L100T vs L100T3

[pierce]

The phase reversal has to do with the type of network and how the filters sum at the crossover point. Without reversing the phase there would be a hole or peak in response at crossover. These are typical 2nd order networks that sum flat electrically at the crossover point only if the middle driver is out of phase.

absolute phase has always bothered me. if all three drivers are the same inversion, relative to the other version, and the crossover networks are essentially the same design albeit slightly different frequency rolloffs, I'm still not sure I understand why + vs - makes any difference.

The midrange and woofer interact above and below the crossover point. No crossover frequency is a hard value, there is content from both drivers above and beyond the frequency. The difference is how the polar response works along with the level of each driver. Subtle changes in the crossover components can have marked results and usually do not agree with calculated values. Many of JBL's designs violate rules for crossovers but that is because they leverage the natural response of the drivers. Many times the designer goes for a typical "sound" versus an absolute flat response.

sure, I've always assumed that crossovers use the speaker as an active element of the circuit, and are designed around the speakers natural characteristics and limitations. The extreme case is the simple 2-way crossover that consists of just a capacitor in series with the tweeter... thats relying on the natural lowpass of the woofer, and that capacitor + the tweeter's voicecoil/inductor comprise a single-pole L-C filter.

The T3 circuit also dampens the high frequency to soften the tweeter response. Metal domes can be a bit harsh so this is an improvement. The midrange is where the music is anyway so the tweeter augments the midrange without overpowering it.

I wonder if I couldn't simply add a small value resistor in series with the 035Ti to achieve some of this rolloff? Although, with my 55 yr old rock-damaged hearing, they don't sound too bright to me

Essentially what the T3 network does is create a "hole" in the transistion between the woofer and midrange by lowering slightly the woofer's crossover point and raising the midrange crossover point while still being centered on 800 hz. . The 2214H has a rise in response above 1.2 Khz so this reduces that peak. The midrange is raised a little and that isolates it a bit from the woofer while tweaking the loudness slightly.

The result is a bit smoother midrange, less shrill highs, and the bass seems clearer. With some tweaking other JBL designs could be helped this way too but we have a production standard that can be duplicated which has the blessing of the JBL design engineers.

There is a host of other things going on too like driver spacing, time alignments, impedance, phase response, edge diffraction, and temperature changes too but in the end the T3 network is a bit nicer to listen to and that is all that matters.

the driver spacing is nearly identical, as are the time alignments (after all, both speakers are planar, and not stepped back like time coherent designs such as Vandersteen). Both L100T and T3 speakers have the same sharp cabinet edges, and the same impedance drivers. I know the T3 is a little deeper, and there was some conjecture on one of these threads that it was a little sturdier (I'd love to see some actual evidence of this, as when I first saw the T3 in a store a couple years after buying my T, I got a distinct impression it was a little cheaper construction, the cabinet felt lighter).

anyways, thanks for the analysis.

[duaneage]

absolute phase has always bothered me. if all three drivers are the same inversion, relative to the other version, and the crossover networks are essentially the same design albeit slightly different frequency rolloffs, I'm still not sure I understand why + vs - makes any difference.
Because drivers have voltage and current lags that affect the interaction with reactive components like capacitors and inductors. The result is a change in output as the frequency changes through the crossover. In actual practice the midrange and/or woofer is complementing the acoustic output of the other driver. If they are in phase in some networks they compliment too much causing a ringing effect at the crossover point.


sure, I've always assumed that crossovers use the speaker as an active element of the circuit, and are designed around the speakers natural characteristics and limitations. The extreme case is the simple 2-way crossover that consists of just a capacitor in series with the tweeter... thats relying on the natural lowpass of the woofer, and that capacitor + the tweeter's voicecoil/inductor comprise a single-pole L-C filter.

Actually the simple 2 way capacitor only design is far from extreme, it's preferred by many. The problem with a simple capacitor is not so much the response (in fact it's simplicity is a major advantage) but the lack of driver protection because too much low frequency information makes it to the tweeter. I don't like woofers run wide open this way because they make lousy tweeters and cause two point sources in the midrange. The L100 is a classic example of this and I have never liked the sound of this system because of the crossover network, or lack of one really.

