what brand of caps?

[babadono]

Michael,

Bob Crites has a tutorial on his website on the rebuild of an AA crossover (using the parts he wants to sell you, of course ). But methinks all his instructions and points should translate into what you are doing. IMHO always SOLDER --- if done correctly it lasts forever.

babadono

[Don Richard]

Make sure you use real lead/tin solder, 60-40 is good. Stay away from the lead-free ROHS stuff, and 4% silver solder has no advantages as far as electrical work goes. It's used for stronger mechanical joints, though not as strong as real silver solder.

[Marvel]

Make sure you use real lead/tin solder, 60-40 is good. Stay away from the lead-free ROHS stuff, and 4% silver solder has no advantages as far as electrical work goes. It's used for stronger mechanical joints, though not as strong as real silver solder.

I've used 2% silver solder for years. It makes a nice clean connection and holdes up well.

The government has a waiver in the lead-free ROHS solder as is tends to deteriorate over time.

Bruce

[colterphoto1]

Thanks for the link, that's pretty much what I'm doing only it doens't look as neat as what Dean has been showing me. I just disconnect one Cap at a time so I can keep everything straight. Did one yestereday, will go out and do a brief A/B listening test with mono source to see if I can tell any difference. The caps in these LSI split tops have been in sealed boxes and look brand new, hope I'm not spending time and money for nothing but figured I'd try it out on these old pro speakers before I tackle everything here one bit at a time. And the screws are brass.

[colterphoto1]

and what's all this about 'break-in' ? Really? Or is that some old techie's tale?

[Panacea Engineering]

Michael,

You might want to Browse through this.....

http://sound.westhost.com/articles/capacitors.htm

[DrWho]

Michael,

You might want to Browse through this.....

http://sound.westhost.com/articles/capacitors.htm

Very good article, but ironically doesn't cover one of the dominant sources of distortion that even gets listed in the capacitor datasheets...try swapping out an NPO ceramic for a Y5V in a high-shelf filter [+o(]

Also, the test circuits he presented are good at making the capacitor not matter (like with a series DC blocking cap). When you start doing filter design, the relative impedances are much closer and the capacitor behavior starts to matter quite a bit. And in a speaker, the voltages and currents are also much higher.

At the end of the day, there is good science behind all of the audible behaviors and it is relatively straightforward to mitigate with good design practice. The easy approach for 99% of the situations is to just spend a little bit more on film caps.

[MikeFord]

and what's all this about 'break-in' ? Really? Or is that some old techie's tale?

It takes a few weeks or maybe a month for our ears to get used to a new sound, so I am a strong believer of "break in" on the perception side.

I am totally mixed on the concept as it applies to most "hifi" devices and materials.

OTOH what is the harm in delaying your judgement on a some audio change?

[MikeFord]

At the end of the day, there is good science behind all of the audible behaviors and it is relatively straightforward to mitigate with good design practice. The easy approach for 99% of the situations is to just spend a little bit more on film caps.

Use a better cap can get real expensive and real big pretty fast.

[erik2A3]

Bypassing is common in power supplies where, as you know, electrolytics are often bypassed by a much smaller value plastic film. Combining different values in crossovers to achieve the required value is common, though I have seen on this forum in the past where it was indicated to be better practice to use a single cap of the correct value. That premise is arguable in a number of ways, one of which being the need for an odd value that can only be obtained by paralleling one or more with another. The other element I have also noticed has to do with what I agree to be improved sonic performance by bypassing or paralleling values, which may have the benefit of lowering overall ESR values, perhaps thereby lending a subjectively cleaner, brighter HF response. There are also times where a spendy film capacitor might not necessarily be advisable, as in the instance where a good quality electrolytic was chosen for a specific reason, as in a woofer circuit, shunt element (where in my view it makes sense not to throw lots of money), or in even a tweeter circuit (a thought most unpleasant to some). There are in fact some very good-sounding, sonically quite transparent electrolytic capacitors. I have read papers or design articles that urged hesitation before an out-of-hand discarding of the specified electrolytic for a like value film type. A few years ago I built a 300B, single ended circuit by a relatively well-known company, from whom I purchased the schematic and some of the parts -- chassis and OPTs primarily, though they don't sell to the DIY community on a regular basis. The designer asked that I PLEASE NOT substitute a film capacitor in either the 300b cathode resistor bypass position OR in the power supply, which is in part my own design (choosing a specific power transformer) using high value, very high quality electrolytics connected in series. This circuit is the antithesis of the usual (often not always deserved, IMO) reputatIon for SETs that are rounded off, dull, euphonic, lacking in LFR. It is extremely fast and transparent, even with speakers in the 86dB/watt range. Some hate the idea of an electrolytic in a cathode bypass position, but in this design it did in fact sound better than with a film, which I admit I tried, anyway. The designer was right.

[erik2A3]

Mike: I agree with you100% about break in being strongly related to the human element and simply the time it takes to get used to a new sound. A good friend wrote me about this the other day as well, explaining that he was probably the part that needed breaking in. Mechanical break in, such as with loudspeaker dynamic driver surrounds and spiders, absolutely, of which I probably have another few hundred hours to go on our new Lowther PM5A drivers. These things take forever....

[colterphoto1]

So then it's not truly a 'break-in' period in which the electrical properties of the device change. I'm having a friend over today for a blind A/B test of an original LSI and one I replaced all caps in. Will report back.

[erik2A3]

I can only speak from my own experience. Some people have great confidence in the perceived break-in period effects associated with passive components like capacitors, resistors, etc., even a simple piece of wire. There are manufacturers who claim such parts or organizations of multiple parts (we often call them "components") MUST be allowed to complete a certain break-in period before said part or component will meet its design criteria in terms of sound. There are also those who consider "breaking in" of a piece of wire (speaker cable or interconnect), dielectric properties of capacitors, and so forth to be generally unfounded. I admit there are times when something sounded better to me after a few days when I became more accustomed to the sound. One crucual "component" (if you will) of playback system that is more often than not completely excluded from the playback chain -- one part I have come to see as the key to the whole system -- and the one which also in my view is probably the least consistent on a daily basis, simply due that specific component's physiology, is the person sitting in front of a collection of transducers and other machinery, trying to determine whether or not the whole thing sounds different one day to the next, and, if so, if the difference is for the better or worse. I remember last year I decided to try a new pair (one per channel) of coupling capacitor from one company that has obviously high overhead costs in terms of the kind of advertising they pursue. There were great reviews all over the internet, though claims that they needed time to "break in.". Despite their rather attractive name (which I suppose some group of individuals got together around a table to discuss and decide upon -- probably much the same way other groups of people at other kinds of companies get together to contemplate best identifying names for room freshening sprays and colors of crayon..."River Canyon Mist" and "Sunset Rust" for the crayon, maybe....the cap I installed sounded really unpleasant from the get-go. And it stayed that way. From day to day and into a couple of weeks, until I got tired of waiting. It had a certain sound that simply didn't improve, though I thought on some days it hopefully began to sound better. I put the original Orange Drops back in and was much more at peace and forgot about it. One capacitor change I made probably more than ten years ago was for Jensen PIO types that were put into a single ended amp. I liked it immediately. They had replaced four Hovland MUSICAPS. After awhile, several months I guess, the PIO types began to sound less, rather than more, transparent. Veiled, actually. And I changed them. Maybe I changed...my moods change. I don't know, I guess I focus more on circuit design now than composistion of capacitors. I know of designers who just stay away from the topic and encourage people who decide to build their amp or preamp , or whatever it might be, to season to taste as far as the cap question goes, but to please use the specified value. I use a mix of motor run and film caps in my own crossovers. The combination sounds good to me. I also know of those who have built or purchased capacitor and cable break in components to help speed up that break in period. And there are people who would never consider the purchase of such an apparatus. What matters in the end, of course, is what one believes in and prefers for oneself.

[erik2A3]

BTW: there is one company about which I know that has in fact focused on the perceived positive effects of bypassing as the very basis of their design: REL Multi-cap, which as their name implies, makes use of bypassing in terms of the composotion of the capacitor (aka self-bypassing). Michael Percy, who has been around for awhile and in my experience is great to work with, I believe has Multi-caps.

[DrWho]

The reason bypassing works in power supplies is because you have
noise at different frequencies. A 100pF capacitor on a power supply has
very little storage capability at 20Hz. Likewise, a big 100uF
electrolytic has relatively much worse ESR - which means its impedance
at the high frequencies is much greater than that of the 100pF ceramic -
even moreso when you take into account the extra lead inductance from
being physically bigger.

I always think it's better to think
of caps in terms of relative impedance, and then look at the power
supply filtering as creating a low impedance path to the ground so that
the noise is common to both sides of the power supply - and then the
parts being powered don't see that noise across their inputs. Looking at
a plot of each capacitor's impedance makes it very clear when you want
to use certain types.

