1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.

    Dismiss Notice

iBasso DX50 MODS

Discussion in 'Portable Source Gear' started by babs, May 27, 2014.
2 3 4 5 6 7 8 9 10 11
  1. babs
    I start this thread  about iBasso DX50 modifications by users
    I will describe my personal intensive DX50 mods journey, beginning 4 month ago
    But, I will begin with a DX50 PCB description (its my personal gess):
    Notes: I didn't find my pictures of the unmodded PCB, so the following pictures are from an already modded DX50, not final

    This is a picture of the top of the DX50 PCB :
    (Warning: tons of technical terms will follow [​IMG] )
    1) HO output capacitors: these 4 capacitors are used to eliminate the DC output voltage (4.5V) from the Headphone amplifiers output, changing these caps is the  first and easier modification to made for improving the SQ.( in this picture there are four Nichicon Gold 100uF 10V caps; good but not enought capacity for the headphones)
    2) HO gain switch: this switch select the gain (low, medium,high) of the Headphone amplifiers
    3) DAC differential amplifier: this is a dual OP Amp, this amplifier sum the differential stereo voltage output of the Wolfson WM8740 DAC, convert it to unipolar voltage, and drive the LO output and the Headphone output amplifiers in parallel
    4) Wolfson DAC: convert the digital music to stereo analog voltage
    5) Flash memory: Flash memory
    This is a picture of the bottom of the DX50 PCB:
    1) LO output coupling caps: these 2 capacitors are used to eliminate the DC output voltage (2.5V) coming from the differentials amplifiers, the signal from the negative pin go to the LO output Jack (via a 100ohms resistor)
    2) HO output amplifiers: these 2 single OP amps, drives the headphone outputs
    This is a very simplified schematic of the analog part of the DX50:
    (Interstage caps are the two small yelllow rectangles near the differential amplifier, more later...)
    To be continued....
  2. babs
    This is the simplified schematics:
    To be continued....
  3. cholero
    Hello babs,
    thaks a lot for your information!
    What I was thinking, looking at your block schematic is that an interesting mod could be to add a dual power supply to have +/- supply (+5V/-5V) for the diff amp and the HO amp (not 0V/+5V). Like that we would get rid of the capacitors. Can you confirm that power supply for the Diff Amp and HO Amp are 0V/5V ?
    Thank you
  4. babs
    Diff amp have a 0 - 5V power supply
    Headphone amps have a 0 -9V power supply
    Changing the actual circuit to a bipolar power supply will be a HUGE task!
  5. musicheaven

    Thanks @babs for starting the thread and more importantly thank you for sharing your experience and know how, this will sure become an interesting and instructive thread.

    Once I'll have a chance, I'll add your thread link to the first page of the main DX50 thread.

    Continue the great work and I am subscribed. :beerchug:

    Edit: your thread has been added to the DX50 main page.
    T.R.A.N.C.E. likes this.
  6. ahmadmfz
    Hi, can you comment on the sound improvements the LME49710 op amp? I'm thinking of doing an upgrade as a diy modder is able to help me do this. Along with the op amp swap what else do you recommend to change? I'll be using the He with my earsonics sm64
  7. babs
    ...I continue my DX50 mod experience
    As I said earlier, the simpler mod to make with DX50  is replacing the output coupling capacitors of the headphone amplifiers (HO output)
    No need to remove the PCB, can be desoldered/soldered on the top
    (fitted with the Nichicon FG on this picture)
    Originally, There is 4 x 220uF 10V capacitors, the make is 'Jovial' (??)
    The dimensions are:  Dia:6.5mm and L:10.5mm  pitch is 2.5mm, but L=12mm can be fitted
    These are aluminum polarized caps
    Minimum voltage for these capacitor is 6.3V (DC bias is 4.5V)
    For the purpose of the test, I remove capacitors and I make a small rig to connect capacitors externally:
    Now, the tests!:
    I put on test lot of capacitors, even some capacitors too big to fit on the case!
    I will resume my finding for some capacitors with PROS and CONS; each capacitor have a distinct sound signature, each capacitor modify the sound and there is no perfect capacitors
    at certain degree, the choice of a capacitor will depend of personal taste, so take my comment with a grain of salt.
    All capacitor have a minimum of 100 Hour of break-in
    All tests with a Sennheiser HD650
    Listed in order of test
    1) Jovial 220uF 10V (original capacitors)
     PRO: easy sounding, a certain warmth, good bass, good mid (voice), unocuous treble
    CON: 'small' sound, restrained soundstage, lack of details and dynamic (veiled)
    2) Nichicon FW 220uF 10V (#UFW1A221MED1TD)
     PRO: easy sounding, open soundstage, good balance between the bass, mids and treble
     CON: can lack precision and dynamic, some haze over the sound
    3) Nichicon FG MUSE 100uF 10V (#UFG1A101MEM1TD)
     PRO: natural sounding, open soundstage, good balance between the bass, mids and treble, some warmth
     CON: can lack precision and dynamic, some haze over the sound, limited bass reach  (OK with HD650)
    4) Nichicon F95 'MUSE' 220uF 6.3V (#F950J227MBAAM1Q2) tantalum SMT capacitor
     PRO: precision and lot of details, good bass
     CON: artificial and hard sound, tiring...
    5) ELNA RFS Silmic II 47uF 6.3V (#RFS-6V470MF3#5)
     PRO: soundstage ok, good balance between the bass, mids and treble, uncolored, same form factor as original
     CON: lack precision and dynamic, haze over the sound, uninvolving...dull (big deception)  very limited bass reach
    6) Kemet T520 220uF 10V (#T520D227M010ATE018) low ESR polymer SMT capacitor
     PRO: precision and lot of details, good bass, uncolored
    CON: artificial sound, lack balance, uninteresting
    7) AVX Oxicap 220uF 6.3V (#NOSD227M006R0100) Niobium oxide capacitor
     PRO: precise soundstage, superb bass and low mid, lot of detail, superb control and precision
     CON: can be colored and dark , the treble is a little recessed
    I have some difficulty to make a choice between the Nichicon FG and the AVX Oxicap:
    - the Nichicon FG is more easy to listen, natural, relaxing; the AVX is more dynamic and involving
    - And the AVX seem somewhat colored to me with a little recessed treble
    But I finally choose the AVX, I can't resist! but its a purely personal choice, it will depend of your sound taste, and your headphone, etc...
    To be continued....
    ranfan and T.R.A.N.C.E. like this.
  8. polyrhythm
    Hi Babs
    Great thread, very intersesting read. Godd stuff. Will be followng this.
    Perhaps you could try doing your mods with the following port. I have installed this and sticking with it for the time being until trying the next official firmware (when it arrives)

