[Tyson]

Hi all,
I built a pair of Elsinore speakers, as detailed in a separate thread here in the lab.  I have figured out a couple of new things about my preferences.  First, I hate ports.  So, I've already deviated from the original design of the Elsi's by plugging the ports.  I lost a bit of bass extension, but gained immensely in musical enjoyment.

The other thing I'm considering changing - they are MTM designs for the top 3 drivers, with the bottom 2 woofers "filling in" the bottom end and dealing with baffle step.  I'll run them this way (as designed) for a while to see how I like it. 

But, one alternative I am considering is just running the top mid-woofer and the tweeter together in an MT design, and using the bottom 3 woofers for just the bass.

OK, finally my question!  I do NOT want to run the 3 woofers in parallel, as it would present a 1.8 ohm load to my amp.  And running them in series/parallel seems to be out since that seems to only apply to an even number of drivers.  So my only option seems to be running all 3 in series.  Question - will this cause a problem for my amp?  Or any other problems I'm (almost certainly) ignorant of?

[Tyson]

Oh and I forgot to mention - I have almost unlimited ability to shape the frequency response in-room since I'm using a (highly modded) DCX2496 and 3 amps in a fully active configuration:  and SS amp on the bass, a tube amp on the mids, and a second tube amp on the tweeter.

[JoshK]

Are they 5 ohm woofers?  8ohm woofers, 3 in parrallel is 2.7ohms.  3 in series is 24 ohms.  This is ok with most tube amps although you do changes some things with respect to load line of the output stage.  It makes the tube amp have a higher dampening factor.  It will likely operate at less distortion as well but will also produce less watts.

That is true for many amp types, that you will get less watts, but otherwise the amp will likely be quite happy.  Someone with a bit more knowledge of feedback in SS amps can chime in on the stability of an amp under a higher impedance, but I think all should be well.

If your tube amp has a lot of global feedback, any compensation circuit in the feedback loop might need retuning, but it should not be critical as most tube amps don't use tons of feedback (they can't). 

An OTL amp would likely be very happy with such a load.

[Daryl]

Hi Tyson,

Having the ability to shape frequency response is great but of little use unless you have the ability to measure the system to see what you need to accomplish.

You not only will need to level the response of each range but contour the cuttoff slopes so that each driver range remains in phase with the others throught their region past the crossover frequencies until their output is insignificant and can no longer interfere with the other driver ranges.

This also requires some understanding of the correlation between response magnitude and phase shift.

From the details of your post it sounds like you are talking of three 8 ohm nominal drivers with 5.4 ohm Re, correct?

Three drivers in series should make an easier load for most amps.

Your amplifier may not produce enough voltage to drive the three drivers to full output when connected in series since the load impedance will then be 24 ohms nominal and the amplifier will only achieve 1/3 it's 8 ohm power rating.

A high quality amplifier likely can drive three 8 ohm nominal drivers in parallel without issue.

Also since your design was intended to be vented and you are running the drivers in the same cabinet volume with the vents plugged your Qtc (system damping) should be too low unless the original design was not optimum (which is also possible since you stated you did not like it with the vents operational).

This being the case a good solution would be to wire the three drivers in parallel and use a 0.8ohm resistor in series with the parallel connected drivers.

This will increase Qtc about 40% resulting in a more extended bass response and your amplifier will see a load on the low side of 4 ohms nominal.

For the 0.8 ohm resistor you can use five 4 ohm / 20 watt resistors in parallel.

Here is a diagram for connecting the resistors...

Of course you will have five 4 ohm / 20 watt resistors rather than the three 5 ohm / 25 watt shown in the diagram and the driver would be your three woofers in parallel rather than the XLS-12 indicated in the diagram. 

[Tyson]

Daryl,
Thanks for the detailed answer, I appreciate the time and effort put into it!

One question - regarding phase, I was planning on using a steep Linkwitz Riley slope for the crossover.  My understanding is that the 4th order and 8th order are both phase coherent, and with this steep of a slope, there's very little overlap in output from the drivers outside their bandpass. 

At least, this is my thinking around my approach for the tweeter/midrange crossover.  I figured I'd take a similar approach with the midrange-to-bass transition as well.  Unless you see a problem doing it this way?

[Daryl]

Daryl,
Thanks for the detailed answer, I appreciate the time and effort put into it!

One question - regarding phase, I was planning on using a steep Linkwitz Riley slope for the crossover.  My understanding is that the 4th order and 8th order are both phase coherent, and with this steep of a slope, there's very little overlap in output from the drivers outside their bandpass. 

At least, this is my thinking around my approach for the tweeter/midrange crossover.  I figured I'd take a similar approach with the midrange-to-bass transition as well.  Unless you see a problem doing it this way?

Linkwitz-Riley crossovers have all sections in phase with each other if that's what you mean by phase coherent.

That condition occurs only if the final slopes of the drivers are what they should be.

When you apply a predetermined slope with an electronic crossover you will get something different than you should because each driver will have it's own unique response shape which is added to that of the crossover circuit.

Thus the relative phase of each driver gets skewed by an unknown value which changes with frequency.

Truth is that most electronic crossovers are not even suitable for the task of being an electronic crossover.

The Behringer DSP electronic crossover and other similar units certainly are suitable for their intended use because of their many adjustment options.

The 8th order slopes will as you say minimize the overlap areas making contouring rolloffs to maintain relative phase much less of an issue.

You likely will be able to get by just using the delay settings for each section.

What I don't know is whether the Behringer comes with a measurement utility and calibrated microphone?

It might even be programmed to remove each drivers characteristic response trends so the applied slopes are true, I don't know?