Page 17 of 17 FirstFirst ... 7891011121314151617
Results 161 to 164 of 164
Like Tree1Likes

Thread: Interested in a USB Audio Amplifier?

  1. #161
    Low Bitrate
    Join Date
    Nov 2002
    Atlanta, GA, USA
    As this thread title is fairly generic, there haven't been any updates in a while, and I'm not ready to launch a specific thread on my project yet - I'll keep adding to this one. I'm still trying to gauge interest and get a feel for desired features. I don't want hi-jack toaster's efforts either. In fact I like to get more information so there's cross over benefits to both potential projects.

    I'm prototyping bits of a design on a bench now. I am no where near completion and I don't expect to be in an early alpha phase until mid-2015 at the earliest. But things look promising. And I plan on fully open sourcing everything at some point. I'm not sure yet if I will build amps for people as it carries a strong liability for damage - both personal property (car catching on fire) and personal injury (hearing loss, electrocution, etc). Regulatory hurdles also stand in the way if I try and sell this as a product. I just want to get it developed first and go from there. I'm sharing here so others may benefit. Send a PM if you want to discuss specific bits in more technical detail. I will also share early schematics / code base one on one as long as there are no IP/NDA barriers.

    Instead of a discrete FET design like toaster's, I elected to go with a fully integrated class-D ASIC. I am starting with a ST Microeletronics FDA4100LV. It's rated at 85W x 4 channels at 4 ohms. It's also bridgeable to 150W x 2 channels at 2 ohms. The intent was to run two amp ICs and bridge two channels for 85W x 6 @ 4 ohm (dash, doors, rears) and 150W x 1 @ 2 ohm for a sub. I designed and built a first cut of of the power board. It sounds absolutely fantastic! However the integrated boost controller isn't that great. And the total power starts to distort a little at max rating. So final true RMS clean power will be de-rated somewhere around 75W per channel at 4 ohm.

    I also bought a mid-range quality Alpine dual driver 10" 4 ohm sub. The bridged amp channels really don't put out enough power to do the sub justice. So... first cross roads. I switched to driving the sub with dedicated Alpine MRX-M55 550W @ 2 ohm single channel class-D amp. It's a descent mid-price amplifier that happened to be on-sale this week on-line. The sub now sounds great and rattles teeth like it should. I have two pair of 6x9 three-way Alpine speakers on my amp power stage running on the bench along side the Alpine amp and sub. The near thousand watt combo fills a small room with huge sound and I can't imagine needing more in most cars. So I'm thinking about dropping the second power stage and just providing line level outs for subs.

    First question. Is there a need for more than 4 channels x 75 W at 4 ohm? Everything is so compact and low heat, there is plenty of room for a second power stage. It would just add cost to get the extra 4 channels. However if someone we're looking to play 5.1 or 7.1 audio sources, I could see where this would be a big deal (I am actually).

    Second question. How desirable is it for me to rework an integrated sub output back into the amp? That is something that is very difficult to do well. I tore down the Alpine amp when it came in today. It is a fully discrete design for both the boost and amp stages. The boost is single stage / single phase and thus carries an enormous coil and nearly 15mF of reservoir capacitance. They use every bit of space in an enclosure larger than what I am targeting and I can see why. It's very well designed, built, and cost conscious. I couldn't do a competitive job at the same price. So I'm not sure I should try - especially at the low frequency end.

    In fact this entire design of mine is not cheap. However I am building it for myself first and the public second if at all.

    For more details of what I either have running on a bench or in a pending board design:

