1. Things you need to know about the new ‘Conversations’ PM system:

    a) DO NOT REPLY TO THE NOTIFICATION EMAIL! I get them, not the intended recipient. I get a lot of them and I do not want them! It is just a notification, log into the site and reply from there.

    b) To delete old conversations use the ‘Leave conversation’ option. This is just delete by another name.
    Dismiss Notice

Arcam and the TDA1541

Discussion in 'd.i.y.' started by fatmarley, Dec 18, 2006.

  1. martin clark

    martin clark pinko bodger

    Yes, it would be non-trivial; maybe a PIC and some code to generate the complementary datastream, followed by reclocking of the output I2S lines. If you're in that far, you might as well do the digital stuff in a FPGA and write your own digital filter too... Not so difficult for someone used to digital stuff but way beyond me!
     
  2. mrdave45

    mrdave45 pfm Member

    yeah, im not sure about trying to correct the signal either. However martins balancing act sounds like it could be pretty special. Yep, the valves (when i get them) will be 6c45pi. Its the diyaudiocraft kit i have. Anyone know what happened to them cos there site is no more.

    From your last comment martin, you're suggesting that the digital filter is beneficial. is it that the saa7220 is just completely naff so its worth ditching the filter for the other benefits. (Im assuming its a dig filter plus oversampler).

    If you can get the pic to recognise the start of each word accurately(hopefully by this point the majority of error correction is done) then coding the complement functions would be a doddle. I defo want to be looking into this. Whats a FPGA. (Uni seems such a long time ago!)

    Happy new year for tonight guys!

    D
     
  3. IDM

    IDM pfm Member

    Hi Ron,

    I once took isolated the I2S in my Marantz CD63, and tried a variety of DAC's including the 1XTDA1543 then I tried 8XTDA1543 and finally 1XTDA1541 using passive I/V (about 70ohms I think) this fed into an SRPP amplification stage using ECC88's. The circuit was taken from Thorstens Thunderstone audio website. From memory the 8XTDA1543 was very exciting sounding with deep bass and reasonable control. The TDA1541 had fantastic mid-band and with certain CD's was superb, however, I thought it was lacking in bass control and depth. I did the TDA1541 mod by using a buffer amplifier close to the I2S and leading the digital signal out of the CD player using 3 lengths 40cm's lengths of 50ohm coax, this allowed me to build the DAC and valve stage off board. I did wonder about trying the 6C45pi tubes but wanted to try them with anode chokes, which I never managed to source. Unfortunately I ultimately killed the CD player trying other ideas (A Mod to far!!).

    I always wondered what it might have sounded if I had avoided the long digital runs. I have now bought the CDpro-2m unit which will allow me to try the TDA1541 implemented properly. I am currently building 3 superregs for the powersupply and cannot wait to get the thing finished.

    One of the issues I have been wondering about is whether to use passive I/V or discrete I/V. How do they differ in sound and do you have any circuit diagrams to consider?

    Cheers and Happy New Year

    Ian
     
  4. ron

    ron Tweaker

    Hi Ian,

    I've built a few different dacs using TDA1541 and TDA1543, TDA1543 is not bad and so much easier to work with but to my ears the TDA1541 is better so I now stick with that.
    I tried stacking the TDA1543's but didn't like the results.

    The valve output stage for the TDA1541 was nice, I've tried Thorstens ECC88 and Lesha's 6C45PI. both are good but I thought Lesha's was better and then went with trying parallel dacs.
    Both times I used passive I/V resistors.

    I now use (in my main dac) Pedja's design which is a AD844 as a common base, not the classic op-amp type implementation, to me this is better than any of the passive I/V and valve based circuits I tried.
    It also uses a Asychronous reclocker on the I2S lines and I now run a DEM reclocker on pin16.
    Jocko's discrete I/V circuit is good too, I'll try and find the details :)
    Maybe You could try the Ad844 comon base as I/V and run the valve buffer on the AD844 pin5 so bypassing that chips internal buffer, I do this but use 2sk170 Jfets, I've never tried it with valves though.

