Uber Quad rebuild project - input please

Discussion in 'DIY Discussion' started by RobHolt, Nov 21, 2010.

  1. RobHolt

    RobHolt Moderator

    Joined:
    Oct 24, 2004
    Messages:
    4,126
    Likes Received:
    9
    On the back of the recent measurements thread I'm proposing a little project here to see if we can improve a Quad 34.

    My views are known but I'll summarise them by saying that I think it sounds fine 'as is' but the circuit is dead easy to mod and there are many ideas out there on what to do.

    My Q34 was given a service last year but that consisted of a new caps only.
    It has like for like value Panasonic PCs in the PSU and some Silmics doing coupling duties in the signal path. That's it - the rest is stock.

    In the interests of science and and cos I'm probably mad :) I'll sacrifice my Q34 on the altar of subjective bodging.

    So obvious place to start is op amps - suggestions please and I don't care how expensive. Nothing silly fast though - no point inviting instability. Oh and J-FETS only please, at least on the input.

    I propose to apply local decoupling at each op amp, so suggest some caps please for the job.

    Many of the coupling caps can be safely removed and replaced with a wire link - so that's an option.

    Op amps used as stock are the TL071 and the Tl072 (dual).

    Suggest away, or if you think this is a daft idea please say!

    Recordings of the before and after will be made at 24/96 for some blind assessment afterwards.
     
    RobHolt, Nov 21, 2010
    #1
  2. RobHolt

    sq225917 Exposer of Foo

    Joined:
    Jan 11, 2007
    Messages:
    1,514
    Likes Received:
    0
    You know the answer, swapping parts is unlikely to bring any great changes in how it sounds, unless there's one uber critical component in the circuit like there is in the Naim 321 cards, feedback and decoupling caps.

    Split the power rails out wherever you can, you can never have too much, local, high quality, wide bandwidth power regulation. Op-amps should only sound different if they do have generally greater performance, and that's unlikely, better decoupling can play a part.

    I think you'd be better off looking at ridiculously closed match pairs for all the components than 'boutique' labels.

    I'd be more interested to see this done with something like a Passlabs B1, very low part count and available for butons as a bare pcb. Cheapest parts and simple power supply vs well matched boutique parts and over the top power supply.
     
    sq225917, Nov 21, 2010
    #2
  3. RobHolt

    RobHolt Moderator

    Joined:
    Oct 24, 2004
    Messages:
    4,126
    Likes Received:
    9
    Good points.

    I think you can have too much local, high quality power though. Consideration needs to given to the PSRR of the active stages. There comes a point were going further is pointless IMO. I know you disagree but you might recall that I performed a blind test inserting a cheap Maplin PSU into a Hicap case and having the listeners prefer it to a Flatcap. Makes you wonder at least.

    We could try recording something through the Q34 v just a length of cable, ie no preamp. That might be a place to start.
    If that shows nothing amiss I suppose there is little point in further bodging.
     
    RobHolt, Nov 21, 2010
    #3
  4. RobHolt

    TonyL Club Krautrock Plinque

    Joined:
    Jun 19, 2003
    Messages:
    1,613
    Likes Received:
    1
    Location:
    Another pink world
    Rob, pop over to Dada Electronics, register, and download their Quad 34 revision guide. The guy behind it is one of the key contributors to the Yahoo Quad Group, runs the Quadspot blog and seems to really know his stuff. I've no idea why it's called Dada Electronics though - no mention of signing 'R Mutt' on the side of the amp or anything like that.

    Tony.
     
    TonyL, Nov 21, 2010
    #4
  5. RobHolt

    RobHolt Moderator

    Joined:
    Oct 24, 2004
    Messages:
    4,126
    Likes Received:
    9
    Read that Tony, and yes he's very clued up on all things Quad.

    Keith Snook is also a good read.
    Dismissed it all previously but happy to give it a whirl.
    34s aren't expensive so picking up another isn't difficult.
     
    RobHolt, Nov 21, 2010
    #5
  6. RobHolt

    Cable Monkey

    Joined:
    Dec 10, 2009
    Messages:
    89
    Likes Received:
    0
    Location:
    West Mids
    Very interested in this project as I was close to getting a 34 before opting for the Pass designed B1. Only input I would offer however is that op amps are the last thing I would change, they wouldn't be my starting point. Clean power would be my starting point.
     
    Cable Monkey, Nov 23, 2010
    #6
  7. RobHolt

    RobHolt Moderator

    Joined:
    Oct 24, 2004
    Messages:
    4,126
    Likes Received:
    9
    You can all hear the results as we go along - if we go along as I'll document the process with 24/96 recording as we go.

    First I'd suggest using a good CD source and comparing straight wire, ie no pre, against the 34 before mods.

