Crossover Roadshow

[Deang]

I thought you might have misread what I'd said. Sorry.

[Guest David H]

Using two of the same caps hooked head-to-head with DC bias reduces the distortion, and can be even lower than with no DC bias.

If distortion can be lower with no DC Bias, what is the purpose? This is not sarcasm.

Dave

[Guest David H]

I thought you might have misread what I'd said. Sorry.

It's all good.

I was hoping Fritz would share a little insight as the to bridge battery Bias as well.

Dave

[djk]

"If distortion can be lower with no DC Bias, what is the purpose? This is not sarcasm."

Read it again, especially the part that says "Single capacitors (as in a coupling cap in a preamp) have higher distortion with a DC bias on them."

And then "The increase in distortion is dependent on the internal symmetry of the construction of the capacitor."

Thus "Using two of the same caps hooked head-to-head with DC bias reduces the distortion, and can be even lower than with no DC bias."

But the main reason for the bias is to avoid the zero-crossing dielectric absorption issue with makes things sound dirty, even though the distortion measures low.

[Guest David H]

Thanks DJK.

I also read through JBL's reasoning for using this configuration, though it seems to me that this really only benefits electrolytic or cheaply built caps.

I guess only benefits is a poor choice of words. Has a greater benefit on electrolytic or cheaply built loose wound caps.

Dennis said, "I'm not sure this is the best way to do it. In use there is a very low impedance from X to Y (X to ground of less than 0.1Ω, and Y to ground of less than 1Ω)."

The battery bridge can be used with a 1k resistor, however it is not required.

The battery bias capacitor bridge seems to have some advantages.

1. Circuit isolation.

2. Without the battery the capacitors are series / parallel therefore do not suffer from increased esr like the series network employed by JBL.

Granted, this will have little to no effect, because the capacitors Dean is using have extremely low esr, and even doubling the ESR is lower than that used in original networks.

Dave

[Guest David H]

Dean, have you had the opportunity to compare the results of the battery bias network to those networks you build with premium components?

If so, what were the results?

Dave

[Guest " "]

I thought you might have misread what I'd said. Sorry.

It's all good.

I was hoping Fritz would share a little insight as the to bridge battery Bias as well.

Dave

I am not aware of anyone else doing the bais using a brdige. I decided to do some test models using the bridge primarily becuase I wanted to eliminate the influences as much as posible of the componets used in the bais process as well as trying to preload both sides of the film with a charge state.

The issue of cumalivtive influences of the compeonets in use stems from measureables such as esr. If you take a standard implementation of a bais, most use two caps with double the capacitance values, and places the bais at the junction. So you get double the ESR amount other things.

Using a bride, you get the series/parallel effect which on one side of the bridge the capacitance ESR does double, but because you have two parallel circuts (both sides of the bridge), the ESR and other properties are halved and brought back to normal.

You still gain double the voltage rating for free

You also gain double the current potential for free

Using the standard approach to biasing...i believe only one side of the film gets charged. That charge is so small that skin effect has to come into play and lead one to question what the state of charge is on the other side of the film. important I think since audio is an AC transmission.

Using a floating soruce of voltage potential allows both sides of the film to be charged...so even if you had some skin effect...the surfaces of both sides of the film would present the AC siganl a preload.

I build a few using both approaches and I like the sound of the bridge approach better.

[Guest David H]

I found a reference to Cycil Bates also had done testing using a series / parallel - bridge capacitor configuration, however I have not been successful finding the articles.

Ref: DiyAudio

starkeyg

diyAudio Member

Join Date: Jan 2002

Location: Sourhern MI

Charging caps?

Hello,

Have been thinking about this and how it MIGHT be making a difference (as far as I can see no accurate blind tests have ever been done by those claiming a difference. And as easy as it would be to extend a switch out that applies and removes the voltage, with no label on it and have people listen and report their favorite switch postion, seems quite easy to do so.

Looking at the articles Cycil Bates did on capacitors - he found that biasing electroylitic caps did improve distortion, but the effect was widely variable and once the optimum voltage was reached actually produces more distortion, not less.

badman

diyAudio Member

Join Date: Sep 2003

Location: Sunny Tustin, SoCal

You presented a statement that Cycil Bates had presented measurements indicating lowered distortion- why ignore that piece of evidence?

Care to comment on the potential improvements offered by a series/parallel connection compared to a single equal value component?

[Guest " "]

I

I can't find a reference regarding the "bridge type" method anywhere. I'm curious where Speakerfritz got the idea. In the thread where he introduced the network and the schematic, Dennis said, "I'm not sure this is the best way to do it. In use there is a very low impedance from X to Y (X to ground of less than 0.1Ω, and Y to ground of less than 1Ω)."

With the bridge type method, notice that the negative side of the battery doesn't connect back to the common connection on the network.

