How good is the crossover in your loudspeakers?

Unlike, say, you and/or I, Erik? 

Danny Richie makes a point of stating that crossovers cannot be correctly done in simulated circuits, but only by using them with the intended drivers that are mounted in the intended enclosure, then doing acoustic measurements. Of course the crossover parts have to be matched left speaker to right, the matching being done electronically.

Now THAT is interesting @westcoastaudiophile. Can you elaborate on crossover component "parasitics?

@bmbmzig: Danny Richie has repeated stated that achieving a desired frequency response (the specs/measurements to which you refer) does only one thing: make the speaker accurate. He over and over has stressed the point that replacing sound-degrading crossover parts with parts that don’t degrade the signal is the way to increase a loudspeaker’s transparency, resolution, fine detail, soundstage layering (depth, width,  height, spacing between singers and musicians), and other audiophile performance characteristics.

For the speakers sent to him that have technical problems (less than optimum crossovers), he designs a new crossover to correct those technical shortcomings. And he also of course builds the new crossover using parts he has found produce good sonic results  For the speakers that already have an accurate frequency response, he just replicates the stock crossover, but replaces the electrolytic capacitors, sand cast resistors, and iron core inductors with audiophile grade parts, including internal wiring, copper binding posts, etc.

People often say (in fact did so just above) "Why not just buy a better loudspeaker, one you like the sound of more?" That’s fine, but what makes you think the better speaker doesn’t also have a "junk"  crossover. All loudspeakers are built to a price point, and using the best crossover parts can increase the cost of a speaker to the point of it not being a commercially viable product. Loudspeaker companies don’t divulge what parts they use in the crossovers, and Wilson Audio "pots" their crossovers so that you can see what’s in them.

The video below---Danny’s latest, posted just yesterday---is an interesting (and unique) case. An owner of a Magnepan 2.6 sent his speakers to Magnepan to have them repaired and refurbished (details in the video). When that was completed he had Magnepan ship the speakers to GR Research, Yes, Magnepan is well aware of the mods Danny offers for their speakers. To hear Danny’s thoughts on the 2.6 (and Magnepan’s in general), you can watch the video.

So if you’re a Magnepan lover, how do you take the advice others have made in this thread, and just buy a speaker whose sound you like more than Maggies? Like what? A "better" model Maggie? Sorry, they all contain the same quality crossovers. At least they did until recently. Magnepan themselves now offer some of their models in the new X Series, which feature crossovers using better parts, internal wiring, binding posts, etc.

And what if you like your current speaker more than any other in it’s price range, but suspect that an upgraded crossover will keep it’s basic sound character, but improve it in the ways I listed in an above paragraph? A GR Research mod is just one alternate to consider.

Yesterday’s GR Research YouTube video is entitled "Is This The Best Deal In High-End Audio? (Magnepan + Mods). It may be, but it has stiff competition from the Eminent Technology LFT-8b, another planar-magnetic loudspeaker. Here’s the video:

https://youtu.be/WjNV8BbZRG8?si=HyRoyMQiMGXhWe1Y

And here’s a video on the ET LFT-8b:

https://youtu.be/Uc5O5T1UHkE?si=yyvaTIj9e-Rc44JL

"Why should I upgrade when I can buy a better speaker?"

That is the title of a video Danny Richie posted a year ago, one I missed. It may answer some of the questions raised in this thread. Here it is:

https://youtu.be/kSOlxLvSR58?si=lE8ci1BobAp2g1QX

Yes, this thread started (and has largely remained) a discussion about loudspeaker crossovers. But now that the subject of open baffle subs has come up (I accept the blame for that ), here’s something that hasn’t yet been mentioned regarding them.

All the open baffle speakers on the market (at least all those I am aware of) have the speaker’s drivers---including the woofer---mounted on a flat baffle, the baffle then mounted on a base. It is common knowledge that one of the penalties of eliminating a sealed or ported enclosure for a woofer is the resulting dipole cancellation phenomenon: without an enclosure to separate the front and rear waves of a woofer, those waves "wrap" around the open baffle on it’s two sides (and top, for that matter), the positive (forward) and negative (rearward) waves meeting on the baffle’s sides (and top), that meeting resulting in a drop off in bass response (+1 plus -1 = 0).

