• State TI Thinks Resolved
  • Locked Locked
  • Replies 13 replies
  • Answers 4 answers
  • Subscribers 32 subscribers
  • Views 867 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

TAS5630B: Propagation delay of class-D analog input amplifier

Zeng.
Prodigy 90 points
Part Number: TAS5630B
Other Parts Discussed in Thread: TPA3156D2, TAS5630PHD2EVM, UCC27284

Tool/software:

Hello

While a typical (audio) application doesn't require short propagation delays (and probably that's the reason why it's usually not provided in the data sheets), we want to use a class D amplifier as an AC current source and propagation delay becomes crucial in our application. There are some power stages (with PWM inputs) that do provide propagation delays, but this requires a PWM generator for a relatively high frequency (we would like to use ~1MHz switching frequency). The questions are

1) Are there analog input short propagation delay class D amplifiers working at ~1MHz? (Short propagation delay ~1usec) 

2) Are there PWM generators IC with good accuracy (<0.5%) for 1MHz and short propagation delay (~1usec)?

Thanks

  • 0
    TI__Genius 9195 points

    Hi Zeng,

    Our audio amps are voltage amplifiers. We cannot force a current.

    Regards,

    Ramsey

  • 0 in reply to Ramsey Foote
    Prodigy 90 points

    Hi Ramsey

    I don't really need an ideal current source, I shouldn't have written it. It's sufficient to be able to generate a desired voltage across the load in our application. 

    Thanks

  • 0 in reply to Zeng.
    TI__Genius 9195 points

    Hi Zeng,

    The propagation delay of a class-D amplifier is largely dependent on the LC filter on the output. Due to this, I do not see how you could use the filtered output and still have that 1usec response time you want. If you only need PWM signals, we do have some analog in class-D amplifiers that have PWM output at 1.2MHz depending on what output power you would need. However, the exact delay between input and PWM output is not data that I have access to and might not have been tested. The TPA3156D2 has 1.2MHz switching and has a phase response plot in figure 10, but this is dominated by the output filter.

    Regards,

    Ramsey

  • 0 in reply to Ramsey Foote
    Prodigy 90 points

    Hi Ramsey

    We have a few class D evaluation modules (like TAS5630PHD2EVM) and measured the delay between the input and output signals and it turns to be ~45usec at 1kHz input signal. According to the simulation of its output filter stage and our load impedance, its delay should be below 10usec at 1kHz. It seems that there are additional contributors to the delay, probably some internal digital or analog filters. My question is basically, what ICs don't have such delay, just a few usec?

    Thanks

  • 0 in reply to Zeng.
    TI__Genius 9195 points

    Hi Zeng, 

    This is not information that we have information about. I would recommend looking at some of our gate drivers as those are designed to have much lower latency.  

    Regards,

    Ramsey

  • 0 in reply to Ramsey Foote
    Prodigy 90 points

    Hi Ramsey

    My hope was that you might be able to pass this question on to someone who does know the internal design of class D ICs. I'm not looking for an exact spec, but rather an assessment for the delay and also a recommendation for ICs that have shortest delay. My guess that this delay originates in some digital processing (might be a decimation filter) and maybe there are ICs that use analog PWM generation that hopefully won't have such delay.

    Concerning your suggestion to use gate drivers, this is what we have been focusing on (more precisely PWM input amplifiers). But in such case we need a PWM generator capable of providing precise PWM input at above 1MHz and of course having a short delay. Designing one from discrete components is a significant challenge (comparing to few kHz frequencies). Using an off-the-shelf IC brings up the same question as before - what's its delay. Do you think that you can get this information about the PWM generators?

    Thanks

  • 0 in reply to Zeng.
    TI__Mastermind 20541 points

    thank you for your question.

    Ramsey will reply you after holiday.

  • 0 in reply to Zeng.
    TI__Genius 9195 points

    Hi Zeng,

    I have asked a few people in the department and no one has this documentation. As you noted before, audio amps are not designed for this and as a result are not tested in this way. I would recommend talking to someone on gate drivers team or converters team. If you are designing a power supply, you should use power supply parts. I have changed the responsible org on this thread to a gate driver group.

    Regards,

    Ramsey

  • 0 in reply to Ramsey Foote
    TI__Genius 14145 points

    Hey Zeng,

    I am one of the applications engineers for our gate drivers. For Class D Audio, a half-bridge gate driver such as the UCC27284 may be a potential candidate. It has propagation delay of 16ns and can switch at 1MHz.

    Let me know if you have further questions.

    Thank you,

    William Moore

  • 0 in reply to William Moore
    Prodigy 90 points

    Hello William

    Thanks for the response, unfortunately it doesn't really help, as this (or any similar) driver requires a PWM generator that most probably will be the one responsible for the delay, unless you can recommend a PWM generator that has negligible (<1us) delay. UCC27284 has also a few more disadvantages, like no integrated switches.

    Thanks 

  • 0 in reply to Zeng.
    TI__Genius 14145 points

    Hey Zeng,

    For a PWM generator, you will need to look into a controller or microprocessor. Also for something with integrated FETs, the motor drives group has some options that are similar to the gate drivers but with the half-bridge power stage internal to the IC.

    Thank you,

    William Moore

  • 0 in reply to William Moore
    Prodigy 90 points

    Hi William

    You don't believe that there can be an analog PWM generator for 1MHz? Do you think that a relatively cheap MCU will be capable of generating PWM signal with high accuracy (<0.5% pulse width error)?

    Thanks 

  • 0 in reply to Zeng.
    TI__Genius 14145 points

    Hey Zeng,

    Yes, there may be an analog PWM generator for 1MHz or a cheap MCU that is capable of that. But, I am the expert on gate drivers so I can't recommend any products for that. If you find something similar on TI.com, you can open a thread with that product and its respective team will help guide you to a correct product.

    Thank you,

    William Moore