Altec ***** Tale

[papereater]

Been reading posts here for the better half of a day. Great group of folks! Here's one.

In 1979-80 I built a pair of enclosures. I had a pair of 420A's (no cones) that I had purchased for $10. Having no idea what I had I called Altec. I spoke to one of the engineers and was invited to visit their Anaheim location. I wanted something I could use for a PA (being a musician) and also use for audio in the home. Being from Southern California the 20 minute ride to Anaheim was no problem. The engineer said I should recone the speakers as 416-8B's then buy a set of N1209-8A crossovers, 908-8A drivers and H-811B horns. He added that when I finished building one of the enclosures to bring it in and they would test it in their chamber. I left with liturature in hand an no idea what I was getting myself into. In those days my shopping list cost just over $900/side plus the wood and hardware.

Well, it's been nearly 25 years and I'm going to do it again. This time I hope they'll look more like funiture than a PA. I'll need all the tips I can get.

What did I end up with the first time around? A set of speakers that has amazed all that have listened.
Not perfect but not bad for a first timer either!

[joyspring]

Also being from So. Cal, I had the opportunity to visit 1515 Manchester Ave. in the late 1970s as well. Altec enthusiastically supported the enthusiasts at that time ;-)

Nice system! Personally, I'd leave them as-is, particularly if you're accustomed to their sound.

Judging from the response/impedance plot, there's not much room for improvement enclosure-wise. Your box volume and tuning approximates an ideal 4th order - I feel that increasing volume and port tuning lower to achieve a drooping response and lower -3dB cutoff (very popular with low-Qts drivers, particularly on this forum) would adversely affect transient attack quality.

Keep the enclosure volume and tuning identical to your present system if you decide to build furniture-grade enclosures.

The only significant improvement I would recommend is a trap filter to tame the response peak by 3-4dB between 1.5 - 5kHz. This is a parallel L-C bypassed by R in series with the HF driver. Though simple in concept, design can be a bit complicated. Approximate values may be computed using the measured impedance of the HF driver through the filter range and standard inductive/capacitive reactance formulae though final values must be fine-tuned by measuring and listening. Implementing this filter may be a bit tricky since the affected bandwidth is so close to the crossover frequency as well though I feel that the effort is worthwhile. You're actually not too far from achieving ruler-flat on-axis 70hz-12khz response!

Another idea would be to fix the time-domain response, either by physically aligning the drivers (which would spoil the flat baffle) or delaying the LF. A passive crossover would employ assymmetric high- and low-pass filters and most likely the addition of out-of-passband low-pass or all-pass filters as well; such a network would be very difficult to design.

Bi-amplifying (using an active, line-level crossover and separate amps for LF and HF drivers) is another possibility. Many high-quality active crossovers have integral continuously-variable LF delay (i.e. - Rane AC22). You may add a single band of line-level parametric EQ to flatten the mid-band response peak as well. You would require two more channels of amplification, however.

Apart from that, the only other recommendation may be an active subwoofer to obtain perhaps another octave below your system cutoff.

Best of luck whatever you decide to do,
BobR

[papereater]

Hi BobR,
Thanks for your input. After reading your post I measaured the enclosure to verify the volume and the port size. Using the chart from design library here I found that I was very close. Over the years I changed the speaker mounting from behind the baffle to the front of the baffle. Without that taken into concideration I was off by about 1.5 sq.in. for the port. I'd forgotten how much mathamatics went into building these.

I need a little clarity here. What is "4th order"?
And, I've seen "Qts" in the formulas, just can't seem to find it again. So, I need a definition too.

Will tuning the port exactly have enough of an effect on the impedance to help with the mid frequency rise? If I remember correctly, the resistance of the air on the coil affects the impedance the amplifier sees.

I looked all over the house trying to find the plans so I could show how I aligned the voice coils in the same plane. The picture doesn't show too clearly that the horn is a separate cabinet that hangs over the lower one by 3.5 inches. What confuses me is the design literature says to reverse the leads to the horn to correct the phase in order to mount the horn on the same baffle as the speaker. How do it know?

I probably won't be bi-amping or using an active crossover so your suggestion of a trap filter would be a good start.

The bass seems to be adequate and is being masked a bit by the mid rise. If I can level the curve out a bit I'd be in heaven.

Sorry to say I'm running a 65 watt Harmon-Kardon solid state amp. I don't use an EQ and I leave the treble/bass level. People have asked me where the sub-woofer is LOL! If I turn the volume to mid-point the neighbors' attic rattles, pictures fall and my wife can hear it in her car (I live in the middle of the block) before she turns the corner. Seriously, what I really like about these speakers is the clarity.

Thanks again for your help.

[Cal Weldon]

Hi Papereater,

Not to step on Bob's toes here, but a 4th order XO is a 24dB per octave unit. Very steep slope. Many parts, big time line loss. Excellent tweeter protection. Superior driver separation.

1st order is 6dB/octave ie: one capacitor on the tweeter and one choke coil on the woofer. Smooth transition from one driver to the next (lots overlap), poor tweeter protection.

2nd order is 12dB/octave with two caps and two coils. Medium qualities all around.

This is very simplified.

Why are you sorry you are running it with a H/K. I've never heard anyone bash their cicuitry. Full discreet, conservative power rating. Why do you think you a 65 watt amp cost so much? H/K is just fine, no need to apologize. Not everyone runs tubes you know.

