivial capacitance, because too much ringing can make the edge accelerators trigger erroneously, and make it oscillate.

    The PCA9306/‌LSF0102 has no edge accelerators, and should work.

    With either TCA9517 or PCA9306/‌LSF0102, the second ADuM on board B should not be necessary.

    For more information about the PCA9306/‌LSF0102, see Voltage-Level Translation With the LSF Family (SLVA675) and Factors Affecting VOL for TXS and LSF Auto-bidirectional Translation Devices (SCEA056).

  • @The TCA9406/‌TXS0102 is not recommended when you have a cable with nontrivial capacitance, because too much ringing can make the edge accelerators trigger erroneously, and make it oscillate.

    What do you mean by "too much ringing" ?

    @With either TCA9517 or PCA9306/‌LSF0102, the second ADuM on board B should not be necessary

    So, the problem can be solved by just removing the ADUM1250 on board B ?

  • The TXS edge accelerator bypasses the internal 10 kΩ pull-up resistor when it detects the beginning of a rising edge. This happens at about 30 % of VCC from what I've heard. When a falling edge is followed by ringing that goes that far, then the TXS will drive the line high. See this example.

    Yes, removing the ADuM1250 (or the TCA9517) should work; you need only one level shifter.