Augmenting the bass range of my speakers (inc. bi-amping)

Hi,

I've decided I'd like to augment the low end output of my speakers, as I feel they slightly lack authority and weight in the bass.

Currently the system is pre and power amp with 3db baffle step correction, feeding 2 way MTM floorstanding speakers using 6.5" Seas coated paper drivers (3mm xmax one way). These are ported and reach 40 or 50hz in-room.

A 15" (260L ported Tempest) subwoofer nestles in underneath from 50hz/45hz (two x-overs to attain 4th order).

This setup is quite good but lacks a bit of slam in the mid-bass range. I don't feel these Seas drivers are up to the task, and would prefer to be relieved of the load, and purely work on midrange.

http://www.seas.no/seas_line/woofers/H571.PDF

It just so happens I have plenty of seemingly suitable 6.5" drivers lying around (100+), which I bought from Acoustic Solutions via Ebay.

I *fiiiinally* got around to having a good go at measuring their impedance curve. My previous excuse not to was my DMM could only read 0.0vac, but I recently bought one that can do 0.000vac. The measurement technique used called for 0.0000vac to be accurate, and indeed the measurements seemed somewhat erroneous. So I calibrated the voltage to 10x the level shifting the decimal place to the left. This gave me believable results.

I left them at home (oops) but the Fs was 43hz, Qts 0.4 or so, and Vas 20L iirc.

My intention is to have a go at installing some of these units in the space allowed by the integrated stand of my speakers. I have modelled them in WinISD pro Alpha and they seem quite useable.

I would very much like to go active, and x-over wherever sounds best. I may remove the baffle step correction circuit and dial in the level of the new bass section to suit. This may be 150hz-400hz I reckon.

By my calculations I have up to 35L of space available.

This is enough for either:

* 2 x 6" drivers ported and giving output to 45hz, or
* 4 x 6" drivers sealed and linkwitz transformed

Option one doesn't appeal too much, because while there may be an improvement in going active and 3-way, it will still be offering similar displacement, and giving the drawbacks of ported loading.

Both designs require about 50-70 watts to reach the likely (guessed) xmax of the drivers at the lowest frequencies they will cover. Bass output should be well over 100db in-room.

Option 2 gives the following benefits:

* a smaller box
* more upper bass headroom
* sealed loading (might sound better) (less group delay)
* opposing driver arrangement for force cancellation
* no port noise considerations or poor control under tuning
* more spread out drivers for more evenly exciting room modes
* series-parallel array for giving 4ohms nominal impedance

Obviously, this is what I'd like to do. Any criticisms/suggestions?

I will upload some winisd screenshots showing the responses, and also pics of the driver and the system in its current state.

Whatever I do, I want to go active and add extra bass drivers, because this will achieve the following:

* more headroom and max spl
* more bass output
* lower distortion midrange and even treble
* a separate bass section allowing room correction EQ (at some point), which won't interfere with the midrange and treble
* a closer match with the output capability of the sub
* a closer bass character match with the sub (hopefully)

Currently the speakers aren't too great for parties or loud home cinema, as they simply don't have the displacement (perhaps not enough amp either). They start to sound a little weak and strained when the dial goes right to the end!

At the moment I'm using about 25w worth of LM3886 (a sort of semi-permenant prototype). I am about to build a better LM3886 with about 50-60w output. This would power either the top or the bottom. In the meantime I have a Roksan Caspian I could use for part of it.

I would mount the psu and circuits for LT and active x-overs in the pre-amp chassis (Rod Elliott Project 88 in 1u rackmount case).

pinkmouse said:

Certainly sounds feasible. I never did get round to getting a driver off you to measure, did I?

Will you still be using the Tempest? If so, I would go for four drivers in a sealed box, forget the LT circuit, and just run them down to 60Hz or so. If you can get a box design with a nice smooth roll off, it simplifies the crossover greatly.

Click to expand...

Hi PM!

Nobody really wanted to measure my drivers for me

I realised I had to just pull my finger out and learn a way to do it myself!

I will be using the Tempest, but I don't think the Tempest will be up to much at 60hz (will be easy to localise as it's on a side wall near the listening seat). Also, sadly, these drivers are not suited to closed box loading with no EQ, as they roll off way too early. I will show a model of this. LT circuit doesn't look too hard to make, even for a beginner in electronics like me.

Thanks for taking an interest.

Hello,

Here is a flow chart making clear what we're talking about doing.

pinkmouse said:

Flow chart! You're not a management consultant are you?

Ok, in your place, the first thing I would do is build a rough test box from chipboard, stick a couple of drivers in it, and do some measuring. Then you know exactly what you're dealing with.

Click to expand...

Certainly not a management consultant! lol

Rough test box, ok, fair enough. I intend to give em a good listen on their own, see if they're capable of the sort of bass I'm after.

I'm not 100% sure about acoustic measurements yet

I did get some rough stuff on the go using some demo software (will look at home) to find my armchair response, and it seemed quite good. I have a Maplin mic and pre-amp that seem to work ok, and a pc, obviously.

Here's the driver I'm gonna use:

Front

Side view

Here's a picture of my right speaker. You can see the sides form an integral stand for the rest of the speaker, but are really unused space. It's ok for up to about 35 litres, or 25 to fit snugly without protruding too much front and back.

Here's one of the drivers in question being tested for resonance - using the added mass method to work out the Vas parameter. Pound coins weigh 9.5g each, so I called it 10 with a blob of blue tac.

Hello,

Here's a screenshot showing two simulations in WINISD. The yellow line represents 4 drivers in a sealed 35L box, with no EQ. The pea green line is the same 4 drivers in a 20L sealed box with a Linkwitz-transform circuit used to improve the response.

Hi PM,

As you might expect, the cabinet size is halved for the same frequency response, but efficiency drops.

2 pairs in sealed isobarik config in 35L is poor, completely overdamped and with little bass, and inefficient.

However, 4 pairs in 35L and ported is interesting - just look at the minimal cone excursion! These are all at 60w input.

This config would take over 500w before exceeding my imaginary 8mm p-p xmax! It could make some very loud bass, but presumably with magnets sticking out and also using 16 drivers (messy!)

And here is the fr. response of green - sealed 20l 4x6 w/LT, yellow - sealed 35l 4 in iso, and grey - ported 8 in iso.