    I also don't think the op amps matter as they are suited to the task even if elderly, but plenty of people swear by better ones. Worth a go IMO if the above reveals a difference in favour of straight wire.
     
    RobHolt, Nov 23, 2010
    #7
  8. RobHolt

    sq225917 Exposer of Foo

    Joined:
    Jan 11, 2007
    Messages:
    1,514
    Likes Received:
    0
    Op-amps, the work of the Devil..
     
    sq225917, Nov 23, 2010
    #8
  9. RobHolt

    RobHolt Moderator

    Joined:
    Oct 24, 2004
    Messages:
    4,126
    Likes Received:
    9
    Clever Devil.
     
    RobHolt, Nov 23, 2010
    #9
  10. RobHolt

    Cable Monkey

    Joined:
    Dec 10, 2009
    Messages:
    89
    Likes Received:
    0
    Location:
    West Mids
    I have a couple of Burson modules in the wings for my next project. Should be interesting.
     
    Cable Monkey, Nov 24, 2010
    #10
  11. RobHolt

    nigelgrant

    Joined:
    Nov 21, 2010
    Messages:
    2
    Likes Received:
    0
    Quad 34 upgrade

    Like you, I find the 34 pretty good, though perhaps not as clear as my Shearne (I have both). There are various mods around - NET audio is supposed to be very good.

    Tutt Technology may also have useful ideas. Dave Tutt is a brilliant audio engineer, formerly a BBC man.

    Nigel Grant

     
    nigelgrant, Nov 24, 2010
    #11
  12. RobHolt

    RobHolt Moderator

    Joined:
    Oct 24, 2004
    Messages:
    4,126
    Likes Received:
    9
    May the Quadfather forgive me, but the Q34 is now well and truly pimped.

    In addition to the low z PSU caps (a little larger in value than before), all electrolytics are now removed form the signal path, all TL071s have been replaced with OPA134 and all TL072s replaced with OPA2134. Local rail decoupling using poly 150nf caps applied to every op amp.

    That is effectively the range of Keith Snook & Dada upgrades, plus the rail decoupling which they don't include.

    The 34 uses 8 single op amps and 4 duals with each stage passing through a 100uf electro cap as standard. I was pretty wary about removing these caps given that dc floating around on pots and switches will often give audible clicks and causes the pots to rustle over time. However the worst I could measure was a few mv (post BB op amps) so it seemed safe to remove them. Interestingly, with no output caps there is <2mv dc offset on each channel.

    It sounds fine and operates just as before, which is a relief.

    Technically it is better in some ways.

    The phono section residual hiss has dropped by 7db and that alone is worth having. The TL072 is quite a noisy op amp though usually fine for line duties. On the high gain phono section the benefits of the quieter BB op amps are clearly audible.

    While at it, I also swapped the TL071 op amps in the Quad 405 for the OPA134s and beefed up the rail decoupling with 47uf rails to ground and 100nf across the rails.
    Noise on this has also reduced though I'd say only 2-3dB but that is expected as the design of the input section means no op amps is silent in this location.

    Not a bad days bodging and all for £35.
     
    RobHolt, Nov 28, 2010
    #12
  13. RobHolt

    felix part-time Horta

    Joined:
    Dec 13, 2003
    Messages:
    757
    Likes Received:
    1
    Location:
    dead
    Nice. The OPA134 is a nice part and staying with FET-input devices means you havent really altered some conditions which might make for obviously-different conditions (such as equiv input noise/current/ resulting dc offsets, RF sensitivity etc)

    Does this 34 still contain the infamous 4066 cmos switches? If so, they're probably the single largest contributor to %THD remaining ... though probably not so bad if only faced internally with a high impedance load. There are alternatives available it you want to go that far!

    PS I'm not familiar with the 34 innards but if any of the opamps use offset trim (ie a pot between pins 1 and 8, usu. centre tap going to one of the voltage rails) absolutely remove it with extreme prejudice. Due to the way it is applied internally such pins on opamps always have much, much worse PSRR than any other and performance *always* suffers if used.

    PPS I must re-commission my AD797-based 405/2 with its added nested feedback...
     
    felix, Nov 28, 2010
    #13
  14. RobHolt

    TonyL Club Krautrock Plinque

    Joined:
    Jun 19, 2003
    Messages:
    1,613
    Likes Received:
    1
    Location:
    Another pink world
    So, does it sound any different? If so you can do mine ;-)

    Tony.
     
    TonyL, Nov 28, 2010
    #14
  15. RobHolt

    RobHolt Moderator

    Joined:
    Oct 24, 2004
    Messages:
    4,126
    Likes Received:
    9
    Mi Martin,

    The 134 and 2134 were selected as the nearest to a modern TL071/2. No point putting super fast parts in the Q34 without completely redesigning the circuit.

    None of the op amps use offset trim.