I wonder if I couldn't simply add a small value resistor in series with the 035Ti to achieve some of this rolloff? Although, with my 55 yr old rock-damaged hearing, they don't sound too bright to me

Sure, an L-pad would be better since it could control the output for the room. In the T3 versions, the shunt resistor remains 20 ohms while the series doubles to 2.4 ohms. The series capacitor changes to reflect the increased resistance seen by the 4 uf cap. If the cap was not changed the crossover frequency would drop as the total load was a little higher than before. Small detail maybe, but if you want to protect the tweeter while keeping it in the proper operating range you have to account for it.

the driver spacing is nearly identical, as are the time alignments (after all, both speakers are planar, and not stepped back like time coherent designs such as Vandersteen). Both L100T and T3 speakers have the same sharp cabinet edges, and the same impedance drivers. I know the T3 is a little deeper, and there was some conjecture on one of these threads that it was a little sturdier (I'd love to see some actual evidence of this, as when I first saw the T3 in a store a couple years after buying my T, I got a distinct impression it was a little cheaper construction, the cabinet felt lighter).

anyways, thanks for the analysis.

Crossover networks can compensate for time alignments to a certain degree, the offsetting of the drivers from center helps with edge diffraction. I doubt the cabinets have much effect on the crossover action, but one never knows. The mid and tweeters are at ear level while sitting, that makes a huge difference compared to lower placement.

I think the L100T and T3 speakers are excellent systems which have huge potential for experimenting and reward. They are still relatively affordable and when you look at the drivers used they are a great bargain. Not everyone likes the full veneer look, I happen to like it very much.

[pierce]

Actually the simple 2 way capacitor only design is far from extreme, it's preferred by many. The problem with a simple capacitor is not so much the response (in fact it's simplicity is a major advantage) but the lack of driver protection because too much low frequency information makes it to the tweeter. I don't like woofers run wide open this way because they make lousy tweeters and cause two point sources in the midrange. The L100 is a classic example of this and I have never liked the sound of this system because of the crossover network, or lack of one really.

FWIW, if you look here http://www.audioheritage.org/vbullet...ad.php?t=24894 you'll see the *original* L100 century had a multi-pole crossover, while the L100A "Late" had a simplified cap-only crossover. Now, the L100's wasn't as complex as that on the T's, notably, the woofer is still running wide open, and if I read that right, the mid and tweeter are both strictly high-passed, but at least its an L-R-C network and not just a C...

anyways, obviously I concur with your eval of the L100T/T3... I own a set When I bought them, circa 1988(?), I took 5 CDs to about a dozen audio parlors and auditioned the a specific track off each disk on each system I was interested in in the $1000-2000 range. The L100T was the first and last system I listened to, and I bought them based on what I'd heard. I think I paid $800 for the as-new demonstrators in this small audio store in my town, after having been to all these fancy places in/around Silicon Valley, listening to Klipsch, Kefs, and all manner of $2000 speakers what I can't remember.

The one pair of speakers I've heard since then that really sounded significantly better to me across a wide range of source materials (and weren't in the ludicruous-$$$ <cough>B&W 801</cough>) were Vandersteen 2c's. Setup right in a big open room, they had an amazing stage presence and clarity whether playing a folk string quartet, or choral harmony or a live rock concert at full tilt.

[duaneage]

The midrange used in the original L100 is worlds apart from what the L100T ended up with. Lower FS, smoother response curve, better cone materials. The L100 essentially was the home version of a studio monitor JBL designed to replace Altec in the studio. The sound they were going for was specific.

[evans224]

I understand some of the discussion on crossover theory. What I understand better is the way that the L100T's sound after having the crossovers updated and modified to L100t3 specs (Duanage does good work!). The result is a much smoother upper/midrange and a more defined. less muddied low end. I suggest you give it a try. I actually have a spare set of L100t crossovers that I am going to put in and I will A/B the speakers.