    • Integrated 4 phase single stage boost supply to 25V
    • Inductor isolation for alternator whine and load dump protection
    • Two FDA4100LV fully differential amps w/ full fault diagnostics and automatic fault cut-offs (each channel carries a +12 volt DC bias)
    • USB audio in/out @ 24-bit / 48 kHz however 96 kHz could be a future possibility
    • USB audio bridge provided by either STM32F415 (168 MHz ARM Cortex M4) or a later still under NDA ARM Cortex M7 at substantially faster speed
    • On board MCU allows for yet to be developed firmware extensions (keep reading)
    • Analog Devices ADAU1446 SigmaDSP.
      • Allows custom audio chains to be developed and tuned to specific vehicle characteristics.
      • See the ADI SigmaStudio page for more detailed information
      • I currently do NOT have plans to integrate the USBi debug dongle into the design.
      • Dozens of drag and drop filters
      • For example 10 channel 16 band parametric EQ easily fits
      • Examples: fade/balance, FIR filters, loudness, compression, cross fade, noise reduction, automatic EQ calibration, time-delay, dynamic bass enhancement, LFE, virtual surround, real surround, etc.
    • Lattice MachXO2-7000 FPGA in the middle for glue. However this could be reduced in scope as the development cycle progresses.
    • Wolfson/CL WM8904 codec providing two microphone inputs with boost and bias as well as 2nd audio zone output (currently being prototyped with an ADI ADAU1701)
    • With the mic paths running through the DSP & FPGA, future opportunities arise for active noise reduction (wind and road) and cabin echo cancellation before forwarding to USB host
    • 2nd audio zone provided by WM8904 pre-amp headphone output via DSP. Offers opportunity for active cancellation of main audio in headphone second zone.
    • S/P-DIF input (is this needed and does anyone want an output?)
    • Second audio codec providing 2 pairs of stereo line-outs for other cascaded amps and one pair of line-ins for auxiliary input
    • 12V switched output terminal providing remote amp on for other amps or antennas (one or two needed?)
    • Amp switching, power supply switching, tuner sampling, MCU & FPGA clocks all source synchronous with common clock and PLL with spread spectrum for low radiated noise
    • < 1mA @ Vbat during quiescent sleep
    • MCU, DSP, FPGA powered over USB from host. Power stages powered from Vbat with remote on over USB
    • I'm considering (but trying not to feature creep):
      • A small LCD for fault reporting and alarming rather than basic LEDs
      • LED bar graph for VU meters to support a desktop version.
      • Driving 2 or more LED NeoPixel (long) chains for cabin lighting controlled by USB, MCU, or audio signal
      • Single wire or dual wire CAN interface
      • Ignition input
      • Resistor ladder input
      • Dash brightness sense/ADC
      • USB suspend/resume on ignition or CAN wake.
      • iPod auth IC - however unlikely due to licensing restrictions. Maybe layout but not populate
    • Milled aluminum case with finned top from FrontPanel Express.
    • Either common automotive connector harness or ATX style connections for vehicle signals
    • Screw terminal blocks for speaker and power connections with quick detach connector blocks
    • Target size 14cm wide, 12cm deep, 4.5cm high (all approximate)
    • Expansion bay for tuner module.

    The last point - tuner. This is probably a long shot to be open sourced - or even released. The products I'm currently testing are under NDA. I'm not sure yet under what conditions or time-line I would be able release anything. However I'm currently working on a single FM/AM/HD/DAB/DAB+ design. I hope to expand this into a dual tuner design in the future with antenna diversity. And if I can work out a good LNA/AGC solution, there are possibilities for more than 2 tuners (think time-shift all your presets). However most of the cutting edge in terrestrial radio tech is still protected IP. I also have a Sirius XM design in the works, however I'm nearly certain that will always be restricted. However as I get things working, I will continually check/research it. No promises. Ultimately a simple AM/FM design would be easy and possible to integrate; maybe based on a well understood SiLabs tuner.

    I'm looking for feedback on these features and any other suggestions (adds or deletes).

    Last question. Why do people want this? I'm not sure I do. The USB part I mean. Most CarPC installs are already pretty bulky. You can get commodity amps from lots of places and there are plenty of USB sound output solutions. I would rather integrate a high end ARM CPU module to this thing and just have it be the entire head unit. But I can see some appeal in having essentially a USB vehicle tailored unit. I'd like to hear your thoughts on why.

    Current status:

    I have the first rev of the power stage built and under test. I have the main DSP prototyped and working on a custom wing board to a standard XO2 FPGA break-out board. Another custom board is on order with the STM32 and tuner to support firmware development and testing. It will plug into the FPGA eval. I have the boost supply running on the bench supplying 25V to the amp stages all from a 12V 50A current source. I have the second rev of the power stage under design with integrated boost.

