To put you guys in the picture a little more, I gave this treatment to a pair of Yamaha NS-20 boxes 10 years ago.
I RC compensated the woofer and tweeter, and except for resonance humps, the box was flat impedence to out past 45 kHz.
These boxes use the same ear-bleeder tweeter as the NS-10, but with this treatment went smoothly and nicely out past hearing range.
I wired the woofer direct, and the tweeter via 1.1111 uF.
This combination cranked big time, and resonated the block of six units that I lived in, and gave THE BEST bar none transient, depth imaging and 3D that I have ever heard, before or since.
I did also try wiring the two drivers directly in parallel, using perfectly non inductive against each other flat, wide conductors fed at the woofer terminals, and this was even better, except for the tweeter resonance causing failure of the fine wires a couple of times, so I abandoned that tact and went back to the series tweeter capacitor.
If I had resonance compensated (RLC network) both drivers, this should have been a good long term success.
This concept relies on the relative admittences of the drivers to do the power sharing, and I found it worked.
Either arrangement was perfectly revealing of AP, so every second track required polarity reversal - this is a bit of a PITA, but the results are fully worth it for full musical enjoyment.
Eric.
I RC compensated the woofer and tweeter, and except for resonance humps, the box was flat impedence to out past 45 kHz.
These boxes use the same ear-bleeder tweeter as the NS-10, but with this treatment went smoothly and nicely out past hearing range.
I wired the woofer direct, and the tweeter via 1.1111 uF.
This combination cranked big time, and resonated the block of six units that I lived in, and gave THE BEST bar none transient, depth imaging and 3D that I have ever heard, before or since.
I did also try wiring the two drivers directly in parallel, using perfectly non inductive against each other flat, wide conductors fed at the woofer terminals, and this was even better, except for the tweeter resonance causing failure of the fine wires a couple of times, so I abandoned that tact and went back to the series tweeter capacitor.
If I had resonance compensated (RLC network) both drivers, this should have been a good long term success.
This concept relies on the relative admittences of the drivers to do the power sharing, and I found it worked.
Either arrangement was perfectly revealing of AP, so every second track required polarity reversal - this is a bit of a PITA, but the results are fully worth it for full musical enjoyment.
Eric.
Hi,
Eric,
Having noticed that you posted this AP thing before using the same frasing, I hope this every other track AP change is just a figure of speech?
I mean, this AP changes from record to record and is particularly present on records having been recorded in various locations/studios where it can depend on potluck if it is reversed or not, right?
From what I can tell from my own record collection this is more of a pop/rock etc. recording prob.
I mean, I never noticed it on any of my classical music recording, or jazz or blues for that matter.
Maybe I instinctively didn't buy the ones I didn't like because the AP was off?
Now there's a thought,
Eric,
Having noticed that you posted this AP thing before using the same frasing, I hope this every other track AP change is just a figure of speech?
I mean, this AP changes from record to record and is particularly present on records having been recorded in various locations/studios where it can depend on potluck if it is reversed or not, right?
From what I can tell from my own record collection this is more of a pop/rock etc. recording prob.
I mean, I never noticed it on any of my classical music recording, or jazz or blues for that matter.
Maybe I instinctively didn't buy the ones I didn't like because the AP was off?
Now there's a thought,
fdegrove said:Hi,
Good...that's where these should be.
Ever tried an outboard x-over?
Some speakers respond well to that kind of mod, others need some x-over tweaking but all in all you gain in clarity since the components aren't rattled by the drivers anymore.
Cheers,
Hi.
Thanks.
Yes.
Yeah, longer connecting wires are rather less of an issue than vibrating crossover components, and yes the change is audible.
Much easier to tweak values too.
PS - I wired one of my boxes with non-directional wire the other day before adding the compensating networks, and that made a very nice improvement too.
Eric.
Positive About Polarity...
We are leapfrogging each other.
Yes, IME compilation cd's are the worst in that just about every second track is inverted.
I also have plenty of rock recordings where most of the tracks are in one polarity, and the top 40 track is inverted wrt the rest.
Jazz, blues and classical are usually recorded in the one polarity for the whole album, but that can still be in either polarity.
For all the reasons above, I consider a polarity selector stage is mandatory in any playback system.
