• State TI Thinks Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 62 subscribers
  • Views 295 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.

UCC23511-Q1: How can I speed up the Rise time of the Gate Driver output (Vout)?

Saad Namou
Prodigy 10 points
Part Number: UCC23511-Q1
Other Parts Discussed in Thread: UCC23511, UCC23513EVM-014, UCC23513, UCC23525, UCC5390, UCC5350

Tool/software:

Hi,

I designed a board using the UCC23511-Q1 Gate driver.  The input is +5V pulses at 8KHz (20% Duty Cycle).  The output Isolated voltage supplies are +17.5V VCC and -5.6V VEE (rated for 1.25A).  The output Vout is driving a very high power (1.7KV & 800A) IGBT (CM800DZ-34H - Mitsubishi Electric).  The final application would have a -800 VDC at the Emitter of the IGBT that would be passed as pulses to the Collector of the IGBT at 8KHz (20%D.C.). 

I ran this test on my LAB bench, but instead using -800VDC, I used -28VDC supplied to the IGBT Emitter.  keep in-mind that the input pulses of the Driver (UCC23511) have a 15ns Rise time, but the measured Rise time of this Driver Output (Vout) was 640ns!!! the data sheet states that the max Vout Rise time is 28ns.  To be fair if I disconnect the IGBT, the Driver Output rise time drops to 80ns, which is still slow. I also noticed that when I connect the Gate of the IGBT to the Driver output, the Driver Output (Vout) rises from 0V up to 5V within 40ns, but then from 5V up to 15V the rise time took 600ns!!!

I bought the Eval board for a similar Driver (UCC23513EVM-014) which has higher driving current, but I got almost the same results. So it can not be my PCB layout.  Then I used an independent large bench power supply to generate the +17.5Vdc and -5.6Vdc, but still I got the same results.

Please can you tell me how can I speed up the Rise time of the Gate Driver output (Vout)?  I need the Rise time to be faster than 28ns (datasheet max spec)...

Thanks,

-Saad

  • 0
    TI__Genius 17486 points

    Hi Saad,

    To supply 6600nC in 28ns, you would need 235.7A. This is impractical. The datasheet rise time is for an unloaded output only.

    Peak drive current is likely a limiting factor with the UCC23511, which is is optimized for low cost and low output current. Are you sure that the UCC23513 did not result in a faster rise time? It does have a 4.5A vs. the 1.5A drive current for the UCC23511. In your schematic above, you have a few extra output loads in parallel that you can try removing to improve your rise time; the LED and the 2.2nF likely are siphoning off some drive current.

    You could try the new UCC23525, but that is the strongest driver with this package footprint.

    If you can use a different package, you should use the UCC5390 to drive this IGBT. Alternatively, you can use a BJT totem pole to buffer the output of the UCC23511 to increase drive strength.

    Finally, you can increase the gate drive voltage to boost drive strength. 20V will charge the gate faster than 17.5V. 

    Best regards,

    Sean

  • 0
    Prodigy 10 points

    Hi Sean,

    I tried everything with some speed improvement but not enough.  Finally I replaced the ISO Driver with UCC5350SBDR‎ (5A/5A) and added MOSFET totem pole (RQ6E040XNT / RQ6E030ATT) on the output of the ISO Driver to Drive my IGBT Gate.  Now I do get the 30ns rise/fall time switching speed that I am looking for.

    I am not sure if  these P-ch and N-ch MOSFET are hefty enough, they are rated 3A and 4A!

    Also why did you suggested a BJT totem pole, is it because BJT has higher current gain?

    Also would the UCC5390 Driver have given me even more switching speed?

    Thanks,

    -Saad

  • 0 in reply to Saad Namou
    TI__Genius 17486 points

    Hi Saad,

    The DC current rating of 3A and 4A is a steady state power rating. You can get much higher short circuit current (Id sat) out of discrete buffers than this.

    A BJT is non-inverting, which makes the signaling easier. Also, they can be smaller for the same output saturation current. But you can use Mosfets if they are cheaper. Smaller Mosfets might have a Vgs limitations, so you can't just use the gate driver to drive the buffer inverter directly.

    UCC5390 has 2x the output stage W/L vs. the UCC5350. It should give you some improvement, although experimentally I have found that the benefit is only significant for very large capacitor loads. Real FETs don't show that much improvement.

    Best regards,

    Sean