DC Bias of Caps?

[boom3]

Well, Infinity did it the mid 90s, and now JBL is doing it on its new flagship, Project Everest DD6600. They're putting a battery across a crossover cap to maintain a bias, and, it is averred, reduce 'crossover distortion'. In this sense, they're talking about the distortion a device shows when it goes from negative going waveforms to positive going waveforms. This used to be an issue with push-pull Class B amps, but most of these operate in a slightly biased mode, meaning they are Class AB.

For modern film caps, I'm not sure how much of an issue this would be. It seems like one of those distortions, if measurable, that is small compared to other distortions and might be, at best, marginally audible. Anybody, everybody, chime in.

[Guest " "]

I've heard and done battery insertions on transistor circuts, but never a cross a capacitor......are we talking crossover as in speaker crossover or active crossover.

I guess next, someone will come along with a push-pull passive crossover that, as a result of the inverted side, will eliminate this switching distortion.

[DrWho]

You can measure the characteristics of any electrical device to find the regions they are most linear. DC Biasing is just an easy to way to get your signal into that region. I've got 3 sections in my text book for a circuits class I'm taking this semester that discusses these issues (yea, I've almost read the entire book before the first day of class....). I don't think recall reading anything about DC bias on capacitors, but I suppose it could be done.

(But don't capacitors ignore DC?)

[boom3]

You can measure the characteristics of any electrical device to find the regions they are most linear. DC Biasing is just an easy to way to get your signal into that region. I've got 3 sections in my text book for a circuits class I'm taking this semester that discusses these issues (yea, I've almost read the entire book before the first day of class....). I don't think recall reading anything about DC bias on capacitors, but I suppose it could be done.

(But don't capacitors ignore DC?)

Capacitors block DC because they store it as an electrostatic charge between the plates, between which there is no direct connection. This is the reason a CRT or othr high-voltage capacitive device can hold a nasty shock for years after power is switched off.

JBL is not a company to take liberties with physics, unlike another well-known company frequently dissed here. However....I still think the notion of making capacitors "more linear" by DC bias is far-fetched. Another point: if you have a battery in parallel with a cap, what does that do to the signal? IIRC, the Infinity design used a very high-value resistor in series with the battery to keep the current level very small.

[Guest " "]

What would a proto-type look like.

Crossover's use bi-polar capacitors, If you had a battery in series with a resistor in parallel with the capacitor, your basiclly adding a very slow capacitor to the circut that would charge during one phase of the signal and discharge during the other.  

Can somebody post a pciture of what they think this would look like.

[sfogg]

Some threads about 'charge coupled' networks on different JBL speakers...

http://audioheritage.csdco.com/vbulletin/showthread.php?t=3427

http://audioheritage.csdco.com/vbulletin/showthread.php?t=43

http://audioheritage.csdco.com/vbulletin/showthread.php?t=8111

Shawn

[WMcD]

It is all very interesting. But I don't buy it.

I'll drag out my favorite home-grown bromide. There is no benefit to solving a problem which does not exist in the first place.

This bias theory "solves" a purported problem. I.e. that the caps suffer some anomaly at zero crossings where applied voltage shifts from plus to minus. And by inference this happens at very small, small voltages.

If there IS a problem it shouldn't be too difficult to show on test bed with an oscilloscope. It should show up on a distortion meter.

There is something to be expected in normal science: The fellow saying there is a problem would illustrate this purported zero crossing problem with a 'scope. Then he would hook up a distrortion analyzer. Then he would he say: Look, look, look. Here it is!

Then he would apply the battery bias. If there is a change, he would say, "Look, look, look, I solved this problem.

And, somewhere in here he would have to show that the distortion (if it is there on the 'scopes) IS audible.

So, are there any such presentations?

Gil

.

[sfogg]

Gil,

I haven't seen any tests like that either. Some people on the Lansing site say they hear a difference (for all that is worth) some say they don't. I don't think it has been ABX tested either way, at least not from what I've seen.

If someone wanted to try this with a Klipsch crossover it would be easy enough to do with a Type A or whatever, esp. if they were building it from scratch. I am kind of surprised Dean has never tried this actually. He won't try glass capacitors either.

Shawn

[Guest " "]

Ok, took a look at the thread.

He thinks he's applying this to the low-pass section.

Problem is he's using a DC power source across bi-polar capacitors, so the circut only works on half the phase, if at all. The circut would charge the battery on during one phase, and discharge it during the other.

If he really wanted to be convincing, he should use two batteries, and two non-polar capacitor's in series, and apply a battery circut to each of the polarized capacitors, at least that way he would have both phases covered.

Maybe he could also use two diodes, two batteries, and use the bi-polar capacitor and apply this to both phases.

This reminds me of the electro-voice tweeter protection circut which uses a set of diodes and a capacitor to store power and then trigger a relay. "Who knew" that if we would have put the EV protection circut in the Low-Pass section we would have solved the crossover distortion problem.

