ALK Universal replacement network price cut

[Al Klappenberger]

Guys,

Good news! I have decided to reduce the pricing on my Universal "A series"
crossover network upgrade!

The new pricing will be $299.00 for the parts kit and $429.00 for a
completely finished and tested set.

Here's why:

There are several reasons for the price cut. One is that this design has
become the "economy" design of my network options since the introduction of
the extreme-slope networks. The main reason is that I have developed methods
of building them that cuts the build time involved considerably. The other
reason is that I am changing several of the parts to less expensive ones. The
Hovland Musicap film-and-foil capacitors I have been using are superb but
very expensive and add very little compared to good quality Metalized
polypropylene capacitors like the Kimber Kaps I will be using from now on.
The capacitors in a filter contribute only very slightly to the total losses
in a filter. The inductors are the real bottleneck. I will never go away from
the Solen "Litz" wire coils I use for the 0.2 and 0.3 mHy inductors. These
are the big advantage. They have 1/3 rd the loss of the same value inductor
wound using solid wire even of a heavier gauge!

I did some comparison measurements on a matched set of Hovlands and a pair of
Kimber Kaps for value and ESR at 1000 Hz and at 10,000 Hz. Here's the
results:

Kimber Kaps:
** cap #1 1Khz: 2.267 uF .057 Ohms * 10KHz: 2.270 uF .032 Ohms
** cap #2 1Khz: 2.285 uF .045 Ohms * 10KHz: 2.290 uF .012 Ohms

Hovland:
** cap #1 1Khz: 2.270 uF .033 Ohms * 10KHz: 2.276 uF .005 Ohms
** cap #2 1Khz: 2.266 uF .020 Ohms * 10KHz: 2.273 uF .0038 Ohms

Notice that there is very little difference between the two capacitor brands
at 1000 Hz, but a factor of 10 different at 10 KHz. The ESR numbers are also
very hard to measure at these low values. The reading jumped around with
every press of the trigger button on my LCR meter. The numbers are just a
rough average.

This says that there may be a measurable difference in the tweeter filter
loss using Kimber Kaps compared to Hovlands but virtually no difference in
the squawker section of the network.

In order to test this I am building two sets of networks. One using Kimber
Kaps in place of the Hovlands in the squawker filter and another set with the
Kimber Kaps in the tweeter filter. The set with the Kimber Kaps in the
squawker filter is shown in the picture. The measurements confirmed that
there is no difference on instrument tests. The tweeter loss is 0.35 dB and
0.55 dB loss to the squawker. This is the same as using all Hovland Musicaps.

I have also replaced the 39 uf Solen FastCap plus the Harmony 1 uF "bypass"
cap in the woofer diplexer section with two 20 uF Solen FastCaps. This cuts
cost a bit more and actually measures BETTER for ESR! This change would show
up as a difference in loss through the tweeter filter at 20 KHz if the
combination was not as good. There was no measured difference.

The 10 Ohm swamping resistor used in this set are the wire-wound "cement"
type rather than the film type I normally use. These are just as good and
were used simply because Solen was out of the film type right now. The tiny
inductance of this type resistor is TOTALLY insignificant at audio frequency.
I will continue to use the film type though as the cost difference is very
small.

I will be building up a set using Kimber Kaps only in the tweeter filter when
I get two more Kimber Kaps. Right now they are back ordered. In the unlikely
event that the tweeter filter suffers significantly from the capacitor change
I will continue to use Hovlands at that position. I will also HOLD THE SAME
REDUCED PRICE and absorb the increased cost since I get the Hovlands at
somewhat lower than "list" price by ordering them in quantity.

The set of networks shown in the picture are already spoken for. I will also
be using up my supply of Hovland caps in the networks I build from now on, so
the next few customers may be getting the Hovlands or Kimber Kaps in various
combinations for the same reduced price.

Al K.

[jwc]

Al this is a great thing. I saw this coming after that long post about capacitors.

Since then, I am building some of your ALK's with Solen only. Using there Film and Foil for the 2.2 uF's. 12 AWG for the low pass Filter. The Litz for the .3 and .2 mH. Weird enough, the only thing not Solen is the 10 ohm 10 watt resisitor. Had to get it from Parts Express because I for got to get it from Solen.

I hope your current thinking "sounds" as good as the logic.

jc

[Al Klappenberger]

JC,

I thought about those new Solen Filem and foil 2.2uF caps. My problem with them is the leads are just not long enough to stretch from the 20+20 uF caps to the 0.2 mHy inductor in the tweeter filter. I got a pile of the bare boards made on a CNC machine and wasn't willing to drill more holes in them! I'm tempted to try them in the ES5000 design someday though.

BTW: Your's right about the capacitor thread. That got my butt in gear to make the pricing reevaluation.

Al K.

[tofu]

JC,

I thought about those new Solen Filem and foil 2.2uF caps. My problem with them is the leads are just not long enough to stretch from the 20+20 uF caps to the 0.2 mHy inductor in the tweeter filter. I got a pile of the bare boards made on a CNC machine and wasn't willing to drill more holes in them! I'm tempted to try them in the ES5000 design someday though.

BTW: Your's right about the capacitor thread. That got my butt in gear to make the pricing reevaluation.

Al K.

so this is the main reason for using kimbers over solen? lead length? or are you assuming there will be a worthwhile change audibly?

[tofu]

i hate how this forum resizes images automatically. if i'm going to download the full version of the thing, i want it to display the damn full version.

