Should I replace the crossover capacitors in my vintage AR 3a speakers?


Am restoring a set of vintage AR 3a's.  Removed original capacitors and checked all three of them fom each speaker (6 total) with an Atlas ESR70 tester.  Every one of them checked out like they were new.  I then checked some brand new capacitors I had recently purchased for another project and they all check out as new (did this to ensure the meter was functioning correctly).  Should I replace or continuing using the originals in the speakers?

beercanshooter

I've recapped similar speakers, most recently, my AR58S speakers.  As in the 3a, it's a 12" 3-way, using dome mids and tweeters.  My 58S's used Unicon branded electrolytic caps, which still sounded very good.  Though, I didn't measure their capacitance or ESR, as both can drift after decades of use.

When recapping, I always mod 1 speaker, and then compare it to the stock speaker.  That allows me to hear any initial differences, and if a balance adjustment needs to be made, I can add small value resistors in series with the capacitors, to compensate.

I do like boutique caps.  Most often using Mundorf and or ClarityCap films.  With this recap, I used Mundorf EVO Oils for the tweeters and ClarityCap CSA's on the mids.  Bypassed NPE's on the woofers.  After the recap, the treble sounded great, both open and airy, but the mids were very recessed and opaque on it's upper end. I let the CSA caps play for weeks, without much improvement.  In the end, I added bypass caps to them, which helped restore midrange balance.  I haven't used ClarityCaps on the mids since.  While good with a bypass cap, I found them too rolled off without.  Now, I typically use all Mundorf caps.  Jantzen and Audyn caps are said to be very good as well, though I haven't personally used those.

Good luck.

Indeed, specifically old elctrolytics. Film to film is usually OK.

What I find amusing is when modders replace old electrolytics wiht film caps, removing 0.5 Ohms of ESR or more in the process and then proclaim what a difference it makes. Well, yeah, it kind of has to at that point, for good or ill. Same with changint DCR in coils without understanding the whole circuit.  When I make speaker crossovers I specifically pick among the various gauge coils to get a specific final sound.  I don't get small gauge coils for cost savings.

Not saying good caps don’t make a difference. Just saying it’s good to know what’s changing.

Good info here from @erik_squires ​​​​​​

Only two things I would add to it:

- The ESR-as-part-of-the-design consideration is most likely to create an issue when a person decides to replace electrolytics with film caps, which have low ESR. As long as you replace same with same you should be OK

- If you’re going to get a Dayton Audio DATS, you’re likely to get the best price from Parts Express since Parts Express owns Dayton Audio

Personally I lean more to the side of leaving well enough alone when appropriate, which is when the factory used high quality components, and those components still test well within spec (of course you always test caps out of circuit). A crossover is not like a Class A amp in which caps are literally being baked to a crisp.

 

Interesting conversation. Have restored a few sets of vintage speakers. Always replace the XO caps. most of the time, don't even test them, just replace. 

Every time it has been a night and day difference for the better! It always opens the speaker up, more bass, cleaner midrange, highs star to sing. 

If you are worried about ESR, try to get caps that don't have a low ESR. Even a cap that might register correct value might have a super high ESR, telling you it needs to be replaced. My last set of speakers a couple caps were in spec, but ESR was about 1ohm. 

From my experience, new caps always read on the low side, old caps are either on the high side, or just at value. If you have a 100uf old cap reading at 100uf or over, it's most likely about done. 

If you are unsure, just do one speaker, compare to the old. if you don't like it, swap it back. 

@erik_squires

Thanks for the info on the ESR differences. Always a good thing to expand my knowledge base on a (very cold) Monday morning. After doing speaker performance upgrades for decades, the typical unexpected sonic difference frrom cap upgrades has been overwhelmingly positive. But, checking the ESR will. now be on our checklist of standard practices.

@beercanshooter

I think there is more to consider here than good vs bad capacitors. Capacitors have gotten much better over the years, and you don’t have to spend $$$$ to get something that sounds MUCH better than the OEM parts. This also applies to resistors, coils -- even input terminals. Our standard practice is to yank all the OEM wire out of the box and replace with "audiophile" grade cabling. Chances are close to 100% that you have good cables from the amp(s) to the speakers. There’s a reason you selected these particular cables. Sound quality? You might see if you can find the same quality cable in bulk to use internally. This could be challenging due to size, availability, difficulty of termination, etc. Just do the best you can. If, the cable is 10 gauge, you might have to reduce the number of strands or conductors to get you at a workable 14 gauge, for example. Yes, even short runs (12") matter.

We’re not trying to paint a mustache on the Mona Lisa here and "mess" with the designer’s intent. Just getting things out of the way, that make them sound worse.

OP:

Honestly the issue I brought up is rare, but when it happens it really messes with people who don't understand why their new fancy caps made the speaker severely change character. 

It's worth doing a full impedance simulation for this reason.  Also helps you understand if your changes are due to changes in ESR or actually better sounding parts. 

Generally, 1990s electrolytics and before were a lot higher in ESR than today's versions, so that's also something to think about when renewing old speakers.

 

Erik

I always test new tubes when I receive them, verify matched sets, .... see where the signal strength is for new compared to my existing used tubes ....

often, the newer parts are tighter tolerances than the older parts, i.e. +/- 1%; +/- 2%; +/- 5%, even back to the 1950’s, +/-10% was fairly common

audible difference? I replaced my capacitors in my AR-2ax crossovers as routine, without testing. I bought standard stuff from Parts Express, the originals were never exotic, and the 3 way design is what sounds so good, the crossover simply diviying up the frequencies as needed for the chosen drivers, renewed woofer surrounds.

Now that I remember, a new capacitor was a dud, found out when the speaker didn’t work. Parts Express replaced it without question, pronto. You would have known before starting the work. I have a multi-meter but never learned how to use it.

You have level controls, to inspect, clean/keep or replace, one pair I replaced, one pair I restored

AR-2ax rehab

new level controls

I have some spare oem level controls a member here gave me, if you find the need

Wear gloves, don’t breathe removing the original fiberglas, new polyfill from Walmart.

Erik, I appreciate your thoughtful answer. I wonder how many people have not thought this thru and replaced the old caps in a crossover just because it seemed like a good thing to do while they were refurbishing an old speaker and unknowingly did more harm than good. Definitely food for thought.

Only as a learning exercise.

Be careful replacing parts that are in parallel to the driver as the old parts may have ESR that is part of the design. That’s actually true for all caps and coils but the one’s in parallel can cause unexpected impedance swings if not kept.

If you go down this path I strongly encourage you to get an impedance test tool like this one from Dayton (as used by some mega brands) or you can build a free one and use Room EQ Wizard. The benefits of this vs. a single cap/coil tester is you use it to test the entire speaker as well as individual components. It will measure uF/mH as well as ESR/DCR. Measure the speakers before hand, and after to make sure all is well.

If things don’t work it’s an incredibly good diagnostic tool. It’s kind of easy to tell what broke, also you can plug the data into a tool like XSim or VituixCAD to do a full electrical simulation.