The frequency at which drop off begins is determined by the dimensions of the baffle. The bigger the baffle, the lower the frequency.at which drop off begins (if the baffle is big enough---like the woofer mounted in a wall---it becomes an infinite baffle woofer. A subject for a different thread).

What the W/M and H open baffle "frames" do is increase the distance between the front and rear of the woofer(s), thereby lowering the frequency at which drop off begins. However, if the enclosed space created by adding side and top panels to a flat baffle is deep enough, a "cavity resonance" is created, which we don’t want. The open baffle frames used by Rythmik/GR Research and Linkwitz are all about 14 to 16 inches in depth, which provides a healthy amount of front-to-rear woofer separation (thereby lowering the frequency at which dipole cancellation commences) while avoiding a cavity resonance within the passband of the woofer system.

Ergo, no penalty to the open baffle sound, with only positive benefits. Another benefit is that the added side and top panels provide increased baffle structural stiffness, hence decreased resonance from an unsupported front baffle. In addition, Brian Ding installs a dipole cancellation compensation circuit in the plate amp (Rythmik model A370) that comes with the OB/Dipole Woofer kit (I believe it is a simple 1st-order/100Hz filter, the filter boosting frequencies below 100Hz at a rate of 6dB per octave). The only people who poo-poo the OB/Dipole Woofer system are those who have not heard it. A Rythmik Audio/GR Research OB/Dipole Servo-Feedback Subwoofer fanboy? Oh, you betcha!

Oops. In my directly-above post, I referred to two styles of frames used in open baffle woofers/subs: the W style, and the M. Actually, W and M are used interchangeably, M merely being an upside-down W. The correct nomenclature for the second style frame is "H". The W/M and H style frames each have their own advantages and disadvantages, but are basically equal in performance and sound quality potential. One advantage of the H style is that it is preferable for a 3-woofer sub, as the W/M is really limited to two woofers. If you look at a diagram or photo of the W/M style frame, the reason for that becomes obvious.

AudiogoN member jaytor has a pair of 2-woofer Rythmik/GR Research OB/Dipole Subs (stacked) standing aside each of his GR Research Line Force loudspeakers, with two woofers facing forward, two rearward. The Line Source is comprised of planar-magnetic midrange and tweeter drivers, with some very serious passive crossovers. Check out his virtual system listing for pics. Look at the size of those Miflex caps!

@texbychoice: Your post brought to mind one other consideration in the passive vs. active crossover matter, one Danny brought up in his video. It is this fact: With a passive x/o installed (almost always) inside the speaker enclosure, to replace the (almost universally used) electrolytic capacitors with high quality film or foil caps will require two things: a lot of money, and a lot of space. The latter type of caps are much, much larger than the former, often too large to be installed inside the speaker enclosure. And the cost of a group of caps (large enough in value to equal the value of the stock electrolytic) can be ridiculously expensive, in some cases more than the cost of a good electronic x/o.

One project Danny did was to replace the active x/o that came with the very fine Linkwitz LX521.4 loudspeaker with a passive x/o. That was no small feat, as the 521 is a 4-way design (the stock speaker requiring four stereo power amps!). By the way, the 521 features an OB/Dipole sub, very similar to the one offered by Rythmik Audio and GR Research. Linkwitz built his OB frame in the "W" style (as did I), while GRR offers plans for both W and M style OB frames for their DIY kit. GRR also has a cabinet maker who offers a really good M frame, available both in flat pack DIY form, or fully built. The best subwoofer on the market? For use with dipole loudspeakers, I say yes! So did Siegfried Linkwitz. Another by-the-way: Linkwitz---like Danny Richie---designed his speakers to have flat measured on-axis and off-axis frequency response.

