advertisement


Removing the 47uf/ 27r for a 1r?

The Captain

The Captain

~~~~~~~~~~
Ive a tangled mess of a modded 32.5 now i cant remember if I should implement the above idea..

As now, the gain 321's have their motherboard tracks cut, so an s'reg each sitting locally, & fed from a diy raw psu. Each other board (6x) has a teddyreg squeezed nearby, this lot fed from a cheapie HP psu.

So the gain boards still have the 27r/ oscon 56uf next to it.. the other boards have the 27r/ maroon 47uf. Should these all be removed for a 1r was it? or just the gain boards? I cannae remember if good idea or what..

opinions helpful, cheers Xapt.
 
With good, quiet regs sat locally you can do what you like, subject only to keeping the regs from oscillating.

You can get away with about 0R33 between an ALWSR into 47uF oscon (or nothing if you stick with a regular electrolytic or *small* film cap e.g. 3uF or less)
You probably dont need a series resistor at all with the teddyreg, becasue of it's relatively high (and resistive) output impedance.

Have a play and see what suits you.
 
That's good to hear. I was just planning an on board TeddyReg version of my 321 & 729 boards and had planned to remove the 27R and replace the 47uF electrolytic with a 2.2uF Wima MKS2.

The cards should squeeze in to a 72, but a 32 is just too tight.
 
so is this sort of a good idea: if I switched the gain 56uf oscon and 27r out for a 3.3uf evox smr and a wire link/ '0.5R' as it were.. and leave the 47uf maroons in the rest of the boards, but just replace the 27r with a wire link?

or no wire links, just gaps?
 
Please report findings with changing the decouplers and resistors in the RC-filter, I am investigating this myself now and would be very glad to hear your results!

/Magnus
 
ok here's a question.. if I were to change summink and it became slightly 'unstable' or it was 'oscilating', what, in sonic terms, do these two terms actually refer to? or.. what does summink that's oscillating or unstable actually sound like? & how without an oscilloscope would I tell if it was anyway?

here these terms often, think wtf..
 
(quick and dirty repost from here)

[it's] oscillation, in an electrical sense. It usually means that an amplifer is being driven into producing an uncontrolled output tone(s). This may be way, way above the audio band, so you can't hear it directly, but the side effects are measurable and audible. Typical tell-tales are unusually warm/hot running (check for high current consumption by the opamp, above the quiescent values given in the datasheet) or odd DC offsets at the opamp output that cannot be traced to DC conditions in the circuit. 'Oscillation' means that the circuit is no longer working linearly (the fundamental requirement of audio circuits!) and the unwanted cyclical noise interferes either directly or by intermodulation with the wanted signal. If it's particularly bad it may produce audible effects - the sense that something is 'wrong' tonally; either bass weight missing, or peakiness in the presence range/treble if you have oscillation happening near the upper end of the audio band.

The causes usually come down to inadequate, poorly-chosen or poorly-placed decoupling, or insufficient (electrical) damping; but layout, parts choice, loading, circuit layout and parasitic loads all come into it. The most cogent and thorough single document I'd recommend reading in relation to using opamps is Linear Technologies' Application note AN47 - High Speed Amplifier Techniques

Note oscillation is a risk anywhere there is gain in a system: this includes voltage regulators! Getting this stuff right makes an enormous difference to output quality, and IMO it's 95% of the problem with uninformed* modification.

*as in an uncontrolled / unmeasured /'always-chuck-the-same (boutique) parts at it' approach. Not that things can't work out, but it's often the case one simply gets lucky, or is already working on a good pre-existing layout..
 
cheers Martin. That's helpful info.. uninformed modification- that sounds hideously like my dog's dinner of a 32.5! thankfully it does sound great, and never gets overly hot, noisy or any obvious faux-pas..

btw, would you have an opinion on the 3.3uf/ wire link instead of the 56uf/ 27r idea?

thanks Capt
 
Agent_Cooper79 said:
Please report findings with changing the decouplers and resistors in the RC-filter, I am investigating this myself now and would be very glad to hear your results!

/Magnus
Click to expand...

Ive been pretty surprised by this switching; so I had an oscon 56uf in the 321's so took those out (destroying tracks in procces- my gain boards have been so many times re-componented some tracks ends are fried off!) for a SMR 3.3uf.. before I thought the stodgy bass was just my minstrels, now the 3.3uf has cleaned things up, rid the bloaty bass for bass you can follow well, slightly leaner allround/ just everything 'clicked' into place soundwise.
 
