[Lost81]

Is it generally a good idea to replace speaker crossover components with higher quality upgrades, while keeping the same values?

E.g. sandcast resistors with Mills or Kiwame Non-Inductive Wirewounds; generic film and foil resistors with Auricaps. Inductors are left untouched.

I have heard "Yes" and "No."

What are your thoughts?

Thanks,


-Lost81

[Levi]

Yes but also depends on your speakers.  I upgraded the crossover parts with same value on my Klipsch speakers and it made a huge difference in high frequency resolution.  My Sonus Faber still sounded better   I would never touch the crossover in my Sonus Faber speakers unless I have spare parts.
Pictured below are the stock vs modified.

[StevenACNJ]

I have rebuilt the crossovers in two different sets of speakers and had huge results.

Here is what I do:

1. Point to point wiring
2. Leave inductors as is.
3. Mills resistors - like for like
4. Sonic Caps from Sonicraft - like for like
5. By-pass tweeter cap with .1uf Mundorf Silver Supreme
6. Goertz Alpha Core MI-1 wire for woofers
7. Audio Consulting 24g solid core silver wire with cotton sleeve for tweeters (VH Audio now has nice cotton sleeved solid core silver wire too)
8. Wonder solder
9. Replace binding posts with Cardas non-plated copper

[byteme]

I did this in my Polk RTA 11t's and also upgraded them to TL's at the same time.  As mentioned above, it made a HUGE difference.  However, these were/are mass market consumer designs.  I'm now using Odyssey Lorelei's and would never DREAM of messing with that shoebox sized crossover!  So, if you want to get a little more mileage out of a middle of the road pair of speakers or a mass market kind of speaker then I say, do it now, it'll yeild huge benefits, however, if you already have something that was carefully thought out and built OR as you say, specifically built to voice a speaker a certain way, then, maybe not.

Pics of the Old vs. New crossover in the Polks

[Hantra]

  I am the ONLY "NO" vote. . .

I suppose it depends on the speakers you have, and how much $ you have in them.  If they are $500 speakers, I'd probably throw some mods to them.  

I know my speakers could probably benefit from some high-end parts, but I'll never do that.  Not only because it will void my warranty, but man, it's just impossible to do it with any degree of certainty what the result will be.  
My luck, I'd spend $500 to mod my crossovers, it would sound like vomit, my warranty would be VOID, and then, even if I put it back the way it was, I would always be worried that I didn't do something right, and then I'd start hearing things, and have to sell them.  

It's just safer not to even go down that slippery slope.

[jeffreybehr]

I disagree strongly with Hantra.  I believe that manufacturers buy the best parts they can AT THE PRICE POINT they're targeting, but that leaves LOTS of room for us to improve.  

I'm in the process of upgrading/rebuilding 5 crossovers at the moment.  
I got NO help from Aerial on the complex CC3B X-over, and the VP of Sales & Marketing actually had the nerve to tell me that they built them to be as good as they know how.  Hey Mr. Kelly--inexpensive Axon polypropylenes are NOT the best-sounding caps that even a reasonable amount of money will buy!!!!!!!!!

Anyway, the CC3B now has auricaps and North Creek Zens in it, with Mills noninductive resistors.  My 4 Kindel PLS-As will have 2 different cap sets--the fronts will be populated with N-C Zens for all series caps and Axons, Rels, and Solens, all bypassed, for parallel caps.  Series inductors are N-C's 14g. MusiCoils, and all resistors are Mills NIs.

The huge difference in cap prices is interesting.  I grew up with the original (Rel) WonderCaps, and then my audiofoolia was in remission for about 15 years.  Now we have TFTF caps costing well over $100 each, and polypropylenes now can be almost $100 (very-large auricaps).  My Kindels have 3 series caps--6, 12, and 36uF.  Auricaps would cost $125, N-C Zens with bypasses about $60, and Solens with bypasses maybe $30...quite a difference, and all are polypropylene-film caps.

BTW, here http://www.altavistaaudio.com/caps.html is an interesting summary Mike Elliot (of Counterpoint and Alta Vista Audio) wrote about cap quality.  Maybe I'll try the TRT Dynamicaps for couplers in my poweramp.

[hanguy]

I have also upgraded the crossover in my Revel F30 and C30 with Sonicaps and Mills resistors. The overall characterics of the speakers did not change. The F30s imaged pretty well to start with and the imaging improved quite a bit after the mod. You can hardly localize the speakers with new caps and resistors.  Details retrieval was also improved as well.