    I've read lots on the cmos switches but i wonder if they don't just have a very conservative spec. I say that as the TL op amps aren't the lowest distortion devices around, and a worse case measured %THD for a whole 34 - so end to end including switches and a string of TLs is 0.02%.
    If there are alternatives that are effectively drop-in I'de consider them - however......

    This was a test really - a reality check if you like. I tend to avoid what I've done here these days and will only mod if i see what i thing is an obvious problem likely to be audible.

    On line (Cd and tape) I think this 'improved' 34 sounds just like the old one.
    On the high gain Radio input (100mv) it sounds just perceptibly less hissy at full pelt.
    Ordinarily there are four op amps in circuit for any given input other than phono (5 op amps). Input buffer, two in the tone and volume circuits and one for output.

    The only real gain seems to be the phono section which is clearly less noisy. That is to be expected when swapping the TL071 for a modern part in a high gain circuit. The phono section does seem to perform a little better so overall worth the money.

    The small and very cramped pcb is well stuffed and doesn't really lend itself to more drastic surgery.

    The 405-2 got lucky in that I had ordered spare 134s and had some nice decoupling caps to hand. Most report that using a quieter op amp on the 405 doesn't improve noise due to the front end design and Johnson noise.
    This is definitely a bit quieter though - not much but audible so has to be worth a £2 op amp.
    It had a complete re-cap earlier this year. I suspect the biggest upgrade for a 405-2 would be a perforated top and bottom cover. The thing is near air-tight and a little air flow would probably work wonders for cap life!
     
    RobHolt, Nov 28, 2010
    #15
  16. RobHolt

    RobHolt Moderator

    Joined:
    Oct 24, 2004
    Messages:
    4,126
    Likes Received:
    9
    Well it works - sort of.

    However what benefits there are seem to be outweighed by disadvantages.
    The most pressing and obvious are that overall noise levels are now higher, not bothersome but a standard Q34 is very quiet.
    Secondly, the PSU probably needs beefing-up as voltages are now somewhat adrift and the solitary negative fixed regulator is running rather hotter.
    Each 134 draws twice the current as TL071 and each 2134 twice that of a TL072. Remember there are 12 of them constantly powered. The higher performance cmos switches work, but again consume a little more juice.

    The upshot is a Q34 that is better on the phono stage for sure, technically and sonically slightly worse on the line stages, and operationally a little 'off', ie the switch leds will sometimes not extinguish entirely even though the signal switching is fine. Rails issues again IMO.

    Worth a go if only to see what is and isn't possible.
    So if you have one, definitely put some modern quiet op amps in the phono stage and perhaps the output op amp. Put some better caps in the PSU and then - stop.

    If you want to do the rest I'd suggest reworking the PSU first. Something for another day and perhaps another year though as I've been frying other fish this week :)
     
    RobHolt, Dec 2, 2010
    #16
  17. RobHolt

    sq225917 Exposer of Foo

    Joined:
    Jan 11, 2007
    Messages:
    1,514
    Likes Received:
    0
    SO basically you've taken a decent sounding old amp and bodged it to make it sound worse. Awesome..... ;-)
     
    sq225917, Dec 2, 2010
    #17
  18. RobHolt

    RobHolt Moderator

    Joined:
    Oct 24, 2004
    Messages:
    4,126
    Likes Received:
    9
    Yes - sometimes best to leave things alone. That's my usual view on designs that work fine 'stock', but I though this worth the experiment as putting it back to stock is easy enough, as is buying another for the £150 or so they cost these days.

    These mods are well known out there as 'improvements' - you'll notice nobody is actually running a full set of bench tests to see what these things do to the basic pre amp performance.

    This modified 34 actually works - its still quieter than an old Naiim Nac for example and on the face of it you've got decent sound coming out of it.
    Look a bit deeper though and all is not well.
     
    RobHolt, Dec 2, 2010
    #18
  19. RobHolt

    Werner

    Joined:
    Jun 18, 2008
    Messages:
    89
    Likes Received:
    0
    Location:
    Belgium
    It is bothersome. Given that a 134 has less voltage and current noise than a TL071 increased noise indicates that something somewhere is now truly wrong.

    I would look at the film caps added as bypass to the rails. I also hope you retained the TL071 that zeroes the power supply?

    --

    A possible upgrade is to bypass the input buffer opamp for CD entirely.

    --

    I once did a comprehensive rebuild of a 44 along these lines: fresh supply caps, input buffer bypass, bypass of a defective 4066, and removal of one coupling
    cap, the output tant. The owner was and is very happy.
     
    Werner, Dec 2, 2010
    #19
  20. RobHolt

    Werner

    Joined:
    Jun 18, 2008
    Messages:
    89
    Likes Received:
    0
    Location:
    Belgium
    A bit scary, that. Not?
     
    Werner, Dec 2, 2010
    #20
Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.