Best type of capacitors for recapping Type A Xovers?

[psg]

What Bob has as a source, amplifiers and

preamps can have a absolute effect on the out come of trying various

components in a crossover. Bob is stating that he uses a switching

device to compare these components. Well if his system is rolled

off and the frequency extremes are softened its very possible that

he and anyone else will not hear a bit of difference switching in

different components since most of the diffence is in the extremes.

Perhaps, but your examples of rolled-off vintage gear are extreme. He has said his gear is flat, right?

NOTE there is more to frequency response then

actual gross output 20hz to 20khz + - 1db !! You can have the

same output across the frequencies and still have softer at the

extremes. Just added this to be clear.

I don't follow. How can it be flat +/- 1dB and still be softer at the extremes?

[NOSValves]

Shawn,

You know what... I just wrote 3 paragraphed and erased them this is just simply a waste. Been through this time and again with you and its never done any good in the past so why bother with it again.

Craig

[NOSValves]

What Bob has as a source, amplifiers and preamps can have a absolute effect on the out come of trying various components in a crossover. Bob is stating that he uses a switching device to compare these components. Well if his system is rolled off and the frequency extremes are softened its very possible that he and anyone else will not hear a bit of difference switching in different components since most of the difference is in the extremes.

Perhaps, but your examples of rolled-off vintage gear are extreme. He has said his gear is flat, right?

NOTE there is more to frequency response then actual gross output 20hz to 20khz + - 1db !! You can have the same output across the frequencies and still have softer at the extremes. Just added this to be clear.

I don't follow. How can it be flat +/- 1dB and still be softer at the extremes?

Go pick out 2 amplifiers of different brands and maybe topology that are tested to be flat + - 1 db 20hZ too 20Khz set up a Sfogg approved ABX switching setup[] I can tell you the two will sound at least some what different. Flat frequency response just means that if the amp puts out say 10V at 1Kz it also does it within 1db at 20 Khz and 20 Hz this is with a single frequency not music (there is no way to test that accurately). It does not mean the reproduction at these extreme frequencies is absolutely honest to the source in fact no amp, preamp or source is absolutely accurate. This is why specs mean almost nothing about what something sounds like in the end. Here is another example using square waves for testing you can have two amps with this flat frequency response one amp at 20 Khz has a square wave response that look like a mountain range across the top the other has a somewhat square looking reproduction with a mild curved corner leading in the the top with no mountain range. Trust me these 2 amps would sound very different but both have the same rated frequency response using the standard sine wave testing.

Craig

[Deang]

Can we talk about inductors now?[]

[NOSValves]

Can we talk about inductors now?[]

Oh my........ now were going to need 84 switches and a triple ABX setup. Help me[]

Craig

[JSharp]

Dean...

they are very often designed to outlandish and questionable

specifications and consequently overpriced, then sold with dubious

claims in an attempt to justify this price...And the case of many of

them, it's more a function of how they're marketed and the volumes

they're manufactured in than it is in any use of superior materials or

better manufacturing techniques.

Nope, not so. Please do a little research O.K., you're giving me chest pains.

http://ldsg.snippets.org/appdx-ec.php3

http://www.capacitors.com/picking_capacitors/pickcap.htm

http://audience-av.com/on_capacitor_dielectric_material.htm

http://www.nutshellhifi.com/library/capacitor.html

http://www.faradnet.com/

http://www.designers-guide.org/Modeling/da.pdf

http://www.audience-av.com/Auricap%20Specifications.pdf

As

you've said, there just aren't that many people building capacitors of

this type anymore so the likelihood of them being built on the same

lines as commodity parts is very high, as is the likelihood the

performance is almost identical.

Uh, that's not what I said.

Sorry about the chest pains.

When you pointed out that three "audio capacitor companies" were using

parts built by the same manufacturer, then pointed out that a couple of

other more common passive component companies didn't build parts in many of the

values needed for crossovers, I assumed you were pointing out that

there were few manufacturers of these devices or that the volumes were low. Do those statements mean

something different to you? If so, then what is it?

It's true by the way, there just aren't that many actual manufacturers

of

these devices nor is the volume large on them comparatively. How many

30uf 250 volt audio caps do you think get sold worldwide compared to

something like a .1uf, 10volt smt0402 ceramic capacitor? Or Bob's GE

motor runs

for that matter? It's many orders of magnitude

different. And there's the rub. Small volume components cost far

more

to make than those that are built in larger quantities even when

everything else about them is completely equal. It follows that cost

has only a

loosely coupled relation to the quality and performance of an

item or the cost of the raw materials

used to build it. We see that last point proven when we notice that

less than $.01 worth of sand can end up selling for $300 when it's made

into a CPU.