I know I said to myself 'must get some of Neil's 321's soon as another track eye end came off! but I've bogded it back for now.
 
Before with the oscon twas a wire link, so just the cap changed.. I reckon it may have been 'unstable' with the oscon.. tho I could be talking utter balls.
 
Cheers Captain

I have recently changed my Oscons to Silmics. What that experience told me was that although fully burned in Silmics may not be utterly perfect in all frequency regions I certainly prefer them to Oscons. My impression, like yours, is that the Oscons seem to cause the circuits to Ring slightly. Very annoying and can really kill the music.

I think I will continue with decoupling the Oscons with group buy Wimas sometime next week. I will report back in my Teddyregs + Silmics = True" thread.

/Magnus
 
Agent_Cooper79 said:
My impression, like yours, is that the Oscons seem to cause the circuits to Ring slightly. Very annoying and can really kill the music.

I think I will continue with decoupling the Oscons with group buy Wimas sometime next week. I will report back in my Teddyregs + Silmics = True" thread.
Click to expand...

If the ultra-low ESR of the Oscons is causing the circuit to ring slightly, adding a film cap in parallel is likely to make things worse - they're even lower ESR! If you want to avoid oscillation/ringing, then try putting 0.3 - 0.5R in series with the cap. In other words, +leg of Oscon -> 2x 1R in parallel -> PCB. That'll add just a smidge of ESR and keep things stable.

You could also try something like 47uF/35V Elna Stargets - that's what Jim specced for the Starfish and they work very well indeed without needing any damping resistors. They're a drop-in replacement, too.

Like someone mentioned earlier (can't remember who) you could try replacing the 27R with a 1R if you're using SRs close to the circuit. I remember LesW (I think) mentioning once that the Naim circuits were quite so well behaved when powered from a very fast (ie. low-impedance) power source like the ALWSR.

Andy Weekes once mentioned that a little bit of magic can be added by replacing the 27R with a small air-core inductor in series with the output of the SR. Edit: here's Andy's words:

Andrew L Weekes said:
It may also be worth trying a damped choke (something like a few uH - 3.9uH for example) with a 1R in parallel, in series with the regulator output (at the regulator). You can try something like 15-20 turns of enammelled copper wire, wound around an 8-10mm drill bit, with the 1R in parallel.
Click to expand...

That quote is taken from here.


Carl
 
Carl, can you remind me; so ive rid the gain boards of the oscons for decoupling, replaced them with 3.3uf smr's.. Ive simply wire links in for the '27r' resistor.

Should I have some sort of resistor there? or can the wire effectively be a resistor/ ok left in?

As Ive said this cap replacement really has worked wonders for the sound- perhaps I had it really wrong with the oscon before tho!
 
Hey Capt.,

There's so many differing opinions on this! I've personally tried a few combinations and eventually settled for SRs as close to the circuits as I could get them and used heavy gauge wire between the SRs and the PCB (about 2" of wire I think). I used no resistors and no decoupling caps, relying on the 22uF Stargets on the SRs (C3) for decoupling. In hindsight that was probably a bit silly and if I was doing it now I'd use 22u for C3, a 1R in series with the SR output, then a 47u Starget on the preamp board. The 1R should decouple the 22u and 47u from each other, a very good thing! Andy took measurements of 10u in parallel with 47u on the output of SRs and found that while transient response was better, the recovery time from the transient spike doubled; I reckon that's the part that sounds like it's ringing.

So my own opinion is that 1R instead of the 27R, and a decent 47uF with a bit of ESR (eg. not oscon or Rubycon AZ or film) is probably a good way to go. YMMV.

I remember trying film caps in place of the 47uF and my first impressions were "wow!" However I soon realised there was something wrong: I put on some electronic music with deep bass (Massive Attack I think) and the bass was so rolled off that parts of the music were actually missing! I whipped those films out pretty quick, I can tell you!

Hope that helps, it's all my own opinion, no science...
Carl
 
cheers Carl- does this generally apply (ie better choice of cap & 1r) to 729's boards with oscons/ 27r's, each fed by a teddyreg? or for that matter, phono boards and 324 boards with maroons/ 27r's/ teddyregs?

or is it only the 321 gains that this thread applies to?

thanks Capt
 
You must log in or register to reply here.


advertisement


Back
Top