I can't stop wondering why you clicked the link. A golden ear should know better what to expect from me.
At least there was one Marshall. I just watched the latest instalment in the Star Trek franchise, they took the aliens down with noise I figure I'm empowered to take down whole galaxies of aliens.
EDIT - Stretching the concept of open mindedness is a good thing. I can't dance either. https://www.youtube.com/watch?v=EGNKgah948s
Last edited:
I love these Russian guys, google translate says they call themselves "My Noise". Reminds me of Joy Division.
https://youtu.be/3VdvJs1WWVU
Destroyer OS, you may click the link, it's not gonna break your precious golden ears.
Edit: link is now correct.
Last edited:
They also made Cherry Bomb a number one hit. I had the pleasure of playing a first press 45 of that for Andy Morris recently. The first time he has heard that from vinyl in decades.
Remember when Shishido first published his SE300b article in Glass Audio? Many round eyes here in the US really did think he was crazy. Then about two years later SE tube amp units flooded the shows.
Last edited:
I've seen more than enough garage-noise music in my life. Some noise music I love. I saw these guys in Portland and loved it - it was in the basement of some weird building I think people lived in, maybe, off of 42nd (maybe? I was riding my bicycle around, and honestly don't really know how I found everywhere I went to).
Guys, can I address another subject? I am in a discussion about the pros and cons of impedance correction for speakers.
The reasoning is that impedance correction makes the speaker look like a resistive and thus frequency-independent impedance. This makes it much easier to design a passive xover when the target load is resistive.
My point however is that because of the impedance correction network, some of the output current from the xover will flow through the impedance corrction network and some through the speaker. So even when from the xover perspective everything is frequency independent, you still have an unknown and frequency dependent division of the signal current, with only part of it contributing to the sound. So you only exchanged one problem for another.
What ye sayed?
Jan
The reasoning is that impedance correction makes the speaker look like a resistive and thus frequency-independent impedance. This makes it much easier to design a passive xover when the target load is resistive.
My point however is that because of the impedance correction network, some of the output current from the xover will flow through the impedance corrction network and some through the speaker. So even when from the xover perspective everything is frequency independent, you still have an unknown and frequency dependent division of the signal current, with only part of it contributing to the sound. So you only exchanged one problem for another.
What ye sayed?
Jan
Weird, I remember only a glowing positive review of the Vendetta in TAS.
Shall I look it up?
It seems like you are making some assumptions that are not clearly stated: Why would division of the signal be unknown? You mean if someone else designed the correction network, so you didn't have a complete model? Also, even if current is divided, why is that necessarily a problem? Wouldn't it depend on speaker efficiency for the current that does flow through it? What about any effects of possible damped resonant current circulating between the correction network and the speaker?
Anyway, it seems like a concrete example might help clarify most or all of the issues.
Absolutely so, Mark.
The key problem, as I remember it, is getting the whole shebang to works as one single entity, including all the parts in the signal path. Often, you get one part to do wonders, but at the cost of another part's performance. You get the impedance just where you want it, but with a say slightly depressed midrange levl, for example. The reason is obvious - you are dealing with several variables all at once with more or less same parts interacting with each other. That "#$%&/() xrossover cost us a btter part of six months' work, but we did it in the end.
B ut when you do get it right, troubles are soon forgottem and all is well in the world.
The key problem, as I remember it, is getting the whole shebang to works as one single entity, including all the parts in the signal path. Often, you get one part to do wonders, but at the cost of another part's performance. You get the impedance just where you want it, but with a say slightly depressed midrange levl, for example. The reason is obvious - you are dealing with several variables all at once with more or less same parts interacting with each other. That "#$%&/() xrossover cost us a btter part of six months' work, but we did it in the end.
B ut when you do get it right, troubles are soon forgottem and all is well in the world.
Jan, I would argue that you do not exchange one problem for another, but rather split the problem in half so that it becomes manageable.
On the speaker side, you adjust the crossover so that it measures straight, and on the amplifier side you adjust it so that the impedance remains constant. The unknown and frequency dependent division of the signal current all happens inside a black box. You don't need to measure it, although you could. The end result is a speaker which can be played from tube or ss gear alike. The down side is component cost and an efficiency penalty.
edit: impedance correction for individual drivers usually makes more sense in order to promote a correct xover alignment. The impedance variation seen in bass drivers on the low end cannot be corrected in any practical way.
