I'm starting to be desperate.
Replacing the 47uF reg output caps with Elna Silmic II brought cold, flat sound to the table. After more than a few hours of playing nothing improved so I put the old caps back in place.
While I was at it, I thought "let's replace those three ugly Nippon chem SXE 2200uF caps that I had put first, wich have the wrong lead pitch and were standing weird on the board besides being ridiculously tall and slim, and let's put three "nice" Nichicon KW instead, it can't hurt, they're audio grade, and at least they have the correct size and lead pitch, they will sit properly on the board, it will look better and be safer".
>> it is a DISASTER. Dynamics are DEAD, everything sounds slow and sluggish to the point of making me believe something is actually broken...
What happened? wrong caps? SXE better in PSU? did I kill the old regulator output caps by putting them back again? Man this is driving me nuts. I almost want to give up. I see people recappîng their B226 with Elna Silmic II all around and be satisfied; for me silmic post reg didnt work, but then what does? I can't believe Nichicon KW is THAT bad. A 35$ Schneider plastic discman would sound better than this
Replacing the 47uF reg output caps with Elna Silmic II brought cold, flat sound to the table. After more than a few hours of playing nothing improved so I put the old caps back in place.
While I was at it, I thought "let's replace those three ugly Nippon chem SXE 2200uF caps that I had put first, wich have the wrong lead pitch and were standing weird on the board besides being ridiculously tall and slim, and let's put three "nice" Nichicon KW instead, it can't hurt, they're audio grade, and at least they have the correct size and lead pitch, they will sit properly on the board, it will look better and be safer".
>> it is a DISASTER. Dynamics are DEAD, everything sounds slow and sluggish to the point of making me believe something is actually broken...
What happened? wrong caps? SXE better in PSU? did I kill the old regulator output caps by putting them back again? Man this is driving me nuts. I almost want to give up. I see people recappîng their B226 with Elna Silmic II all around and be satisfied; for me silmic post reg didnt work, but then what does? I can't believe Nichicon KW is THAT bad. A 35$ Schneider plastic discman would sound better than this
...😀 How can you know what is good or not in this pcb as you swap all by half a kilo of caps each time !
Re read the advices 🙂
Re read the advices 🙂
I do not know the definition capabilities of your system which would reveal such differences in sound which you describe, but in my experience, sonic differences between capacitors are always in the range of nuances - audible but never night and day.
I can confirm that adding capacity to any position in B226 actually kills the sound, specifically dynamics and transparence. Also, KW is a fine capacitor so not likely the cause. A few hours is not sufficient burn-in time for Elna Silmic, so they say.
I can confirm that adding capacity to any position in B226 actually kills the sound, specifically dynamics and transparence. Also, KW is a fine capacitor so not likely the cause. A few hours is not sufficient burn-in time for Elna Silmic, so they say.
Well, with SXE in the power supply and old original frako caps post regs and silmic iI output coupling it sounded good. With same config but silmic II post reg it sounded more flat and cold. Since few people told me "leave the old reg output caps", I put them back. So basically the change I hear now must be due to the swap from Nippon chemicon SXE to Nichicon kW...?
It's been playing a bit longer now, and it seems to be better. Maybe that's my main fault: not giving time for the caps to burn in.
I didn't add any capacitance... Same values as originals.
I do have a very revealing system, for sure. Maybe I should just stop for a few days and just listen if the sound changes.
I remember once I had a vintage SAE Mark One preamp fully recapped by professionals and at first when it came back from the technician I was soooo disappointed, it sounded like s**t with no coherence, exposed treble, and no bass. Took weeks to sound as good as excpected.
Have a much better preamp now, but the SAE was sweet, with the right music.
What are the values & voltage of both KW & SXE.
It's not KW is a bad cap, but each cap may not be "universally" right each time. It would be too easy !
I believe you're on the good path as it's a Learning curve. You know you have already a good conf with the SXE at powersupply position.
