[Bert]

Hi everyone,

As promissed, the specs of my speakers.

Satelites:
65Hz - 18500Hz
Xover @ 2000Hz
4Ohm
Bass:
40Hz - 120Hz
Xover @ 110Hz
4Ohm

Can't wait to connect my AKSA to them...

[AKSA]

Having just negotiated another Oz Christmas, here's a few pointers:

1.  A Christmas Lunch and a Christmas Dinner gives the opportunity for two big meals and two different venues.  This is important as the typically dysfunctional Australian family creates such horrific friction at the first gathering that a valuable opportunity for reparations is required at the second...

2.  Often the day is very hot, typically over 30C, although this year it was about 24C in southern Australia, a welcome respite.  This encourages feverish drinking, of course, merely to keep cool.  More food is required to meet this exigency, of course.  

3.  The Red of the Father Christmas costume suits us well.  In fact, the colors were chosen by an advertising man in New York back in the thirties for an ad campaign for the Coca Cola company, and the colours have stuck!  There's not too much European about Santa Claus......  

I hope all enjoyed a tasty and warm Chrissie dinner, and that the New Year is a good one for all of us!

More later........

Hugh

[Mark_Walsh]

Good Afternoon All,

(This year Christmas Dinner for us was the nice mix of classic warm roasts, with entrees of prawns, oysters, and "bugs".  Last year we had fresh scallops in their shells and threw them on the barbie.  Nice and hot with a 38 degree day out there, but so scrumptious.)

Given that I have just finished a 3 year long kit subwoofer amp with active crossover (by heatshrinking everything and screwing on the lid of the case), I was disheartened to read the two Subwoofer/Crossover threads in the AKSA forum, with some of the writers suggesting that active crossovers will introduce loss of spatiality to my main speaker system.  I have always wanted to try some spatiality at home.  

One suggestion was to split the line level signal from the source: one limb to the AKSA and the other to the subwoofer amp, presumably matching up normal "roll-off" of the main speakers with the subwoofer crossover frequency and volume.  So far, so good, but the advance on that came from Hugh Dean who suggested accelarating the roll-off of the main speakers with a passive high pass "crossover" in the circuit.  How is this done?  Is an appropriately sized and specified capacitor in series with the speaker what is done, or is it more complicated than that?

Any help appreciated,
Regards,
MW

[andyr]

Good Afternoon All,

... Given that I have just finished a 3 year long kit subwoofer amp with active crossover (by heatshrinking everything and screwing on the lid of the case), I was disheartened to read the two Subwoofer/Crossover threads in  ...

Mark,

Could U possibly describe in more detail your 3-yr-long-kit-subwoofer project.  I am in the process of building a GK-1RP but the *next* project I take on will probably be subs!  Maybe email me directly on .

Regards,

Andy

[PSP]

Hi Mark,
Using two outputs that I built into my Aspen TLP preamp, I am currently running as follows:

Preamp (TLP)-->0.033uf RelCap RTX-->55w AKSA/N-->Sats

Preamp (TLP)-->Marchand XM-9 XO (f3=100Hz)-->Sub amp-->Sub

The 0.033uf cap should roll off the lows going to the sats amp at 6db/octave and 100Hz.  The caps, by the way, are placed between Cardas GRMO male RCAs and Vampire panel mount female RCAs, shielded, and covered with heatshrink... the whole thing is maybe 3 inches long and goes on the preamp output RCAs.  That way, I can easily listen to the system with the caps in or out of the signal path.  When I listened to this (caps in place), there was still too much bass coming from the sats (which are GR Research Paradox 3 floorstanders), so I've stuffed the ports.  Since the Marchand XM-9 is exactly 180 degrees out of phase at the XO frequency, I hooked up the sub "backwards"... red sub lead to black amp binding post, black sub lead to read amp biding post.  Later, I confirmed by listening that this was the right configuration.

This is a very nice sounding system.  It is very significantly better than "running the sats full range and bringing the sub in from below"... that setup sounded quite good too, but once I tried rolling off the mains with a combination of inline cap and stuffing the ports, I would not go back.

By the way, I spent about a week tweaking the Marchand XM-9 (with Marchand's reccomended opamp upgrade, Nichicon Muse caps in the crossover and power supply, Schottky diodes, shielding the power supply, ferrite beads on the power supply leads, etc. and I'm sorry to report that using the XM-9 to high pass the mids and highs shaved off just a little (but clearly noticable) detail from the sound.  When I heard that, I immeadiately returned to the inline cap setup.... never got around to listened for imaging.  I suppose, just to be sure, I should go back sometime and listen to the XM9 high passed to the sats with the ferrites removed from the XM-9 power supply leads.  The systems sounds really good right now, so I'm not inclined to mess with it for a while.  Besides, I have a GK-1R to build...