    Plenty of things to try out in this hobby:D
     
  5. S-Man

    S-Man Kinkless Tetrode Admirer

    After reading this I thought I had better dig one of my Arcam Alpha+'s out and have a dabble. I have mostly given up trying to sort the wayward bass on this CDP!
    The +5V supply has its own 317 feeding 2R2 then a 33uF oscon in // with a ceramic. I tried shorting the 2R2 but the sound was worse (probably due to the 317 not liking the low ESR oscon). I'm not a fan of opamp based regulators so probably won't try an SR here.
    I then tried decoupling -5V to -15V using a 10uF 35V ZL. Again the resulting sound was worse. I notice that Arcam has decoupled -15V to analogue ground and -5V to digital ground and I reckon that the extra cap messes up the grounding.
    Some time ago I completely reworked the grounding as recommended by Guido Tent. I reversed this mod because it ruined the sound.

    So I still haven't sorted the Arcam's fat bottom. I wish I could, because except for the bass it's quite a decent sounding player.
     
  6. ron

    ron Tweaker

    Try the DEM reclocking, it tightened my bass up nicely! also I found the double crown dac to have a better bass than the standard A and S1
     
  7. S-Man

    S-Man Kinkless Tetrode Admirer

    Can you point me towards a concise description of how to do the DEM reclocking (never quite grasped it from several zillion posts on DIYA).
     
  8. ron

    ron Tweaker

    I'll try and find it:) some people had a few problems with noise when fitting it but mine seems ok so far
     
  9. ron

    ron Tweaker

  10. martin clark

    martin clark pinko bodger

    Quite possibly - it worth a try though. If you want to persist, try an smallish film cap instead (say 0.1-1uF); I did play with cap types a bit here and found some quite audible differences, which might be down to leakage currents.

    I've no experience with the Arcam, but having had a look at the schematic, IIRC there are some largish series resistors (22R or so) in the supply lines to the output stage opamps. It looks like Arcam went after low noise at all costs, but I would definitely try linking these out, (or at least reducing markedly to 1R or so) which should tighten things up nicely.
     
  11. ron

    ron Tweaker

    I've tried quite a few different caps across the -15v and -5v, as Martin says the types here make a difference.

    Ones I've tried are 0.1uf Mica,0.1uf RTX,10uf Oscon SG,10uf Oscon SP,10uf ZA and a Oscon SEP.
    I personally liked the 10uf Oscon SP best, decoupling caps for +5/-5/-15 are 100uf ZL's bypassed with 10nf + 1R in series.
    the only place I use Oscon is the SP series across the -5 and -15v pins.
    The modified Flea supplying the +5v works well in my unit, this is run from the supply that feeds the other regulation

    The Flea XO has its own transformer
     
  12. IDM

    IDM pfm Member

    Are there any problem or issues with using IC holders with the TDA1541A? I would hate to hard solder my double crown chip as I might want to play with it in other configurations in the future.

    Cheers,

    Ian
     
  13. ron

    ron Tweaker

    I've used those gold pin ones with the little clips at the end which lift the chip out when you press them.
    They didn't give me any problems
    [​IMG]
     
  14. S-Man

    S-Man Kinkless Tetrode Admirer

    Thanks for the advice (Martin and Ron). I will give the DEM reclocking a go at some point in time.

    BTW, I also have a Rotel 955BX which uses the same chipset. The layout looks better than the Alpha+ in that it has a ground plane and is all on one pcb. It doesn't sound as good as the Alpha+ though, I guess I should start by snipping those pesky muting trs out!
    I always intended to build a proper I/V and output stage for this because there is loads of room in the chassis.
     
  15. IDM

    IDM pfm Member

    Hi Ron,

    I have been looking at Pedja's design for the I/V conversion and have a rather simple question. I understand that you have to adjust the trimmers to set the offset nulling current source, Pedja mentions the figue of 1mV. My question is where do you measure the 1mV across? Sorry for asking such basic questions.

    Secondly have you upgraded the suggested AD844 power supply regulators?

    Cheers,

    Ian
     
  16. fatmarley

    fatmarley Registered User

  17. ron

    ron Tweaker

    Hi Ian,

    sorry for the late reply:)
    I measure accross the 1n8 (c120/123 in the schematic) cap when adjusting the nulling.
    Basically its pin2 -in which is input and pin3 +in which goes to ground, I can get mine to stay around 0.6mV
     
  18. fatmarley

    fatmarley Registered User

  19. martin clark

    martin clark pinko bodger

    Matt - add 4K7 across the 390R on the 5v reg, that'll get you close enough. (-14.9v / -5.01v)
     
  20. ron

    ron Tweaker

    yes, sticking a 4k7 in parallel with that 390R resistor bringing it down to 360R should lower the -5.31v down to 5v near enough spot on
     

Share This Page

  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