    vaggbaba likes this.
  9. babs
    I continue now my DX50 mods description:
    Before a go through the Headphone amplifier replacement modification, i will describe an easier and maybe more interesting modification: the DAC differential amplifier replacement.
    As I said earlier, the DAC differential amplifier is a dual OP Amp, this amplifier sum the differential stereo voltage output of the Wolfson WM8740 DAC, convert it to unipolar voltage, and drive the LO output and the Headphone output amplifiers in parallel.
    All the sound signal pass by this amplifier, its maybe the principal element the sonic signature of the DX50.
    For someone who use only the LO output of the DX50 (for an external Headphone amplifier for example), this is the only mod which be necessary to raise the SQ of the DAP
    Placement of the amplifier:
    (pin 1 is at the top left on the picture)
    This amplifier have the following implementation:
    - 5V unipolar power supply
    - Dual Op Amp, 8 pin Soic case, standard pinout
    - Must be rail-to-rail output type
    - Input and output Bias at 2.5Vdc
    In addition, due to the usage of this amplifier, the following characteristics are important:
    - Low input voltage noise
    - Medium to high speed (more than 10MHz GBW is a minimum for me)
    - Low power to preserve battery
    - Low THD
    The problem is: is not really possible  to put a socket to test lot of  OP Amps, and make a choice between sound signatures
    So I choose an high quality audio amplifier which I already use with great success  on a previous modification on a desktop DAC: the TI LME49721MA
    - Designed for high quality audio
    - Input and output are Rail-to-Rail
    - Medium speed: GBW = 20MHz
    - Low noise: 4nV/sqrtHz
    - Low power : 2.15mA
    - vanishing small THD:
    And the more important ''specification'' : I know they sound very good!, extremely neutral and musical
    With a SMD rework station (basically a controlled temperature hot air gun), I remove the old Amp and put the new one
    Because I cannot exchange the amp to compare the sound of the old and the new one, the following evaluation is 100% psycho-acoustic stuff...
    So the new amplifier seem to cleanup the sound; the soundstage seem more open and precise, voices seems smoother
    I have no need to concentrate on microscopics improvement; the improvement seem absolutely evident for me
    Another worthwhile modification, in the same circuit, is the replacement of the OP Amp decoupling Cap (see picture, yellow rectangle), placed just over the Amplifier I just changed.
    The original capacitor is a standard 100uF 6.3V tantalum in a 3528 case
    I replace it with a 22uF 6.3V Oxicap (AVX #NOSB226M006R0600)
    I didn't expect any real improvement for this modification, it was just for the fun, but I was really surprised by the result: better control over the bass and midbass, more body...recommended
    ( yes, I replace a 100uF capacitor by a 22uF one; seem incorrect... but in this kind of circuit, the ESR of the capacitor is usually more important than the absolute capacitance)
    Next: the Headphone amplifiers....
    To be continued....
    ranfan likes this.
  10. polyrhythm
    Hi Babs
    When you say treble is recessed, do you meen a HF roll off or that there is a very slight a dip in high frequency level with no loss of detail?
    Seems like a very simple and easy mod to do, and that partnumber is readily available. I wonder what Capicitor the DX-90 uses on HO? any one know?
    Look forward to reading your Headphone Opamp upgrade.
  11. babs
    "treble is a little recessed" mean: the treble seem lower in overall amplitude than the mid and bass
    There is no loss of details, in fact, there is lots of details
    If I remember, the pure frequency response is flat to +/- 0.5dB, but you can't use the frequency response as a mesure of SQ or sound signature
    DX90 seem to have no coupling capacitors at all, this is the main advantage of this DAP in my opinion
    The modification is not so simple with the Oxicap than with the Nichicon; you must add some wires to solder them on the thru-hole pads
    Thanks for the commentary
  12. ahmadmfz
    Is it possible to swap out dx50 op amp with the amp that dx90 uses?
  13. babs
    The OPA1611AID single Op Amp can be used as HO drivers on DX50, but without any clear advantage (specwise) over other good audio OP Amps (like LM49710)
    Maybe the sound signature can be different?
    The BUF634 high current buffers cannot be used on a DX50
  14. ahmadmfz
    Yeah I was also thinking of an LME49710. Can you describe how much better is the sound?
  15. head-hi
    I'm tempted to do a mod to the LO on the DX50, but will wait for more on this thread before cracking it open. I'm not a fan of the SQ after trying many combos of amps and HP's. Rockbox nor the 1.29 MSE update helped. I hope babs will keep this going.
2 3 4 5 6 7 8 9 10 11

Share This Page