E-MU 0404 USB mod wiki - diyAudio
Go Back   Home > Forums > Wiki

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Default

E-MU 0404 USB mod wiki

Contents


[top]Motivation

I decided to start this wiki to keep track of all info we can find about how to improve the E-MU 0404 USB (and possibly the Tracker Pre) sound interfaces.
Main modification topic is here

[top]Component identification and function

The contents of this section should be to identify on-board components and their function.

Detailed pics of EMU0404 are HERE

Image of the DAC section


Component identification (pic found on the web, thanks to the guy who made it)

[top]Observations

  • Opamps U5, U16, U33, and U39 get Vcc=6.14V and Vdd=-6.46V. All opamps seem to be on the same rails.
  • Vdd on these opamps has nasty high frequency ripple of about 6mV p-p
  • Vcc measures about 2mV p-p
  • Pin 5 of the TPS6755 supplies the negative voltage Vdd, and has the exact shape ripple seen on the opamp Vdd; interesting that the TPS7655 works at a switching frequency of 160kHz and the ripple frequency seems right about that. If anything, this would be where I'd start, by replacing the Vdd supply.

[top]Power converters

PHKI = TPS62200DBV IC

  • High efficiency sot23 step-down dc-dc converter, 2.5V-6V input, adjustable output 0.7V to 6V, 1MHz PWM
  • Datasheet

  • Part ID: U42
  • Nearby: R310 code 66D is 475K, R313 code 90C is 84.5K
  • Vout=0.5 * (1 + R310/R313)=3.31V
  • DAC digital supply; also goes to the ADC (AKM5385)

  • Part ID: U23
  • Nearby: R207 code 26D is 182K, R208 code 66D is 475K
  • Vout=0.5 * (1 + R208/R207)=1.82V
  • Seems to supply the scary BGA chips...


    PHOI = TPS61040DBV IC

  • part ID: U10
  • Low-power sot23 dc-dc boost converter
  • 1.8V-6V input, adjustable output voltage up to 28V
  • up to 1MHz switching frequency
  • up to 400mA
  • nearby: R99 code 39B is 2K49, R97 code 94C is 93K1
  • Vout=1.233 * (1 + R97/R99)= 47.33V
  • Is this the 48V phantom power??? When the TPS61040 is rated to give only up to 28V? There must be some trick with the nearby MosFet 4850...
  • Datasheet

    PGVI = TPS73001DBVR IC

  • Low-noise, high psrr, RF 200-mA low-dropout adjustable linear regulator
  • Vin up to 6V, adjustable output voltage 1.22V-5.5V
  • PSRR 68dB at 1kHz, line regulation 0.05 %/V, load regulation 5mV (delta Vout % / delta Iout, up to 200mA), output noise voltage 23uVrms (TPS73018) from 200Hz to 100kHz
  • Datasheet

  • part ID: U35
  • nearby R283 code 94C 93K1, R300 code 47C 30K1
  • Vout=1.225 * (1 + R283/R300)=5.01V
  • Vin is from U43
  • seems to supply the AK4311 chip (digital interface)

  • part ID: U25
  • nearby R225 code 94C 93K1, R230 code 47C 30K1
  • Vout= 1.225 * (1 + R225/R230) &= 5.01V
  • Vin is from U43
  • DAC Analog supply; also feeds the ADC (AKM5385)

    S28B = LM2704 IC

  • Part ID: U43
  • Micropower step-up dc-dc conveter with 550mA peak current limit
  • 2.2V-7V input voltage, up to 20V adjustable output voltage
  • Nearby R312 code 01D is 100K, R314 code 243 is 24K
  • Vout=1.237 * (1 + R312/R314)= 6.39V
  • Positive supply to the Op Amps. Also used as input by some other regulators.
  • Datasheet

    6755I = TPS6755I IC

  • Part ID: U28
  • Adjustable invertind dc-dc converter
  • Input voltage 2.7V-9V, output voltage limited to Vout≤(12V - Vin)
  • 160kHz fixed-frequency PWM controller
  • This supplies -6V, the negative rail for the OpAmps
  • Datasheet
  • [top]Opamps

  • Note: In the Emu 0404 usb the opamps get +/- 6V on pins 4-8; but the levels don't seem very symmetric (+6.2V;-6.0V).
  • Note: In the Emu 0404 2.0 (White edition) levels are not symmetric too, but they're a bit different (+6.0V;-6.3V)

    15532=NE5532 (?)