For about a week I’ve been down with a serious bout of a rare form of migraine headache known as a "cluster" headache (I started getting them forty years ago). Imagine getting a bolt of lightning to your brain, and waiting for the next one to hit. Sometimes it’s in a matter of seconds, other times minutes. My clusters usually last about three days (most often starting just above one ear, making it’s way across my head to the other ear), this time the longest ever. The prescription drug I take usually helps, but this time didn’t. People have been known to commit suicide to end the pain of their clusters.

Anyway, catching up with this thread has been for the most part delightful; lots of great comments from knowledgeable, informed audiophiles. As for the others, oh well.

As others have already said, upgrading the crossover in your loudspeakers can be done without resorting to buying a kit from GR Research. And if you like a deep hole in the frequency response they may produce (due to two drivers being out-of-phase at the x/o point, or as in the original version of many of the Klipsch models, the woofer and tweeter not even reaching each other until their respective outputs have dropped way below the mean response of the speaker), that is of course your right. But to call that defect a "voicing choice" is just silly. What it really is, is poor engineering. Klipsch corrected the poor x/o filters in the Mk.2 versions of some models. That is not a matter of a natural evolution (or tighter parts values control), it is correcting a mistake. It was obviously done in response to Danny Richie’s evaluations.  

There is no doubt that Magnepan’s introduction of the X Series versions of their models was also made in response to Danny Ruchie’s evaluation of a few models. I mean, Magnepan has been making all their models with the same crossovers since 1970, the X Series upgrade appearing only after Danny’s videos aired.

Here’s something to consider: I don’t know the prices, but it could be that the GR Research upgrade kit for the, say, Magnepan 3.7i, might be about the same price as the cost to get the X Series version instead of the standard 3.7i. But here’s the deal: Magnepan uses all better parts in the X crossover, but the x/o filter characteristics are no different from those in the standard model. It in no way "corrects" the problems Danny Richie found in his examination of the 3.7i.

What problems, you ask? The same problems John Atkinson found in the last Stereophile review of a Magnepan he did, decades ago. Magnepan has never sent another speaker to the mag, Wendell Diller saying it was because Magnepan’s can’t be measured like other speakers. Both Atkinson and Richie found the drivers played "over each other", a result of the shallow x/o filters and the chosen crossover frequencies. That creates serious problems of comb filtering, a phenomenon known to speaker designers for many, many years. It’s a testament to the quality of the Magnepan planar-magnetic drivers that the speakers sound as good as they do in spite of the flaws in their crossovers!

The beauty of the old Maggies (like the 3.6) is their series crossovers. With a good active crossover (like the First Watt B4 I mentioned earlier), you can create your own filters. I don’t remember what filter characteristics Danny came up with for the 3.7, but with a x/o like the B4 you can try 4th-order low- and hi-pass filters at 400 Hz. If you don’t like it, try something else.

Or, you can just buy a pair of Eminent Technology LFT-8c’s. A single push/pull planar-magnetic driver (magnets on both sides of the Mylar diaphragm) for 180Hz up to 10kHz, a ribbon tweeter for 10kHz up, and a dipole woofer for 180Hz down. Wendell Diller has been insisting forever that a monopole woofer "does not work" with a dipole loudspeaker, and Magnepan has been working on a dipole woofer system for a number of years. Why so long? No need to wait any longer, Eminent Technology already has one.

Or even better, add to your dipole loudspeakers a pair of the unique OB/Dipole Servo-Feedback Woofer that Brian Ding of Rythmik Audio and Danny Richie collaborated on. Open Baffle (2 or 3 12"  woofers in an open baffle frame), dipole output, and servo-feedback control of the woofers. It will play up to 300Hz, unique for a sub. It comes with a plate amp that contains a dipole cancellation compensation circuit, and all the controls you want and need, and in the analogue domain.

Now THERE is a ridiculous statement! Ignorance is bliss.

Dear readers, fellow audiophiles and music lovers, I invite you to compare the tone, attitude, and most of all information offered in the above post by devinplombier to that of Danny Richie in the above video. Draw your own conclusions.

@chuck: Your question of "does Danny measure these speakers in an anechoic chamber?" suggests that you have not actually watched many of his videos. If you had, you would know the answer to that question.

