LARRY VAN WORMER'S RS LX5 PRO MODIFICATIONSAbout a year ago I bought a couple of LX5's from Radio Shack after reading glowing reviews on the list (for the tweeter) and in print. This was my first forray into the audiophile (and I mean that in a good way) realm, I was quite happy with the speakers initially. As soon as I added my 1259 sub, I noticed the boominess of the upper bass coming out of the LX5's, and the icky way that they exagerate sibilant (or "s" type) sounds. I'm not sure why, but it seems like you get one really good driver in your system (anywhere) and everything else sounds like crap in comparison. It turns out that the sibilant mishandling is caused by nasty breakup modes of the woofer, and these can be tamed by use of a higher order xover. Mod 1: Stock drivers with crossover changeLarry Van Wormer has performed extensive characterization and optimizations on the LX5, and he was kind enough to send me the schematic of the xover he developed for them -- 3rd order electrical and 6th order acoustic for both tweeter and woofer. Click here for the instructions in MS Word format. For those of you without MS Word, click here for a plain text version of this mod. This document also discusses enclosure treatments. Here is the crossover from the document:
Tweeter crossover:
(Note that the tweeter phase is reversed)
Neg-------C1----------+------C2-------+ C1= 7.2 uF
| | + C2= 10 uF
L1 Tweeter L1= 0.3 mH
| | -
Pos-------------------+---------------+
Woofer crossover:
L2= 1 mH (same value as orig,
Pos---------L2----+----L3-----+-------+ but I recommend aircore)
| | | L3= 0.52 mH
| C4 | + C3= 22 uF
C3 | Woofer C4= 38 uF
| R1 | - R1= 1.5 ohms
| | |
Neg---------------+-----------+-------+
Crossover Components:
(If anyone wants to use other types of components, of course, go ahead.)
C1 can be the original component, with a 1 uF polypropylene bypass
C2 should be at least a cheap polypropylene, such as Solen or Bennic
L1 can be a cheap aircore, such as Madisound's cheapest line.
L2 and L3 should be aircores, at least equivalent to Madisound's Sidewinders
C3 and C4 can be something cheap, such as Bennic bipolars
R1 can be a cheap power resistor, 15 watts is plenty.
The high order is very desirable, both for reducing woofer resonance at higher frequencies, and because the drivers are not time-aligned (higher order xo reduces the overlap region). This xover even reuses the existing xover components in the LX5. Kudos to you Larry on this great design!
Mod 2: New SEAS woofer with crossover changeI used Cool Edit to create warble tones to check the bass response of the LX5 via close miking with my modified RS SPL meter. The tests showed a broad upper bass resonance, even with the ports plugged and the cabinet stuffed with poly fluff. This indicated to me that the cabinet was too small for the stock driver, and way too small for a ported alignment using this driver (which is how the stock LX5 comes, BTW, go figure. Larry recommends *really* stuffing these babies if you keep the stock driver, see his mod instructions above).Not one to leave well enough alone, I decided to replace the stock woofer with a driver that would work better in the given volume. Someone on the list (I'm sorry, I don't remember who) said that the SEAS MP14RCY/P (5" driver with phase plug @ $38.20 from Madisound) was a drop-in replacement for the stock woofer in the LX5. I checked the dimensions, plugged the numbers into Blaubox, and sure enough the cabinet was actually somewhat large for the SEAS driver, which should be good down to ~100Hz in a sealed box. Fine with me, this left room for the new xover components and double-dynamatting of the interior, and still plenty of "experimental" volume that could be used to get the sealed volume just right. Reducing volume is *not* a problem (just ask any theoretical topologist worth his Klein bottle: it can be sent off into the fourth dimension via solid fibrous hexahedrons inside the cabinet). The SEAS drivers came and I mounted them in the LX5's (I used new screws to mount them: #8 - 1/2" hex head sheet metal screws.) This was my first experience with such a smooth midrange, and I was very impressed. The sibilance problems were completely gone, and the difference in smoothness over the stock drivers was so pronounced that it was somewhat difficult