[he46570]

OK, folks.  I stepped into the unknown, and replaced the paper-in-oil capacitor in my pair of Ref 3A De Capo i's.  Just for your reference, I actually have the plain De Capos, but did the tweeter mod from Reference 3A to upgrade them to the "i" version.  Has everything except the fancy sticker on the back.

As you may know, the existing tweeter cap is branded "Illusion", and is a can-style paper-in-oil capacitor.  It appears to be of modern construction, despite the retro packaging.  Others have noted they appear in ASL as coupling caps, but to my knowledge the Ref 3A is the only speaker to implement these caps.

I thought at length as to what capacitor I wanted to replace it with.  I have had good experiences with Hovlands, though they can sound a little strident at times - something I very much wanted to avoid in the De Capo given its simple first order configuration and high sensitivity.  I also considered Jensens, both aluminium and copper foil versions.  The copper foil caps were prohibitively expensive (nearly $100 a cap for 5uF), and there has been some criticism of the aluminium foil caps as sounding soft and rolled off in the high frequencies - again something I wanted to avoid.

I then discovered the Dynamicaps, made by Wonder.  You can read more about them here.  What little feedback that was on the web was very positive.  Besides, I love dynamics  

Now, the cap in my Ref 3A is rated at 5.17uF.  Note that caps generally have a tolerance of +/-10%, but the 5.17uF was written on the outside of the cap.  no doubt someone measured and matched the caps before assembly.  I wanted to closely replicate this value to avoid the changed operating point as a confounding factor in the comparison.

The Dynamicaps I received were matched, and measured at 4.98uF each.  So I needed to make up the difference with a bypass cap - I had some 0.22uF Jensen copper foil caps handy.  But hold on a sec - 4.98 + 0.22 is 5.20, higher than the 5.17.  Sure, but I'm not sure if 0.03uF makes that much difference.  I confirmed that the difference was 0.03uF using a capacitance meter.


This shows the Dynamicap with a paralleled 0.22uF copper foil.  The 0.22uF was actually made for high voltage coupling purposes, thus it is about the same size as the Dynamicap.  Should still be OK for this application, though.


Here we see the existing paper-in-oil capacitor secured to the internal brace inside the speaker.


Replacing the capacitor simply meant removing the leads from the existing capacitor, and hooking up the replacements.  Note I used a 2% silver solder that I love using, and my trusty Weller soldering iron.


Because I didn't know whether the upgrade would be better or not, I left the original paper-in-oil caps in situ, and didn't permanently secure the replacement caps.

So, how does it sound?  Firstly keep in mind I have only about 5 hours of playing time on these caps.  I suspect break-in will be a significant factor.  Also, I was unable to do blinded or double-blinded trials, so this is entirely my completely biased opinion.

The first disc I tried was the new Hilary Hahn Bach Concertos SACD (DG 474-639-2).  This disc has fantastic resolution of Hahn's violin as well as the LA Chamber Orchestra.  It's a good test for stridency - since violin can often portray a strident tone, certain tracks on this disc did indeed sound strident with the original cap in place.  Well, score one for the upgrade.  Hahn's violin had all the presence and expression it had previously, but reduced the stridency of the original cap.  I certainly don't think this was achieved through muffling/lower resolution.  I breathed a sigh of relief.

Christy Baron's Steppin' SACD (Chesky SACD227) is a also a great recording - intimate, shimmering cymbals and Christy's voice which has great tone and dynamics.  Playing the first track, "Will It Go 'Round In Circles", made me stop and check whether I had connected the speakers out of phase.  A huge amount of ambience had appeared out of nowhere.  The speakers were not wrongly connected, and I attribute all of this change to the upgraded caps, possibly the bypassed copper foil.  Christy's voice was more intimate than I have heard before, and the cymbals floated in space without any hint of harshness.  

Time for one more disc.  This time, Anne-Sophie Mutter's Carmen-Fantasie (DG 437 544-2).  If you don't have this disc, run, don't walk to your local store - great recording, even better performance.  The opening part of track 6 is a fantastic test of macrodynamics - Pablo de Sarasate's Fantasy from his Carmen opera has huge dynamics, with the whole weight of the orchestra coming down with Mutter's violin.  It's also a great test of decay, since after the fff attacks from the orchestra, they delay to almost nothing and you can distinctly hear the fade to black and the ambience of the concert hall.  The Dynamicap lives up to its name - strong dynamics coming through, though it doesn't sound out of control or flabby.  The decay is vastly improved - you can litereally hear the delay last several seconds as the sound resonates around the Vienna concert call.  One point of concern, however - with this PCM recording, you can hear some of the PCM harshness come through more readily than with the Illusion cap.  I'm not sure if this is due to increased harshness from the cap (I think less likely given the Hahn test above), or just the cap revealing more of the recording.