    Next update will likely switch to a dedicated thread.
    Last edited by eeguru; 12-01-2014 at 04:45 PM.
    Save virtual trees....

  2. #162
    Maximum Bitrate
    Join Date
    May 2012
    Grandville, MI
    I can see two different paths.. And Different threads are cheap...

    If you are integrating a USB sound solution and class D amplifier this makes it much easier to accomplish things on a PC. I would strongly suggest using a DAC solution instead of a simple sound card solution. The cost difference shouldn't be much but the flexibility would be significantly more. Personally I think you should stick with the USB solution even if you tie it to an ARM unit to make a self contained unit. ARM units have USB ports too... Make it a true dual purpose unit. With a DAC unit build in instead of a generic sound card you can control what frequencies and any delays one might need for each output. Would be an awesome setup.

    If you have a NDA on the setup you can likely sell the unit but can't disclose the programming aspects of it. You could contact the company about your intentions. Generally the population is able to order small amounts as "development" items.

    Another option you could look at is having an expandable module setup. On the master unit you can have the dac/sound card and setup a data bus and power bus. Then you set it up to have as many modules as one could hook to it. Probably with a design power limitation. (So maybe 3 2 channel amplifiers and 2 1 channel amplifiers might be the limit... ) This would allow for easy expansion and if someone just wants a 2 channel amp they can do so or if they want 6 channels of stereo and 2 of subs they can do that as well. You might even release a module that has 2 4 volt audio outputs on it with an amp turn on so someone could add a subwoofer amplifier of their own and still benefit from the other channels. I think having plug in modules works because you could release it with a 4 x 45 watt amplifier and if someone wanted a 4 x 75watt instead they could go that route or could add a sub amp on it if they so desire. Locking yourself into one design forces you to have to have more than one variation if you want to be able to power more than one setup.

    I think it is a good idea just depends on how you want to make it work.


  3. #163
    Low Bitrate
    Join Date
    Nov 2002
    Atlanta, GA, USA
    In terms of technical specifics, I'm not sure what you mean in terms of "DAC solution" vs "sound card solution". The only time the signal goes analog (other than to an external amp) is inside the FDA4100LV at the error amplifier. The DACs are fully contained within the amplifier ICs.

    48kHz/24bit PCM audio gets shipped down the USB data path via regular isochronous transfers according to the USB Audio Class Specification. It's queued and output via digital I2S to the FPGA, routed to the DSP, routed out of the DSP and direct to the amplifier packages. The MCU implements a composite device with two audio class payload end-points - one out for speaker audio and the other in for mics or tuner or maybe both. There is also a control end-point that is partially defined to allow for extended features (path setup, programming, etc).

    The current physical layout is already composed of 3 to 4 boards - power supply and amps, status and front connectors, control board, and optional tuner board. If I had a version that dropped the second AMP IC, I would just depopulate that circuit and the 3+4 phases of the boost supply. I'd rather not break it up more than that. It adds cost and failure points.
    Save virtual trees....

  4. #164
    Low Bitrate
    Join Date
    Nov 2002
    Atlanta, GA, USA
    I have not had much free time to work on this in the past few weeks. However when I hit the next significant milestone, I'll restart the thread on another more active forum. From the day-to-day new posts I see overall, this forum is effectively dead.
    Save virtual trees....

Similar Threads

  1. Small and simple USB audio amp anyone?
    By JC-S60 in forum User Innovations
    Replies: 134
    Last Post: 04-10-2011, 10:09 AM
  2. USB Audio codec just isn't no sound
    By ramrod in forum BoomzBox HD
    Replies: 21
    Last Post: 07-14-2010, 09:13 AM
  3. diy usb audio with optical in / out
    By Raas in forum Off Topic
    Replies: 5
    Last Post: 05-01-2008, 04:03 AM
  4. No audio from line input for sirius usb adapter
    By Ryan711 in forum Car Audio
    Replies: 4
    Last Post: 09-08-2007, 08:07 PM
  5. PCI Audio vs. USB Audio vs. Onboard Audio
    By waltwalt in forum Car Audio
    Replies: 12
    Last Post: 07-26-2007, 09:36 AM


Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts