Best type of capacitors for recapping Type A Xovers?

[BEC]

Craig,

Guess you will have to help me with what you mean with that one.

Bob

[DizRotus]

There once was a man from <?xml:namespace prefix = st1 ns = "urn:schemas-microsoft-com:office:smarttags" />Burton

Of his opinions he was always certain

When they passed out the humor

He thought they said, Tumor

So he hid behind a curtain.

[NOSValves]

Bob,

I mean your nit picken. You can't possibly think that someone from HH Scott when the big SS push was on was going to tell you. "Oh tubes are better and this SS amp is a POS but you should buy it anyway"

Craig

[Deang]

Hee hee. I've never pitched a new GE can, but it was fun pretending I have. The sound I hear are from the old nasty ones I toss. After a while I do manage to collect a couple of handfuls worth of the GE motor runs, and they sit until I find the time to build something pretty with them. I think the GE cans provide a very rich and detailed sound at lower volume levels, but makes the speaker too bright/aggressive at higher volumes. I like the Type A's with them for late evening listening. I just think the Auricaps and Kimbers provide the best balance at all volume levels. When I build for myself I go broke and use PPT Thetas. The difference in sound between an Auricap and Kimber to a Theta is a hair's width -- but I can hear it.

"Magic caps" are mil spec polypropylenes using virgin film and specifically designed for AC applications. I'm grateful that companies exist that purchase and stock these capacitors so "audiophiles" and hobbyists can obtain them in small quantities. I'm not likely to ever be in the position where I can order directly from a factory that requires a minimum order of a thousand per value.

The impact to the sound from the use of different capacitors is sometimes overstated, and I've certainly been guilty of this myself -- but there are differences, and they aren't that difficult to hear. I'm sure if I was tested using a system I'm not familiar with and music I don't know well -- it would be difficult.

For the most part, I've lost a good deal of interest in the stock filters -- I think they fall short compared to the constant impedance filters. Debates about capacitors aren't as interesting to me as the changes in sound caused by the circuit.

[NOSValves]

Where I come from for something to be funny it actually has to be funny[] I didn't see anything humorous in Bob's reply or his follow up reply. No big deal either way.

Craig

[NOSValves]

Hee hee. I've never pitched a new GE can, but it was fun pretending I have. The sound I hear are from the old nasty ones I toss. After a while I do manage to collect a couple of handfuls worth of the GE motor runs, and they sit until I find the time to build something pretty with them. I think the GE cans provide a very rich and detailed sound at lower volume levels, but makes the speaker too bright/aggressive at higher volumes. I like the Type A's with them for late evening listening. I just think the Auricaps and Kimbers provide the best balance at all volume levels. When I build for myself I go broke and use PPT Thetas. The difference in sound between an Auricap and Kimber to a Theta is a hair's width -- but I can hear it.

"Magic caps" are mil spec polypropylenes using virgin film and specifically designed for AC applications. I'm grateful that companies exist that purchase and stock these capacitors so "audiophiles" and hobbyists can obtain them in small quantities. I'm not likely to ever be in the position where I can order directly from a factory that requires a minimum order of a thousand per value.

The impact to the sound from the use of different capacitors is sometimes overstated, and I've certainly been guilty of this myself -- but there are differences, and they aren't that difficult to hear. I'm sure if I was tested using a system I'm not familiar with and music I don't know well -- it would be difficult.

For the most part, I've lost a good deal of interest in the stock filters -- I think they fall short compared to the constant impedance filters. Debates about capacitors aren't as interesting to me as the changes in sound caused by the circuit.

Dean,

Come on I know you even keep the old motor runs and polish them for reuse[]

Craig

[BEC]

Dean said: "I think the GE cans provide a very rich and detailed sound at lower volume levels, but makes the speaker too bright/aggressive at higher volumes."

Dean,

And what quality of the cap would you suggest might explain that observation.

Bob

[Erik Mandaville]

Bob:

I have sent you an autoformer-related PM.

Thanks,

Erik

[BEC]

Good Evening Erik,

You have a reply.

Bob

[Deang]

"And what quality of the cap would you suggest might explain that observation."

I don't know.

[JSharp]

"If it measures good and sounds bad, it is bad; if it measures bad and sounds good, you have measured the wrong thing."

- Daniel R. von Recklinghausen,

former Chief Research Engineer, H.H. Scott

I don't think anyone disagrees with that quote. Now if I were trying to

be snide, I'd ask how business has been of late for the H.H. Scott

Company. But I'm not, so I won't.

The question I'm interested in answering is more basic. Why is it, that

when we take a group of people and have them listen to an assortment of

equipment built with simple passive components, and a number of those

components are replaced with other components that have substantially

similar measured electrical characteristics, that the listening group

can't tell the difference without first knowing the identity of the

components that were changed?

