1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.
  2. Where we are now: The site is back up and running! The character set encoding issue has been fixed so old posts should now look ok. I am currently working on themes/skins etc and I will be deleting the temporary green skin very soon as it has many issues. I will do a very quick variation on the stock blue skin first just to get a light-grey background and san-serif font, and I will set this as default. Later this week I will hopefully add a professional third-party skin in pfm colours (life is way too short to do this crap myself, and I've found one I really like than needs next to no tweaking). We are getting there slowly...
    Dismiss Notice
  3. May I please request people check and update their email address if it is out if date. I ask as I’m getting countless bounce errors as the server fails to send notifications through etc. I’ll eventually figure out how to send these to a black hole somewhere, but it makes sense to address it at source as quite a few folk are obviously not getting the thread and conversations notifications they have set in their preferences.
    Dismiss Notice
  4. 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

Flea and other mods to CDi...

Discussion in 'd.i.y.' started by Dowser, Apr 14, 2010.

  1. Dowser

    Dowser Learning to bodge again..

    So, Graham's #2 now has an S1 DAC, plus 10nF decoupling between I/V opamp +/- pins, plus a 10uF Wima between DAC -15v & +5v pins, and Flea OP2 is used to power SAA7310 directly. I also re-positioned the Flea to allow the 1541 and 7220 to breathe...this means a longer run between Flea OP and chips;

    [​IMG]

    Comparing #1 & #2 - subjectively, I now easily prefer Graham's #2 (sorry Graham, I reckon you are wrong on the R1 DAC :)). However, I also realised I am running different mains cables (stock and big thick jobbie) and ICs between #1 & #2 and pre - Linn ICs from #2 & stock mains, some unkown IC and thick mains cable from #1 - swapping them sees me prefer Graham's subjectively even more.

    Enough for tonight now, I'm just going to listen to #2 - I am dosed up on pain killers after surgery, so any thoughts I have are unreliable at best currently :)

    While at it, I traced out some of the regulator strategy used in the CDi, I will update the other thread in here tomorrow to look for the best way forward to provide an isolated supply to the 7220.

    No worries Graham - thank you for giving me the opportunity to get both running again - much easier when you have something to compare a faulty item with :) From what I've heard so far, I need 2 negative ALWSRs so I can also power the output stages of #1 with them too!

    Richard
     
  2. Dowser

    Dowser Learning to bodge again..

    Time for surgery to CDI #1 - I'm going to bolt 2 dual secondary torroids out the back of #1, and use centre tapped full wave to power a Flea and provide a dedicated supply (initially) to the existing Naim LM317 regulator for the SAA7220.

    As per my other thread here, I'm going to try a few different options with listening inbetween - first up will be CRCRC of 12k uF and 20 ohm for the flea (previously tested), and 20-22k uF and 2-3 ohm of CRCRC for the dedicated 7220 supply.

    Mock up looks like this (note - I previously already moved both Naim bridges to the right, so I have more space here than stock);
    [​IMG]

    I will mount veroboard here to terminate external transformer secondary windings, and C RC snubbers (once my CheapoModo arrives from the US to calculate optimum values!), before wiring on to the capacitors. I will dead-bug to diodes onto the cap pins;
    [​IMG]

    Caps have an M5 thread for mounting - shame it's so large, but I can fit 3 in the CDI easily - 2 for the SAA7220, and one for the Flea with a couple of lower value caps dead-bugged on;
    [​IMG]

    Richard
     
  3. Dowser

    Dowser Learning to bodge again..

    A little progress tonight - this will be the C RC snubber PCB (once I have received my CheapoModo kit to measure secondaries & calculate R required :)) for 2 dual-secondary windings, from which I will run twisted/plaited wires to diodes and caps around 10-12cms away;
    [​IMG]
    [​IMG]

    Transformers hanging off the back - they are just too big for me to fit the Naim cover from back to front, so it will have to be front-back carefully;
    [​IMG]

    Richard
     
  4. Dowser

    Dowser Learning to bodge again..

