Bringing 828 cabs up to snuff...

[Alien_Shore]

Yep, I went with a Hiraga hybrid. I'm hooking this one up today and I'll see what it sounds like.

I'm trying to follow along here on this filter design.

L1: 6.6mH, C1: 18 uF -> 500 Hz 12 dB LP

C2: 19 uF, L2: 3.3mH, R2: 70 ohm -> 500 Hz 12 dB HP

L3: 3.3mH, C3: 10uF, R3: 12 ohm, R4: 50 ohm, R5: 5 ohm -> 6dB notch (2KHz-4KHz)

Is this what you ended up with?
I built a similar notch filter for my 802s, and wound up using a variable resistor for R3, so I could dial the notch for different sources.

[alancohen]

Alan,

It's really nice to see how your A5's have come along ( somewhat reminds me of Panos saga ) .

To the best of my recollection, you're the only poster ( that I can remember ) who has built up this unique filter ( though I'm sure it's really a memory gap on my part ) .

Much appreciated, Earl. It's been a treat putting it all together. A task that would and could not have happened without everyone here.

It's hard to believe that not many more people have tried to duplicate Mr. Hiraga's system with all the glowing talk about it. It certainly doesn't cost too much to try it out.

I'm really enjoying how smooth it sounds. I'm even getting a reasonable stereo image. The speakers haven't disappeared, but they're blurred a bit. Varying speaker placement is not much of an option considering the size. I think if I could get them a little closer together it might help. But all in all, they are much more than I ever expected.

- - - Updated - - -

I'm trying to follow along here on this filter design.

L1: 6.6mH, C1: 18 uF -> 500 Hz 12 dB LP

C2: 19 uF, L2: 3.3mH, R2: 70 ohm -> 500 Hz 12 dB HP

L3: 3.3mH, C3: 10uF, R3: 12 ohm, R4: 50 ohm, R5: 5 ohm -> 6dB notch (2KHz-4KHz)

Is this what you ended up with?
I built a similar notch filter for my 802s, and wound up using a variable resistor for R3, so I could dial the notch for different sources.

Yes, I used the 6.6mH on the 16 Ohm 515s and 3.3s on the 8 Ohm 288. I'm not sure that was kosher, but it made sense to me.

If R5 is the attenuation matching resistor, would I increase the value to attenuate more HF or decrease it?

[GM]

Sounds like they need more toe-in. As a general rule, their axis need to cross at, or in front of, the LP's 'sweet spot', though where along this axis sounds best overall must be found empirically due to the speaker's particular polar response that in some apps can 'throw' the sound behind the LP. I was somewhat surprised at how much toe-in changed when I switched from a ~70 Hz mid-bass horn to just dual baffle mounted drivers.

GM

[alancohen]

Sounds like they need more toe-in. As a general rule, their axis need to cross at, or in front of, the LP's 'sweet spot', though where along this axis sounds best overall must be found empirically due to the speaker's particular polar response that in some apps can 'throw' the sound behind the LP. I was somewhat surprised at how much toe-in changed when I switched from a ~70 Hz mid-bass horn to just dual baffle mounted drivers.

GM

Here's a question before I start tuning in toe-in: What is the most expedient way to align the 805/288s and 515Bs? Since I have the 805s in nice little boxes, I can slide them back and forth.

[Alien_Shore]

Yes, I used the 6.6mH on the 16 Ohm 515s and 3.3s on the 8 Ohm 288. I'm not sure that was kosher, but it made sense to me.

If R5 is the attenuation matching resistor, would I increase the value to attenuate more HF or decrease it?

In the parallel contour filter design I'm familiar with, the parallel resistor in the tank circuit (R3) controls how much attenuation you get at the resonance frequency, because R3 acts as a bypass from the tank; higher values of R3 = more attenuation. With the addition of R4 and R5 in series/parallel configuration, it appears that Hiraga was going for an L-Pad attenuation effect for the driver. So I suspect that R4/R5 need to simply be tuned as you would an L-Pad to get the amount of overall attenuation for the driver's response, after it's been shaped by the HP and contour filters. Perhaps Earl can chime in and confirm...

[GM]

What is the most expedient way to align the 805/288s and 515Bs?

http://audioroundtable.com/misc/Altec_AN-9.pdf

GM

- - - Updated - - -

If R5 is the attenuation matching resistor, would I increase the value to attenuate more HF or decrease it?

What are you trying to accomplish?