Yet when given the identities and other information about the

components, the group is likely to argue at length about the sonic

differences among them.

Just a guess on my part, but I don't think the answer will come from

engineering texts. Books on psychology might be more applicable. Or a

field study where we interview an assortment of carnies, magicians,

salesmen, politicians, faith healers, etc...

[NOSValves]

Jsharp,

I agree with you and this is my point about taking a week or two of careful listening to truly come to a conclusion about any change in your system not just coupling caps in a X-over. The mind can really play some serious tricks on you. This is why I mentioned the distortion test that a few of us were taking in another thread. I found it near impossible to repeat my score after the first run of the day. If I took the test first thing in the morning I found I scored very good. If I repeated it anytime later in the day especially back to back the score fell by a good margin. Setting up switch boxes for quick changes IMHO just doesn't work especially when both crossovers are of reasonably good quality and the same electrical circuit your just not going to pick out any major differences. Besides like Dean and I mentioned many of these differences are much more apparent when you notch up the levels and try to get some live like reproductions. To me this makes perfect sense the more you pump into the cap the more its weakness will be magnified. To me lessor quality caps become very harsh at higher levels. I think trying to test for a reason is a impossible task.

Craig

[BEC]

What kind of voltage level here are we talking about? Perhaps 1 volt through the tweeter cap for low level and 3 volts when being pushed hard? You think that somehow effects how a 370 volt rated cap performs? Need to hear some science behind that one.

Bob

[JSharp]

Jsharp,

I agree with you

and this is my point about taking a week or two of careful listening to

truly come to a conclusion about any change in your system not just

coupling caps in a X-over. The mind can really play some serious tricks

on you. This is why I mentioned the distortion test that a few of us

were taking in another thread. I found it near impossible to repeat my

score after the first run of the day. If I took the test first

thing in the morning I found I scored very good. If I repeated it

anytime later in the day especially back to back the score fell by

a good margin. Setting up switch boxes for quick changes IMHO just

doesn't work especially when both crossovers are of reasonably good

quality and the same electrical circuit your just not going to

pick out any major differences. Besides like Dean and I

mentioned many of these differences are much more apparent when you

notch up the levels and try to get some live like reproductions. To me

this makes perfect sense the more you pump into the cap the more its

weakness will be magnified. To me lessor quality caps become very harsh

at higher levels. I think trying to test for a reason is a impossible

task.

Craig

Craig -

There's probably more going on than can be heard instantly with some

form of simple ABX switching. But I think some people use that as a way

to sell $5K speaker cables and then claim a person should "let them

break in for X hours before you decide" thus getting around the fact

that people can't differentiate between those cables and 12ga zip cord

from Home Depot. So even though I agree, there's probably a time factor

in order to hear small but real differences, I think that conclusion

gets misused.

And here are two things I can't reconcile -

It's hard for me to accept that all these minute sonic differences are

just some form of groupthink that large numbers of people are infected

with. People with different experiences, with different equipment, in

different places, and at different times.

But when we can't measure any substantial differences electrically,

people have an impossible time determining what's been changed, if

anything, unless they have some prior knowledge of the changes.

Another <shrug> I honestly don't get it.

--

Jim

[Deang]

http://www.hometheaterhifi.com/volume_11_4/feature-article-blind-test-power-cords-12-2004.html

http://www.pcavtech.com/abx/abx_data.htm

I think the problem I have with the process is that it typically puts people in an environment completely at odds with what they are used to. At home -- you have a room, speakers, and gear that form a symbiotic relationship, and work together to create a sonic signature that one becomes very familiar with. There is a Gestalt Effect of sorts, and small shifts in signature are relatively easy to hear. There is also the issue of the music one normally listens too as opposed to using material that one has little if no familiarity with.

For example, I'm very familiar with how my car drives. I run it hard, and I can tell pretty easily when it's time to do some work on it. It's a combination of things: how it shifts, feels through the steering wheel as the RPMs go up, the sound of the motor when I downshift, etc. Now, it's still running very good by most people's standards, but I can tell performance has fallen off a bit. Someone else could borrow my car for a couple of days and not even notice or be aware of the things I noticed. I could change the plugs and oil, give it to them again for a couple of days and I seriously doubt they would even notice the difference. Like most analogies, this one is not perfect, but I'm using it to show that with a high level of familiarity comes heightened awareness and sensitivity to small changes.

In the context of networks and their passive parts -- my listening "tests" are done using $2000 tube monoblocks, a $1200 tube linestage, and a $1500 player. Compared to what some are using I suppose it's a modest system -- but it's very resolving, open, and liquid. Now, if someone is going to say that there is no discernable difference in sound between my gear and a $399.99 receiver coupled to a $129.99 CD player from Best Buy because ABX testing said so -- then I'm going to assume either the testing method is flawed, or the listeners have hearing damage. In the same vein, I have to wonder about any results or conclusions drawn from listening tests done using a system of the latter type.