Outside of the questionable and unproven audio differences between

something like the GE caps and some of the specialty parts, this is my

real problem with a

lot of the high end audio products. You're not necessarily paying for

anything

superior, you're often paying for nothing more than manufacturing

inefficiency because of design criteria and tight

specs that aren't proven valuable, and low manufacturing volume.

So while I thank you for the links, I don't see one study that shows

any correlation between the more obscure characteristics and

construction techniques of capacitors and the audibility of those items

in use. But maybe I missed it? That is the subject of this thread isn't

it? While you're looking for those, hopefully you'll do some

research into manufacturing cost vs. volume curves in electronic

component manufacturing so you can see where the costs come from in

specialty component manufacturing. Hint - it's mostly not quality,

performance, reliability, or any of the other things people might want

you to think it is...

--

Jim

[Al Klappenberger]

"I don't get it? Why would you use Solens in your high buck ES crossovers and Hovlands in your less expensive universal? I mean if they make no difference why the heck would you spend 20 times more for those 2 locations on the universal networks? Wouldn't it be pertinent to omit the Hovlands and pass the savings onto your customers? "

Greg,

There is several answers to this. (None of them are very good though!):

1 - This was my first design and I want to pull out all the stops! I was also not very aware of alternitives at the time.

2 - Snob appeal!

3 - The design is very simple compared to the ES networks and I still hope that the design might have an edged using super quality caps.

The truth is, I have always stated in my downloadable design info that you could save some money by using Solen caps instead. I really think the overkill is those 1 uf Harmony bypass caps. I have already stoped using the one around the 39 uF cap in the ones I build. I am using two 20 uF FasCaps instead. The ESR (Q) is actually better wit two 20 uFs. The 6.2 + 1 uF could be a single 6.8 uF FastCap instead. I suppose I should give the customer cap options like Dean does!

Al K.

[NOSValves]

"I don't get it? Why would you use Solens in your high buck ES crossovers and Hovlands in your less expensive universal? I mean if they make no difference why the heck would you spend 20 times more for those 2 locations on the universal networks? Wouldn't it be pertinent to omit the Hovlands and pass the savings onto your customers? "

Greg,

There is several answers to this. (None of them are very good though!):

1 - This was my first design and I want to pull out all the stops! I was also not very aware of alternitives at the time.

2 - Snob appeal!

3 - The design is very simple compared to the ES networks and I still hope that the design might have an edged using super quality caps.

The truth is, I have always stated in my downloadable design info that you could save some money by using Solen caps instead. I really think the overkill is those 1 uf Harmony bypass caps. I have already stoped using the one around the 39 uF cap in the ones I build. I am using two 20 uF FasCaps instead. The ESR (Q) is actually better wit two 20 uFs. The 6.2 + 1 uF could be a single 6.8 uF FastCap instead. I suppose I should give the customer cap options like Dean does!

Al K.

Al,

Thanks for the great straight forward answer. Although I would prefer the Hovlands [] I guess I'm a snob LOL!

Craig

[BEC]

Craig,

What I said or meant to say is "reproduces the audio specturm within 1 db or so". This would mean all the frequencies at least from 20 hz to 20,000 hz measured with a good sweep generator for an input and a good specturm analyzer for an output reading device.

Bob

Bob

[NOSValves]

But isn't that still one at a time sine wave or am I mistaken? Wish I had a spectrum analyzer maybe some day in the future. You should pump some square waves through you system if your really curious. See how well they can be produced 20hz to20khz

Craig

[BEC]

Craig,

You might describe it as a sine wave with continuously changing width (frequency) and the rate of change being adjustable from very quick to very slow. I also use noise when looking at power bandwith. That might be described as all frequencies all the time inside the bandwith being tested.

Bob

[BEC]

Dean wants to talk about inductors. Well we can start that discussion off with a bit of a test. Those who have easy access to their crossovers can try this. Use just half of the switch arrangement I described earlier and connect it so that it can alternately short or not the inductor in the woofer circuit. You could also just use a single test lead with alligator clips on each end to short the inductor. This test should show you the maximum difference you could possibly have in the absolutely most extreme conditions. When you see how little difference it makes to the sound whether it is in the circuit or not, you will see the difficulty in determining whether one inductor is audibly better than another.

Bob Crites

[tofu]

"I don't get it? Why would you use Solens in your high buck ES crossovers and Hovlands in your less expensive universal? I mean if they make no difference why the heck would you spend 20 times more for those 2 locations on the universal networks? Wouldn't it be pertinent to omit the Hovlands and pass the savings onto your customers? "

Greg,

There is several answers to this. (None of them are very good though!):

1 - This was my first design and I want to pull out all the stops! I was also not very aware of alternitives at the time.