[Al Klappenberger]

Tofu,

The main reason is the lead length, but I think Kimbers are a bit better than Solen FastCaps at a modist cost increase. The Kimbers still give the network SOME snob appeal witout the big price tag of the Hovlands. I see no reason why the DIY builder couldn't do very well with Solens. They are good caps, or I wouldn't use them at all! You would just need to splice the leads longer or cram everything closer together. I just don't like doing that.

BTW: Using Microsoft Explorer, all I do is left-click on a picture and it lets you download a picture at full size.

Al K.

[Deang]

Just put your cursor on the pic and click!

[Deang]

Everything I've read about Litz says it's used in high current applications to reduce power losses due to skin effect at VERY high frequencies. I understand that you measure higher 'Q' with Litz at the frequencies we care about -- but what I don't understand is why the AC component, or inductive reactance isn't a factor. I mean, there is series resistance -- and the solid air cores have considerably less than the Litz air cores.

I've dropped back to using a single cap at the primary position. It's a voodoo thing.

[Marvel]

Just put your cursor on the pic and click!

That's great and I do it all the time. However, if you upload a pick

that should be able to fit in the frame/table correctly, and it shrinks

them smaller automatically. The old way wasn't really that hard to work

with, and gave us a lot more options. (end rant)

Bruce

[Al Klappenberger]

Dean,

The usual use for litz wire is in RF inductors at frequency in the range of 100 KHz to, maybe, 1 MHz. It improves the Q by providing many surfaces to carry the current. The theory is that skin effect makes current flow move more to the surface of a conductor as frequency increases. HOWEVER: At audio frequency the skin effect is not a factor. I really do not know why it makes such a dramatic improvement in Q at 6 Khz. My personal THEORY is that it acts like the laminations in core of an iron core inductor by breaking up edy currents that are induced in the core. I suspect somehow current trys to flow around the wire rather then through it. I can't figure how though! The reason Litz wire isn't used at frequencies higher than a few hundred KiloHertz is the distributed capacity between strands. That is also not a factor at audio. At low audio frequency (woofer filters) only the DCR of the wire seems to count. Litz wire don't help any. It only startes to be a significant factor above about 3 KHz. In the case of the .2 and .3 mHy inductors in my Universal network the Litz inductors have "Q" values around 40 to 50 at 6 Khz compared to less than 15 for solid wire. Good caps, on the other hand, range over 1000.

BTW: I agree about the display of graphic images. If it's not biger than the "Client area", it should not be compressed.

Al K.

[jwc]

Dean. Are those some Super AA's? If so, where is the low pass inductor? Maybe its not. Those yellow caps are exactly the same size.

Also. Can the T2A be substituted for the Universal 3619 for your Super AA?

jc

[Deang]

Yes, "Super AA's" for lack of anything else better to call them. Those caps aren't really the same size.

You can use a T2A, but you'll have limited attenuation choices since there is no 'X' tap.

Right, no low pass on two of those networks -- can you figure out why?

[Deang]

Al,

I'm putting in an order to Solen on Monday. I'm about ready to build up my '"Super ALK 350's", so I guess I'll build them both ways and see what happens.

[jwc]

Yes, "Super AA's" for lack of anything else better to call them. Those caps aren't really the same size.

You can use a T2A, but you'll have limited attenuation choices since there is no 'X' tap.

Right, no low pass on two of those networks -- can you figure out why?

oohhh Dean. I was just getting ready to pop a beer can and you pimped me. Hmm. Is it so the user can biwire or biamp and you will place the low pass inductor on a seperate board. Or just simply put on another board and tap the input leads to it.

jc

[colterphoto1]

Yes, "Super AA's" for lack of anything else better to call them. Those caps aren't really the same size.

You can use a T2A, but you'll have limited attenuation choices since there is no 'X' tap.

Right, no low pass on two of those networks -- can you figure out why?

oohhh Dean. I was just getting ready to pop a beer can and you pimped me. Hmm. Is it so the user can biwire or biamp and you will place the low pass inductor on a seperate board. Or just simply put on another board and tap the input leads to it.

jc

Yes, or I was thinking naturally about my split LS cabinets which might naturally be actively biamped and that portion of the xover would be superfulous.

Michael

[Deang]

Very good Michael -- yes, it's for AK Series users who already have a good low pass section sealed up inside their bass bin. They just have to swap out the top section and reconnect the jumper coming from the woofer section.

[D-MAN]

Oh, man - NOT the low pass-in-the-back-chamber thing![]

DM

[D-MAN]

I've been thinking about how to hide the crossovers (being that they just seem to get bigger and bigger) for the naked horn look, i.e., no top cabinet.

First, the BC is OUT! never in a million years would I do that.

Ok, so that leaves what? Down the back of the tailboard - that's one idea.

That would be one custom crossover, that's for sure. Now how to protect it from being banged up?

The other is a separate (nice-looking) box that sits between the amps and the cabinets. Means long wires, and if you guys don't think wires matter, try that! Well, it could be done.

Anyway, just thinking. It's a problem, that's for sure.

DM

[jwc]

Very good Michael -- yes, it's for AK Series users who already have a good low pass section sealed up inside their bass bin. They just have to swap out the top section and reconnect the jumper coming from the woofer section.

So it is basically like the second example I made in which you tap from the inputs to a low pass filter. In this case, you are using the existing low pass in the bass bin. So does this AK series use a 2.5mH inductor for the low pass? I thought most of the newer low passes were at least a second order and had an inductor value of around 4.0mH.

[Deang]

I'm not too big on anything in the chamber either -- I think it's a bad idea all the way around. But it's already in there, and it's a good filter.

That's right JC, 4mH plus cap (140uF for the AK-2, 100uF for the AK-3).