The purpose of JBL E145

[frank23]

This is what Zilch on the forum measured for the tuning frequencies:
http://www.audioheritage.org/vbullet...l=1#post216106

I've measured and documented 4507 tuning several times in these forums.

Here's my recollection, but y'all can look it up to be sure:

4 ports open = 40 Hz, the stock tuning for 2225H, E140, and similar.

One port closed = 34 Hz, stock 4430 tuning.

Two ports closed = 28 Hz, extended bass or sub duty, which is how I run them mostly....

So I'm running the E145 with one port closed in the 4507, so tuned to about 34Hz.

[gibber]

This is what Zilch on the forum measured for the tuning frequencies:
http://www.audioheritage.org/vbullet...l=1#post216106


So I'm running the E145 with one port closed in the 4507, so tuned to about 34Hz.

That two-port 28Hz tuning you like less is very similar to the 30Hz i have here. Instead of changes on the ports my cabs have, i just tried using a beefy transistor amp and gave it a 10ohm resistor in series to my (early version) K145's ca 9ohm coil dc resistance. New Qes' equals (Rg + Re) / Re * Qes, so Qes goes up to 0.52 or so and hence new Qts is around .49 (see Qms in tables earlier in this thread).
Result is a more "normal" bass sound, the mid is similar as before, and subjective speed is slightly less, imho not too much so to offset advantages in the deeper register. Mid output subjectively is down by 2dB (should be 3 dB but the higher Qts with series-R maybe reinstates some of the premature-but-gradual rolloff that a driver of Qts=0.23 w/o series-R has).

[frank23]

But then what happens to the amp damping the cone when you put a 10 ohm resistor in between?

[ivica]

But then what happens to the amp damping the cone when you put a 10 ohm resistor in between?

As resistor being 'in series with the driver" the amplifier would "see" less load, , so less current, so "more comfortable"

regards
ivica

[gibber]

But then what happens to the amp damping the cone when you put a 10 ohm resistor in between?

Qts influences what we hear frequency/impule response-wise. Qts expresses the composite of electrical and mechanical damping near resonance. So even if the amp dampens less, there is still the same mechanical damping of the cone. I consider Qts=.23 for K145 after re-gaussing (E145 is similar) as too low from what i heard: "fast" as you rightly said, but overdamped, lean bass. Some drivers work well having such low Qts, but their fs/Qts (or fs/Qes) ratio is lower.

One way around it would be to mod the speaker or box mechanically in order to alter damping; or do it through electrical changes. The 10 ohm resistor will give the amp less control over the coil (coil impedance divided by feeding impedance equals "damping factor"). I reduced damping factor to about unity. So the amp is less able of damping the cone, intentionally so in my case, as i deemed the sound overdamped.

If you deem a driver Qts' of 0.49 a bit high for reflex applications, you may want to optimize the resistor for what are called "B4 maximally flat" alignments. According to literature, in a lossless enclosure this is achieved at a driver Qts of around 0.38. Any higher Qts (or Qts' in my case) and you get some amount of undesired ripple in your response. From the tables earlier in this thread you can see Qms of my old version paper-surround driver is around 2.5. Your E145 will be similar, maybe 3.1. Using Qts' = Qms*Qes' / (Qms+Qes') gives me a "target" Qes' of ca .44 to achieve Qts' just shy of .38. Choose Rg' as Qes' / Qes * Re - Re. Note that Rg' is the addition of amp, resistor and cable resistance. For many transistor amps and for short runs of cable, it's often safe to assume zero ohms for both. Then, 4.7 ohm or 5.1 ohm are then the E6 resp E24 std resistor values for targeting your E145 at Qts' 0.38. Try it out, it's easily done. Or try a 10 ohm resistor first and then parallel a second 10 ohm after listening in for a while

Ralph

[frank23]

Thanks for the thorough explanation above btw. I have yet to try and add the 10 Ohm. Recently I had a pair of Event Opal studio monitors on loan. They are fabled for their low end, which is enormous for a speaker their size I must say. But when I hooked up the E-145's again. Man, that is so much better. The 145' reach much lower, and give much more weight to the low end. I can't imagine them cutting of fairly high as you say. The Opals are supposed to go to 35Hz in room, but the E-145 in the 4507 have much more weight to them.

