On Friday, 9/28, Grumpy (Dave) came by loaded with an arsenal of testing equipment to get to the bottom of the "hole" issue of the depressed midrange of the 2235 and more specifically, 2205J recently recharged and reconed as a 2235 by Orange County Speaker (OCS).
Sitting drinking my morning coffee, I looked out the windows at about 9:00 a.m. and Dave is walking up the walk. We had talked about the date and time, and not hearing anything (Dave was out of town), I wasn't sure this was going to happen. But Dave was right on time and it did.
First order of business was load software onto my laptop for the test. Dave brought the Parts Express Woofer Tester 2 along with his Mac laptop with the necessary software to determine the differences and possible reason that I was getting a "suck-out" in the 600-900 Hz range. I'm sure that Dave will chime in and tell us exactly what was done, but I know that I needed a bunch of nickels to perform some of the testing.
What we found that was OCS used a legitmate JBL 2235 cone and all parameters supported the JBL tag that was still on it. But the BL, reportedly a measure of the magnetic strength in the gap, was down around 17-18 whereas the 2235 specs call for 20. Dave brought two real 2235s along and the measured BLs were 22-24.
This difference (~15% off spec) could account for some roll off, perhaps, and anyone is welcome to chime in here for an educated opinion. Has anyone had a 2205 recharged and measured the BL and if so, what were your results?
While woofer testing was typically performed in Dave's calibrated lap, all in cabinet testing was done with an L200 cabinet. Most tests were performed with one port blocked. Dave's graphics clearly showed the benefits of blocking one port, although at those frequencies the result is less audible than visual in the plots.
Replacing the Alnico-framed 2205/2235 with a new 2235 in the L200 cabinet did alter the response somewhat, but after many measurements and mic placement changes, we decided to move the speakers relative to the walls. Dave feels that sidewall reflections appear to be the main suckout culprit in the 500-900Hz octave. The real 2235 did exhibit a response similar, though not identical to the 2205/2235 and in these cabinets, in this environment, and using these crossovers did have the dreaded dip in response (that most people don't seem want to acknowledge ).
Scooting the systems closer (inches) to the rear wall helped smooth the response a bit elsewhere.
The day quickly passed and we headed off to Ruby's for lunch. We got back to the house and continued the tests going through various combinations of crossover settings, and slight movements of the speaker and mic placement, and after reinstalling the 2205/2235, testing the right and left speaker for similarity.
BTW, Dave's tests did prove what I asserted all along that a 16 ohm resistor in series with the woofer did smooth out the troublesome 600-900 area to a large extent! (Actually we used a 20 ohm in this case.) I know that many here said that this is strickly taboo (putting a resistor in series with the woofer), but there it is! The really weird thing is, it hardly had any effect on the woofer's volume! Dave has the plots.
When I reconnected the speaker to the amp, because the speaker had now been moved back to the wall (was pulled out) I couldn't see the wires and inadvertantly put the + on the red and - on the black (wrong for my system). I fired up "You're So Far Away" (on Brothers in Arms) and Dave instantly said "something's wrong". He instantly knew the polarity was wrong and the Yamaha's self test agreed.
Later when testing the speakers for similarity, we found differences in the plots because the right tweeter was out of phase. I phased them using a sound level meter looking for the louder volume for a tone near the crossover point. But that portion of the spectrum is suseptible to combing, so that small differences in frequency have large differences in volume. Anyway, Dave's plots clearly showed the difference and both speakers track extremely well together now both in volume and frequency.
According to Dave, the speakers are +/- 3dB from ~40-14KHz (albeit 1/3 octave) and that's still nothing to sneeze at. Some by-ear tweaking of levels for the listening position made the system pretty smooth to listen to (the point, after all). L-R matching for frequency response tracking is within 1/2dB almost throughout the useful bandwidth, according to Dave, in no small part due to my carefully measured speaker placement and fairly symmetrical room.
I looked at the watch and it was now 4:00 p.m. Time for Dave to head back to the family. A lot of data was obtained and there may be some suggestions out of this to make the "keeper" crossovers work even better to fill the hole.
What a great day! Thanks for stopping by Dave!