Active Crossover/Biamp for 604 duplex?

[joyspring]

Ooops,

In my previous posting, I mentioned that the distance between voice coils of the 604 is approx. 7.2cm; this is for the 604-8K with the ferrite magnet structure!

The 604D with its longer Alnico magnets and returns would require more time delay between HF and LF.

BobR

[marcuswilson]

Ihgin,

The posts on this topic are all really informative and interesting and a few different approaches expressed.
I am not as committed to the pure paths of many of you as my needs are different. I do not use 604s at present as I require speakers with a higher output and do not like systems where the 15" driver is used above 500 Hz. The horn is too small to be used down there. You can see a picture of my current speakers on my website, they have an upgraded JBL active crossover, Bose 1800 on the bass and a Threshold 200 for mid/high with the stock JBL passive.
I suggest you wind your own inductors, the only thing you need is an inductance bridge. Before owning one I used to measure inductance with a multimeter and a pot, by feeding 50Hz (60Hz in your case) from a small transformer into the inductor in series with the pot and adjusting the pot until the voltages were equal across each element. The inductive reactance at that frequency should then equal the pot's resistance. Best to have a good RMS reading meter. You can get inductor winding data from many texts, probably the ARRL radio amateur's manual would be a good start.
I could help by providing details of how to wind those particular inductors, but you will learn a lot if you do it yourself!

[joyspring]

Hello again,

A question and a few points...

Just to clarify - is the adjective `sand' (as in `sand circuits' and `sand filters') a reference to silicon semiconductors? It appears to be considering its juxtaposition to `tubes.'

billfort and lhgin: A multitapped 1.5mH inductor is far from unobtainable; perhaps the simplest means of obtaining a multitapped inductor is to purchase a single-tap 1.5mH inductor, a small bottle of varnish and a brush. Carefully unwind the inductor (if it's not potted) and have a friend with a function generator and 'scope measure the inductance.

Once you've found the right spot, solder, coat the junction and all tested areas with varnish, let dry, rewind and you're done.

As an alternative to measuring actual inductance, measuring DC resistance (using the ratio of DCR values listed on 604xover.pdf) is also viable though less accurate. Physical variance (in compliance, MMS, etc.) between the two 604s will probably swamp absolute accuracy of these inductors but consistency between channels is desirable.

billfort: In the US$300-$600 price range, active crossovers will have adjustable RC delay rather than digital delay. IMHO, good digital delay is far preferable; well-executed DSP exhibits linear phase response (equal group delay at ALL frequencies). Now that sampling frequencies are much higher than the Nyquist and bit depths exceed 20-bits and as embedded MPUs become faster and cheaper, we'll see a lot more of this type of DSP prior to final D/A conversion very soon.

RonSS: Asymmetric crossovers are not all that unusual and higher orders may be used to increase power handling, reduce the effects of out-of-band resonances/impedance problems or even flatten/shape on-axis frequency response or overall power response. Altering filter Q can help systems sum flat at crossover.

Altec was not particularly well-regarded for its dividing networks - most were strictly textbook B2 filters. A comparison of the stock Altec 604-8G crossover and the UREI 828/811 is an especially stark comparison...

Not only does the UREI fix the time-domain issue but also flattens the inevitable increasing amplitude of the 38cm driver above 1kHz (because of increasing directivity index, something unavoidable with the large driver size).

Hope this adds some insight to this lively forum ;-)

BobR

[lhgin]

Hi guys

BobR,

Is it possible for me to series a 1.35mH and .015mH coil to form a tapped coil as in the ML crossover?

Marcus,

Could you please point me to your website as well the coil winding side? It'll be very interesting to learn and perhaps if it's easy and economical enough, I'll just wind myself instead of getting from the suppliers.

All,

I've found a source for quite reasonable components and he could supply all the coils and resistors in the crossover except the tapped coil. Check this out:
http://www.chokes.com

I've emailed ElectraPrint for the tapped coil too. I wonder if he's able to do it or not?

rgds, lhgin

[AllanCT]

Anyone know what they make the glass out of that is used to fabricate tubes?

I would doubt any delay would be needed with an active crossover and a 604 series.

[David Deratany]

I will be most appreciative if someone can tell me how much time delay I should start with in this instance:

604 and 811B with mounting flanges aligned for both 802D and 902A (not at the same time)

Same for 511B horn.

