First Order Crossovers: Pros and Cons

Viridian: All drivers exhibit some type of roll-off that is mass related. Too heavy of a cone for the capacity of the motor structure and suspension and you lose high frequency response. Too light of a cone for the capacity of the motor and suspension and you lose low frequency response.

These aren't crossovers, they are electrical limitations that are brought about by the mechanical design traits of the driver. Given that the entire cone is excited relatively equally across the audible bandwidth, the individual sections of different material types found in the driver are all sharing the load equally.

As far as the ringing of the driver goes, much of this has to do with Ohm's poor choice of type and quantity of damping material inside the cabinet and the horribly attrocious speaker cable used. Fix these two things and you go a long way towards having a much better speaker. Just remember that Ohm used different voice coils on the A's & F's, different frames / support structures, different types / quantities / placements of damping material. As such, each speaker system might be different and would require individual amounts of TLC of different types and in various areas.

Once you've got all of that straightened out, then all you need is an amp that can produce gobs of current and can deal with a high level of reflected EMF and you'll be even better. You have to remember, just as the amp directly modulates the voice coil in this design, the reflected emf from the driver directly modulates the output of the amp. There's no loss or "buffer" from ANY type of a crossover on this design. Not all amps are load stable. As such, some will change tonal balance, distortion characteristics, harmonic structure, transient response, etc... as the drive levels and amount of reflected EMF varies.

Now factor in that this is a very low impedance design, so your "high bias" Class AB amp will be switching over to Class B mode extremely rapidly. In effect, 10 watts of Class A power will only get you about 2.5 watts of Class A power with the F's. If you've got a typical Class AB amp, your 2 watts of Class A power will get you about half a watt of Class A power before switching over to Class B output. Now you get to listen to the constant crossover distortion that is so prevalent in most amps. This is one of the reasons that very high bias AB or straight Class A amps tend to sound so much "sweeter" and "cohesive" i.e. less crossover distortion.

Given that these speakers are appr 82 dB's in sensitivity, unless your amp is of a VERY high bias design, you'll be operating in Class B mode for the majority of listening. Needless to say, this is a very demanding speaker and most amps simply aren't up to the task.

For sake of reference, the amps that i'm runnning with my F's are rated at 250 @ 8 and 500 @ 4. I've conservatively guessed that i've got appr 800+ wpc @ 2 ohms, which the F dips noticeably below. According to the manufacturer of this amp, the first 50 watts are Class A @ 8 ohms, leaving us with 25 watts of Class A @ 4 ohms and 12.5 watts of Class A @ 2 ohms. This keeps the amp from constantly crossing into Class B operation during normal listening levels. I've performed quite a few mods to these amps i.e. cleaned up the signal path, pulled appr 4 feet of wire from each amp chassis by re-routing the internal cabling, added bypass caps as needed, etc... All of the internals are wired with Cardas cabling too. For the record, the Cardas was installed at the factory, not by me.

As was posted in a previous thread, Bill at Millersound considers the Ohm A's and then the Ohm F's to be the finest speakers he's ever heard once he's done modifying them. Coming from someone that has repaired / modified everything from Wilson's to garage sale specials, this is pretty high praise. Then again and based on what Bill told me over the phone, he changes the voice coil, the spider and the foam surround. In effect, the only thing that remains stock is the actual cone of the driver.

Bill says that this brings the nominal impedance up quite a bit, making it an easier load on the amp. The sensitivity is also drastically increased, resulting on a further reduction in strain on the amp. Transient response is also improved due to less driver sag and keeping the voice coil in the magnetic gap of the motor structure. This in turn reduces the ringing since the efficiency of the motor is increased, therefore increasing control over the cone itself.

Other than that, i've got a very large set of E-stat's with dipolar woofers, large 5 driver four way towers, time-aligned monitors with subs, several different horns, etc... The F's, even without Bill's "re-design", are still my favourite speaker. They just do some things that no other speaker that i've ever heard can do. That is, once i found some amps that could do them justice and tweaked the tar out of them : ) Sean
>

Drubin: Those aren't Walsh drivers, they aren't omni's or a point source, there is a crossover network and they are two way designs. The only thing "Walsh" about them is the name. Sean
>

After you hear a true full range point source omni driver, anything else sounds less like music and more like a speaker. This is true regardless of the crossover slope used : ) Sean
>

Drubin: I don't know of any that are currently in production. Even those that come close ( German Physiks, MBL, Huff, etc...) use an additional driver of conventional design ( woofer ) to cover the bottom end. The original Walsh's are very time consuming to build the driver and require tons of hand labor. Three different types of cone material have to be formed, bonded together and internally damped to duplicate the Walsh driver as found in the Ohm F. That's why it has been called the most difficult dynamic driver to rebuild. Sean
>

Drubin: One set of my F's are "smoked" and awaiting a "full tilt" driver rebuild as per Bill at Millersound's suggestions. The other set is packed up. I am currently down to using what was primarily my HT system for both music and movies now. After relocating, i hope to have everything up and running, but i'll have to install a new AC system. My latest purchase ( 450 lb iso transformer / voltage regulation / surge suppression system ) should be here next week or so for that house. Only problem is, i don't know where i'm moving to as of yet. I may end up becoming a Hoosier and moving out of the "Land of Lincoln" if things keep going the way they are here. Illinois is becoming more and more of a "police state" as each day goes by.