    A bit more progress, would be finished if not for idiocy :)

    Drilling more holes;
    [​IMG]



    10uF Wima across -15v and -5v pins of TDA1541A, and 100uF decoupling vIn and vOut of the -5v regulator (spot the mistake...I didn't before powering it up...);
    [​IMG]

    Cut track for dedicated supply to SAA7220 regulator;
    [​IMG]

    Graham's modified Flea wired in (with input filter changed to 1k7 and 100uF - I also fitted a 10uF Wima where there was a tiny purple electrolytic before...only after removing did a realise it was a Sanyo Oscon...I nearly refitted it). Again, I have gone for a longer signal run to allow space around TDA and SAA chips to help with cooling;
    [​IMG]

    Different new smoothing caps from planned - that M5 nut on the ROEs will make refitting the outer cover a little hard;
    [​IMG]

    Two new supplies wired in, cabled to still allow me to flip the PCB if needed;
    [​IMG]
    [​IMG]

    Populated snubber PCB (I went for C RC with 100nF, 10nF and 57R pending delivery of CheapoModo to calculate correct value of R);
    [​IMG]
    [​IMG]

    On initial test I realised I had wired transformers to wrong rectifiers/smoothers...DOH! Big job, because I had also wired Flea supply bypassing power switch, but SAA supply to the power switch.

    Finally get it correct, Flea supply good, SAA supply input good, just moving probes to output to see if 1 ohm resistors were OK...and BANG! As can be seen above, I fitted the 100uF caps to input and output of the TDA's -5v regulator the wrong way round. Stoopid!

    Hopefully just the cap on the input, and reg (& my dual crown TDA!) is still OK - I check tomorrow.

    Richard
     
  5. Dowser

    Dowser Learning to bodge again..

    Ah bugger - not working properly... Not sure if related to cap explosion yet or not.

    100uF cap I fitted wrong way round on input to TDA -5v LM337 completely exploded - the aluminium can has scratch marks on it where it grazed other components on circuit board as it was encouraged to leave the rest of the cap at high velocity (first time I have ever unintentionally fitted a cap round the wrong way and had it explode - I used to do it for fun as a young technician of course, but normally wired in the corner of someone else's workbench :)).

    Having said that;

    - -5, +5 & -15v supplies to TDA fine
    - new SAA7220 supply with CRCRC is dropping around 1v; 14v at diode, 13v after last cap (starts around 12.5v, increases and settles to 13v after 30 seconds). Output of LM317 to SAA7220 is at 4.9v.

    However, both LM317 5v reg for SAA7220, and the SAA7220 itself is getting too hot - hot enough to burn my finger, certainly hotter than before, I can only run it for 2 or 3 mins at most.

    Other than that it reads a disc and plays fine - I briefly tried connecting audio out and no signal though (I really hope I have not killed TDA - boohoo!).

    Basic checks so far, all I see is that the Flea outputs 5v pk-pk when only the Flea is powered, but that this drops to just under 1v pk-pk (and gets a bit more jagged) when I apply power to rest of machine (and new dedicated supply for SAA7220). See piccies below.

    Next step will be for me to disconnect new dedicated SAA7220 supply and reconnect as standard. Not tonight though - a bit pissed off at this stage ;)

    Any ideas or advice from previous experience?

    Cheers, Richard

    Flea output when only it is powered;
    [​IMG]

    With rest of machine powered up;
    [​IMG]
     
  6. martin clark

    martin clark pinko bodger

    Something really, really not right with the 7220/psu; there's no way noth reg and chip should run that hot, when they didn't before. (SAA7220 runs warm, but comfortably so IME; 35-40degC tops) Oscillation, or you've killed something inadvertently I fear.