Cheers,
Cal

[papereater]

Hi Cal,

Thanks for the explanation.

I like my H/K, but I've always liked analog/tubes. I'm one of those that still own a turntable and a collection of vinyl. I wanted a McIntosh MC2205 in the late 70's. I got a wife and a couple of kids instead.

As the late John Lennon said, "Life is what happens when you're busy making other plans."

[joyspring]

Hi Papereater & Cal,

By 4th order, I was actually referring to the enclosure `alignment'; sealed enclosures approximate second-order whereas vented (port tuned) enclosures approximate 4th order amplitude and phase behaviour.

Thiele, Small, Benson, et al. quantified the mechanical attributes of loudspeaker systems in electrical filter analogues and that's been the basis for modern loudspeaker and enclosure design since the 1970s. Differing `alignments' are described by their electrical counterparts (Bessel, Butterworth, Chebyshev, etc.).

Qts is total driver `Q', the sum of mechanical Q (Qms) and electrical Q (Qes). `Q' is the inverse of damping and in simple terms is the magnification of driver resonance.

Generally speaking, drivers with light cones and powerful electromagnetic motors (high Bl) have low Q whereas drivers with heavy cones and low Bl have higher Q. Magnet weight may be an indicator - large magnets usually (but not always) denote low Q.

You will notice one thing with low-Q driver-based systems; with a standard B4 (4th order Butterworth) or QB3 (quasi-Butterworth), the system -3dB point is rather high, usually > 2x driver resonance. As I alluded to in my first post, many attempt to offset this by using a larger enclosure and lower port tuning and live with a drooping response. For best transient performance, I recommend retaining your current box volume and port tuning as it's so close to optimum for your 416-8Bs as it gets.

So far as a trap filter is concerned, here's a starting point I determined from the response and impedance plot you provided; this trap would fit between the crossover HF output and the HF driver.

First, capacitive and inductive reactance:

Xc = 1/(2*pi*C*f)
Xl = 2*pi*f*L

where

Xc = capacitive reactance
Xl = inductive reactance

C = capacitance (Farads)
L = inductance (henries)
f = frequency (hz)

For a center frequency of 2500hz, approx. impedance of 12.7 ohms through the affected bandwidth:

<pre><font face="courier">

+-- C --+
| |
o------------+-- L --+----------------o
| |
+-- VR --+

o--------------------------------------o

C = 5.0 uF
L = 0.8 mH
VR = 25 ohm, 5 watt wirewound

</font></pre>

4.7 uF and 5.6 uF are the closest standardized values to 5.0 uF.

At VR = 0 ohms, the filter is bypassed and the peak remains. As VR goes up, the trap has greater effect. Where the control should be parked is subject to measurement.

And on that topic, since you probably do not have access to an anechoic chamber ;-) you may attempt to find someone who has a `quasi-anechoic' gated measurement system to reduce the effects of environmental reflections.

Before I go on, the values are approximate and I'll make no guarantees about performance. Also, just obtain high-quality 5% polypropylene caps and 16-18 AWG air-core coils, nothing fancy; you have greater variance and more egregious compromises elsewhere in your system. These are loudspeakers after all. Also, never mind the silly silver wire that seems so prevalent with audiophiles.

I'll concur with Cal on your solid-state Harman-Kardon amp; they make excellent products!

Solid-state sounds fine and nearly all high-quality amps (solid-state or tube) sound identical below clipping. In fact, the only really appreciable audible difference between high-quality amps nowadays is noise floor.

Certainly tube amps have a more gradual transition into clipping but that's really beside the point; if you're ever going into clipping, you need a bigger amp. Leave the distortion for musical instrument amps. In this day and age of 24-bit/96khz digital, 16-20db+ peak-to-average DVD-A/SACDs, the viability of 8 watt Single-Ended Triode amps is questionable.

Additionally, solid-state has one huge advantage over vacuum tubes - the performance remains consistent over time whereas vacuum tubes eventually degrade. You can guess which type of amp I use in my studio...

So add the trap filter and don't fret over your H-K amp; use the money you saved not purchasing inane, non-engineer-developed `boutique' vacuum tube gear, silly cables, etc. and purchase a whole bunch of CDs/DVD-As/SACDs and tickets to live performances instead.

BobR

[joyspring]

Cal,

So far as passive crossovers are concerned, 4th order does incur a significant amount of loss and is probably not advisable.

Active crossovers obviously overcome the insertion loss ;-) and many high-quality 4th-order examples abound.

BobR

[papereater]

Hi BobR,
I know you said not to get fancy with the components for the filter. I was planning on going to my local electronics store, buy the stuff and put it together during lunch. Haha! I didn't know what types of coils were used so I did some surfing and was surprised at the price differences.

Do I proceed with plan "A"?

Thanks again for your help.

[joyspring]

Papereater,

The prices do increase markedly as the wire gauge increases... what you're paying for is the copper!

Personally I would choose the 16 or 14 AWG as anything larger has diminishing returns so far as lowering DCR is concerned...

You actually may want to purchase parts for only one channel for now but also obtain the next values up and down for both the cap and coil. You could then substitute values until you achieve the sound (or measurement) that you prefer...

Best of luck!
BobR

[Cal Weldon]

Hi Bob,

I should have read your post more thoroughly. I just went off on some rant about the XO when you were discussing the box tuning. lol

I agree that by the time you get into a 24 dB XO you really should be going active.

Happy listening,
Cal