Crossover Help Requested for Paul Carmody's Carrera

[audiojerry]

I was hoping for some expert advice on crossover parts for the Carrera. I want to use the best quality where they would make an audible difference.
The choices on Meniscus differ from those on Parts Express. Meniscus has Mundorf; PE has Jantzen, as examples.

For inductors L1-L3:
should I go with all air-core?
is iron core the right choice for L-3?
is it better to use a larger gauge? is 14 better than 16?
is L2 the most important when considering parts quality?
what would be the best choices for inductors?

For capacitors C1-C4:
are C1 and C2 the most important when cosidering parts quality?

For resistors R1-R4:
is metal film the obvious choice over ceramic?

It seems like Dayton offers good quality crossover parts, but I'm willing to spend more if they result in sonic improvement.
Does using Mundorf Supreme or Mundorf oil capacitors in C1 and C2 make sense?
Same question for Mundorf copper/nickel resistors.
What's the best choice for L2?

This question comes from ignorance and curiosity: are any connections that connect positive to negative considered shunts?

Thanks - I appreciate your help!

[DeZZar]

This will be a can o' worms in terms of the level of personal opinions that will be offered up here. You are going to get directly conflicting advise. Beware of anyone telling you it doesn't matter, they are simply the other end of the snake oil spectrum.


Personally, I can recommend Jantzen - definitely a step up in sound quality in my opinion. Mundorf would be similar.

C1 and C2 are your critical caps - spend the money here.

You can save some money on the parts in shunt (positive to negative) but ultimately everything in the crossover is in the signal path.

Higher gauge coils will offer less resistance but be a lot larger and more expensive. Low resistance on the woofer network here is likely a good thing providing a little more output in the bottom end.

I wouldn't put any iron core inductors in this network because as its a two way there isn't really anywhere to hide. I might only introduce them in 3 way+ where they are used at very low frequencies.

Resistors should be the wire wound type like Mills resistors or similar.

Of course the big caveat here is that the difference these higher quality parts can make is relative to the overall quality of your system.

[LOUT]

It might be more accurate to think of the differences between part brands and types as slight XO tweaks to make things slightly brighter or darker and/or slightly change the knee shape...these aren't necessarily "better" or "worse", since it depends on what you're hoping to change.
I'm pretty sure some of the cheaper resistors and capacitors also have a tiny amount of inductance (which is sadly unlisted in their specs) which may slightly help or hurt the end-result response. Solidcore inductors typically have much lower resistance compared to Aircore, which can be good or bad depending on where they're used.

Some of the really thoroughly measured or listening-tweaked designs may call for particular parts because those happen to be what makes the overall speaker measure/work the absolute best regardless of brand or type hype.

There are some inherently negative things to look out for; like NPE caps having a lifespan where they'll eventually need to be replaced, and solidcore inductors having power/soak limits and you shouldn't place them right next to the speaker magnet nor other magnetic parts, but I think the rest of it all depends on what changes you want to make.



More importantly, WHAT about the Carrera's sound would you like to improve or change?

[Chris Roemer]

The only critical coil (DCR-wise) is the main one on the woofer. To keep voicing the same, you should match whatever DCR it was designed with. (I don't see a parts list by Paul, but a #14 or #15 air-core for the 1.5mH is what I'd use.)

You might not want to afford the 22uF cap on the woofer's LCR notch, but using a (cheap) npe here will probably cause the notch to "drift" over time. Not good.

When I first got my WT3/DATS, I measured MANY 5w/10w "cheap" (Xicon - or whatnot) wirewound resistors (against the more expensive Dayton Audio "non-inductive" variety). What I found was that the DA resistors were NOT "non-inductive", but had very low inductance; but so did the cheapies - about the same, actually. I used them more often than not 'cause Dayton's "gold" lead coating is prone to cracking (and falling off - esp. when bent).