  • Part ID: U1, U2, many others
  • Internally compensated for gain 1
  • Slew rate 8V/uS, en 8nV/sqrtHz, GB 10MHz, Vo 5mV
  • Possible replacement AD8599, en 1nV/sqrtHz, Vo 10uV, GB 10MHz, slew rate 16V/uS, thd -105dB or 0.00000056
  • Datasheet

    2068=JRC2068=NJM2068

  • Part ID: many
  • Unity gain bandwidth
  • THD 0.001%
  • Slew rate 6V/uS
  • Datasheet

    2114=JRC NJM2114

  • Part ID: U5
  • Headphone driver opamp
  • Datasheet
  • [top]Other


    4850 IC

  • Near inductor of U10 PHOI
  • N-Channel power MOSFET.
  • Used with the DC/DC converter U10.
  • Datasheet

    15D diodes

  • Transient Voltage Suppression diodes
  • Each input and output line has one of these
  • Succesfully substituted with 1N914 diodes in series (see datasheet for pinout)
  • Datasheet


    OWNK Muting N-JFET transistors

  • Each output line goes through one of these components marked "OWNK" (couldn't find this reference anywhere)
  • Found out the purpose of these: they mute lines when board turns on!
  • No change in voltages on pins when changing Direct Out mode
  • Sources are all wired to GND while drains are wired to line output
  • Gates seems to be supplied by negative voltage (varies between -5.5V to -3V in different measures). At startup this goes to +4V. Immediatly after it drops to the negative voltage above
  • If you're getting VERY low output volume, most likely these are burnt. Just unsolder them and it will work like never before (plus you'll get rid of some parasitic capacitance)
  • [top]Post DAC Output Stage

    The output stage has a summing filter first (opamp U22), very similar to what's given in the AK4396 datasheet. This is followed by a volume potentiometer, then a buffer (opamp U32).
    The output of the buffer goes to the output jacks and also to another opamp (U39), in inverting unity amplifier configuration, that produces the inverted output for the balanced (!!) TRS
    output jacks. Shown are left channel circuitry through main and headphone out. Diodes and JFETs other than the HP out aren't labeled with the right PCB name.

    Schematic:


    (Right channel takes the "B" of each opamp.)
    Eagle 6.1.0 Schematic file
    Original schematic drawn by pilli

    [top]Mod categories

    [top]Power supply

    Filter Capacitors

    You can swap out some of the filtering caps and putting new higher valued ones. This will (maybe) improve filtering. The DC/DC converters present in the PCB doesn't seem
    to allow much changing in value though. Going too much up with value, will bring in an endless reset loop of the EMU.
    Basically what you can do is:
  • Change the 3x10uF supply filtering caps near the 4 above with higher valued caps. Somebody reports 470uF to be the maximum accettable value which won't make the board unstable (reset loop)
    ADDITIONAL NOTE: it seems that even 220uF value will make the board unstable, especially when EMU has been turned on for long periods. 100uF may be a good value (it's 10 times bigger than original, anyway)
  • Change the big 100uF 63V cap (seems to be orange in EMU0404 1st model, not the white one) with something higher (I put 220uF). Keep in mind that this is for 48V, so you'd
    better swap it with 63V still.

    Don't even give a try changing the 2x220uF near U5 HP driver, because 470uF will make the board unstable.

    Output stage Op Amps supply

    An "easy" thing to do is to provide a linear regulated 6V supply for the output stage opamps (U22, U16, U33, U39) with a dedicated power supply.
    The 6V rails reach the opamps through two jumper 0 Ohm resistors, components R149 and R157. If you remove these two, you can attach a new 6V supply.

    It is too tricky to solder on these tiny pads though. You'd better use the two local capacitors C103 and C109 pads, which are through-hole electrolytics. You'll need to attach your ground nearby on a pad nearby too.

    Notice that the headphone driver U5 receives the same rails but through two other resistors, that are 1 Ohm, R83 and R74. These are larger, of course, because that's a low impedance driver (headphone). These have local elcos too, C30 and C39.

