• AVR Freaks

Hot!No new audio chip ?

Author
Jan Audio
Starting Member
  • Total Posts : 53
  • Reward points : 0
  • Joined: 2018/09/24 08:12:24
  • Location: 0
  • Status: offline
2018/09/30 08:57:52 (permalink)
1 (1)

No new audio chip ?

Hi, i am using the DSPIC33FJ128GP802 for the audio DAC.
This DAC is to noisey, when will there be a new chip for audio without the noise ?
I know i can buy a external audio DAC + ADC, only they are not in DIP avaiable.
I also would like 16 bit audio ADC on board.
Instead of 40 MIPS i would like 400 MIPS.
All in DIP package please.
How long does it take ?
 
thanks
#1

11 Replies Related Threads

    du00000001
    Just Some Member
    • Total Posts : 2781
    • Reward points : 0
    • Joined: 2016/05/03 13:52:42
    • Location: Germany
    • Status: online
    Re: No new audio chip ? 2018/09/30 09:59:54 (permalink)
    0
    400 MIPS in a DIP package? Maybe forever. (Answering: "How long does it take ?")
     
    Obviously you know little about the physical limitations of semiconductor devices, not to mention things like propagation delays (and all other sorts of delays), crosstalk etc.
    Might be much easier to denoise the "noisy" DAC available...

    PEBKAC / EBKAC / POBCAK / PICNIC (eventually see en.wikipedia.org)
    #2
    Jan Audio
    Starting Member
    • Total Posts : 53
    • Reward points : 0
    • Joined: 2018/09/24 08:12:24
    • Location: 0
    • Status: offline
    Re: No new audio chip ? 2018/10/01 02:47:11 (permalink)
    0
    Thanks.
    Why does STM32 have 400MHz in even a smaller package ?, dont they have the same limitations ?
    #3
    du00000001
    Just Some Member
    • Total Posts : 2781
    • Reward points : 0
    • Joined: 2016/05/03 13:52:42
    • Location: Germany
    • Status: online
    Re: No new audio chip ? 2018/10/01 03:07:16 (permalink)
    0
    400 MHz only works from a cache or eventually a very fast RAM - never from flash.
    Once there is a cache miss, these MCUs are slowing down significantly. (Once you read the fine print, you can discover hints about this).
    And branching makes them slump (due to long pipelines).
     
    While ST states "maximum theoretical performance of the Cortex-M7 core" when executing from flash or external memory, I somewhat doubt that due to physical limitations (and expect this to be only valid for cached code executing linear code).
     
    Generally, smaller packages have some advantage due to shorter signal lines (bond wires etc.) - at least when it comes to accessing external memory.

    PEBKAC / EBKAC / POBCAK / PICNIC (eventually see en.wikipedia.org)
    #4
    Jan Audio
    Starting Member
    • Total Posts : 53
    • Reward points : 0
    • Joined: 2018/09/24 08:12:24
    • Location: 0
    • Status: offline
    Re: No new audio chip ? 2018/10/01 07:48:13 (permalink)
    0
    These chips are all flash right ?
    Is that not always executed from flash then ?
    #5
    Alpha Whisky
    Super Member
    • Total Posts : 51
    • Reward points : 0
    • Joined: 2014/10/02 07:05:22
    • Location: 0
    • Status: offline
    Re: No new audio chip ? 2018/10/01 08:56:13 (permalink)
    0
    If you're seeing a SNR much worse than 90dB from that device then it is most likely down to your circuit and/or board layout, not the dsPIC, and a different DAC wouldn't produce much improvement without a better board layout.
     
    Given that you're using a DIP package are you using stripboard or breadboard? Fairly decent audio can be achieved that way, but it's a bit of a black art. Good results are easier to achieve with a well laid out multilayer board with at least one ground plane.
    #6
    Jan Audio
    Starting Member
    • Total Posts : 53
    • Reward points : 0
    • Joined: 2018/09/24 08:12:24
    • Location: 0
    • Status: offline
    Re: No new audio chip ? 2018/10/01 09:19:26 (permalink)
    0
    No, i build many synthesizers, the noise is from the DAC, others have the same problem :
    https://www.microchip.com/forums/m479331.aspx
    #7
    du00000001
    Just Some Member
    • Total Posts : 2781
    • Reward points : 0
    • Joined: 2016/05/03 13:52:42
    • Location: Germany
    • Status: online
    Re: No new audio chip ? 2018/10/01 09:58:45 (permalink)
    0
    The ST Arms have comparably large RAMs as Arm code that shall execute at highest speed has to either be cached or copied to RAM prior execution.
    One of the many mysteries of the Arm implementations. Even (uncached) nom. 120 MIPS Arms from NXP (stating 0 wait states) start to falter when branching, calling subroutines etc. (These chips have/had flashed that couldn't exceed 30 MHz.) Be careful if you find a "128 bit wide flash interface": this is one of the signs for inappropriate (slow) flash architectures.
    Looking at the 'H7 RAM architecture, you should notice that the overall RAM is somewhat partitioned into "special function blocks": not every RAM cell is the same.