You make an argument against Danny by theorizing "I am sure some of the components have deteriorated or drifted. Also, manufacturing has gotten much better. Components are held to tighter tolerances and measuring equipment has gotten much better." These are what we call specious arguments., the answers to which dispel your theories.

Have you watched the video I just posted above? You really should, it’s full of a lot of accumulated speaker design wisdom. If you haven’t watched it, your opinion is of limited credibility and gravitas. IMO.

It’s not a question of whether or not a person "believes in Danny", but rather if what he says holds up to serious scrutiny. We are all free to make that assessment for ourselves.

For those who are truly interested, here is Danny Richie’s response to Andrew Robinson’s video regarding crossovers. The clips from Robinson’s related videos simply reveal that he (and his off-camera wife/girlfriend) is not a serious audiophile, and is in no way qualified to be considered a hi-fi reviewer/critic. IMO, as always.

https://youtu.be/OSCMw-lGwok?si=VX39SlTp3XRfN_0-

The link @ned1000 posted above is the second video of a 2-part series Danny Richie made on the pair of Wilson Audio Watt/Puppy Series 8 loudspeakers a customer sent him.

Here’s part 1:

https://youtu.be/Tma9jFZ3-3k?si=nJTc2qlLVhmK3WVR

And here’s part 2:

https://youtu.be/pIt2pcQvf6M?si=RzZbPLKteA4DqhaZ

I imagine there may be a fair number of negative responses to what Danny has to say in the videos. All the usual arguments against Danny, with the added factor of Wilson being such a revered brand. All I can say is: I don’t consider it a given that David Wilson or his son are better loudspeaker designers than Danny Richie. In my opinion these two videos actually make the case for the exact opposite. You are of course free to disagree.

You may not be aware of the fact that Steven Stone gave GR Research "The Best Sound Cost No Object" award at the 2013 Rocky Mountain Audio Fest in The Absolute Sound. In one of the two videos above Danny talks about the sound he heard in the Wilson Audio room at RMAF.

A good number of hardcore Maggie owners tear into the crossovers that Magnepan installs in their speakers, especially the guys who frequent the Planar Speaker Asylum.

One easy mod is to replace the stock binding posts with ones from Cardas; the Cardas posts fit in the stock holes with no problem. While they have the back plate off the speaker, they bypass the fuse block (it contains parts made of ferrous material!), which may easily be accomplished by simply moving the round tin connector on the end of the internal connecting wire from the fuse block to the binding post, no soldering required. Of course soldering the wires onto the posts is even better.

The crossovers in most loudspeakers contain compensatory parts---parts needed to deal with problems inherent in the speaker’s drivers. Maggies don’t; Magnepan crossovers are simple "textbook’ filters (1st, 2nd, 3rd, or 4th-order. That’s 6, 12, 18, or 24 db/octave).

Owners of earlier Maggies (X1.6)---which have parallel crossovers (X1.7’s have series crossovers)---have the great option of bi-amping their speakers. Hook up the output of your pre-amp to the crossover, and use one amp for the woofer, another for the midrange/tweeter drivers.

I have a pair of Magnepan Tympani T-IVa (the precursor to the current 30.7), which came with an external crossover. But it is inserted between a single power amp and the Tympani’s, one of the crossover’s output jacks going to the two separate bass panels, the second set of outputs to the midrange/tweeter panel. That crossover contains two filters: a 3rd-order low pass at 250 Hz, and a 2nd-order high pass at 400 Hz. Running the Tympani’s in this manner is of course not bi-amping, and ignores one of the main benefits of bi-amping: removing bass frequencies from the signal the midrange/tweeter amp "sees." Removing them allows for more power to be available to the midrange/tweeter drivers (bass frequencies use up most of an amp’s output capability), and with lower distortion (thanks again to those darn bass frequencies).