Anyway, if you're interested I can post my results as the caps progressively break in.

Robert

[mcrespo71]

Cool stuff, robert!  Missed you at the NY Audio Rave, but hope to see you soon.  I'd be interested in break in impressions.

Michael

[he46570]

Thanks Michael.  Sorry I couldn't make it... I have mid-terms this week, and I was frantically studying for macroeconomics.  Come to think about it, I should be studying right now  

In case people are wondering, my setup is as follows:

* Sony SCD-777ES SACD player (modified with clock, op-amp, cap upgrades)
* Integrated self-made SE 2A3 amp (ECC99 direct-coupled to 2A3, Tango output trannies, pure silver-wired point-to-point)
* Reference 3A De Capos
* 24" deluxe Target speaker stands
* Cables are self-made 5N rectangular core silver in thin-wall Teflon sleeve[/list:u]

[cyounkman]

Robert,

Thanks for posting! Sounds very interesting indeed...

Anthony will be thrilled to hear this, I'm sure; and this will likely precipitate more melted solder on my dining room table.... Anthony?

[wongstein]

Well Chris, you should be the thrilled one, to know that your speakers can sound even better, for relatively little money.  I, on the other hand just feel a little vindicated after your rather blase response to my original suggestion.  Thanks for your post Robert, and kudos to you for your adventurous spirit and electronic savvy.  I'd love to hear more about the sound after a couple hundred hours of play.

I bet that changing the trimming cap would make some noticeable difference in the sound (if not as much as the main cap), and considering that these smaller value caps are cheaper to swap around it may prove to be very worthwhile.

As for you, Chris, there'll be no soldering at your place until I hear something like "Hey Anthony, maybe you were right about that cap changing idea after all ... I'm sorry I ever doubted you ... your so insightful and intuitive and ... " etc etc.  A few minutes of that, and the solder may flow once again.

Anthony

[mcrespo71]

^^           ^^

It's posts like Anthony's that will forever keep me on the Reference 3a forum!

Michael

[wongstein]

I'm glad you liked that, Mike, and agreedly, I'm not leaving until I'm asked to leave, in which case I'll go get some De Capos just to come back and wreak more havoc

Of course, I'm really just kidding around about feeling vindicated.  Now that Robert with his astute cap choices has blazed a path for De Capo owners everywhere, it's a totally different game, whereas the prospect of doing the cap mods without a known fruitful starting point was a lot more daunting.   I say starting point, but I think the caps he's chosen have thoroughly excellent reputations, whereas I've heard either little or sometimes disparaging things about Ilusion caps, so it's another question altogether whether it can get "better" beyond his upgrade, or once again "just different".  Of course, we're still waiting on Robert's verdict.

I wonder if the caps values are matched at the Ref 3A factory to the particulars of each tweater ... I mean, I wonder if anyone can go get 5.17 uF caps or if one is best served getting as close to the value that came with their own De Capos.  I just double-checked the pictures of Chris' caps in my http://www.audiocircle.com/index.php?topic=6093">previous post an noticed that, yes there's a sticker that says 5.17 on it.  I never stopped to ponder what that meant  


Cheers,
Anthony[/url]

[mcrespo71]

Hey Robert,

Would you be willing to help me mod my De Capo I's with this new cap?  That is, if you think it is superior to the cap you had in your previous pair.  I'll be happily living with the De Capo I's for at least a few more months.  My Dad doesn't want to ship the Quads to me, so if we do make a trade, it will be during the HE 2004 hear in NYC.  He may be willing to drive out from MN to make the trade.