From my experience, I would say it is well worth the money and effort spent on the crossover upgrade of commercially oriented speakers, like Revel and Polk.  The original component selection is based on a price point instead of sonic consideration. For example, Revel uses an electrolytic cap in series with the mid-range driver which is a waste of the quality of its excellent mid-range drivers IMO.

Mike

[JoshK]

I think if you are worried about screwing something up the simple answer is to pull out the original crossover (snip either side, pull out, set aside) and rebuilt a new one with superior parts, point to point, taking great care to follow the original design.  Install this new crossover and then evaluate, if it messes something up, put the orig crossover back in.  The only thing that will have changed is a couple of solder joints nothing else.  As far as warranty, just don't tell anyone at the company.

[Carlman]

I like Josh's approach and I agree to a degree with Hantra.  I have never done a crossover upgrade other than the wires to it and from it.  There is no degree of certainty in replacing these parts to me.

I now have a pair of experimental speakers, though that I'm thinking of doing this kind of stuff to.  A pair of KEF Coda 70's that I can't seem to get rid of... They need a few tweaks to sound good and maybe a few more to sound great.  So, I may experiment with the crossover and see what the deal is...

However, in doing so I feel I'd need to know more about what I'm doing.  So, it'll be a while before that experiment.

-C

[Hantra]

I think if you are worried about screwing something up the simple answer is to pull out the original crossover (snip either side, pull out, set aside) and rebuilt a new one with superior parts, point to point, taking great care to follow the original design.

Good idea!

Thing is though, when the COST to the distributor is $1,500 on your tweeter/mid, you can bet your ass the distributor will be inspecting your solder joints.  

[JoshK]

Not if you desolder the wires from the XO and do a good job.  

[markC]

Had very positive results in swapping out sand cast for mills resistors, (same value), and what appeared to be Dayton caps for Sonicaps, (same value). Left the inductors alone. Much clearer presentation in the top end. First time I questioned the highs in my SS amp.

[Rick Craig]

Is it generally a good idea to replace speaker crossover components with higher quality upgrades, while keeping the same values?

E.g. sandcast resistors with Mills or Kiwame Non-Inductive Wirewounds; generic film and foil resistors with Auricaps. Inductors are left untouched.

I have heard "Yes" and "No."

What are your thoughts?

Thanks,


-Lost81

In some cases it's not a good idea to swap parts. Inductors vary in DCR (d.c. resistance) and this can greatly affect the transfer function of the circuit. Electrolytic caps also have higher DCR than film caps so if you replace them you may create response problems. This can happen with the minature electrolytics often used in low $ commercial speakers - I had this occur when I modded a $500 pair of B&W's.

Resistors are a safe bet to replace. Most sandcast resistors have higher inductance and can cause a slight rolloff around 18-20K.

If you really want to improve things it's better to have a new crossover designed in many cases. This will have a greater impact on the sound than just swapping out parts.

[jeffreybehr]

This week I received all the parts and built the new crossovers for my Kindel PLS-As.  Bill Kindel helped me redesign (= simplify) the new ones.  Used North Creek Zen and Harmony series caps and MusiCoils, Mills noninductive resistors, and bypassed Axon polyprops in the lo-pass section.

Installed them today including soldering the speaker cable to the crossover components*.  WOW!!!!!!!!!!  The old crossovers had good parts** in them, but I think the sound quality went up another half value on the 10-scale.

Some cautions expressed above are correct.  For instance, BK had me add a 0.1-Ohm resistor in series with a new polyprop cap that used to be an electrolytic--he said the series resistance of the cheap cap was part of the network.  

However, you still shouldn't be afraid to upgrade caps in your crossovers--after all, they're the least-perfect 'passive' component we use in our systems.


* Soldering the speakercable to the first crossover device eliminated a cable crimp and a mechanical joint and is well worth the minor inconvenience, IMO and in my system.

** I built these almost 20 years ago, with polyprolylene caps by Solen, Rel, and Wonder, all bypassed with Rel polystyrenes and/or TJA Teflons, and all air-core inductors.  Reused the original sand-type resistors.

[mld277]

Great thread!  I have a pair of Advent 4's and I suspect that they would benefit from some new caps,etc.
  Anyone have actual experience with these?