    NOTE: 6V seems a bit low. I would try 12V.
    Personally I wouldn't mind this mod because DAC resolution is fixed. Having higher supply voltages over opamps will give better resolution for the opamps themselves. But then, the wave coming from the DAC will still have AKM5385 resolution. As pilli told me in a PM, using 9V didn't gave him much results (needs to be confirmed) (FixMetal)

    Linear regulated +5V PSU

    The other common mod is to create a linear regulated PSU with a 7805 and some bypass transistor. Note that you'll have to dissipate a great amount of heat. Some pics on the
    web shows a fan over a heatsink for getting more heat out of it, but that way will most likely inject fan noise into your board (because of DC motor supplied by your regulator).
    Keep in mind that you'll need a (at least) 1.2A transformer and a 12000uF input filtering capacitance to get a 5% ripple before the regulator. Spike smoother caps on the
    diode bridge will surely improve the thing.
    This mod won't (of course) eliminate the side effect of frequencies injected by DC-DC converters INSIDE the board. You should note that if the 160KHz (as supposed above) is the dominant noise in this board, it's 8 times over the audiable threshold! Even harmonics will go UP to 160KHz, not down, so there's really no chance for you to ear this (not even for your dog, I swear)
  • [top]Opamps


    Not every opamp seem to be a good compromise being replaced instead of the original ones. I tried LT1364 in place of U33 and all I obtained was short between Vcc and OutB of the opamp. Same as above, I couldn't change U3 nor U7 with NE5532 without having HUGE pops, crackles and s**t over the 2 inputs too. There must be some decoupling/bias stuff that won't let some higher spec opamp work well.

    [top]DAC Decoupling caps


    It is up to you to decide wheter get rid of these or just raise the value . Many says that without these caps you get a better bass response and a more organic overall sounding board. The fact is you'd inject some DC into your speakers. BUT, there's a but. Since the EMU can't handle passive speakers, any that you will use will have some electronic between the board and the speaker itself, making your line out signal strong enough to drive the speaker. Any of these electronic stuff has surely a DC decoupling capacitor before whatever the signal goes in. Actually you don't really have to think about DC being injected into speaker.
    If you're not enthusiast with this, you can always replace the 4x47uF decoupling caps with higher valued ones, like 220uF.

    [top]Fixing left channel peaks

    For somewhat reason Creative made a layout with lines prone to DC-DC converter and USB datarate noise. The left channel line is near an inductor, which of course induces noise to the line. Thanks to Sidiy on the forum, there's a fix for this. Instructions goes like Sidiy had
    this done. Topic's here and here
    Quote:
    The fix is obvious.
    1. Lift R151 and R167
    2. Cut the traces near the pads of R123 and R128 as shown, making sure not to go too deep or to chip some neighboring components.
    3. Solder a pair of wires, twisted a couple of times, to rework the connection. On the right side I chose the pads of R108 and R110 which belong to the same original nodes while for the left side I used the leftmost pads of R151 and R167. I did not want to pass the wires over the ADC to C101, not optimal in my opinion. Make sure not to reverse the phase.
    Pic of the area


    Traces shown


    Step-by-step mod


    Note: be very gentle while lifting the two resistors as the pads are real weak


    Contributors: iko, elingeniero, pilli, FixMetal
    Created by iko, 3rd March 2010 at 07:53 PM
    Last edited by FixMetal, 12th May 2013 at 02:18 PM
    Last comment by FixMetal on 12th May 2013 at 11:20 AM
    13 Comments, 13,762 Views
    Page Tools
    Search this Page


    Posting Rules
    You may not create new articles
    You may not edit articles
    You may not protect articles

    You may not post comments
    You may not post attachments
    You may not edit your comments

    BB code is On
    Smilies are On
    [IMG] code is On
    HTML code is Off
    Trackbacks are Off
    Pingbacks are Off
    Refbacks are Off
    Hide this!Advertise here!

    New To Site? Need Help?

    All times are GMT. The time now is 03:05 AM.


    vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
    Copyright 1999-2014 diyAudio

    Content Relevant URLs by vBSEO 3.3.2