For the record, my impetus for pursuing this approach came while listening to Jens' system and then discussing things over a wonderful dinner with Jens and his lovely wife.

Because I respect Hugh so much, I'm inclined to believe everything the man says.  I have built three 55w AKSA Nirvanas and one channel of a 100w AKSA (my sub amp) and in three (four?) years and hundreds of e-mails he has always been spot on about the sonic impact of some crazy idea I had... the man knows how to get sound out of silicon and glass.  Still, people who have heard the Linkwitz active designs have been very enthusiastic, imaging included... so I think it can be done and very well, but perhaps it needs to be done with specific drivers in specific geometries, all compensated just so...

good luck... my advice is to stay cool, have fun, listen, and tweak slowly.  Be a student of audio, sound, and music.  Relax... this is like fine cooking, it takes time and subtlety to climb the mountain.

Peter

[Mark_Walsh]

Dear Andyr

Why 3 years to build??  Not because of anything clever.  'Twas a kit.  The Toroid transformer failed within 0.5 seconds of powering it up ... 5 months to replace ... months to re-install ... thought I had done something wrong in soldering it up ... months to diagnose that NOTHING was wrong ... months to clean up and close the box.   The ... 's represented other activities such as windsurfing, helping kids with homework, keeping wife happy , etc.   Sorry to disappoint you


Dear PnP,

This is exactly the advice that I needed ... a cookbook on how to go about this project.

1. Presumably RelCap RTX is a brand?  
2. And the capacitor is in SERIES with the circuit to the speakers? (I have trouble with TLA's* ... it took me about 45 minutes to work out that HT was home theatre, and not High Tension, when reading another AKSA thread.)
3. How do you shield the whole caboodle**?


Regards,
Mark

*     Three Letter Abbrebiations
**    Shebang; etirety; entity; construction.

[AKSA]

Mark,

I believe Peter (PSP) has explained wonderfully how he has implemented a passive rolloff of the low frequencies for his main speakers by using a quality capacitor (an RTX made by Reliable Capacitor Company of Nevada) of suitable dimension (in his case 0.033uF) at the input to the main amp, and in place of the present C1.

In conjunction with a nominal 43K input impedance, a 0.033uF input cap on an AKSA gives an impedance matching this 43K at 112Hz.  (43000/Xc, where Xc = 159000/(0.033uF x 112Hz).  Where the Xc is equal to the input impedance at 112Hz, the input to the amplifier is now halved, ie down 6dB.  At 56Hz, it will be down a further 6dB to 12dB below reference, and at 28Hz is down 18dB.  This may not be quite enough;  we really want about 24dB down by 28Hz because we have a sub and don't want to overwork our main speakers, so Peter's excellent idea has been to block off the port on his mains, and thus cripple the LF behavior of these speakers.  The upside is that the speakers will very likely deliver a better sonic performance now that resonance has been taken out of the mix, and the rear of the cone has constant, sealed box loading.

Shielding the amplifier is best achieved by mounting it in a metal enclosure, and keeping all signal leads away from the power leads.  The amp is not so critical;  the preamps are much more touchy.

My apologies for not answering this earlier.  I do like your fetching use of TLAs.  You clearly have a sense of humor best delivered with a small umbrella.  I think TLAs are marvellous, but for the essence of artful obfuscation, I suggest FLAs....

Cheers,

Hugh

[PSP]

Hi Mark,
Hugh explains almost exactly what I did.

To date, I have not yet replaced AKSA amp C1 with 0.033uf caps.  That will be a good idea someday, when my setup is more settled and I plan not to experiment on the amps for a year or two.  As it is, I have left C1 at its stock value and have placed the 0.033uf cap (to roll off the lows) at the input RCA of the amp... plugged into the amp input RCA so I can easily remove it for AB comparisons.  Since I have three AKSA 55Ns, this setup allows me to easily swap amps... to check the effect of some tweak, or to allow me to pull an amp for a month or two while I rework the layout, etc.

The reason for putting the cap--yes, in series--in front of the power amp that feeds the main speakers (and not between the amp and the speakers) is as follows:

C = 1/(2 * pi * f * Z)

C = capacitance in farads
pi = 3.14159...
f = the desired roll-off frequency, Hz
Z = input impedance of the system you want to filter, in ohms.

If you choose to put the filter in front of the amp, you plug in the amp impedance (43000) into the equation.  If instead you choose to put the filter in front of the speaker, you plug in the speaker impedance (6 ohms??)... as you can see, you can use a much smaller cap (thus, you can afford a much higher quality cap) if you put the filter in front of the amp.  In addition, speaker impedances are often wildly frequency dependent so good calculations require measurements and probably some trial and error as well.