[LOUT]

When I first got my WT3/DATS, I measured MANY 5w/10w "cheap" (Xicon - or whatnot) wirewound resistors (against the more expensive Dayton Audio "non-inductive" variety). What I found was that the DA resistors were NOT "non-inductive", but had very low inductance; but so did the cheapies - about the same, actually.

Do you happen to remember roughly how high/low the inductance was?
For example, was it as high as 0.03mH+ or was it a decent ways below that?

[audiojerry]

DeZZar, your thoughts help reinforce what I was already thinking.

Chris, your recommendation on L2 cements it for me.

LOUT, "WHAT about the Carrera's sound would you like to improve or change?" That's a valid question, but since I never heard what Paul originally created,
I have no way of knowing what I would like to change. Because I have invested heavily in my audio equipment, which is highly resolving, I want the speakers to be as transparent and neutral as possible. I don't want the bass to dominate the character of the speakers, and I want the tweeter to produce clean and clear without highs without bringing attention to itself.

I know that some audio equipment may need a speaker to compensate for inadequate bass or an edgy top end, or something else, but I am not approaching this build in hopes of correcting a weakness in my system or because I am not pleased with the sound. It's exactly the opposite. I could not be happier with my system. I am just looking for a fun endeavor with the objective of learning and seeing if I can build a diy speaker that can compete with my $4500 ProAc Response D2R. The D2R is similar in size to the Carrera with a ribbon tweeter. It has exceptional neutrality and silky smooth highs. I will be fun to match them up in a head to head competition, and I would be thrilled to find the Carrera fairing favorably or maybe even coming out on top. My last diy was Paul's Amiga floorstander, and it is an impressive sounding speaker. If I did not have the ProAc, the Amiga could be my go-to speaker.

Thanks to each of you so far. It really helps alot!

[djg]

There's always tone controls and EQ. Room effects might cause significant shortfalls in your final listening experience.

[JRT]

Why do you have two threads open on the same subject?

[fpitas]

Stereo needs two speakers ;)

[djg]

Again, ask Meniscus. Even though Mr. Carmody did not put much detail in his website writeup, it's very possible he worked up a preferred XO parts list with Meniscus, waiting there for you. Or perhaps Meniscus has a preferred XO list. They are speaker enthusiasts.

I have found Meniscus faster on the phone than email.

[Chris Roemer]

WT3 data:
ohms, Le1k, Le10k, "value" ,type
5.06n .08mH .02mH "5.1" DAG (Dayton "Audio Grade") - advertised as 2% "non-inductive" at the time
5.01n .09mH .02mH "5.1" DAG
4.92n .09mH .02mH "5ohm" 10w 10% (quite cheap - easily < $1)
4.90n .07mH .02mH "5ohm" 10w 10%

2.01n [email protected] "2ohm" DAG
1.98n [email protected]
1.99n [email protected] "2ohm" 10w

3.0n [email protected] "3ohm" DAG
3.0n [email protected] "3ohm" DAG

16.0n [email protected] "16ohm" DAG

Essentially, after (not too long of) a while, I stopped measuring XO parts w/DATS (eXcept for npe caps).
ALL parts were well within tolerances (even the "cheap" 5w/10w resistors - advertised as 10%, were within 2% - like the DARs),
and I never really saw any need to pay for upgraded parts (Dayton's pp caps (5%, or 10%?) were within 2%, just like the co$tlier "2%" Dayton poly caps).

[LOUT]

Wow, thanks for taking the time to find and type the list.
Nice to hear even the cheaper stuff can often keep tight tolerances.

I'm a little surprised to see the mH being that high (I saw some of the higher-end stuff claiming 0.0007mH, though I don't think frequency was given), but I think the ~0.03mH+ only really mattered if it was around 10khz+ so it's probably fine. And by "mattered" I only mean it appeared to make a 1db difference while smaller amounts wouldn't even amount to that....so still very small and what you measured was all better than that up high.


Have you found the stuff like the Dayton NPE caps drifting pretty noticeably relatively fast (less than 5-10years) or are you mostly double-checking on older gear to see when things need to be replaced?
I've heard the XO NPE's within the last decade or two are a lot more stable than older versions, but I don't know if that was a random opinion or from someone who was testing.