I also recognise that most speakers are hopeless at reproducing consistent polarity over the whole audio band, and this is a large reason for many listeners to say that they cannot hear it - IOW, most listeners have never heard speakers that are phase coherent over the audio band, and these don't play any recording properly.
Eric.
We are leapfrogging each other.
Yes, IME compilation cd's are the worst in that just about every second track is inverted.
I also have plenty of rock recordings where most of the tracks are in one polarity, and the top 40 track is inverted wrt the rest.
Jazz, blues and classical are usually recorded in the one polarity for the whole album, but that can still be in either polarity.
For all the reasons above, I consider a polarity selector stage is mandatory in any playback system.
I also recognise that most speakers are hopeless at reproducing consistent polarity over the whole audio band, and this is a large reason for many listeners to say that they cannot hear it - IOW, most listeners have never heard speakers that are phase coherent over the audio band, and these don't play any recording properly.
Eric.
Hi,
LOL. Sure looks like it.
As for the rest: yes, yes and yes, I agree.
Fair enough although more and more designers are paying attention to this, the best are often the simple FR speakers.
Even when some present some deviation from the truth or are really FR, when put in the right box or used in OB they often offer superlative imaging, something you can never achieve when phase problems are present.
Compensation networks can work miracles here but shouldn't be considered as a cure all.
Cheers,
LOL. Sure looks like it.
As for the rest: yes, yes and yes, I agree.
Fair enough although more and more designers are paying attention to this, the best are often the simple FR speakers.
Even when some present some deviation from the truth or are really FR, when put in the right box or used in OB they often offer superlative imaging, something you can never achieve when phase problems are present.
Compensation networks can work miracles here but shouldn't be considered as a cure all.
Cheers,
fdegrove said:Fair enough although more and more designers are paying attention to this, the best are often the simple FR speakers.
Even when some present some deviation from the truth or are really FR, when put in the right box or used in OB they often offer superlative imaging, something you can never achieve when phase problems are present.
Compensation networks can work miracles here but shouldn't be considered as a cure all.
Cheers,
Yes I agree about single FR driver giving the best depth imaging.
The point of the Yamaha experiment was to run the woofer FR, and bleed the tweeter in to add the missing high highs, and this worker a treat and did not spoil depth imaging at all.
I agree that compensating networks won't polish a turd, but they can make the smell less bad.
Other benefits of course with a NFB amp is lower overall system distortion, useable power goes up and the amp (SS) runs cooler, and drivers don't get cooked.
Those Yamahas were rated 40Wrms, 80W music and I belted the hell out of them with 150Wrms/side and no breakges with the tweeter series cap fitted.
Eric.
Bricolo said:
Yes, it's just basic AC theory. Let's look at a simple example, a woofer with an inductance-like rising impedance. It's fed, in my simple example, by an inductor. As frequency increases, the voltage across the woofer will start rolling off, but that rolloff will stop (i.e., the frequency response will level off) when the speaker's inductance starts becoming significant with respect to the inductor feeding it.
Now, put an RC series network across the woofer to flatten out its impedance. The frequency response of the voltage across the driver will now roll off as before, but will continue rolling off forever. The frequency response of the woofer will now droop more at the high end compared to the non-RC-network version.
Bricolo said:
Decent design software (I use Calsod, which is old and clunky, but works extremely well) will ALWAYS take into account the driver's complex impedance curve. Of course, you have to know that curve, which is why I'm forever saying that to pull off a successful speaker design, you MUST have means of measurement.
Cheap And Easy.
The RC network, otherwise known as a snubber network, functions to quench the energy generated by the VC when in overshoot or resonating condition, and this helps it to track the drive waveform better.
This less erratic behaviour translates to an improved smoothness in the transduced sound, and this is the sonic benefit.
IME this can also allow full range operation, or lower order low pass crossover slope.
This will not cure all drivers, but is well worthwhile the experiment imo.
Eric.
The RC network, otherwise known as a snubber network, functions to quench the energy generated by the VC when in overshoot or resonating condition, and this helps it to track the drive waveform better.
This less erratic behaviour translates to an improved smoothness in the transduced sound, and this is the sonic benefit.
IME this can also allow full range operation, or lower order low pass crossover slope.