In the Tympani owners manual, Magnepan actually encourages owners to bi-amp their speakers, using an active crossover. I have the great little active crossover First Watt used to offer, the B4 (get it? ), which provides 1st, 2nd, 3rd, and 4th order filters in 25 Hz increments from 25 Hz to 6275 Hz. The B4 is completely discrete, no OpAmps or ICs. With the B4 a Tympani owner may not only enjoy the benefits of bi-amping, but also exactly duplicate the speaker’s stock crossover. AND enjoy the benefits of not putting the signal through the stock Maggie crossover.

By the way, Magnepan offers their flagship model---the aforementioned 30.7---in both Standard and "X" Series versions. Rather than spending the money on the upgraded version, I would instead choose to spend it on a good quality external crossover and second power amp. Unfortunately, the 30.7 contains a series crossover, not a parallel one. So to bi-amp it one must perform internal surgery. Not for the faint-of-heart! Not a problem with the Tympani’s---they all had parallel crossovers. In 1973 I bi-amped my first Maggies---the original Tympani T-I---with an ARC passive x/o and two ARC amps.

Like some others here, my current main loudspeaker---the Eminent Technology LFT-8b---employs a midrange driver (a push-pull planar-magnetic design) with no crossover in it’s passband: 180Hz to 10kHz, with 1st-order (6dB/octave) high and low-pass filters at those frequencies, handing off to a sealed dynamic woofer for frequencies below 180Hz and a ribbon tweeter for 10kHz up.

For those who want improved bass response, the Rythmik Audio/GR Research Open Baffle/Dipole Servo-Feedback Woofer (unlike all "normal" subs, it may be used up to 300Hz) may be used in place of the LFT-8b’s stock woofer (the LFT-8b includes dual pairs of binding posts, one for the woofer). Magnepan has been working on their own dipole subwoofer for a few years now (Wendell Diller has long insisted that non-dipole subs "Do not work"---Wendell’s exact words---with a dipole loudspeaker). Magnepan owners: no need to wait for it!

The OB/Dipole Woofer includes a plate amp that provides all the controls the Eminent Technology LFT-8c does (the 8c uses the same planar magnet-drivers as does the 8b, but a dipole---though non-open baffle---woofer in place of the 8b’s monopole), but operates in the analogue domain. The 8c operates digitally.

Excellent post, @deep_333.

It is only in crossovers which produce poor driver phase relationships (or other technical problems) that Danny Richie does a x/o redesign. If he finds no such problem(s), he merely puts together a parts package that replicates the stock x/o, but with audiophile grade high end parts. He takes the stock crossover out of the speaker enclosure, showing all to see what it is comprised of. Watch a few of the videos, and you will see the proof that even many expensive speakers use poor quality parts in their crossovers. Iron core inductors, sandcast resistors, electrolytic capacitors, binding posts with ferrous parts, etc.

The X Series versions of some Magnepan models offer much the same crossover parts package as does Danny’s Magnepan upgrade kit. From the Magnepan site:

"The general idea of the X Series is to take the existing design of a given model and improve it..."

"How is that achieved exactly? Better capacitors. Better coils (inductors). Better resistors. Better wiring. Better connections. Better materials." These were all covered in Danny’s video on upgrading the Magnepan model send to him by a customer, and included in his upgrade kit.

Magnepan continues:

"Okay it’s better on paper, sure, but how does it sound? The short but truthful answer is that it sounds better! A lot better." 

Here’s one of Danny’s videos on the subject:

https://youtu.be/8IQ4t1Y1mxo?si=shhz3kYWZID3IX9f&t=1

@spenav: Yes, he has.

To be more specific, he has found some speakers with crossovers that need (iho) no "correction", but can be improved by using better parts (capacitors, resistors, inductors, coils, binding posts, wire, etc.) of the same electrical value. If you do a search through all the GR Research videos on YouTube, the title of some episodes gives a hint that the speaker under review met with his approval. He has praised the designs of Andrew Jones, for instance. There are some others, but yes they are in the minority. But remember, people send Danny their speaker because they themselves find fault in it. No owner of a Magico is going to do that!

In other cases he has advised the owner that the cost to "fix" their speaker is not cost effective, and they would be better off starting over.