[Chris Roemer]

I THINK "modern" npes are better than the "vintage" ones.

I re-capped my Dynaco PAT-4 preamp (circa 1973?) after I got my WT3(/DATS) about 10 (or 15?) yrs ago.
If it had 40 npes in it, a full 25% of them were "far off" spec.
W/my WT, I could run an impedance sweep (on-screen) and you could see all the "nominal" -looking cap curves, and the BAD ones either ran high, low (sometimes nearly FLAT) or had VERY weird curve shapes. In MY mind, any npe caps from the '70s/'80s PROBably have had about a 25% failure rate (and I'm not talking moderate "drift", either).

A "good" 700uF curve (below) is the orange, bottom-most one (the bad one looks like a vertical line).
The pink curve is a good 320uF (6v) - (there's a bad one at the upper right).
The green curve is a good 50uF, w/a good 5uF above it (yellow line).

[djg]

700uf "flatline" DOA.

[audiojerry]

Hey Chris and LOUT, I don't mean to interrupt your discussion - please continue, but I wanted to snatch back my post for another question.

After purchasing the xover components and ofc wire, if I decide not to do it myself, would anyone be willing to offer or recommend someone to assemble it for a fee?

[djg]

If you're buying your components from Meniscus, they do a very good job assembling XOs. They can indulge your tastes for however you want it. They assembled my Continuum XOs and I was very happy.

[fpitas]

Electrolytic caps in general are far far better than they were even 20 years ago. Their use by the millions in switching power supplies drove research and competition to deliver a long lasting product that is sealed properly and withstands high temperatures. In addition the esr is a fraction of the old vintage values.

[mattsk8]

I THINK "modern" npes are better than the "vintage" ones.

I re-capped my Dynaco PAT-4 preamp (circa 1973?) after I got my WT3(/DATS) about 10 (or 15?) yrs ago.
If it had 40 npes in it, a full 25% of them were "far off" spec.
W/my WT, I could run an impedance sweep (on-screen) and you could see all the "nominal" -looking cap curves, and the BAD ones either ran high, low (sometimes nearly FLAT) or had VERY weird curve shapes. In MY mind, any npe caps from the '70s/'80s PROBably have had about a 25% failure rate (and I'm not talking moderate "drift", either).

A "good" 700uF curve (below) is the orange, bottom-most one (the bad one looks like a vertical line).
The pink curve is a good 320uF (6v) - (there's a bad one at the upper right).
The green curve is a good 50uF, w/a good 5uF above it (yellow line).

Interesting, thanks for sharing. Do you think modern NPE caps are actually better or was it age that made the vintage caps worse? Short of saving your data and the caps you swapped in for another 45 years there probably isn't a good way to figure that out.

[mattsk8]

Hey Chris and LOUT, I don't mean to interrupt your discussion - please continue, but I wanted to snatch back my post for another question.

After purchasing the xover components and ofc wire, if I decide not to do it myself, would anyone be willing to offer or recommend someone to assemble it for a fee?

Fairly sure Meniscus assembles xovers.

[djg]

Yes, they do. Very neat work.

[mattsk8]

Hey Chris and LOUT, I don't mean to interrupt your discussion - please continue, but I wanted to snatch back my post for another question.

Not sure if your post here was sarcasm, but everything they were discussing is very pertinent to your original question. If you've searched much at all on your query, you know that the "great capacitor debate" is one that has raged on for decades with little to no evidence that suggests much audible difference. I'm not making the statement that "caps don't matter", but I have been in blind A/B comparisons with expensive caps vs cheaper ones. They do sound very mildly different, but one wasn't necessarily "better" sounding.

All that said, IME the red Audyn caps are great. If I'm spending in an attempt to perfect things, I really like Mundorf Supreme caps. Will Mundorf caps make your speaker vastly superior sounding to the same speaker using Audyn caps? I doubt it but it would be a fun comparison to listen to.