This will not cure all drivers, but is well worthwhile the experiment imo.
Eric.
I see it as somewhat related to putting a power factor correction cap across a lagging power factor load like an induction motor or a welding trasformer. The cap supplies the reactive VA's instead of the power source and so the load appears resistive instead of inductive. Not entirely the same with the speaker situation but sort of similar...SY said:
Hi Graham,
Yes the compensation network makes the total load appear as resistive to the amp.
This has benefits more particularly for a NFB type amp because the propagation time through the amp stage is not infintessimly short, and this sets up dynamic ringing between amp and load.
A resistive load means no electrically stored energy to fling back at the amp, and cause dynamic correction that can never be perfect.
Eric.
Yes the compensation network makes the total load appear as resistive to the amp.
This has benefits more particularly for a NFB type amp because the propagation time through the amp stage is not infintessimly short, and this sets up dynamic ringing between amp and load.
A resistive load means no electrically stored energy to fling back at the amp, and cause dynamic correction that can never be perfect.
Eric.
Bricolo said:
It's not a matter of linearity in this case, it's a matter of following the Buttorworth rolloff. Without the Z compensation, it won't. Now, that is something which must be considered in tandem with the driver's acoustic response to determine whether it's desirable or undesirable; drivers DO have non-flat frequency responses.
In crossover design, the terminating impedance of the driver and its interaction with the crossover components can be a very useful thing in shaping frequency response. Trying to make things act like they do in textbooks is a futile project.
SNUBBER...
Hi,
Eric,
Under certain fault condition the impedance correcting network CAN in fact work as a snubber.
It is often used in OTL type amps ( valves or MosFETS) where it serves a dual purpose: impedance correction and snubber.
Although I am pretty sure it originally wasn't seen as a potential hazard snubber though....so I feel you do have a point there.
Cheers,
Hi,
Eric,
Under certain fault condition the impedance correcting network CAN in fact work as a snubber.
It is often used in OTL type amps ( valves or MosFETS) where it serves a dual purpose: impedance correction and snubber.
Although I am pretty sure it originally wasn't seen as a potential hazard snubber though....so I feel you do have a point there.
Cheers,
Bricolo,
Zobels are placed across the terminals of the driver to make the driver appear as a resistive component instead of a reactive one.
If there were no crossover, the zobel would do nothing except modify the impedance. Since there is a crossover, the change in impedance vs. the unZobeled driver will cause the resulting response to be different with an otherwise identical crossover.
It is quite possible to design a speaker with spreadsheets and formulas found on the internet or in cookbooks. Many people do this, but it is not the best way. When using the simple formulas, you must select drivers with a wide region of overlap in response and select a crossover point somewhere near the geometric center point of that overlap (at least for 1st order crossovers) and you must make sure to modify the impedance of the drivers so that it appears flat for a couple octaves either side of the crossover point. This method still leaves out the horizontal offset between the acoustic centers, which can cause a tilt in the polar parttern depending on crossover design. All these factors mean that it is easier to design speakers with software.
If you want to build speakers that measure flat and sound neutral, without a lot of time spent tweking, you must measure and simulate. The good news is you can get measurement and design software for free - the program is called Speaker Workshop.
Zobels are placed across the terminals of the driver to make the driver appear as a resistive component instead of a reactive one.
If there were no crossover, the zobel would do nothing except modify the impedance. Since there is a crossover, the change in impedance vs. the unZobeled driver will cause the resulting response to be different with an otherwise identical crossover.
It is quite possible to design a speaker with spreadsheets and formulas found on the internet or in cookbooks. Many people do this, but it is not the best way. When using the simple formulas, you must select drivers with a wide region of overlap in response and select a crossover point somewhere near the geometric center point of that overlap (at least for 1st order crossovers) and you must make sure to modify the impedance of the drivers so that it appears flat for a couple octaves either side of the crossover point. This method still leaves out the horizontal offset between the acoustic centers, which can cause a tilt in the polar parttern depending on crossover design. All these factors mean that it is easier to design speakers with software.
If you want to build speakers that measure flat and sound neutral, without a lot of time spent tweking, you must measure and simulate. The good news is you can get measurement and design software for free - the program is called Speaker Workshop.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.