Last edited:
Sloppy language from my side.
What I mean is that without impedance correction the design of the passive xover is difficult because the load at the xover, which must be known for the design process, varies strongly with frequency. So that makes the xover design harder.
If you apply impedance correction, the design of the xover becomes much easier because its load is frequency-independent. But now part of the xover output is shunted, frequency dependent, through the impedance correction network and doesn't do any 'work'. I would think that this somehow needs to be taken in account in the xover design so that is what I mean by 'you exchanged one problem for another'.
But maybe I am seeing black where it isn't. Just wondering.
Jan
Effectively, it sounds like what you describe as not doing any work, is being described by Vacuphile as a loss off efficiency. So, would you consider loss of efficiency to be a problem? Maybe it would depend on exactly how much of a loss of efficiency would occur in a given case?
Yes, this subject generates much discussion, much is uninformed from lack of unexperience.
Yes, this is the ideal, excepting deliberate FR tailoring.
Yes, it's called a conjugate load.
Sure, some energy is 'dumped' into the network in parallel with the driver during energising/change and this is what makes the driver/network combination appear resistive to upstream components and amplifier.
During overshoot condition magnetically stored energy is 'dumped' locally providing driver local damping and never makes it back to the crossover or amplifier.
In much typical consumer grade amplifiers this lack of return energy is of useful benefit in terms of eliminating a host of ringing artifacts due to NFB, lousy power supplies and reactive cable\connection+reactive load.
This also causes cooler operation of output stages.
The non obvious trick is to mount the compensation components right at the driver terminals so as to avoid circulating currents between crossover stage and driver.
Four flex wire kelvin connected drivers could be of advantage.
.The RC shunt network also absorbs RFI picked up in driver and cabinet wiring.
You are only losing the reactive (imaginary) currents.
Nope, you just cured some problems.
Dan.
Yes but my point is that this loss of efficiency, or whatever, is frequency dependent because the driver Z is frequency dependent.
Jan
The driver sees a voltage across the terminal and responds to it the same as always. The amplifier sees a load and delivers a voltage to it as long as the current is available. The correction networks see a voltage and change it to heat.
The only difference the driver sees is when it is making back emf is that in effect the connection cable loss to a source impedance is changed.
Now if you have designed the crossover network to allow for the connecting cable's frequency dependant losses then you will have a difference and a loudspeaker that changes behavior depending on loudspeaker cable and amplifier characteristics. Of course that would lead to....
Jan I'm getting into speakers myself.
The woofers impedance spikes is always low due to the FS (resonate frequency) of the driver. Usually it's so low that the crossover doesn't interact much with your dampener. If you increase the value of the dampener say on the woofer, then it tends to suck some of the higher frequencies out (as in above resonate frequency). The correction may be as simple as a slightly smaller first inductor on the woofer's section.
For tweeters the resonate frequency usually isn't even in the playing spectrum, it's down below the crossover point. This means that you can dampen it even more because you want it to roll off better. The is actually normal practice for a lot of tweeters that otherwise won't roll off as fast as they should.
My question would be, if the amplifier doesn't have feedback does it then get some variation in the output of the speaker, if it requires a hair more power for all of your wizardry inside the crossover?
The woofers impedance spikes is always low due to the FS (resonate frequency) of the driver. Usually it's so low that the crossover doesn't interact much with your dampener. If you increase the value of the dampener say on the woofer, then it tends to suck some of the higher frequencies out (as in above resonate frequency). The correction may be as simple as a slightly smaller first inductor on the woofer's section.
For tweeters the resonate frequency usually isn't even in the playing spectrum, it's down below the crossover point. This means that you can dampen it even more because you want it to roll off better. The is actually normal practice for a lot of tweeters that otherwise won't roll off as fast as they should.
My question would be, if the amplifier doesn't have feedback does it then get some variation in the output of the speaker, if it requires a hair more power for all of your wizardry inside the crossover?
Last edited:
- Status
- Not open for further replies.