Michael
M

[brucegel]

I hope this will be received with the right spirit, which is to forward everyones knowledge about sound.This thread strikes an odd chord and I will explain why.First there seems to be a confusion in the community about what part of the chain is causing "issues".The source, either your cdp or your cart/arm/table are where 99% of your issues lie.I use the capitole cdp mk1 version and I can assure you that there is no stridency in violin recordings and I have heard hundreds.I work for the SEATTLE SYMPHONY so I hear live instruments daily in the backstage/rehearsal rooms and I know what real sounds like.Can you change a system for the better by replacing speaker caps...of course you can but that isn't the actual issue about your system.I am not advocating that everyone go out and buy a 5000 to 8000 dollar cdp, that would be foolish (I gladly leave that to retail) but there are tubed cdps that cost less that would add distortion of the pleasant kind to your systems and restore the smoothness so often sought among us.As a part time recording engineer I can offer one further insight.When it comes time to find a particular sound for the group you are recording for example a choral group with 3 or four violins and cellos plus percussion the first and most important consideration is  which hall or church or studio will give the best results and this is analagous to your source component.Then which microphones and preamps to use and you are still in the realm of source except now its electrical and not physical.Once you have put this on tape or hard drive it has been forever changed and the monitors used for mastering no matter how high end are changing it further still.The ears of us all are the only true arbiter and thats personal...one mans etched is another mans detailed is anothers shrill...and so it goes.I think it's cool to fiddle with mods I just think we could fiddle in a different area than whats being talked about here.

[he46570]

Bruce,

You raise some good points, however I must disagree on your assertion that upstream components produce 99% of "issues".  I don't know where this figure came from, and I'm sure many people would agree the speaker itself contributes to the sound in a great degree.  In addition, besides looking at component contributions in an individual fashion, there are well-known interactions between components that influence the sound - the best described is between amplifier and speaker with factors such as input impedence curve, phase shift, output impedence curve, frequency response, harmonic distortion, power capabilities etc all have a role to play.

My second objection is I feel the cap plays a far larger role in the sound than you give it credit for.  After all, it is the only component in the crossover, and is in complete series with the signal.  Having played with speaker caps in the past (extensive experimentation with Lynn Olson's Ariels), even in a complex 3rd order crossover, series caps have a marked difference on the sound.  What more for a 1st order.

Also, I'd like to clarify the 'stridency' issue.  I am in no way saying the stock Ref 3A is a 'strident' speaker - far from it.  I think it is more balanced in the treble than many other speakers I've heard.  However, I have noticed on that one particular track that the sound was unnatural, thus 'strident'.

To my ears, at least, the stridency did indeed improve with the cap change.  So, what am I actually hearing?  Did the new caps subtract something from the signal generated upstream, thereby adding distortion and coloration?  Or did the caps reduce distortion relative to the Illusion cap?  Who's to say... I am judging the improvement through solely subjective means.

What matters, though, is I feel the sound has improved.  Makes me feel happy, therefore is good for me (maybe not for you).  Sure, maybe if I buy a $10k source, it would solve the stridency issue without te cap mod.  But that is a moot point if the cap itself is causing the stridency - as to what is true, it's hard to say.

Oh - Michael, I would be quite happy to help you mod your speakers!  It might be interesting to do an A/B comparison between your stock speakers and mine first...

Robert

[wongstein]

I hear what you're saying Bruce, but within cost constraints, this mod may be the most economical way to either achieve a higher level of sound or as I believe you're suggesting, just improve component matching.  It's the aversion to opening things up and mucking with them that prompts us to just go out and replace boxes, and that's an expensive proposition almost every time, and harder and more expensive to reverse after the fact.

Robert's observations about the sound improvements are consistant with HP's comments in TAS, when stating unambiguously his preference for the Multicaps in the older version of the ASL Hurricane, over the paper/oil caps that they used later, reportedly also Ilusion caps.  The Multicaps that were used (and can still be ordered in a Hurricane "HP" version) were probably metalized polypropalene caps just like these Dynamicaps.  Coincidence?

Anthony

[mcrespo71]

I'd like to just chime in as well.  Robert used to own his own company that made low powered tube amps- Audio Genesis, so he isn't just a moron with a soldering iron- he does have some credentials to be making changes like these.  Now if I did it- I am a MORON with a soldering iron- wouldn't know what caps to put in, etc.  Besides, he can still put in the previous caps if he doesn't like the sound.  The mod is not irreparable.  I do agree that sometimes people can go crazy with mods and perhaps disrupt the delicate voicing a designer put into a product (e.g., a lot of the mods for Rega tables completely screw up the way Roy Gandy designed the table- I should know- I wasted $ on a few of them), but then there are mods that can take a product that was designed to a certain price point up to the next level.  It's just important to figure out what you want to change.  Since Roberts system is so simple and the fact that he made his own amps and cables, this cap change seems like a logical idea.  We'll see if it works out that way.

Michael

[brucegel]

Fair enough fellas... sounds like you all have a firm grasp on what your doing.Maybe reference 3a will keep an eye on what your up to.

[he46570]

Don't get me wrong, Bruce.  I think you raise valid concerns.