    PEBKAC / EBKAC / POBCAK / PICNIC (eventually see en.wikipedia.org)
    #8
    du00000001
    Just Some Member
    • Total Posts : 2781
    • Reward points : 0
    • Joined: 2016/05/03 13:52:42
    • Location: Germany
    • Status: online
    Re: No new audio chip ? 2018/10/01 10:08:11 (permalink)
    4 (1)
    Re: noise
    You've got an ENOB of about 9 - which is what you may expect from a mixed-signal chip. (About the same applies for the ADCs and the non-audio DACs.) This is the price to be paid for having a multitude of contradicting requirements (fast flash with good data retention vs. fast digital logic vs. precise analog periphery vs. ...)  somewhat fulfilled by a single production process.
    If you need beyond 9 effective bits, attach an external DAC by means of SPI!
     

    PEBKAC / EBKAC / POBCAK / PICNIC (eventually see en.wikipedia.org)
    #9
    Jan Audio
    Starting Member
    • Total Posts : 53
    • Reward points : 0
    • Joined: 2018/09/24 08:12:24
    • Location: 0
    • Status: offline
    Re: No new audio chip ? 2018/10/02 02:14:34 (permalink)
    0
    Effective number of bits ?, this DAC is not a 9 bit DAC,
    it is a 16 bit DAC, with 14 bits accuracy.
    Is there something i dont understand ?
     
    + there is a special capacitor for the DAC on pin 27 & 28, is that not enough ?
    #10
    JPortici
    Super Member
    • Total Posts : 692
    • Reward points : 0
    • Joined: 2012/11/17 06:27:45
    • Location: Grappaland
    • Status: offline
    Re: No new audio chip ? 2018/10/02 02:44:29 (permalink)
    0
    Jan Audio
    Is there something i dont understand ?



    Yes. Resolution != ENOB.
    How did you measure that it has 14 bit of accuracy? the datasheet only mentions 61dB SNR which translates to 9.84 ENOB
     
    Given that i can't find any graphs or other parameters on the DAC i wouldn't expect it to be any better
     
    EDIT: Yes, i see that in the DAC chapter it mentions 14 bit and 90dB SRN but the electrical characteristics at the end of the datasheet don't agree with that
    post edited by JPortici - 2018/10/02 02:46:08
    #11
    du00000001
    Just Some Member
    • Total Posts : 2781
    • Reward points : 0
    • Joined: 2016/05/03 13:52:42
    • Location: Germany
    • Status: online
    Re: No new audio chip ? 2018/10/02 03:00:43 (permalink)
    0
    There's something you do not understand...
     
    The ENOB is derived from the SINAD ( which is <= SNR - 61 dB @ 96 kSPS (Table 30-46) ): see
    https://en.wikipedia.org/...e_number_of_bits  for a discussion of this.
     
    For silicon revisions up to 3002, not even this value is reached.
     
    Delta-sigma converters are bitches.
    And it doesn't help either that you've got 3 output signals from a single converter - with a differential voltage <= (+/-) 2 V. Adding a low-pass and/or reducing the sampling rate might somewhat improve the situation. Beyond this, a precise analysis of the noise by means of a spectrum analyser might give some indications on how to reduce the noise.
     
    And no - the capacitor on the analog supply is not "enough". It's more like a minimum requirement, but not necessarily capable to guarantee the datasheet specifications as PCB layout, power supply "quality" and alike have a significant influence.

    PEBKAC / EBKAC / POBCAK / PICNIC (eventually see en.wikipedia.org)
    #12
    Jump to:
    © 2019 APG vNext Commercial Version 4.5