I was more & less sure than the S II will give nthing at the outputs of the regs, but what says W Studer is right too, this caps needs a little break-up, maybe the heat of the soldering affects the cap for a little time, but I don't know really if it's an audio myth or not (some serious guys noticed).
According my experience S II is only good each time as DC output cap, putting two in reversed polar position will add an extra of transparency which can be good but not each time ! that's why a tunning is a low and borring work.
If it's always better & more secure to fit with the good size of cap on a pcb, sometimes ears preference goes towards a cap which can have the bad size and bad pitch. Simply try to glue it with a reversible glue like the transparency one given by electric glue guns. It's about security and avoiding vibrations. the weight of a cap which doesn'ton the pcb is not a good thing but may be fixed to avoid the weight only on the leads (for big caps).
Have you a photograph of the previous blue cap at DC output ?
What is the source of your caps ? Have you a cheap capacitor checking tool (capacitance value) ? You don't need really one as far your are not working on a filter where the values have to be the good ones. But It's always interresting to measure the value of the cap you swap in relation to its marking. Most of elctrochemicals caps have a 20% precision range.
Try also to avoid silver soldering which need more heat !
It's not KW is a bad cap, but each cap may not be "universally" right each time. It would be too easy !
I believe you're on the good path as it's a Learning curve. You know you have already a good conf with the SXE at powersupply position.
I was more & less sure than the S II will give nthing at the outputs of the regs, but what says W Studer is right too, this caps needs a little break-up, maybe the heat of the soldering affects the cap for a little time, but I don't know really if it's an audio myth or not (some serious guys noticed).
According my experience S II is only good each time as DC output cap, putting two in reversed polar position will add an extra of transparency which can be good but not each time ! that's why a tunning is a low and borring work.
If it's always better & more secure to fit with the good size of cap on a pcb, sometimes ears preference goes towards a cap which can have the bad size and bad pitch. Simply try to glue it with a reversible glue like the transparency one given by electric glue guns. It's about security and avoiding vibrations. the weight of a cap which doesn'ton the pcb is not a good thing but may be fixed to avoid the weight only on the leads (for big caps).
Have you a photograph of the previous blue cap at DC output ?
What is the source of your caps ? Have you a cheap capacitor checking tool (capacitance value) ? You don't need really one as far your are not working on a filter where the values have to be the good ones. But It's always interresting to measure the value of the cap you swap in relation to its marking. Most of elctrochemicals caps have a 20% precision range.
Try also to avoid silver soldering which need more heat !
Last edited:
Update:
Okay so yesterday I wrote how bad it sounded (and it really, really did sound bad), but that was right after putting the brand new KW caps in.
What I did then (as everybody in the building was sleeping) for the next 4 hours, I listened a few of my favourite CDs trough headphones (but not using the B226 headphone jack, I used the preamp's jack), playing a favourite track, then do something else and leaving the CD to play for a while, then come back and listen again, then change disc, repeat the process... (albums used: MALA-Mala in Cuba, Rufus Wainwright - Poses, Feist - Let it go, Michel Herr & Jack Van Poll - A tribute to Belgian jazz, Serge gainsbourg- Balade de Melody nelson)
it became quickly evident that it's a total nonsense to put fresh, never used caps in, listen for 10 minutes and decide based on those ten minutes if the new caps are a good move or not (remember, this is my first ever own recap job).
The most extreme example, the first CD was Mala in Cuba and there's this track, "changuito", where the electronic bass drops so suddenly and strong that if you're not aware you might jump off ur seat. Well, at first it really did sound like "flop...", like the PSU was struggling to follow the demands. But at second listen, after the whole disc played, it sounded good.
My guess is, for some reason, KW caps might need more than a few minutes to form, and they also might need a lot of burn-in time, wich seems to be a constant with "audio" caps, if I trust what I read here and there on the net.