Viridian: I hear what you are saying about each individual "layer" of material within the F and / or A, but this is completely different than how a whizzer cone works. The whizzer cone has a completely separate level of output / radiation pattern / frequency response than the main cone. On the other hand, the segmented Walsh design uses multiple types of material that pass signal from one into the other. If any one of the segments didn't pass signal into the other in a relatively uniform manner, you would end up with a loss of surface area. When you lose surface area, you end up with the associated reduction in low frequency extension. As anyone that has ever had a decent running pair of F's or A's know, weighty bass is definitely not their problem.

As to the Threshold 400A that you had, i looked up my notes on this amp. As most of you know, i'm a pretty vocal supporter of Nelson Pass designs, so what i have to say about this amp may come as somewhat of a shock.

First of all, the 400A is definitely not a stable amp. That is, it changes frequency response / transient response / distortion characteristics as impedance varies does so to a pretty noticeable degree. Sonically, the bass is solid so long as the amp isn't pushed. Once you start to throttle the amp, the bass gets soggier. This has to do with the lack of power supply reserve ( not enough filter capacitance ). While the warmth region and lower midrange are pretty decent, as frequency climbs, the upper mids and treble regions were noticeably forward and bright. This tends to highlight any type of lean recordings, exacerbating the problem. With most SS preamps, this amp would tend to introduce a very noticeable sibilance and glare problem.

Once into the top octave, the treble response fell off sharply. This was not as noticeable as one would think though due to the high frequency emphasis just below this point. A definite lack of brass shimmer to cymbals with more of a pronounced "raw & steely" sound.

Due to the increased upper midrange / treble output, transients seemed to be "in your face". The harmonic structure seemed slightly disjointed because of the lack of balance between the lows and highs.

Much of this could have been DRASTICALLY improved by working on and improving the power supply. After all, everything that the amp is capable of ( or doesn't do well ) starts in the power supply. This is why so many aftermarket modifications center around cleaning up / stiffening ( increased reserve ) / improving speed in the power supply of gear. If you've read any of the overly long AC based threads that i've posted, i've commented time and time again about how important the proper design of the power supply is.

As a side note, Nelson Pass has an article about how important the power supply is on his website. As to the Dayton-Wright's, Nelson also used these speakers to test various loudspeaker cabling on. He mentions these speakers in another article on his website, as they used to send some amps into convulsions. Using such a reactive / high demand speaker provided an excellent test bed to see how an amplifier would respond to changes in the load that it saw via altering the speaker cables & their electrical characteristics. As mentioned in that article, some cables allowed the amp to drive these speakers with no problems whereas other cables caused the amp to shut down. As such, there's an obvious audible and measurable difference there i.e. sound from the speakers with some cables and no sound with other cables : )

As far as ribs go, i'm trying to cut down on eating meat. When it comes to good barbecue though, i don't know how i could turn down either beef or pork. That is, until i get serious about my health and dining habits. My midsection is starting to protrude and as hot as it gets in the summer, i'll only be miserable if i don't lose some weight. Either that or convert some of that "mid section speed bump" into muscle on other parts of my body. Can anyone else relate??? : ) Sean
>

Off-axis lobing causes non-linear frequency response i.e. added peaks and dips at various points along the curve. The more consistent the dispersion characteristics of the drivers as angle of radiation is altered, the less noticeable the lobing will be. The differences in baffle related diffraction also come into play here, so speakers with a more advanced baffle design will suffer less from this also. I haven't seen pictures of all of Roy's speakers, but i do know that he's very aware of baffle related problems and seems to pay attention to such things. Sean
>

Good thread and great technical contributions from Roy and Karl.

Good points also brought up by Seandtaylor99 i.e. parts quality has improved in terms of speaker technology, but how well that has been implimented in most designs may be another story. Just getting some of the basics right in older designs places them miles ahead of newer designs using higher quality parts in many cases. It is too bad that some of the "speaker industry giants" aren't around now to take advantage of the better quality drivers that are available to use now.

Sean's comments about "vintage" electronics is also true i.e. i've often said that older products with updated componentry can many times outperform newer products for a LOT less money. Sean
>

Some great points from Golix and i agree with some of the points that he's making here too. This is the reason that i love my modified Ohm F's, warts and all, and why i've said what i have about them. You've got one driver that is both phase and time coherent, covering the entire audible range with great bass weight and a phenomenally spacious radiation pattern. Other than that, and as i've mentioned before, any other attempt at loudspeaker design becomes extremely complex with multitudes of trade-off's involved. Juggling the trade-off's boils down to the personal preferences of the design engineer and the individual buying / listening to the speakers. As such, it is a no-win argument, just a discussion of various beliefs and preferences. There is only one way to achieve specific levels of performance, and at this time, even that approach has limitations. Sean
>