[orthobiz]

My Dahlquist DQ-10 woofers were blown, had not listened to them for 10 years. Sent them to Regnar (surviving the ashes of Dahlquist's American demise) at www.regnar.com and had them redo the speaker binding posts, recone the woofers, resurround the midwoofers, refinish the wood, recloth the grill and, oh yes, rewire and reevverything the crossover. The speakers sound great, but I did not do just "one" thing so who knows if it helped or was even necessary? I'm still in listening paradise.

biz

[avahifi]

I read this thread with interest, because nobody seems to know about the two most important tools necessary when starting a parts upgrade, namely a precision ohmmeter and a precision capacitor meter.

Parts are NOT the value that is marked on them, they are the value that they actually measure.  If you replace a 10% tolerance capacitor with another, you can actually make a 20% difference in value change.  Will you hear that as you change the way the crossover works by 20%?  You betcha!  Is it because you put in better sounding parts?  Duhh.

Measure the part being removed, and replace it with another within 1% of the same value to eliminate those obvious (at least obvious to me) inadvertant circuit changes.  Of course that means you need to have a large stockpile of each value to be changed to have a chance of matching the removed parts, oh well.

Oh by the way, while you are at it, match the parts values on both speakers within 1% too.  Kinda nice to have two speakers that are the same.

Oh yes, you need to know what we told B&W years ago (and they took notice of what we told them when the 801 Series II speaker did not work) that two large bar magnets in a row become one bar magnet, and a bar magnet with two coils wound on a magnet is called a transformer and what you put in one coil you get out the other  -- in the case of the 801, 100% coupling at 200 Hz between the woofer and the midrange crossovers. We fixed it by moving one of the inductors off the board, they finally fixed it with a new board layout (but gave us no credit).. Oh well again.  The point, no inductiors in parallel with each other are allowed in a speaker crossover.  Each at right angles please.

Finally remember that a loudspeaker cabinet is not a nice place for microphonic parts, such as large soft film capacitors.

Likely you can make a major improvement to your speakers just by placing a padded patio stone or bag of lead shot on top of them.  Ah, but thats too simple, and too inexpensive.

Frank Van Alstine

[jeffreybehr]

I read this thread with interest, because nobody seems to know about the two most important tools necessary when starting a parts upgrade, namely a precision ohmmeter and a precision capacitor meter.

Parts are NOT the value that is marked on them, they are the value that they actually measure.  If you replace a 10% tolerance capacitor with another, you can actually make a 20% difference in value change.  Will you hear that as you change the way the crossover works by 20%?  You betcha!  Is it because you  ...

Frank, you are indeed a very smart fella and a talented, experienced audio designer and I'm just a 'know-nothing' hobbiest, but your statement infers that speaker manufacturers MEASURE the caps and inductors they use and intend to install, say, a 135uF shunt cap in that woofer circuit rather than the 150uF it's marked.  Maybe a few hi-end outfits do, but I can't believe the majority do.  They use 10%-tolerance parts because the values being off by as much as 20% simply doesn't make that much difference in sound quality.  

But I'll ask--how close should we be getting in our replacements?  1%? 2%? 4%?  And what happens when that 12uF hi-pass cap in one box measures, say, 10.9uF and in the other, 12.8uF?  If we follow your advice, we'd be building the 2 crossovers with 2 different values.  Or maybe we should be building stereo crossovers that are matched within, say 5%?  Sounds good to me.  

I think using capacitance and resistance meters is a great idea, and I do indeed do that, but I use them to insure that the parts I'm installing are within tolerance and especially that the composite caps I have to build are the right value.

Or do I completely misunderstand your point?

[avahifi]

The point is that when you are changing parts without exactly matching measured values, you are making audible uncontrolled circuit changes.

The object is to control all variables as closely as possible.  Changing circuit operation when changing parts means variables are not being controlled. Sonic observations that "this part sounds better" are not valid as long as other variables not controlled can make bigger sonic changes than the variable you think you are listening to.

Frank Van Alstine

[avahifi]

By the way, in building our Biro L/1 speakers, the crossovers are designed for the target (printed) value capacitors used, and then each capacitor is measured and those not really close to the target value are simply not used. It is nice to produce loudspeakers that all come out the same.  The final test (before burning in on music) is matching the speaker to a standard model on wide band white noise.  Its so easy to hear anything different this way.

Frank Van Alstine