    With the 7220 not powered up, yes the flea has to deliver a handful of mA - the SAA7220 logic inputs when unpowered 'look like' a diode or two, to 0v; so the clock signal in that state isn't necessarily awry :)

    (the dac is quite robust IME, since so much of it is current logic - ie mostly 'looks like' a current sink/source on the supply pins and doesn't much care about absolute DC voltage, subject to working at all. I've one which withstood +12v on the +5v pin for some weeks...no damage, measured the same as ever! Philips often ran them on +6v in their own players etc)
     
  7. sq225917

    sq225917 situation engineer

    Drop that 100n to 15n and you'll be closer to ideal for c-rc
     
  8. Dowser

    Dowser Learning to bodge again..

    Thanks - I do not see oscillation (but then also have never seen oscillation...so do not know what it looks like :)). I reconnect the stock 7220 power rail as next step.

    I thought Mark used 100nf and 10nF in his CheapoModo - and I just happen to have a whole bunch of them already. The kit arrived yesterday from US, so need to check. But all that can come once I get the damn player working again :)

    Richard
     
  9. Dowser

    Dowser Learning to bodge again..

    Damn - 7220 and reg gets too hot reverting back to stock supply too :( I assume I must have damaged something when the 100uF cap exploded.

    Looking at the cct (from a Marantz CD60), there is not really much external to the saa7220 itself that could be causing it, but I will have a quick check later this evening.

    On the bright side, better a faulty 7220 than 1541!

    Richard
     
  10. Dowser

    Dowser Learning to bodge again..

    Faulty SAA7220, normal temp now and audio again after replacement - question is, because of blown cap, or my DIY supply?

    I check supply again and try reconnecting it later. I at least confirmed that the SAA7220 I replaced in Graham's CDI #2 wasn't faulty - not by fitting to CDI #1 I hasten to add, I took one from a working CD50 I have, then fitted Graham's into that (single sided PCB :)).

    Richard
     
  11. Dowser

    Dowser Learning to bodge again..

    Hmmm - need help. This is the stock supply (shared rail between 2 LM317 delivering a dedicated 5v for SAA7220, and then another for other digital chips;

    Voltage at input to LM317;
    [​IMG]

    Scope for noise (100mV AC);
    [​IMG]

    I then took the DIY supply I built, put a 47 ohm 5W resistor between it and ground (just under 400mA load) - DC was just over 14.5v, scope trace;
    [​IMG]

    Is that oscillation?

    I then changed the CRCRC on the raw DIY supply - replaced the 1 ohm resistors with wire links. With same load I then get 16v DC and this;
    [​IMG]

    A bit confused now - I did something very similar in my Marantz CD60 without issue - the only difference was that I only used one transformer winding and a bridge rectifier, whereas in CDI #1 I have used centre tap and full wave, so I'm effectively connecting CDI ground plane to the centre tap of the transformer. Is that causing an issue?

    Thanks, Richard
     
  12. Dowser

    Dowser Learning to bodge again..

    C*nty-F*ck-Buckets! With the testing I did above last night I killed another SAA7220! Trying it tonight and no audio out, plays fine - chip was not overheating this time - check output from 7220 to 1541 and it's all messed up (comparing with CDI #2 on same shelf). Input to 7220 from 7310 was at least similar to #2, but still a collapsed clock signal.

    So, out came Graham's original 7220 from the CD50 I fitted it to last night, and it is now in CDI #1 and all is well in the world.

    I have now completely disconnected the new transformer's dual 12v secondaries from the rest of the player while I ponder...and seek advice :)

    All I had was the centre tap of the transformer connected to the central Naim earth point last night, and then connected the 14v from the new supply, via a 47 ohm resistor, to the central earth point to measure things.

    Even when I wrote it last night, I had a nasty feeling that directly connecting a centre tap winding to the main earth point was probably not a good idea, and I assume this is what has caused the issue.

    While I have done it before when powering a Flea, the Flea PSU earth return is separate from the main Naim CDPs earth return. And when I made a dedicated MTPR 5v supply for the SAA7220 in my CD60, it was only using one secondary winding and a full bridge - so there were diode junctions separating the transformer for the CDPs main earth point.

    I'd really like to know the technical reason behind it though - can anyone confirm?