2 - Snob appeal!

3 - The design is very simple compared to the ES networks and I still hope that the design might have an edged using super quality caps.

The truth is, I have always stated in my downloadable design info that you could save some money by using Solen caps instead. I really think the overkill is those 1 uf Harmony bypass caps. I have already stoped using the one around the 39 uF cap in the ones I build. I am using two 20 uF FasCaps instead. The ESR (Q) is actually better wit two 20 uFs. The 6.2 + 1 uF could be a single 6.8 uF FastCap instead. I suppose I should give the customer cap options like Dean does!

Al K.

reading this one post made skimming through the entire 8 pages of this thread worth it for me.

... and here i was about to put down $80 for the four hovland caps.

edit: well this brings the cost down to a total of $143 for caps/inductors (with a higher guage wolen woofer inductor) and an extra $60 for the two auto transformers.

[sfogg]

Bob,

"You might describe it as a sine wave with continuously changing

width (frequency) and the rate of change being adjustable from very

quick to very slow. I also use noise when looking at power bandwith.

That might be described as all frequencies all the time inside the

bandwith being tested.

"

Is your FFT a two channel model?

If it is and it has a transfer mode a *very* cool test you can do is

'Y' the input into the device under test. The other side of the Y you

run into channel A (or whatever is used as the reference in transfer

mode). From the output of the device under test you tap into that and

run that into channel B.

The transfer mode will show how B differs from A. IOW, if B is

'perfect' you will see a straight line on amplitude and phase response.

If B is rolled off you will see that on the screen.

Because the transfer mode compares A against B you don't need to use test tones for this test. It can be done with music.

Shawn

[BEC]

Shawn,

Yes, I have a two channel Wavetek 5820B Specturm Analyzer and an HP 3325A Synthesizer/Function Generator. The Analyzer directly diives an HP pen plotter. I bought them used of course. I guess the new cost of the two of those may have been over $30,000. The Wavtek is the only bench analyzer I have seen made specifically for audio testing. With its previous owner, it made a living for them setting up sound systems in auditoriums, theaters and concert halls around the counry. I even have its flight case they carried it in. It only goes up to 50 Khz, so great resolution is possible in the audio region.

Bob Crites

[NOSValves]

Craig, You might describe it as a sine wave with continuously changing width (frequency) and the rate of change being adjustable from very quick to very slow. I also use noise when looking at power bandwith. That might be described as all frequencies all the time inside the bandwith being tested. Bob

That might be described? I think its just a singal sign wave that sweeps back and forth. I like Shawn's idea better (did I just say that). I myself would know pretty much all I need to know from a simple square waves to at least know some what what your amplifier or preamp does at the frequency extremes. Some what old world but gets the job done.

I still say test equipment can not possibly give the entire picture when it comes to actually playing music. If test equipment could do it all it would simply take all the fun out of this stuff anyway.

Craig

[BEC]

Craig,

Looking at square waves will give you a rough approximation of what you can see with great detail on a spectrum analyzer.

Bob

[sfogg]

Bob,

Try the transfer mode sometime on the analyzer if it has that. It is

very slick... Al was the one who told me about it on the Spectral

Dynamics equipment.

"Yes, I have a two channel Wavetek 5820B Specturm Analyzer and an HP 3325A Synthesizer/Function Generator."

Nice choice in function generator... I have the same model. That older

HP stuff is fantastic, I have a couple of other HP pieces too.

My spectrum analyzer is a Spectral Dynamics SD-380 with a couple of

different input cards for it. The one I normally use is a two channel

model that gets up to 40kHz, the other input card is only a single

channel (no transfer function) but it goes to 100kHz.

Shawn

[DizRotus]

Bob,

The caps and K33 arrived yesterday safe and sound.

Nothing I've read in this lengthy thread has caused me to second guess my purchases from you. I'm eager to collect the La Scalas from Michael Colter on Monday and then get them ready for delivery to the high school.

In refurbishing the La Scalas, I'll try to be as scientific as possible and change only one variable at a time. Ultimately the AA networks will get the new GE caps you sent, and the "new" Eminence Gamma 8 ohm woofer in one of the speakers will be replaced by the K33.

I want to try to hear and discern any differences between the old and new caps and between the Gamma and the K33. Which would you change first? My plan is to listen to them as they are for a while, then change the caps and listen some more, and finally swap out the Gamma for the K33.

On the other hand, perhaps getting the woofers to match prior to refurbishing the networks will allow any differences presented by the cap change to be more apparent. Whadayathink?

[BEC]

Neil,

I would get the woofers squared away first.

Bob