In practice, you often flip the equation around:

f = 1/(2 * pi * Z * C).  Then you plug in the value of the various caps you can find to see what frequencies you can get.  You can put two caps in parallel (sum their values) to get values in between the standard values.  Remember that caps usually are specified in micro farads, this is farads x 10^-6).  You can re-write this equation as follows:

f = 1,000,000/(2 * pi * Z * c), where "little c" is expressed in microfarads...

This is lots of fun, yes?

Peter

[PSP]

Mark,
You also asked about shielding...

First of all, I bought matched caps, so the roll-off would be approx the same for both channels.  This cost $0.50 for each cap.

I built my inline filter by wiring each cap lead to the RCA center pin.  So:

RCA female (hot)--|cap|---RCA male (hot).... it's a good thing that the cap is present or this post would be x-rated  
RCA female (earth)---copper wire, in teflon tubing---RCA male (earth)

Then, I slipped a length of tinned copper braid (1/2 inch diameter, from Michael Percy) over the entire thing and ran a drain wire from the braid to RCA earth at one end only.  I made the braid shielding short enough so that it wouldn't quite cover both RCAs... you want the shield to be earthed at one end only, the other end floats... this to reduce ground loops and hum.  Then I put some heatshrink over the whole mess.  This is not a super-attractive device, but it gets stuck into the back of the amp, mostly out of sight.

Good luck,
Peter

[andyr]

Mark,

I believe Peter (PSP) has explained wonderfully how he has implemented a passive rolloff of the low frequencies for his main speakers by using a quality capacitor (an RTX made by Reliable Capacitor Company of Nevada) of suitable dimension (in his case 0.033uF) at the input to the main amp, and in place of the present C1.

In conjunction with a nominal 43K input impedance, a 0.033uF input cap on an AKSA gives an impedance matching this 43K at 112Hz.  (43000/Xc, where Xc = 159000/(0.033uF x 112Hz). ...

Hugh,

Sorry to butt in here but, in the interests of "Correctness" - Factual, not Political   - I must pick you up on your statement "Where the Xc is equal to the input impedance at 112Hz, the input to the amplifier is now halved, ie down 6dB."

Using the equation which Peter provided:

C = 1/(2 * pi * f * Z)

C = capacitance in farads
pi = 3.14159...
f = the desired roll-off frequency, Hz
Z = input impedance of the system you want to filter, in ohms

yes, you get f=112Hz if Z is the 43Kohms input impedance of an AKSA.  And if Z is the 3ohms of my Maggie ribbon, you get f=2KHz for C=26.5uF.

However, a cap in series is a 1st-order Butterworth HP crossover; such crossovers specify the -3dB frequency as the "crossover frequency" ... not the -6dB point.  I believe only Linkwitz-Reilly crossovers specify -6dB points, because these are designed to sum flat across the crossover region ... summing two "half-power points" gives you unity!

Speakers having Butterworth crossovers specify "spread" crossover points - ie. the two -3dB points are different frequencies (for instance, the specs for my Maggies say: Base LP @ 300Hz, Mid HP @ 500Hz).

Those people trying to match a sub with the 1st order HP roll-off of their main speakers explained above should not use the same frequency (112Hz in the above example) as the "crossover point" for their sub.  They need a lower -3dB crossover point for their sub ... probably around 80Hz.  If they used 112Hz, this would give them a hump in this region   .

Regards,

Andy

[Jens]

Dear Peter, Hugh, Andy,

Andy is right, of course - which is also why Peter had to stuff his ports, as this made his speakers roll off at a higher frequency.

And this works more or less the same, whether you use caps in front of the amp or the speakers (provided we're still talking 6 dB x-overs).

The trick is to try and have the midrange roll off acoustically at the same point as the x-over frequency. This is better than moving the sub x-over frequency down, because with a combined acoustic and electrical roll-off you can actually get a roll-off that is quite close to a second or even third order L/R x-over. I use this "trick" in my speakers and can assure you that it is a good one.

Obviously, it is easier to make this work when you design new speakers ....

Cheers,

[Mark_Walsh]

Dear PSP, Andy, Jens and Hugh,

This is exactly what I needed to know ... complete with formulae so that I feel as if I know what I am doing.

(The only formula from school which came to me last night as I was dozing off to sleep was "X = 2.pi.f.C" and I cannot even recall what that equalled - impedance through a capacitor?  So much for my electronics knowledge.  That is what we have amplifier designers and kit marketers for.)

Most excellent regards,
MW

[AKSA]

Andrew,

Thank you for correcting my mistake.  

Coffee tomorrow?  I believe I owe you one.......

Cheers,

Hugh