So far so good, listening trough headphones is one thing, today i will listen some more, but I will wait a bit before making any further judgement.
It doesn't sound anymore like something is broken, and that's enough for now, I can give the caps some time to burn in.
I'll give some further updates of course, but what I can already tell, doing this first ever recap, is:
1) doing the actual desoldering / soldering on a PCB is easier than I expected
2) choosing the right cap is less straightforward than I expected.
Here's a link to "Mala - Changuito" 🙂 the whole album really puts severe demands on any audio PSU.
https://www.youtube.com/watch?v=v1dhv7HTKck
Last edited:
If something sounds genuinely really bad after a component change then one or more of the following may be true:
1. the new component is faulty, so it is not functioning as it should
2. the new component has more/less stray inductance/capacitance than the original so the circuit is oscillating (probably at an RF frequency - you won't hear it, but you may hear the resultant audio distortion)
3. the new component is too large (in physical size) so it has added stray capacitance
4. there is now too much stray inductance due to sloppy wiring
Also possible:
5. the old component was faulty, but you had got used to the previous bad sound so now you don't like the genuine improvement
Any non-pathological change in sound due to swapping one component for a different one of the same value and equally suitable dielectric (etc.) will be so small that in almost all cases it will be inaudible. It will never be 'obvious', 'night and day' etc.
Note that a regulator output is one of the few places in an audio circuit where component strays could have a big effect due to potential instability or resonance. Any effect seen here has absolutely no connection whatsoever to any real or imagined behaviour when the same component is used for normal audio coupling or decoupling.
1. the new component is faulty, so it is not functioning as it should
2. the new component has more/less stray inductance/capacitance than the original so the circuit is oscillating (probably at an RF frequency - you won't hear it, but you may hear the resultant audio distortion)
3. the new component is too large (in physical size) so it has added stray capacitance
4. there is now too much stray inductance due to sloppy wiring
Also possible:
5. the old component was faulty, but you had got used to the previous bad sound so now you don't like the genuine improvement
Any non-pathological change in sound due to swapping one component for a different one of the same value and equally suitable dielectric (etc.) will be so small that in almost all cases it will be inaudible. It will never be 'obvious', 'night and day' etc.
Note that a regulator output is one of the few places in an audio circuit where component strays could have a big effect due to potential instability or resonance. Any effect seen here has absolutely no connection whatsoever to any real or imagined behaviour when the same component is used for normal audio coupling or decoupling.
Hi DF96,
Doest the point 3 is about the pitch leg or is it the number of foil turns when you are talking about size ? Sometimes there is close caps with same capacitance but different package : one widther the other longer ! sometimes same pitch but the ESR changes a little ? What should we look at on the datasheet in relationto your input ?
I like the point 5 : so true 🙂
Doest the point 3 is about the pitch leg or is it the number of foil turns when you are talking about size ? Sometimes there is close caps with same capacitance but different package : one widther the other longer ! sometimes same pitch but the ESR changes a little ? What should we look at on the datasheet in relationto your input ?
I like the point 5 : so true 🙂
Last edited:
While I have to agree with you and the scientific laws behind that true statement, on a subjective level, I can tell you having used the same amplification and speakers for years, that it is very revealing of differences that might be way more subtle, or perhaps in some cases inaudible, on a more classic system.
I'm using speakers based on the Klipsch LaScala, with aftermarket crossovers made of Solen Musicaps and litz wire inductors (ALK "Universal" crossover for the Klipsch Heritage range), updated tweeters, updated horns for the mid compression drivers, fed by a heavily modded Trends TA10.1 used as power amplifier where every original cap, resistor, potentiometer has been replaced with high-grade audio equivalent (not my job, obviously, there's a crazy modder who sells those "as is"), trough a very rare but good Belgian Korn&Macway SP100 preamp (seen in Germany as the sonic equivalent of Treshold Fet Ten), all connected by VanDenHul interconnects.
i can attest the extreme "magnifying glass" effect of that combination, as I hear everything, even the support on wich electronics are sitting, with ease.