Presuming the bridge type method even works, if you started with the batteries in the networks, and then pulled and discharged them, there's still a decent chance you ended up listening to them charged anyways -- because the longer a capacitor remains charged, the harder it is to keep it drained.

how would this be measured...best would be to use a measurement tool that injects a signal at the frequency of operation....measuring with DC wouldn't make any sense since DC signals do not pass thru a functional capacitor...some measurments done correctly would always be useful. In an bridge bias circut, there s no chance of short since the voltage potential would never be seeking a ground path...it would only be seeking it's inverse polarity which would be on the other side of the film and never at ground.

I don't think the grounding of the battery in a typical bais solution has any bearing on bais functionality.......floating battery bias solutions were very popular during the late 70's as a way to force a class A/B output stage into full class A biasing....there was no need to ground the batteries used .

I'm not sure I understand why I would want to put the battery out and used them with a hopefully expired state of charge....forgive me...this must be something the forum as a whole learned something about and I am just not aware of it.

anyway, different strokes for different folks.....bais anyway that you like....if your of the balanced signal crowd...the bridge approach might appeal to you....if you just want to get a sense of what biasing is all about at an entry cost..you can go the route of the major manufactures who can provide a bias solution at a low costs.

I will some day build a simple amplifier with end to end bridge type battery biasing....certainly would want the batteries and circuts floating....can you imaging the havoc if you tried to share batteri bias grounds in an amplifier circut...the same potenitail for contamination due to capacitor leaks concern me using common grounds...but maybe I'm just reading in to things too much.

[Guest David H]

speakerfritz

Thanks for posting, its nice to have the additional input.

Dave

[Deang]

I also read through JBL's reasoning for using this configuration, though it seems to me that this really only benefits electrolytic or cheaply built caps. I guess only benefits is a poor choice of words. Has a greater benefit on electrolytic or cheaply built loose wound caps.

"...I have never heard a capacitor type that didn't improve (or change) including the nearly perfect teflon variety." - G Timbers

It should produce the same level of improvement regardless of the dielectric type. The operative word being "should" - you very well might be right however.

The battery bridge can be used with a 1k resistor, however it is not required

Capacitors leak DC, some a little, some a lot. The resistor value is tweaked based on the amount of calculated leakage. For a PIO capacitor, the value is 470Kohms. Some even use two resistors, one on each side of the capacitor. We really don't want DC from the battery leaching into the network. The resistor serves three purposes: protects the drivers, extends battery life, and allows for the use of a single battery to charge the entire network.

Dean, have you had the opportunity to compare the results of the battery bias network to those networks you build with premium components?

I wouldn't expect a metallized or film type to sound like a paper in oil or wax impregnated stacked film, regardless of what you do to them. Those capacitors sound really different to me, almost ethereal.

I used a small pair of Jamo monitors which already had Bennic metallized polypropylenes in the crossovers. This made them good candidates. I only modified one, and compared it to the other. It was interesting. I didn't notice a tonal shift, what I noticed related strictly to the soundfield, which became more airy and expansive. It was enough to make me want to pursue it further.

The goal is to provide much of the what the better film types provide at a greatly reduced cost. This "testing" is technically in Beta, a purely subjective analysis. I really just wanted the opinions of those who have nothing invested in this. I'm hardly objective when it comes to any of this stuff, as I have pretty strong opinions. This is why I use people like Michael (Thaddeus) and Mike Stehr (and a couple of others who prefer to remain anonymous). My choices in what I build and sell are to a large degree determined by the validation I receive from those who are willing to listen and provide feedback. That was really the idea behind the roadshow, I just wanted to know what people thought. I suppose it's also true that I wanted to provide an opportunity for people to hear something different, to give them some experience so they would have something tangible to hook into because of all of the crossover discussions. So, to a large degree, this is still about those things. All I really want to know at this point is what does it bring to the table and is it worth doing. I believe it is, but I need the validation.

Using the standard approach to biasing...i believe only one side of the film gets charged. That charge is so small that skin effect has to come into play and lead one to question what the state of charge is on the other side of the film. important I think since audio is an AC transmission.

This doesn't make any sense. You connect the positive side of the battery to the midpoint of all of the capacitors, and then you connect the negative side of the battery back to common. This completes the circuit, and charges all the capacitors in the network. This idea that are there are two sides of the film that need to be charged is erroneous.

In an bridge bias circut, there s no chance of short since the voltage potential would never be seeking a ground path...it would only be seeking it's inverse polarity which would be on the other side of the film and never at ground.

I would like to know more about this hypothetical "short".

I think the supposed benefits derived from the bridged method are easily offset by the increase in materials and labor. If the build begins to approach the cost of a build using high quality film types, then there is no point in doing it - might as well just use film and foils. There is also limited room on the board, and with the chosen method I can easily retrofit any existing crossover. Again, the goal is to approach the subjective performance of a film and foil build for pennies on the dollar. I was able to do Carl's crossovers for less than $200.00, and I have an email from Thaddeus that would cause a riot if I posted it. I also consider Greg Timbers and the engineering staff of JBL to be trustworthy and reliable. In a nutshell, I'm pretty sure they know what they're doing. The web is flooded with positive reporting related to this implementation.