Thank you very much

I

[RonSSS]

Bill F,

Good post, I like your thinking.
I have a pair of 904-8A's which are very similar to the 604-8K.
These come with a crossover at 1500hz, and is 18db/oct on the high end and 12db/oct on the low end. They must have had good reasons for doing this. Also, the upper end has dual shelving controld for the mids/highs. It's a bit crude, but it works to help with the hot spots in the 2-4Khz range.
I tried biamping with home made 24db/oct LR "sand" filters. I used identical amps top and bottom.......8W PP 6B4G Heath W5's.
I never got it to sound correct. Matched levels and all that jazz.
The problem was either with the slopes, or the sand. I eventually gave this up and went back to the stock unit with upgraded caps.
The horns on these pups are pretty revealing, and I suggest to all of you that you don't throw away the stock crossovers......it's always good to go back when puzzled by your "upgrades".

BTW....Marchand will sell you the schematics for their x-overs, and the ckt isn't all that complicated.

Ron

[joyspring]

Hi lhgin,

Sorry for the long post but here goes!

Regarding using two disparate coils - this arrangement would not have the mutual inductance that the multitap would have. I don't know whether this is critical in this network's design but if you're going through the effort to build it, you may as well stick to the spec ;-)

It's been a while since my electricity and magnetism classes but I remember that M=K*SQRT(L1*L2) where M is the mutual inductance, L1 and L2 are the inductance of the individual coils and K is the coefficient of coupling. That the coils are in series and that I cannot estimate a value for K probably wouldn't help.

Here's a set of inductor design formulae that may help, I found these in Dickason's `Loudspeaker Design Cookbook':

R= desired DCR, H= core height (inches), d= wire diameter (inches), L= inductance (microhenries), N= # of turns and r= core radius (inches)

(A) suitable for large cores

r=H
H=SQRT(L/(5590*R))
N=3.94*SQRT(L/H)
d=(0.841*H)/SQRT(N)

(B) suitable for smaller cores

r=2H
H=SQRT(L/(6170*R))
N=2.61*SQRT(L/H)
d=(0.738*H)/SQRT(N)

This should roughly concur with the formulae at Marcus' recommended site.

Best of luck whether you decide to wind your own or can find an outfit that will fabricate these for you. Ah yes, other outfits that stock inductors, capacitors, etc. are:

Madisound
Solen

AllanCT: Humourous post re: silicon dioxide ;-)

Regarding the necessity of time offset correction in the 604 series; this was a contentious subject when the UREI Time-Aligned(tm) series was introduced in the 1970s. The physical offset in a 604 is indeed far smaller than a system that would definitely require delay: i.e.- separate horn (511B/811B) and 15" driver mounted on the same baffle. Since the 604 is essentially a point-source, this phase misalignment has much less effect on power response than it would in a standard two-way system.

However, the square waves and cosine pulses generated by the UREI TA monitors are soo much better than compared to those of a standard 604.

True, great measurements don't make great sound but in my experience with both my UREI 811As and stock Altec 604s, these measurements are definitely reflected in better transient response and less long-term listener fatigue.

BobR

[marcuswilson]

Bob,

Good on UREI for time aligning the 811, 813s. But shame on them for destroying the sound with light bulbs!
A local studio bought a pair of 813s to replace their EV Sentry IIIs many years ago. The studio's acoustician spent all night setting them up and in the morning told the owner to send them back as they were way out of spec. The owner didn't return them and just lived with them. I preferred the EVs! Many years later the studio owner was in the US and broached the subject with the boss of UREI, who said that they hardly ever met spec as the Altec drivers varied so much but they couldn't afford to do anything about it at the time so just sent them out. Hardly anyone noticed!
lhgin,
There's a little house icon beside my posts, it is a hyperlink to my website.
It's probably worth learning to wind coils only if you're going to do a lot of it in the future.
I design mine and get the guys with the machines to wind them in quantity.
There's a great article by A.N. Thiele (Small's mate) in the AES journal Vol48 Nr12 Dec 2000 about air cored autotransformers, including how to wind them. Check out your local library.
M

[David Deratany]

Just as I thought my system was reaching the end of its road to improvement, you come along with this.