As for manufacturing and selling his own, he does, in two forms:

1- As stated above, he sells loudspeaker kits (subs too) for DIY enthusiasts. He supplies the drivers and crossover parts, the customer builds the enclosure and assembles and installs the crossover. By the way, Rythmik Audio also offers their subs in both factory assembled and DIY kit versions. The plans for the F15HP enclosure call for a 4cu.ft. box, while the factory built version uses a 3cu.ft. one. You can build the enclosure in any manner you want,  as long as the internal volume is correct.

I built my pair of 4 cu.ft F15’s with double walls of MDF and Baltic Birch ply, and braced the Hell out of them (a 1.5" square brace every 6 inches, front-to-back, top-to-bottom, and side-to-side (I copied the honeycomb bracing in the Salk subs, which coincidentally used the Rythmik Audio sub kits inside Jim’s beautiful enclosures). Another fault Danny finds in most loudspeakers is insufficiently-braced enclosures. Tekton, anyone? 

2- Factory built loudspeakers and subs. Danny Richie and Brian Ding of Rythmik Audio collaborated on some subwoofers; GR Research sells the Rythmik Audio F12G, the G used in reference to the company name. The G version of the F12 incorporates a paper cone version of the 12" woofer, and Rythmik sells the plain F12, the cone of the same woofer being aluminum. Danny prefers the timbral character (and lower moving mass) of paper over aluminum, Brian the stiffness of aluminum.

GR Research sells only the F12G sub, while Rythmik Audio sells many other models. Danny and Brian collaborated on a remarkable model, the unique Open Baffle/Dipole/Servo-Feedback Woofer. THE sub for all dipole and planar loudspeakers. It was that sub that led me to GR Research.  

Danny has for a number of years now offered his speaker (and sub) kits as assembled and finished products, the work being done by a couple of cabinet makers he partners with. Those cabinet makers also offer the Danny Richie-designed enclosures (the plans for which come in the kit) the GR Research DIY kits require, selling them as "flat packs"---the enclosure baffle, top and bottom, rear and side panels packed in a carton. All that’s required are some woodworking clamps, wood glue, and paint or veneer. Not for your average audiophile, obviously. Not everyone is an @erik_squires. 

The above responses are not uncommon. Loudspeaker voicing, spectral and timbral characteristics, the basic sound of a loudspeaker company’s designs is understandable and to be expected. Richard Vandersteen has a sound he goes after, as did David Wilson. Everyone claims to be trying to achieve the "accurate" reproduction of acoustic instruments and voices, yet every designer makes a line of speakers that makes recordings sound different from that of other designers.

Danny Richie has addressed the above arguments (the term not used in it’s pejorative sense) in some of his other videos, and does so again in this one. His argument is: Would a designer think to himself "I know what would make for a good sounding loudspeaker; I’ll operate two drivers in such a way that they are out-of-phase at the crossover point where the output of the two drivers meet, the result of which is a 12dB hole in the loudspeaker’s frequency response?" Danny says "No, no loudspeaker designer thinks that."

If you look at John Atkinson’s measurements of the Eggleston models that have been reviewed in Stereophile, you will find the same "hole" in the frequency response Danny did when he measured the model a customer sent him. Is a frequency response hole (12dB down from the speaker’s midean output) a loudspeaker voicing choice, or a design fault? In this video Danny Richie gives you his opinion. You are of course free to disagree with it, and even like the sound of an Eggleston speaker.

I myself have never heard one, but I find the topic of loudspeaker crossover design an interesting and important one. Is a 12dB hole in a loudspeaker’s frequency response a "Problem that doesn’t exist"? The "corrections" Danny came up with for the Klipsch models sent him by customers have been incorporated into the Mk.2 iterations of those models by Klipsch themselves. The crossover ideas Danny suggested and offers for Magnepans are now offered in Magnepan’s own "X Series" upgraded versions of some of their models. Are the X Series versions a solution to a problem their standard versions don't have? Is Magnepan cynically catering to a gullibility they know some audiophiles fall for? C'mon, you know Magnepan better than that!