My second half, who likes music a lot but is completely new to hearing music trough a proper audio system, sometimes points out "there's something changed with he sound?" as minor mods have been taking place.
It's not to flaunt as "how great" my system is (i adore it but some audiophile would most likely hate it), it's just the added combination of everything working at very low power, with the extremely low noise floor of the flea-powered Tripath power amp, the very good transparency of the preamp, good cables, and compression drivers based 104dB sensitivity speakers.
None of the changes I've heard so far trough this Revox recapping process were, to my ears and trough my system, subtle.
That's also my experience ! With a good layout, the ending result goes from horrible to good in relation to the choice of caps !
And sometimes theory is not working with the experimentation, more exactly this is the opposite : some expérimentations makes lting some theories as the 100 experiment which is not working like the 99 before !
It's too expensive for the brands but some rare to optimize their devices with this cap process : most build it like a fridge : if the scopes and the reliability is good, and if we can make cheaper to rise the margin, the Ockham razor : this is the good choice !
Experience is also important then it comes to tweak with caps, but engineers are not too much trained at this, but the more open minds (T. Loesch, P. Rogic and many others luckilly... and guess : their stuffs sounds good !)
And sometimes theory is not working with the experimentation, more exactly this is the opposite : some expérimentations makes lting some theories as the 100 experiment which is not working like the 99 before !
It's too expensive for the brands but some rare to optimize their devices with this cap process : most build it like a fridge : if the scopes and the reliability is good, and if we can make cheaper to rise the margin, the Ockham razor : this is the good choice !
Experience is also important then it comes to tweak with caps, but engineers are not too much trained at this, but the more open minds (T. Loesch, P. Rogic and many others luckilly... and guess : their stuffs sounds good !)
Hmm.
24 hours later, about 10 hours of playing.
Has it "improved"? yes and no.
There's definitely a lack of attack, and voices, drums, mid frequencies are too laidback and spread all over the space between the speakers instead of being upfront and focused as it's always been (with some variations) on my system, including with the stock Revox, and with the Revox with SXE PSU caps.
I could never get used to that sound. I like voices to be clearly delineated in the center.
Also, there seems to be a very slight distortion occuring on high pitched, complex sounds such as choral masses, and treble is rather splashy and reminds me of some cheap speakers with a physiologic curve and tweeters the size of a dime.
I will listen some more but I doubt things will improve to such a large margin...
Here' the PSU schematics again. I have a hard time finding out wich filtering cap feeds the analog section, and wich feeds the TDA1541 and SAA7220 chips. If I knew, I would keep the Nichicon kW for the analog side, and put the Nippon chem SXE for the digital side... if they are fed from different caps, that is... based on the schematics I would guess C5 goes to the DAC and SAA, but I'm really not sure...
24 hours later, about 10 hours of playing.
Has it "improved"? yes and no.
There's definitely a lack of attack, and voices, drums, mid frequencies are too laidback and spread all over the space between the speakers instead of being upfront and focused as it's always been (with some variations) on my system, including with the stock Revox, and with the Revox with SXE PSU caps.
I could never get used to that sound. I like voices to be clearly delineated in the center.
Also, there seems to be a very slight distortion occuring on high pitched, complex sounds such as choral masses, and treble is rather splashy and reminds me of some cheap speakers with a physiologic curve and tweeters the size of a dime.
I will listen some more but I doubt things will improve to such a large margin...
Here' the PSU schematics again. I have a hard time finding out wich filtering cap feeds the analog section, and wich feeds the TDA1541 and SAA7220 chips. If I knew, I would keep the Nichicon kW for the analog side, and put the Nippon chem SXE for the digital side... if they are fed from different caps, that is... based on the schematics I would guess C5 goes to the DAC and SAA, but I'm really not sure...