I mean, normally the choke and parallel resistor is for tonally balancing the horn's response with the woofer's at the LP. This usually means the extreme HF is rolled off too much, so a bypass cap is paralleled to allow some portion of the horn's output to be let through unattenuated; so for initial voicing, the cap should be left off and using cheap pots to dial it in the resistor is a good plan, then replace with fixed resistors before experimenting with different cap values to get as much top end 'air' as practical and don't be surprised if the values are [way] off from someone's touted design.

GM

[Earl K]

In the parallel contour filter design I'm familiar with, the parallel resistor in the tank circuit (R3) controls how much attenuation you get at the resonance frequency, because R3 acts as a bypass from the tank; higher values of R3 = more attenuation. With the addition of R4 and R5 in series/parallel configuration, it appears that Hiraga was going for an L-Pad attenuation effect for the driver. So I suspect that R4/R5 need to simply be tuned as you would an L-Pad to get the amount of overall attenuation for the driver's response, after it's been shaped by the HP and contour filters. Perhaps Earl can chime in and confirm...

I'll see if I can muster up something coherent for ya ( tomorrow ) .

Today is BBQ day !

[Earl K]

I'm trying to follow along here on this filter design.

L1: 6.6mH, C1: 18 uF -> 500 Hz 12 dB LP

C2: 19 uF, L2: 3.3mH, R2: 70 ohm -> 500 Hz 12 dB HP

L3: 3.3mH, C3: 10uF, R3: 12 ohm, R4: 50 ohm, R5: 5 ohm -> 6dB notch (2KHz-4KHz)

Is this what you ended up with?
I built a similar notch filter for my 802s, and wound up using a variable resistor for R3, so I could dial the notch for different sources.

Hi Mike ( Alan ),

EDIT :

(i) One should run an impedance curve for this network ( connected to the HF driver ) before deciding how best to attenuate the HF section any further .
(ii) Quick calcs, indicate the HF section dips into dangerously low impedances ( less than 4 ohms ) in the upper octaves .


FYI : Here's how I see the filter breaking apart ( into it's fundamental sections ) .



- Zone 3 ( the Tank filter / notch filter ) has added R4 as a "Q" modifier ( for the notch ).

- Zone 4 ( because it's just a single parallel resistor ) will alter the load impedance ( due to the lack of a "build-out resistor" ) if R5 is arbitrarily changed .

Some Overview :

- Inline parallel filters ( Tank type ) will always return to very low ( approaching zero ohms ) self-impedance at the far extremities of the filters bandwidth .
- As a result of that action, the ( total ) HP filter would seem to drop the load impedance down to below 4 ohms ( & a bit lower ) from 20K on & higher .
- Dropping lower than 4 ohms, may destabilize some amps / sending them into HF oscillation .

Alan's dilemma ;

- If he needs more overall horn attenuation ( to balance the horn & woofer sections ) then he could add an inline resistor ( even a variable 8 ohm pot ) between the junction of R5 and the driver's input terminal .
- If he needs more mid-range notch / then he should follow your advice and install a variable 30 ohm pot ( or maybe an 8 ohm POT with a 8 ohm inline resistor ) in place of R3 .

- Under no circumstances would I simply reduce the value of R5 ( to achieve more attenuation ) .

- Anyways, that's how I see this all working ( & playing out ) .

- I'm on holidays ( starting tomorrow am ) for the next 11 days ( so I can't really be of any help on this, for the immediate future ) .

[Alien_Shore]

Thanks for the tutoring, Earl. I see now what you mean about R4 adjusting the contour filter Q and not being part of the overall driver attenuation.
This is a tricky little circuit. I don't have the 288, but I can prototype it using a small format driver and look at some impedance curves to better understand it.

[alancohen]

Big Thanks Earl! I'm trying to get my brain around what this filter is doing, but it's a little over my head.

I really don't 'need' more HF attenuation; they really sound as smooth as I could possibly ask for. If anything my room could use some more damping. I have still have a few hard reflective spots in the wrong places.

I've been playing a large variety of music and don't feel a need to fine tune in the least. This just might be a case where I have to stop fiddling with them and just enjoy them.


So to recap this project:

828 cabs only - $162 and a drive to Queens

515B woofers - $400 + $125 recone for one

288-8Ks - essentially free. I got them with the A6s for $750 and sold the N1285-8B XOs for $250 and I'm selling the 8256s for $600 including 4 3156s and the MRII564s for $125.

805 Tar-filled horns were the big ticket at $1200 including mounting boxes

Hiraga XOs about $100

Misc bracing, damping and paint supplies $75

Rounding it all off I'm into a pair of A5s for a little over $2K. Not too bad.