Capacitors Redux

[Aadams]

I think I am not alone in realizing this forum has a cycle of topics that ebb and flow through the years.  For this reason I thought I would research the archives before asking a question about efficacy of capacitor replacement in my AR 58s pair which have the original capacitors and sound identical and OK to me.  I came across this thread from 13 years ago that still seems very relevant , at least to me.  I discovered what I was thinking had already been expressed by others years ago.  I thought I would share. Enjoy   

ADAMS

 

 

[AR surround]

My experience in recapping five pairs of speakers is that replacing electrolytic capacitors in kind does not always guarantee that the voicing will remain "as original."    I surmise that latter day electrolytics may have lower ESR than the original caps.   Also, the ESR of electrolytic caps varies with frequency and getting the curve of the replacement caps to match the original as-new caps might be quite a challenge?

A specific example is a pair of Boston A70's that I preemptively recapped with identical value electrolytics.   The recapped speakers sounded terribly bright and I had to add a 1/2 ohm series resistor to achieve acceptable sonic results.   I ultimately got the sound back to what I liked by adding a 0.01uF Dayton F&F bypass cap across the electrolytic.

[Stimpy]

I'll play along, as I want to recap my newly acquired AR58s speakers, too.  But, unlike my AR90's, the non-Callins caps, in the 58's, seemed to have aged better and still sound good.  Good enough, that I may not do the planned recap.  At least not yet.

Also, I was surfing recap threads last night, and a AR58s owner, at the DIY Audio forum, stated that he almost ruined his 58 speakers, by using polypro's in the recap.  It changed the voicing too much.  So, he went back to NPE's.  But, as AR surround mentions, even NPE's aren't 100% given to not cause audible changes.  I suppose that as long as one knows they may have to play with series resistance too, there shouldn't be too much of a shock, once the new speakers are fired up, and sound is different, after a recap.

[Aadams]

I was struck by what K Kantor said then seconded by T Tyson---------- to paraphrase------- if it doesn’t sound broken but you want to change the sound you are better off experimenting with an equalizer.  In 2004 his preferred approach to capacitor replacement was replace only the ones you know are malfunctioning and with the same type.  

My 58s pair have all the original 5 Unicon caps each.   The caps don’t look bad, no swelling or leakage.  For me, attenuated with an equalizer they are Ok, with zero attenuation they are unbearable.  I run all my ARs in the same way and they all sound OK TO ME.

I am not a tinkerer.  I don’t want to do a preventive recap when the speakers are working satisfactorily.  But, if I did it and got a different sound I would try to equalize it out until they sounded right again.

Adams

[Stimpy]

You should swap the stock 58s crossovers, with crossovers from the AR91/92 (it's the same crossover for either speaker).  Those crossovers are exactly the same as the 58s; capacitor and coil values, and crossover frequencies, with the addition of mid and treble level switches.  They typically use Callins capacitors, instead of the Unicon brand.  So, a recap would still be in order.  But you would gain some adjustability, with the level switches.  You can 'usually' find them fairly cheap on eBay.  Though like everything else AR these days, the pricing on the crossovers has jumped too.

Oh, I truly admire Ken Kantor.  But sometimes he's too much the engineer.  I recap for longevity as well sound quality.  'Cause to me at least, in my experience, recaps are very audible.

 

[Stimpy]

5 hours ago, Aadams said:

I was struck by what K Kantor said then seconded by T Tyson---------- to paraphrase------- if it doesn’t sound broken but you want to change the sound you are better off experimenting with an equalizer.  In 2004 his preferred approach to capacitor replacement was replace only the ones you know are malfunctioning and with the same type.  

 

One other thought, keep the signal chain simple.  The simpler the better, as simpler usually sounds better.  I worked in audio for years, doing audio and Rom formatting, and as well as audio mastering.  I used very good EQ's for work, out of necessity.  But, I'd never use an EQ in my signal chain at home.  Too complex a device, and the extra signal cables can easily color the sound.

A similar story.  I upgraded the stereo system in my wife's car.  Good stuff.  Alpine head unit, and Boston Acoustics Pro separates for speakers.  I put that in first, for sound, until the rest of the system arrived.  Soundstream Class A amps, a Coustics electronics crossover and a Alpine EQ and CD changer.  Well, the simple head unit and Boston speakers sounded awesome when I put it in.  So, we were super excited for the remainder of the system to arrive.  When it came, I got everything installed, and tweaked, and it sounded like crap!  All detail lost.  No energy.  Very bland, no matter how I adjusted it.  I even went back to my dealer and paid them to adjust it further.  No go.  They were disappointed too.  Only thing I could figure, was the added gear was too colored and complex, with all the extra electronics and cables, and it took away from the sound of the simpler set up.  Lesson learned.  Only what's needed.

[ra.ra]

2 hours ago, Stimpy said:

....crossovers from the AR91/92 (it's the same crossover for either speaker).  Those crossovers are exactly the same as the 58s....

This is not fully accurate. The AR-91 and 92 use the same tweeter and midrange drivers, so yes, those portions of x-o circuit are identical - - - but correspondingly, the two models employ different woofers and therefore have dissimilar components in the woofer circuitry.

4 hours ago, Aadams said:

The caps don’t look bad, no swelling or leakage.

