|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
60.241.77.110
Howdy there! I'm fooling around with some fairly modern commercial speakers which (AFAICT) have some of the drivers connected in series. I've never seen this before, or even heard of it. But why not? Assuming you can weasel your way to calling it an 8 ohm speaker without laughing too hard, what are the drawbacks here?
I think the Bottlehead Straight 8 used a series parallel configuration, at least at some version of it...? But I can't turn up any discussion about this subject outside of guitar amps, car audio and so on. Would appreciate a link or resource if you can recommend anything.
My speakers are four driver, three-ways and the two mids are series wired (or seem to be!)
Thanks in advance for any thoughts about this :)
Follow Ups:
The Horned Heil is an upgraded single driver speaker, utilizing a Heil AMT up top. I haven't upgraded mine, but here's some information on it, including a link to the source.
Hope it helps.
"The two sections have a complementary shift - when woofer impedance changes, it effects both low and high pass sections, and changes the Q of the filter sections. The result is quite obvious - unlike a parallel crossover, the response remains flat regardless of a shift in the woofer (or tweeter) impedance. If both change in any direction, the same thing happens. In theory, this means that the series network is almost immune from impedance variations in the drivers.
One warning about the first order series network.....it seems that some digital amps see the inductor as a short at low freq. and can cause the amp to spit out enough DC to cook a woofer. I've only seen this happen with digital amps of various types."
Another series crossovers speakers:
There's nothing inherently wrong with series connections. That's how I wired my wife's sound system in her studio with a mono converter and two speakers adding up to a 16 ohm load which I figured would be better for her receiver than 4 ohms. And yes, the Bottlehead Straight 8s are wired in a series/parallel method with 2 speakers in series and those 2 wired in parallel to the next 2 to maintain an 8 ohm load, but it's been a while since I built a pair so I don't recall the exact methodology.
-Rod
There is nothing inherently wrong with series connection of drivers, and if the net result meets the designer's goals for system integration and impedance and sensitivity, then in that case it's the right thing to do.Myths unfortunately are perpetuated about series connection degrading the electro-magnetic damping (electrical Q, or Qes). If we were talking about wiring a resistor in series with a woofer then yes the damping would be degraded (electrical Q increased), but we're not - we're talking about wiring one fully powered voice coil in series with another. The additional magnetic field strength increases the electrical damping by the same amount that the additional series resistance reduces it, so damping (electrical Q) stays the same.
I use series connection in some of my models, including my best ones. I have measured carefully and neither series nor parallel connection change the electro-magnetic damping (Qes).
See my reply to BigguyinATL downthread a bit for more debunking of myths.
Duke
Me being a dealer makes you leery?? It gets worse... I'm a manufacturer too.
Edits: 10/01/14 10/01/14 10/01/14
Maybe it's because I work in the electrical industry - I keep a simpler view of it.
Coils in parallel.
Voltage is the same across coils. Current doubles when you add 2nd coil and impedance goes down by 0.5 x Z. Differences in coil impedance result in different current magnitudes and phase angles. Voltage across parallel branches has to be the same.
Coils in series.
Current is the same through both coils. Voltage halves across each coil because the same voltage is being applied across 2 x Z. Differences in coil impedance result in different voltage magnitudes and phase angles. Currents within series branches has to be the same.
This is why I don't buy into the idea that series connections have inherent disadvantages over parallel. Differences in coils and how nature "balances" this out results in changes in voltage OR current depending in connection type. The point is, different coil impedances, mechanical properties - aka any TS parameter differences - are going to show up whether you have parallel or series connections.
Closely matched drivers are always better.
Cheers,
Presto
The last attempt I remember to popularize series crossovers was Kimber's DiAural crossover. A patent was granted to Eric Alexander for it. Ray Kimber had a prototype pair of speakers that he was demonstrating widely about 15 years ago. They licensed it to Aperion Audio, who sold lots of speakers with DiAural before switching back to conventional crossovers sometime in the mid-2000s. DiAural was also in the Roman Audio Centurion. Others may have picked it up as well, but I'm not sure.