Edited March 27, 2014 by DeanG

[Thaddeus Smith]

and I have an email from Thaddeus that would cause a riot if I posted it.

i believe I've already exceeded my riot inducing comment quota for the year.

regardless of the distortion or the thing-a-ma-jig or the blah-blah-blah.. i still enjoy these battery biased networks and ultimately that's what matters.

what i think is difficult, or maybe the delta we're ultimately trying to solve for here, is how to describe these audible differences on a product page so that customers can efficiently choose between high, medium, and low priced offerings.

[djk]

Lots of foggy reasoning in this thread.

The only reason I mentioned the distortion thing was if someone had read the Cyril Bates articles and didn't realize that the battery bias connection is not the same test method used by CB.

"Looking at the articles Cycil (sic) Bates did on capacitors - he found that biasing electroylitic caps did improve distortion, but the effect was widely variable and once the optimum voltage was reached actually produces more distortion, not less."

That is an incorrect summary, and does not apply to the method used in the crossover networks.

"cheaply built loose wound caps."

Incorrect, its primary benefit is with regards to dielectric absorption and the zero-crossing of the signal. Caps must have good construction to sound their best, too tight is just as bad as too loose, but lead termination is far more important.

"Circuit isolation"

Use a 3.3MΩ resistor then.

"the capacitors are series / parallel therefore do not suffer from increased esr like the series network employed by JBL."

The ESR of the 2x larger caps is lower, so the end result is comparable to 4 caps in series/parallel, but minute differences in ESR are not the point at all.

"So you get double the ESR amount other things."

Incorrect, see above.

"You also gain double the current potential for free"

Incorrect, current potential is the same as a single cap, and even if it was free (it's not as you have to buy and mount 4x capacitors). Current is limited by the ESR and the surface area of the part, it's strictly a function of heat dissipation.

"...i believe only one side of the film gets charged"

Beyond incorrect. How can only one side be charged? It's not possible. Remember, the amplifier output impedance (less than 1Ω) is the return path for the bias.

"which became more airy and expansive"

The bias improves the low level detail, the quite passages are less noisy, making the detail easier to hear.

Using the bridge method you still need a resistor. If the caps are not perfectly balanced, and at all frequencies, current will flow through the battery.

Remember, this is not about distortion or ESR, it's all about dielectric absorption and the zero-crossing of the signal.

Edited March 27, 2014 by djk

[Guest " "]

"I think the supposed benefits derived from the bridged method are easily offset by the increase in materials and labor. If the build begins to approach the cost of a build using high quality film types, then there is no point in doing it - might as well just use film and foils. There is also limited room on the board, and with the chosen method I can easily retrofit any existing crossover. Again, the goal is to approach the subjective performance of a film and foil build for pennies on the dollar. I was able to do Carl's crossovers for less than $200.00, and I have an email from Thaddeus that would cause a riot if I posted it. I also consider Greg Timbers and the engineering staff of JBL to be trustworthy and reliable. In a nutshell, I'm pretty sure they know what they're doing. The web is flooded with positive reporting related to this implementation."

yes...the web is full of many things. Battery bias'ed, or charge coupled has got the interest of a lot of DIY folks. what is interesting....collectively...and even on this forum...folks are removing the battery because once charged...for the current music session...they want to break their ground reference connection that exist as a result of the traditional bais or charge approach. with a bridge implementation...there is no need to remove the battery for the listening session once charge....that's because in bridge implementations there is no ground reference...and obviously....audiophile DIY folks feel the need to remove the ground reference for a reason.

so sure...lets watch JBL and the like to see if they ever recognize the curiosity folks have with breaking their ground reference by removing the battery and see what JBL does so we can follow suit.

I think since folks are breaking their ground reference...that speaks for itself as to the limitations of what I would call a half phase implementation of charge coupling. I did not tell them to break their ground reference and apparently since folks are doing so....theres something there.

Maybe someday, the folks on other forums who are breaking their ground reference by removing their batteris will figure out you can implement a bridge design and not have to fiddle with it anymore.

cost is certainly a factor....my time has cost...and after a few beers..I don't want to fiddle with anything..much less trying to unsnap a 9V battery or two.

so sure....half phase battery baising or charge coupling is good....and as the folks who have found out removing the battery to break their ground reference is better (else they would not do it)....and certainly...if you can spring the cash and have the space for the parts...a bridge charged coupling implementation takes this whole thing to a new level since there is no ground reference.

I did not understand this battery removal thing earlier today so I googled to see what folks were up to. I think they are on the right track...and someday....they will be buzzing about bridge implementations.

congrats to the klipsch forum for being able to assess this technology long before the folks on other fourms can spell it.

on to you has been provided a battery bias or charge coupled implementation that does not have a ground reference.

may the force be with you.

[djk]

[Deang]

I have absolutely no idea what you're talking about.

[Guest David H]

I prefer pepperoni.

[Deang]

Fritz said, "May the force be with you", and it does look a lot like a Star Destroyer. "Impressive, most impressive."

[mike stehr]

I have absolutely no idea what you're talking about.

Neither do I, but I have this sudden urge to break my ground reference.

[djk]

"but I have this sudden urge to break my ground reference"

After all, gravity is just a theory.