to get used to at first. Close miking showed no resonance problems up to a couple of kHz, and only a couple of dB variation from 200Hz up to the cutoff frequency (1/3 octave warble tones used here). The frequency response plot in the Madisound catalog is about as ruler flat as they get, and this was confirmed through my (somewhat crude at this point) measurements. I finally figured out the speaker testing program "Speaker Workshop" from Audua and started using it heavily. I also constructed several inductor and capacitor substitution boxes suitable for use in crossover design, as well as a wand-type microphone, and a new preamp for it. With this test setup, I tried Larry's xover (the one designed for the stock LX5 woofer) and found it to be not too horrible even for this new Seas woofer. The tweeter level was down somewhat compared to the new woofer, though. Larry says this woofer is more efficient than the specs would lead you to believe (most manufacturers include the baffle step in their efficiency numbers, leading to an automatic +3 dB SPL increase). Larry had been eyeing this Seas driver for a while, and suggested that I mail him one so that he could measure it and design a new xover for it. I dropped everything and sent him one. Larry is a really great guy, and helped me immensely by sending me his intermediate results and patiently answering my questions regarding them. I would set them up on my stuff and fiddle around with the values, observing the results. He even suggested a method for xover design that worked very well. Through these exercises, I was able to debug my setup, as well as gain an appreciation for the true art of xover design. Larry uses LAud and CALSOD for his work, LAud to capture the "model" or frequency response of the drivers, and CALSOD to do the actual xover design in. I'll be getting a copy of CALSOD soon, and hope to get Speaker Workshop to the point where it will kick out a useable model. Success! Larry has finished with the xover design, and it appears below in his own words:
I have been tinkering away with the crossover design, and have
settled on the following as my recommendation. As with all such things,
it's a compromise:
- good spl flatness
- quite good phase match
- decent impedance
If I aim to get any one of these perfect, I compromise the others too
much, in my opinion. As it stands, this is really pretty good. I've
listened to it, and as well as I can tell from the one speaker, it
sounds just fine.
I will be polishing up my description of the process over the next
couple of weeks.
Woofer:
pos-------R1---L1---+---L2---+--------+
| | |
| C2 | +
C1 | Woofer
| R2 | -
| | |
neg-----------------+--------+--------+
R1 = 1.0 ohm
R2 = 2.7 ohm
L1 = 1.2 mH
L2 = 0.4 mH
C1 = 16 uF
C2 = 25 uF
Tweeter:
pos------C3-----+--------+
| |
L3 | +
| Tweeter
R3 | -
| |
neg-------------+--------+
C3= 10 uF
L3= 0.3 mH
R3= 2.0 ohm
Larry recommends Madisound Sidewinder inductors (or similar quality air core)
for L1 and L2. C3 should be a decent quality poly cap. The rest of the
components can be somewhat lower grade (bipolar electrolytics and cheaper air
core inductors) if desired. Typical DCR's are already figured into the
inductors.
Here is my measurement of the above xover with the new SEAS woofer in the LX5 enclosure (via Speaker Workshop and capacitor and inductor substitution boxes):
This is a rather crude splice done at 500 Hz of the close miked and gated 1 meter measurements. Note that the response is down 6 dB at 100 Hz. I haven't attempted to adjust the internal volume yet, and this may extend the low end response. Larry undoubtedly has a better curve for this, since his equipment is much better, and he is controlling for confounding factors in a more systematic fashion. The roll off around 10 kHz is probably due to my soundcard.
Parting thoughtsI admit that this is a somewhat expensive mod to a cheap speaker, and I did it mainly to get out of constructing boxes (live in an apartment at the moment). Total cost: $75 for each LX5 (on sale) + $40 per new woofer + $20 per new xover = $135 each. Kind of high, but sometimes you have to go with what you already have. In for a dime, in for a dollar, I alway say.I'll post more when I get the components for the woofer circuit and have a chance to give the speakers an extended listen.
|