Sorry, but this is daft! Spoil a good regulator with extra output resistor and keep an expensive low ESR cap??
Crazy.
Use an el cheapo regular electrolytic and it will automagically have a bit of ESR and you don't need the extra series R. Win-win.
Jan
Simply physical size. Big surface area means big stray capacitance.Eldam said:
A system which is sensitive to minor details of components is a symptom of problems, such as being on the verge of instability or picking up too much RF.KanedaK said:
You should be able to read through schematics if doing this kind of work. Anyway, here it is:
+12/-15V
D9-D12
C11,C12
IC4, IC5
C13-C16
Feeds the analogue stage, SAA7210 (+12V), TDA1541 (-15V), relay K1 (+12V)
+5V
DZ1
C4
IC1
C2
C3
Feeds the standby circuit (digital)
+10V
DZ1
C4
Feeds the transport
+5V
DZ1
C4
IC2
C6
C7
Feeds SAA7210, TMS4416, SAA7220, TDA1541, HEF4054, AD7528 and servo electronics
-5V
C5
IC3
C9
C10
Feeds TDA1541 and servo electronics
-10V
DZ1
C5
Feeds the transport and servo electronics
+12/-15V
D9-D12
C11,C12
IC4, IC5
C13-C16
Feeds the analogue stage, SAA7210 (+12V), TDA1541 (-15V), relay K1 (+12V)
+5V
DZ1
C4
IC1
C2
C3
Feeds the standby circuit (digital)
+10V
DZ1
C4
Feeds the transport
+5V
DZ1
C4
IC2
C6
C7
Feeds SAA7210, TMS4416, SAA7220, TDA1541, HEF4054, AD7528 and servo electronics
-5V
C5
IC3
C9
C10
Feeds TDA1541 and servo electronics
-10V
DZ1
C5
Feeds the transport and servo electronics
In my experience, the B226 does not suffer from these issues. I wonder where do all these problems and differences described here come from.
I ditched the Nichicon KW, put some nippon chem SXE back in there (C5, C11, C12), remember I put the KW because the size fitted better and it looked better and I thought audio caps HAVE to be better, but I prefer by far what I hear with the NC SXE. Punch is back, and voices are back to the front! No more syrup! No more diesel motor!
Maybe I should have let some more time for the KW to burn in, but frankly, I couldn't stand the laidback midrange and the soft bass. I doubt that would have improved by much
anyways, I have an awful lot to learn, for sure, and who knows what's gonna be next? Change of the AOPs perhaps? I'd prefer to gain some refinement by swapping the old NE5532 than pouring some syrupy caps into the mix...
Maybe I should have let some more time for the KW to burn in, but frankly, I couldn't stand the laidback midrange and the soft bass. I doubt that would have improved by much
anyways, I have an awful lot to learn, for sure, and who knows what's gonna be next? Change of the AOPs perhaps? I'd prefer to gain some refinement by swapping the old NE5532 than pouring some syrupy caps into the mix...
In the ebay store of these guys (no affiliation, information purpose only):
[FONT=Arial, sans-serif][FONT=Times New Roman, serif][FONT=Times New Roman, serif]S.A.R.L. ELECTROBREIZH
Le Vau Ballier
5 Rue des Compagnons
22960 Plédran
FRANCE[/FONT][/FONT][/FONT]
Sorry, I forgot to thank you! 🙂
I can read a schematic, to some extent (to some decent extent for someone who never ever studied electronics), but somehow there was something unclear about this one; I have the service manual of another CD player, for example, where you have the schematics of the PSU section then for each rail it is duly written: "to DAC", "to servo", etc... here with the Revox, all you have is little numbers (wich I thought correspond to some pin numbers on the "decoder" schematics, but I was unsure...) anyways, as I said, there's so much I need to learn. Taking a functional CD player and putting fresh caps, same value same rating, was SUPPOSED to be a straightforward operation 😀
- Status
- Not open for further replies.