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.

UCC27211: 400ns-off output when 40ns-off input

Part Number: UCC27211
Other Parts Discussed in Thread: UCC27282,When HI input signal is 40ns off, I assume HO output is 40ns off or nearly 40ns off. But, I found that HO output is often very long off (400ns off). Usually, HO output is 40ns off. Is there a way to get rid of this random very long off? Datasheet 7.9 says, minimum input pulse width that changes the output is MAX 50ns. When input pulse is less than 50ns, dose this IC work wrong?
  • I cannot find Rich Text Editor! I cannot upload my oscilloscope waveforms!
  • Hi, TkMk,

    Sorry you are having difficulties.

    Maybe this presentation might help? I don't read Japanese, so I don't know, but see page 13... 

    https://e2e.ti.com/support/site-support/f/1024/t/896945

    You can also set your default in your e2e settings to default to the Rich Text Editor. Click on your picture/name in upper right hand corner, settings, and then select "Enhanced" as your default editor.

  • Hello TkMk,

    I see a previous comment about not being able to upload the waveforms. It will help to see the waveforms to see the behavior you are describing. In my E2E environment I see the option of attaching files and also inserting pasted material from Word, which is the W. Usually one of those options should work to attach image files.

    Can you provide information regarding the operating conditions such as VDD voltage, operating frequency and the MOSFET parameter Qg?

    Regards,

  • Hello Don,

    Thanks for your reply. Now I am using Rich Text Editor.

    I changed "text" to "HTML" in my e2e settings.

    Regards, TkMk

  • Hello Richard,

     

    Here are my waveforms.

    Ch5, orange color, is HI-VSS

    Math2, purple color, is HO-HS

    Ch6, blue color, is LI-VSS

    Ch8, green color, is LO-VSS

     

    The VDD voltage is about 15V – 19V.

    The operating frequency is not stable, since these are pulse-generator shutting-down waveforms.

    The MOSFET Qg is 35nC max.

     


     


  • Hello TkMk,

    Thank you for the waveforms. I do see the behavior you are questioning. It is hard to see what exact pulse widths of the negative going HI input are showing the difference in the HO response to the shorter width low to the wider width low pulses in question.

    Is there some way you can get a pulse width indication of the HI showing the expected HO response and the longer HO low response.

    I can see the HI looks a little shorter when it looks like the HO goes low for a longer period.

    This may be getting close to the minimum pulse response mentioned in the datasheet. Is this behavior an issue in your system, or more of a question of why this is happening.

    There is a newer 100V driver which does have specifically improved narrow pulse response, although the datasheet parameters look similar. The UCC27282 does have very good narrow input pulse response which you may want to consider,

    Regards,

  • Hello Richard,

     

     

    I attached the numerical data (xlsx file). It is 1.6ns sampling, you can see the exact pulse widths.

    Do you find any clue?

     

    This IC behavior may be an issue in my system. When power MOSFET switches, the IC may get short impulse noise. If this IC outputs long-off, there will be PWM error. If this IC outputs long-on, there will be MOSFET shortcircuit.

     

     

    Regards, TkMk

    TkMk_waveforms_data.xlsx

  • Hello TkMk,

    Thank you for the numeric plot data.

    But I do have some questions, it is not obvious to me to see what is going on.

    Is the Math2 the voltage on the HO-HS? It looks like 19.2 to 19.5 on many entries.

    What is the VDD in this application, make sure VDD and HB-HS is within the datasheet recommended range.

    The last columns M I assume this is 1 when there is a difference in the HI and HO-HS column, to highlight where we should look, although there will be a difference during the propagation delay. I am not sure what the Filtering M is, does this flag the long time durations of the HO-HS low compared to HI low?

    Also I want to confirm the plots you attached ealier showing the normal HO-HS low pulses and the long HO-HS pulses are at 200ns per division. Can you confirm?

    Regards

  • Hello Richard,

     

     

    At both image and xlsx, Math2 is ch3 - ch4, that is HO - HS.

    HO-HS is about 19V, it is true that out of datasheet recommended range.

    Usually, VDD is 15V. But it seems to be temporary 19V when shutting-down.

    It is under absolute maximum rating, I hope the IC still works.

     

    XLSX filtering M is just for people to find long time difference between HO-HS and HI.

    It is actually summing column M. When there is long time difference, the number will be large.

     

    You see 4 plot images. First and second are 5us per division. Third and fourth are 100ns per division.

     

    Regards, TkMk

  • Hello TkMk,

    It is not very straightforward to determine the couple of items I would like to confirm, I did look at the strings of data and looking at the M column to determine the extended HO-HS low amplitude.

    Can you point out the HI narrow low pulse width where the HO-HS response looks as expected, and also the HI narrow low pulse width where the HO-HS has the extended low level pulse width. I would like to confirm if the pulse widths are getting narrower when the output response does not look correct, or wider in pulse width (low going).

    Regards,

  • Hello Richard,

     

     

    I added columns and plots to xlsx.

    I found that,

    the HI narrow low pulse width where the HO-HS response looks as expected

    is more than 28ns,

    the HI narrow low pulse width where the HO-HS has the extended low level pulse width

    is less than 26ns.

    So, the input pulse width is narrow when the output response does not look correct.

     

    (2020-07-23 and 2020-07-24 are Japan national holidays. My next reply will be on 2020-07-27.)

     

    Regards, TkMk

    TkMk_plots_added.xlsx

  • Hello TkMk,

    Thank you for confirming the pulse widths where the HO operation is as expected, and there is extended low pulses.

    You may want to consider testing the UCC27282 which does have improved narrow pulse response, the datasheet can be found here: https://www.ti.com/lit/ds/symlink/ucc27282.pdf?ts=1595442568623&ref_url=https%253A%252F%252Fwww.ti.com%252Fsitesearch%252Fdocs%252Funiversalsearch.tsp%253FsearchTerm%253Ducc27282

    Is the response to the narrow pulses such as 24ns required, or is thin only in the case of possible input noise concerns?

    If this narrow response is not required, and it may be a noise concern, you should consider adding a small R/C filter on the driver input pins.

    Regards,

  • Hello Richard,

     

     

    I have to modify the title. It is 400ns-off output when 24ns-off input.

    In my case, narrow response is not necessary. So, I will make UCC27211 not to receive narrow pulse or narrow noise.

     

    My final questions are:

    What I found is HI narrow-off pulse. How about the HI narrow-on pulse or LI narrow-off pulse or LI narrow-on pulse? Does extended output pulse happen in these cases?

     

    Is the datasheet value “minimum input pulse width that changes the output” shows the limit that the extended output pulse never happen?

    I saw the value is 50ns for UCC27211, 20ns for UCC27282.

     

     

    Regards, TkMk

  • Hello TkMk,

    When I inquired about the possible response of the UCC27211 the behavior that had been confirmed was regarding the narrow positive pulses but the discussion confirmed that narrow negative pulses will have similar behaviors.

    The UCC27282 does have a improved specification and actual behavior is improved over many half bridge drivers. The specification implies in one regard that the output will actually respond to the narrow input pulses but in practice also includes the output being close to the input pulse width.

    If you can avoid the narrow input pulses which I thought would be the case in your application, the UCC27211 will be OK for your operation. If you want to explore a device with better narrow pulse behavior I recommend the UCC27282.

    Confirm if this addresses your questions, or you can post additional questions on  this thread.

    Regards,

  • Hello Richard,

     

    I found the same post unintentionally. 

    UCC27211 Issue with Short Pulse Widths

    e2e.ti.com/.../625040

     

    Your answers and this past post resolved my issue. Thanks a lot.

     

    Regards, TkMk