Getting back to speaker level crossovers,
it's still the same impedance perspective, but this time you're making
frequency dependant voltage dividers across all the various parts and
the speaker itself. It is ok to have a non-ideal capacitor impedance
just as long as it stays constant - when that is the case, the shape of
the impedance sets the total system frequency response - in some cases,
it may even be more ideal to have a shape that isn't perfectly textbook
(since our speakers are anything but flat). However, the second that
capacitor impedance starts changing with voltage or temperature, then
you start creating distortion. You've also got the shelf-life affecting
the freqency response slowly over time. Whether or not that distortion
is audible is a totally different matter, but it's certainly something
that can be measured way beyond the limits of our hearing acuity.

All
that to say, the impedance (energy storage) of the capacitors in a
speaker are very close to that of the speaker itself - especially at the
corner frequencies where the capacitor will have its largest influence
on the sound (so if you want to know what to listen for, find musical
content happening nearing the xover frequency - the only downside is
that all the other artifacts from having two dissimilar speakers, the MF
and HF, playing the same range will usually make it very difficult to
hear, and also make it very listening position specific unless you can
get outdoors into the far-field). When you go to a bypass capacitor -
say two caps with half the value each - then the impedance of each cap
is double (half the energy storage) - even if the ESR was cut in half,
it is such a small percentage of the total impedance at that frequency.
Say your cap is effectively 8ohms at 1kHz, then an ESR change from
40mohm to 20mohm is only a 0.02dB change - and that's just for the
linear constant expected behavior. The non-linear difference should be
more like 0.001dB - both are levels totally swamped by the
non-linearities of the speaker drivers, which are lucky to be within
1dB.

That is why it always makes way more sense to put more money
into the drivers than into the xover components - and if you go to an
active xover, then you don't have to deal with any of that and you can
improve the clarity in the xover region by adding time-alignment. That
difference is easily several orders of magnitude greater than a perfect
capacitor versus an old stock capacitor.

At least that's where I
sit on the issue today - I'm still open minded to hear about situations
where the numbers indicate that there should be a greater difference. I
believe in a causal universe so I believe anything audible should have a
known reason behind it. The more I learn at Shure, the more I realize
just how much capability we have to model all of these behaviors - makes
for some interesting improvements too when the math works out, and then
you can just do it on command within any design - that's when you know
you're understanding how something behaves.

[colterphoto1]

Thanks for all that but I'm waaay behind you guys. We're talking basics here. Roger came over and I was ready to proudly show off my A/B comparison of LSi splits with new caps on ONE side, with source material in mono. Well, after I got them wired correctly (that split AA network doesn't wire up like you may think - the 'standalone' LF OUT terminals are what lead to the INPUT on the network for starters, it's a disaster. Seems there must be something terribly wrong with one or more speaker components so I'm back to forgetting about the damned network revisions and taking all of the components out and testing them side by side. One mid driver and/or one or more tweeters must be wonked up. One cabinet set sounded great with the new network, and the other rather dull. Then swapped networks and the new network one sounded like hell colmpanred to the other. Seems I have one good network and maybe one nice set of components here. Drat. Time to shelf this project again for several months. This stuff is so frustrating some times.

[erik2A3]

Mike: So you're with Shure now? That's really great -- I knew you were headed for something like that; thanks for sharing your thoughts and expertise on this. I agree bypassing can work in power supplies, but like seemingly everything, there are those who are pro bypassing in PSUs, and as just as many against. Where I have found a difference in power supplies is honestly not so much in the DC ripple filter, but on the AC end prior to rectification. And in this case as a high frequency shunt to ground to ameliorate HF oscillation/hash associated, in particular, with solid state rectifiers feeding high values of filter capacitance. As far as signal level bypassing, I've corresponded many times with others about bypassing of cathode resistors in certain stages of amplification, which to me can greatly improve not only frequency response but, as of course would be obvious, gain as well. But there are designers who would far sooner opt for what might be described as a more "pure" approach, particulalrly if the alternative makes use of a dreaded electrolytic. In such cases, I have have also seen a trend to bypass such an electrolytic with a small percentage of the total capacitance in a plastic film type. Also on a signal input side of things, a very small value capacitor (ceramic is FINE!) can be used to bypass an insulated RCA ground tab to ground at a very close point on the chassis -- the purpose here, again, to funnel out noise in the form of RFI before it gets into the audio. So I think they do have their place, but as you suggested, it helps to know the hows and whats of the practice.

[colterphoto1]

I give up.

[Guest " "]

by-passing caps, like everything else in audio, doesn't make a difference except when it does make a difference......to qualify my statement...here are some pic's which were posted here on this forum a few years ago...and what I get out of the pic's is that the results of the by-passing varies depending what types of caps you use in the by-passing pairs. What I got out of this was that PIO and polycarbonate pairs had the best results. So if your are by passing caps, take note of the type of cap it is and the pair member indicated and the expected result.

[Guest " "]

...........