What does particularly concern me, is that Ref 3A goes to great lengths to use this particular capacitor inside the speaker.  I'm sure polypropylenes like the Dynamicap or Hovland are more accessible to them, and are made under tighter tolerances.  From what I know of new-production paper-in-oils, I believe the Illusion cap would actually be quite expensive.  Lastly they are a pain to work with due to their size and weight.  It would be great to get their opinion on this.

Nevertheless I think a high quality polypropylene is a better choice here.  I've had some experience with the MIT Multicaps (RTX series) in small signal coupling applications, and they sound great - haven't heard them as speaker caps, though.  As these puppies break in, I'm convinced the Dynamicaps are worthy of their reputation (and that copper foil isn't doing any harm either).

Thanks for the endorsement, Michael!  We should definitely get together soon - maybe in a couple of weeks (I'm heading to San Fran next week over spring break).

Robert

[mcrespo71]

Oh - Michael, I would be quite happy to help you mod your speakers! It might be interesting to do an A/B comparison between your stock speakers and mine first...

Yeah, that would be very interesting.  After you put some hours on those caps, I 'd be interested.  I will admit I do have some trouble controlling the treble on the De Capo I's in my all reflective room.  Most of the time, the treble is AMAZING- so airy and detailed- but there are times when it crosses the line.  Perhaps the cap can help with those bad times and still allow the amazing clarity of the De Capo I tweeter to shine through.  I'll let Robert make that call after he has put some time on the caps.

Oh, it would also be interesting to compare my stock De Capo I's with the stiffer bracing to your De Capo with the tweeter upgrade and cap change before doing any mods.  Plus, I want to hear what the De Capo I's sound like on 2.5 W!!!!!

Michael

Ryan Adams Gold CD

[mcrespo71]

It would be great to get their opinion on this.

Hey, fearless moderator, Younkman!  Why don't you be a good leader and get Tash's opinion on this?

Michael

Ryan Adams Gold- CD

[Paul L]

I am not so sure about getting "better" by modding in this level of equipment especially speakers like Reference 3a.  What you hear is the "difference" it may in your system.  It may be better in your particular system by may not be in another system.

We learn this experience when we fine tune our LW-1.  Every component contributes its own sonic signature.  The way to find the most neutral one is to try the modded equipment in as many as systems with wide diversity of character.  Expensive components do not necessary sound better in certain design or suit the application.  It is the balancing act that becomes an art when desiging the final product.  This is in the manufacturer perceptive.  As a cusumer, you can mod to suit your system and your taste but that does not mean to make the speaker better.  It is just like my daugter buying a T-shirt and she cut the sleeve and cut the length to suit her taste and to cope with her 'friend' association.  She is not buying the original T-shirt anymore.

[he46570]

I'm very familiar with this - I feel a synergy does exist between individual parts within a component.  In my amplifier designs I often balanced the sound by not only adjusting values/operating points, but also types of resistors, caps, transformers and inductors to properly 'voice' a device.  In general the simpler the device (especially for 2-stage single-ended tube designs), the more sensitive it is to part changes.

My theory is this applies to speakers too, the Ref 3A a prime example.  Now that I have subjectively heard a benefit from the part change, I cannot assume the old part was flawed or the new part is better because it is more expensive (I don't even know if it is more expensive...)  However, I think it makes the speaker work in a better way, given my existing set-up.

Individual results may vary, as some might say.  But trying to philosophize will hold back the adventurous spirit!  

Robert

[Terry]

OK, folks.  I stepped into the unknown, and replaced the paper-in-oil capacitor in my pair of Ref 3A De Capo i's.  Just for your reference, I actually have the plain De Capos, but did the tweeter mod from Reference 3A to upgrade them to the "i" version.  Has everything except the fancy sticker on the back.

Thanks for showing how it should be done. Now, can you tell me how you removed that woofer? I attempted to do this and could not get the driver to budge when I applied gentle pressure with a screwdriver. Please describe the method you used. Thank you!

Terry

[he46570]

Here's a sure fire way.

Unscrew both the woofer and the rear banana terminal panel.  The rear panel should come off pretty easily.  I'd then suggest placing the front of the speaker on a flat surface (make sure nothing's pressing on the tweeter).  

Press downwards on the woofer while you provide upwards force on the speaker box (you can pivot it from the top of the speaker).  The woofer should then fall out.

What's holding it in, is there is a small gasket around the woofer which is supposed to prevent air from leaking from around the mount.

Hope this helps.

Robert