[pierce]

Duneage, you mod these crossovers for folks? Interesting. Actually, I'm pretty good with a soldering iron and have a weller temp controlled soldering station, so if I can source the parts, I could do it myself (I'm on a real tight financial leash this year, my wife has been unemployed from her tech-writer career about a year, a kid goes to college next year, work is downsizing, eeeek).

Think I can get away with eyeballing the 'type' of the parts in my existing crossover, then ordering the alternate values from digikey or mouser?

It strikes me that the inductors will be the hard ones, do you custom (re)wind them ? for instance, in the woofer section, L1 goes from 3 to 3.2mH, C1 from 40 to 44uF, and the R goes from 39 to 41 ohm (via 2x82ohm in parallel). in the mid, C3 goes 17 to 15uF, L2 0.5->0.6mH. In the tweet, C7 goes 3 to 4uF, and R4 goes 1.2 to 2.4ohm.

Do you think the R and/or C values could be tweaked to keep the L values the same?

[duaneage]

I did evan224's crossovers, I have a spare set I can exchange with. I did not change the shunt resistors across the woofer because they are so close in value to each other tolerances pretty much make them the same. I replaced the electrolytic capacitor in the midrange circuit with a high quality poly cap and used poly caps for the high pass as well. The resistors in the mid and high pass were changed as well.

The inductors I sourced were too big for the board so I extended the two coils with appropriate gauge wire and verified the values to within 1%. Then I reassembled the boards and tested them in my L100T speakers as well as bench checked them to make sure all was well. I also replaced some connectors with new ones because they were a little corroded.

I'm toying with the idea of offering this as a service, I can do single day turnaround because I have a set to core with. I need to determine what the parts and time costs are and see if it is worth doing. I spent about 1.5 hours on Evan's networks doing the work and another hour testing and evaluating them.

So a short answer is yes, there is no reason why the crossovers could not be modified provided you source replacement parts, have a selection of coil wire handy, the proper tools to solder with, and a means of testing the inductors (or replacing them) to make the switch. Along the way an enterprising engineer might want to replace all the electrolytic s with better caps for durability and quality.

[cosmos]

I'm toying with the idea of offering this as a service, I can do single day turnaround because I have a set to core with. I need to determine what the parts and time costs are and see if it is worth doing. I spent about 1.5 hours on Evan's networks doing the work and another hour testing and evaluating them.

Interesting.. have you gone any further with this idea? I have a set of L100T that I am toying with upgrading...

[duaneage]

Interesting.. have you gone any further with this idea? I have a set of L100T that I am toying with upgrading...

I have worked on streamlining the process a bit and identifying suppliers for the parts. I guess the real questions is what levels to offer for this. I could replace the electrolytics with different value electrolytic capacitors for an economy upgrade but most people prefer high quality capacitors instead. Replacing the midrange and high pass caps with polys and simply adding a poly cap to the low pass is a step up in performance.

Replacing all the electro caps with poly would cost quite a bit more and there may be space limitations on the boards, the poly caps are larger by a significant amount. This might require new circuit boards and that gets complicated.

Finally I've been asked about a CC biased version. I could make a CC version of this network but it would definately require a new board, all new capacitors, and additional parts as well. This would be the Lexus level and probably cost as much as a pair of T3 speakers used.

So if I buy large quantities of the parts and keep it simple I can do 3 levels of upgrade.

1. Simply replace the caps with different values, rewind and test the inductors, new resistor for the high pass and test the results.

2. Poly caps for new values, same tasks as above.

3. Poly caps all around, same tasks as 1 and 2.

Once I get an idea of costs and benefits I'll post something. In the meantime contact me by PM if you want to do something and we can discuss.

[JSF13]

I understand some of the discussion on crossover theory. What I understand better is the way that the L100T's sound after having the crossovers updated and modified to L100t3 specs (Duanage does good work!). The result is a much smoother upper/midrange and a more defined. less muddied low end. I suggest you give it a try.

I'll second that. I just picked up Heather James L100T's with crossovers upgraded by Duaneage and I have to say, having owned L100T's before, that I totally agree with the above. Any of you out there with L100T's should definitely think about doing this upgrade. You won't be sorry.