While I am unwilling to become an advocate to further stimulate the rampant automatic re-cap madness, a visual inspection of older caps is only the first and most obvious step in the evaluation process. Although it is a PITA, removal and measurement seems to me to be the only legitimate litmus test for assessing capacitor accuracy and compliance.

4 hours ago, Aadams said:

For me, attenuated with an equalizer they are OK.....  I run all my ARs in the same way and they all sound OK TO ME.

This is indeed faint praise. I would think that the AR-58s (as well as your other AR models), when working properly, would generate slightly elevated adjectives or adverbs from the AR aficionados that tend to occupy this forum. 

[Stimpy]

18 minutes ago, ra.ra said:

This is not fully accurate. The AR-91 and 92 use the same tweeter and midrange drivers, so yes, those portions of x-o circuit are identical - - - but correspondingly, the two models employ different woofers and therefore have dissimilar components in the woofer circuitry.

Thanks ra.ra.  I stand corrected.  I apologize for the misinformation, as I wouldn't want to propagate incorrect data to our fellow AR enthusiasts. 

I guess, since the AR91's and AR92's use the same crossover board "blank", I incorrectly assumed the parts were identical as well.  (What a putz).  But, the AR91 crossovers do match the AR58s crossovers, other than the level switches.  So, it could work as an 'upgrade', in order to add some measure of adjustment to the AR58s speakers, if desired.  I can live with that...! 

My downfall.

 

[Aadams]

8 minutes ago, ra.ra said:

This is indeed faint praise. I would think that the AR-58s (as well as your other AR models), when working properly, would generate slightly elevated adjectives or adverbs from the AR aficionados that tend to occupy this forum. 

What a subtle prod.  Actually I am quite satisfied. Judgement of good sound is so subjective, it was perhaps a lame way of saying my system settings might not be satisfactory to others.  For me it is enough if noise does not intrude, distortion is sufficiently low that instruments and voices sound as expected, with plenty of dynamic range.  

I remember reading that proper measurement of caps required removal and because there is no audible problem I have not touched the 58s circuitry.  I once replaced the caps in a 2ax when I had it open for pot cleaning. The task was trivial and afterward I thought they sounded a little smoother in the mid-range. For me the 58s, having so many caps would be a PITA as you say, so I don't want to open them on spec.  My equalizer adjustments are very mild.

Adams

[ra.ra]

1 hour ago, Stimpy said:

...since the AR91's and AR92's use the same crossover board "blank"...

Yes, without examining this more closely, it could easily be assumed that all of the 91/92 x-o components are identical because of that board "blank" which serves both models. However, not unlike the differences between the AR-5 and AR-3a which share identical tweet and mid, the low-end crossover components of the 91 and 92 are in fact not the same due to the significantly different woofer drivers employed.  

 

[Stimpy]

Yea, I found that out last night, after I finally looked up the various schematics.    There are definitely parts differenences between the 91 and 92.  Duh.  Sorry, and thanks again.  .

[DavidR]

Pretty sure the only difference is the 92 board uses an 80uF (no 40uF) cap and the 91 board uses a 100uF and a 40uF on the woofer.

[Stimpy]

There's a coil difference too. The 92 uses a 2.85mH inductor, while the 91 uses a 1.37 coil.  Though, I was surprised to see the difference in capacitors, with the 92 not using the 40uF cap.

[DavidR]

I believe that's the woofer zobel circuit. Larger driver = larger cap

[Pete B]

I say always recap small value NPEs because their chemistry is such that they do not 

last reliably past 10 or 20 years.  You need to test them at voltage, not just small signal.

This is undisputed in the electronics industry, sure your equipment might work even at

30 or 40 years old but you are taking a huge risk and it is probably NOT working optimally.

Military specs are to only store electrolytic caps for 10 years max.

More of my thoughts on this:

http://www.classicspeakerpages.net/IP.Board/index.php?/topic/9255-smaller-advent-woofer/&do=findComment&comment=115979

And how to properly test older caps at working voltage:

http://www.classicspeakerpages.net/IP.Board/index.php?/topic/9255-smaller-advent-woofer/&do=findComment&comment=116424

Old leaky caps will pass bass into tweeters and mids, they are a risk not worth taking.

Subjective tests have to be done blind.

 

[Aadams]

24 minutes ago, Pete B said:

Subjective tests have to be done blind.

Thank you.  Maybe even double blind if you can get some help.

Adams

 

[Pete B]

An old thread:  http://www.classicspeakerpages.net/IP.Board/index.php?/topic/1431-capacitor-myths/

Paul S. Barton (founder of PSB) on break in:  http://www.classicspeakerpages.net/IP.Board/index.php?/topic/1431-capacitor-myths/&do=findComment&comment=63255

 

 

Tom Nousaine on break in:  http://www.classicspeakerpages.net/IP.Board/index.php?/topic/1431-capacitor-myths/&do=findComment&comment=63254

I had links to Ken Kantor's comments on break in but most in that thread are broken due to one of the site updates.

Here are Ken Kantor's comments, there were more but the links are broken:  http://www.classicspeakerpages.net/IP.Board/index.php?/topic/749-speaker-break-in-myth-or-fact/&do=findComment&comment=52320

 

Wish I could edit my own threads so that I can fix the links.

 

[Pete B]

I measured a Collins cap here back in 2005, it was defective before reforming it:

http://www.classicspeakerpages.net/IP.Board/index.php?/topic/1411-ar-11-collins-capacitance-increase-needs-reforming/