    I've ordered 4 new 7220s from a hopefully reputable source in China ($5 each with postage against Little Diodes £28 each plus postage...I can source a brand new 1541A for that here in CH!) - via AliExpress, so can easily claim money back if fake...but up to a 4 week wait.

    Currently listening to CDI #1 with just the dedicated supply for Graham's modified Flea as the only mod (well, except for wet-tants in output blocking duties), and I am asking myself why I bother with further mods - it sounds fantastic :)

    Richard
     
  13. Dowser

    Dowser Learning to bodge again..

    Wow - #1 now sounds better than #2 - why? Dual versus single crown, modified Flea, or center tapped Flea supply in #1 (vs bridge in #2) - no idea, intend to find out though :)

    Interestingly, earlier even after replacing DRAM and SAA7310 in #1, it's lm317 was still running as hot as the dedicated one for the 7220. I replaced the 3 old blue tants around this 317 while doing recent work and temp is now back down lower that the 7220 one.

    In Graham's I now notice both reg and 7220 are running hotter that those in #1 - tant replacements coming up all over too I think!
     
  14. Dowser

    Dowser Learning to bodge again..

    Hmm - I left it powered overnight, and the Flea has failed. Seems to be base-collector open circuit on the input filter bc550c fitted (rather than stock bc447c). Will a ztx751 do as a replacement pnp transistor?

    Thanks
     
  15. martin clark

    martin clark pinko bodger

    Any NPN will do, especially with the lowered input R in the filter BC547 or sim, any generic equivalent you have.
     
  16. martin clark

    martin clark pinko bodger

    OK, this and the failures of SAA7220s - is your add-on transformers actually ok, or is it a bit 'leaky' from AC mains - say large inter-winding capacitances..? Is all well on the secondary side.. measuring output rails, prior to their interconnection with the player's 0v system..? beware of sending rectification currents through nominally 'quiet' 0v side planes/connections.

    (I've a Flea powered-up for most of a solid 11yrs now, the CD2 player it feeds likewise, and no such failure!)
     
  17. Dowser

    Dowser Learning to bodge again..

    Yes, agreed, something not right.

    I ran similar mods in a Marantz CD60 for 12 months or so no problem - key difference being I only used a single secondary winding and a bridge rectifier for each supply rather than centre tapped and full wave.

    The Flea I originally fitted my my CDI back in 2015 was using center tapped and full wave though - and in that player the SAA7310 & DRAM failed 6 or 8 weeks later. However, I lifted the whole transformer/PSU/Flea out of the CDI and fitted it to a Woodside WS3 where it has ran successfully since March until I removed it from service a few weeks back.

    Weird. I do some careful measurements again before connecting all back up.

    Richard
     
  18. trancera

    trancera pfm Member

    Gosh I ain't logged in for eons, but reading this just thought I'd make a daft comment.

    I have a flea in my Arcam 5 from the original gb kits (thanks Martin)

    Anyway, a while after installing and enjoying some weeks I noticed something odd. When the Arcam was switched off the transport spins to I think full speed and sits like that.

    Yes, I had tapped into the mains before switch for the little xfr by accident, idiot, but never did get round to working out why it was somehow powering the motor!

    (likewise I was concerned re center taps etc)

    Anyway you never know might be daft comment that generates a good idea.
     
  19. tenlizardus

    tenlizardus New Member

    Congratulations for your determination and my thanks for your detailed documentation. I big lesson to me indeed: after read your work I will try to fix my damaged (by me) Marantz CD94 MKII, 7 years ago (until now he´s in coma inside a box).

    Cheers,
    Antonio
     
  20. Dowser

    Dowser Learning to bodge again..

    Good luck - if you fail, sell it to me :)

    I have many other things going on currently, tax returns and trying to make a highly modified Impreza Turbo pass its first roadworthy test since I built - not as simple as UK or IRL, every fecking mod needs a piece of paper certifying them...I'm busy hiding those I don't have the paper for :)

    It will be a few more weeks until I come back to the CDIs, currently happily listening to Graham's #2.
     

Share This Page