[djk]

DC bias is used on capacitors to reduce the effects of dielectric absorption.

The dielectric absorbs some of the charge, and then it releases it when going through the zero crossong point of the signal. An expensive capacitance meter can measure this. All one needs to do is measure a capacitor with poor DA vs one with good DA. An electrolytic vs a polypropylene is a good example. Listen to the two. If you can't hear the difference, then there is little point in paying for expensive caps, and no point in your trying battery bias of same. The less DA a cap has, the 'blacker' the silence between notes in quiet passage will be, and the 'cleaner' it will sound.

The DC bias shifts the crossing point away from zero. This is important because there is no signal at the zero point to mask the 'grunge' of the dielectric releasing itself as it switches polarity.

DC bias was seen in early Infinity speakers because of the low crossover points to the midranges, and the low impeadances dictated large value capacitors (electrolytics). Electrolytics also last longer with a polarizing voltage across them, in addition to sounding better. Battery life is essentially the shelf life of the battery, unless you have a leakey cap (which needs replacement).

A well respected EE and amplifer designer on DC bias on good caps:

Auricaps

Post #1

Wondering if anyone has much experience with Auricaps? I have been testing them for some time and would like to get others experiences with them. Maybe ask a question or two as speaker crossover caps. Steve

Quite good - strong bass for their size.

Sound muddy until they have a decent DC bias across them. This is fine for the majority of tube circuits, of course.

Cheers,

Hugh

__________________
Hugh R. Dean
Aspen Amplifiers P/L
Melbourne, AUSTRALIA
www.aksaonline.com

Of course, the difinitive article on caps is: Jung, Walt & Marsh, Richard N.; "Picking Capacitors", Audio, Feb./Mar. 1980
(but I can't seem to find a current bookmark for it. Has instrument tests with and without DC bias)

For reference, here is the old JBL K2 diagram:

http://manuals.harman.com/JBL/HOM/Technical%20Sheet/K2S9800-9800SE%20ts.pdf

As you can see, only one battery is needed, no matter how complex the crossover gets. The basic idea is to double the cap value, and pull it up to the battery with a 2.2M resistor.

[Deang]

The Jung/Walt/Marsh documents are on Richard's website, but he's applying some changes to the site so they're not available right now. www.capacitors.com

Interesting stuff. So why the application on the low pass section but no application of the technique on the high pass?

[djk]

"So why the application on the low pass section but no application of the technique on the high pass?"

The JBL uses it on every cap in the whole network. I'm not sure about the old Infinity models.

I've seen people compare cheap caps with bias vs expensive caps without bias, and they prefer the bias set-up. Remarks usually follow along the lines of 'cleaner' and more 'detail'. I am unwilling to pay the price for the better Mundorf and their ilk, so I can't speak for any improvement to something like that (the cost to double not only the value of the caps, and the total number would be unreal).

I can see it with lower priced caps. You could pull the batteries afer one year and put them in your smoke detectors for the next year, no extra cost that way.

[mark1101]

Sounds like another gimmick to get away with cost reduced electronics.

I can tell you I have Auricaps in just about every amp and speaker I own. They are not muddy, and have one of the shortest break in periods of any caps I have tried. I think they are excellent in almost every regard. I don't know where people get their information. They certainly must not be spending their cash and actually trying them. I would love to hear the application where an Auricap sounds muddy.

Edit: I guess you can infer that I don't use batteries on them....[]

[boom3]

DJK:

"The DC bias shifts the crossing point away from zero. This is important because there is no signal at the zero point to mask the 'grunge' of the dielectric releasing itself as it switches polarity."

So...what does this "grunge" look like on a scope? I really think that with any good film cap, this 'grunge" would be below the noise floor of the electronics, not to mention the listening room.

[Guest " "]

I'll have to try this on a crossover network that already has series capacitors.

[tigerwoodKhorns]

I was just reading up on this last night when I found an article that I had downloaded on my hard drive. Before charge coupling, JBL used very small film and foil caps as "bypass" caps for the transition zone.

Don't dismiss what JBL is doing just because they are non-klipsch. The guys on the heritage forum are very knowledgable and JBL makes very nice stuff.

Here is a capacitor question. I am re-capping a pair of B&W speakers (DM3000's - small towers). I can go for Solens at abotu $40 or Sonicaps at about $160. Are the Sonicaps worth 4 times the price?

[Guest " "]

Tigerwoodkhorns

Do you have a copy of the article?

btw, I'm not dismissing this, I think it will be easy to try, especially if you have xovers with caps already in series such as 4th order and 6th order xovers.

If it rates up there with by pass caps, then all the more reason to try it.

[tigerwoodKhorns]

I think that you have to double the value of the caps and then put them in parallel to maintain the same value of the original cap. The article is a press release but has a very simpel diagram on how to do it.

[tigerwoodKhorns]

.

[tigerwoodKhorns]

..