[gibber]

It's sacrilege, i know. Would still be interested to hear what those that jumped their own shadow have found out, though
New year day is often pretty boring - why not get hold of a pair of resistors. Can be pretty low wattage with high SPL cones like E145, unless you go for PA sound levels

Enjoy the fireworks tomorrow
Ralph

[ivica]

It has been posted here on this forum that the E145 T/S values as published by JBL cannot be correct as they do not match up like they should according to the T/S formulas. There probably was a typing error in the Vas.

......!

I think for E145 it has to be Vas=427.7 lit

regards
ivica

[4313B]

I think for E145 it has to be Vas=427.7 liters

That is what we arrived at.

It's sacrilege, i know.

Mark Gander and I went over this back in the late seventies with respect to his overdamped 124A aka 2203 and 136A aka 2231 designs, Qts = 0.14 and 0.19 respectively, along with various series resitor values. In the end it was determined that the DCR of the low pass inductor coupled with careful volume/tuning ended up being the best solution. 4315 volume/tuning and 4331/4333/4343/L300 volume/tuning along with their respective passive networks and whatever EQ was going to be applied in the Studio.

One of the hallmarks of the traditional JBL transducers was their low Qts. Arguably the 124 and 136 went a tad overboard.

In any case, examine the low pass filter in the DD55000, which used a variant of the E145. Unknown would be the DCR values of the coils.

[frank23]

I finally tried this today, after having tinkered with it on and off over the past few weeks.

With a measurement microphone and pink noise I measured 40Hz in comparison to 100Hz. With a 4.4Ohm resistor in series with the E145 40Hz was at -14dB compared to 100Hz. Without resistor 40Hz was at -13dB compared to 100Hz. Of course due to the resistor the total level was different, but its the difference between 40Hz and 100Hz that we are looking for here. I could see no improvement in lf extension when inserting a resistor. in fact extension suffered a little. The E145 was in a 4507 5cu 150l cabinet. I think I had it with all 4 ports open, but I was tinkering with that also, so I can't recall that exactly, might even have been closed :-)

I know, 1dB is not much, but the measurement was repeatable, it definately showed that there was no gain in LF extension with the resistor in place with the E145 in the 4507 cabinet.

[frank23]

It may sound strange regarding the T/S data, but I am now using the E145 in a closed 4507 cabinet. I moved it forward 30cm, and the bass sounds very fast. I probably miss the lowest octave, but the in-room bass extension is much better than I expected and is very controlled with great energy. Organ is just fine. Maybe home theatre fans would miss a bit of the thunder of a helicopter flying over or so, but I listen to music.

I crossover to the 2123 quite low at 100Hz, maybe the rising output from the E145 fills in the missing 2123 output between 100-200Hz and makes it a nice match that way, I'm not sure. Experiments with lowering the Qts through resistors were not satisfactory for me, so they are hooked up straight to the amp again.

[frank23]

I have seen the E145 frequency response curve as measured by Greg Timbers many times, but have never seen the version I found here, with the 2nd and 3rd harmonics added in the graph (note: raised by 20dB).



Would anyone know what the vertical scale here is? If it is 2dB for every horizontal line, then this E145 is about the best low-mid driver ever with a 2nd harmonic at about -70dB at 150Hz. It then betters the current 15" from the DD66000. If it is 1dB for every horizontal line, distortion at 150Hz is still low at -45dB, but not really special.

[Earl K]

I have seen the E145 frequency response curve as measured by Greg Timbers many times, but have never seen the version I found here, with the 2nd and 3rd harmonics added in the graph (note: raised by 20dB).



Would anyone know what the vertical scale here is? If it is 2dB for every horizontal line, then this E145 is about the best low-mid driver ever with a 2nd harmonic at about -70dB at 150Hz. It then betters the current 15" from the DD66000. If it is 1dB for every horizontal line, distortion at 150Hz is still low at -45dB, but not really special.

I'm quite confident that the displayed 1700hz breakup peak is not 16db high / I'll believe 8db .

I'll support ( the opinion ) that the total vertical scale being displayed, is 40db .

Keep in mind that the chart shows 9.2 volts ( as the applied power ), so it's still very clean ( IMO ) .

[frank23]

And if the noted voltage is correct, it should probably say "10W@1m", not "1 W@1m" as it says now. It looks like the "0" was removed. I have no idea how distortion scales with level though.

[Earl K]

And if the noted voltage is correct, it should probably say "10W@1m", not "1 W@1m" as it says now. It looks like the "0" was removed. I have no idea how distortion scales with level though.

Here's a ( very ) handy online calculator for ( decibel > THD distortion conversions ) .