[NOSValves]

Bob,

I just did some quick testing with the system wound up pretty tight. The woofer will experience peaks in the 8 to 10V range, The Mid driver experiences almost the same. The tweeter has much less and rarely exceeds 3V. The problem with trying to test for this IMHO is that these voltages are multiple frequencies at once rising and dropping faster then my Fluke digital meter can even record. I'm not sure how you could duplicate that situation with test equipment? Same things goes for tuning and designing an amp. I can't just go by the test results because there not a perfect representation of the load music presents.

Craig

[sfogg]

Dean,

"if someone is going to say that there is no discernable difference in sound between my gear and a $399.99 receiver coupled to a $129.99 CD player from Best Buy because ABX testing said so -- "

Fundamentally ABX testing can *not* prove there is no difference between A and B. You can't prove a negative. This is important to understand.

What it can do is prove a positive (IOW yes there is a difference between A and or that during that test the listener could not tell the difference between A and B. It is important to note that just because someone failed that test that doesn't eliminate the possibility of someone somewhere else passing it.... or even the same person passing the test a day later. That is why it can't be used to prove a negative. Anyone that says otherwise doesn't understand what ABX testing can and can not do.

For example if you claim to hear the difference between two caps an ABX test could be used to determine if that is the case. If you pass the test it proves you hear the difference you claim.... it also proves that in at least some sitatutions there are differences in sonics between caps. Perhaps 99% of the population can't hear the difference but you proved you can.

However, if you fail the test the only thing it shows is that you couldn't determine the difference between the caps. It doesn't mean Craig (or whoever) couldn't under a different test.

This is why WRT ABX testing the burden of proof is on those making the claims. Those that doubt the claims can't prove a negative, it isn't possible. Only those claiming a positive can actually prove it.

"I think the problem I have with the process is that it typically puts people in an environment completely at odds with what they are used to."

So do ABX testing in your own home in your own system. With an ABX box you *can* literally extend the test over days or even weeks if you so desire.

However... if you do that you will likely find that you are *far* more able to determine small differences in sound when you listen to a familiar pieces and are able to switch back and forth between A/B/X quickly for immediate comparisons. Auditory memory is extremely faulty... rapid comparisons help to cover for that.

As an example of how bad auditory memory can be try the PCABX program you mentioned above. Use the SPL level comparison files or make your own with a digital audio editor program. Start off with 3dB level difference files and set the switching time to a very very short delay. You should be able to pass this test with absolute ease. Now set the delay between switching to something higher... say 10 seconds. You might find your scores are no longer as good as when there was no switching delay. I forget what the limit of the PCABX program is for delay but then set it to a minute and see if you can still pass the test.

Then try repeating the differing time with a smaller level difference... maybe 1dB.

You will likely find at some point you will be able to pass the test with little to no delay between switching and won't be able to pass the test with a longer delay between switching. And that is knowing exactly what to listen for as the difference between A and B.

"There is also the issue of the music one normally listens too as opposed to using material that one has little if no familiarity with. "

Most ABX tests the person under test chooses the material listened to. And often they are done in their own system as sort of 'prove it' tests for wild claims they have made with regards to whatever.

Shawn

[Deang]

"What kind of voltage level here are we talking about? Perhaps 1 volt through the tweeter cap for low level and 3 volts when being pushed hard? You think that somehow effects how a 370 volt rated cap performs? Need to hear some science behind that one."

Not that it matters all that much in this context, but the Type A network driven with 100 watts will send about 15 watts to the tweeter -- about 10 volts.

Just a thought here, but maybe it's more closely tied to how the ear responds when you yank all of the resistance out of the circuit.

Frequencies for maximum sensitivity of human hearing

The hearing curves show a significant dip in the range 2000-5000 Hz with a peak sensitivity around 3500 -4000 Hz. This is associated with the resonance of the auditory canal. There is another enhanced sensitivity region at about 13,500 Hz which may be associated with the third harmonic resonance of the auditory canal. The high sensitivity region at 2-5kHz is very important for the understanding of speech.

Show entire loudness curve set

Index

Hearing concepts

HyperPhysics***** Sound

R Nave

Go Back

Auditory Canal Resonance

The maximum sensitivity regions of human hearing can be modeled as closed tube resonances of the auditory canal. The observed peak at about 3700 Hz at body temperature corresponds to a tube length of 2.4 cm. The higher frequency sensitivity peak is at about 13 kHz which is somewhat above the calculated 3rd harmonic of a closed cylinder.

Index

Loudness concepts

Hearing concepts

HyperPhysics***** Sound

R Nave

Go Back

[Deang]

Thanks for the post Shawn -- I understand.

[psg]

100 Watts of what? Noise?