As I understand it, the main difference between DiAural and a typical series crossover, and the claimed advantage, is that the DiAural is capacitor-less. It had some disadvantages, such as needing a relatively high power resistor in parallel with the LF driver which hurt efficiency, and not being any easy way to equalize the drivers, e.g. to notch out peaks or baffle step compensation.
A series connection also doubles the inductance that both speakers see, so - if used for woofers - or better mid-woofer - the low pass filter may be unnecessary. Also note that placing the woofers in series pretty much distroys the idea of damping for either woofer and the resonance peak in impedance will be seen by both woofers.
"The hardest thing of all is to find a black cat in a dark room, especially if there is no cat" - Confucius
"A series connection also doubles the inductance that both speakers see, so - if used for woofers - or better mid-woofer - the low pass filter may be unnecessary."Wrong, because while the series inductance is doubled, so too is the series resistance, so the drivers' inherent inductive rolloff characteristic is unchanged. The same goes for parallel connection - the frequency response stays the same.
"Also note that placing the woofers in series pretty much distroys the idea of damping for either woofer..."
Wrong, and in fact the effective system damping of the speaker/amp combination increases.
While a woofer's Qes (electrical damping) goes UP (damping is reduced) as series resistance is increased, it also goes DOWN (damping is increased) as the BL (magnetic field strength) is increased. With series connetion Re is doubled, but so is BL, so Qes STAYS THE SAME.
As a thought experiment, think of a single woofer... the first half of the voice coil is in series with the second half. Does the one destroy the damping factor of the other? No, because BOTH halves come with their share of magnetic field strength. This is electrically and magnetically analogous to two identical woofers wired in series.
Now let's bring in a high output impedance (low damping factor) amplifier. Let's say it's a specialty tube amp with no global negative feedback, and its output impedance is 4 ohms. Into an 8 ohm load, that would give the amp an effective damping factor (input impedance divided by output impedance) of 8 ohms/4 ohms = 2. But into a 16 ohm load. the effective damping factor is 16/4 = 4! So by going with series wiring, we have doubled the effective damping factor of the amplifier/speaker combination.
(This doubling of effective system damping factor would hold true regardless of the type of amplifier used or its output impedance, but the impact on the system's frequency response is greatest with high output impedance amplifiers, so that's why I chose such for my example.}
"...and the resonance peak in impedance will be seen by both woofers."
True but irrelevant. Recall our thought experiment where the first half of the woofer's voice coil is in series with the second half. The whole impedance curve is doubled in value, so the doubling of value at system resonance is of no greater or lesser consequence than what's happening at any other part of the impedance curve.
* * * *
I normally wouldn't be so in your face when someone is mistaken, but your profile says you're an acoustic engineer who works for Bruel and Kjaer. Imo your position as an engineer with one of the most respected companies in its field holds you to a higher standard than the average audiophile.
Duke
Me being a dealer makes you leery?? It gets worse... I'm a manufacturer too.
Edits: 09/30/14 09/30/14 09/30/14 09/30/14 09/30/14 09/30/14 09/30/14 09/30/14 09/30/14 09/30/14 09/30/14 09/30/14 09/30/14 09/30/14 09/30/14 09/30/14 09/30/14 09/30/14 09/30/14 09/30/14
Re. Amplifier Damping factor related to "driver control" by a low output impedance, not Qts.
And you are right about the inherent LP: filter being unaffected by the series connection - I must been thinking too fast on this one.
"The hardest thing of all is to find a black cat in a dark room, especially if there is no cat" - Confucius
"Re. Amplifier Damping factor related to "driver control" by a low output impedance, not Qts."I think we're almost on the same page here, but Qts IS "driver control". And amplifier output impedance modifies Qts by modifying Qes, which one of its two constituents (along with Qms).
The amplifier's output impedance modifies Qes just as any series resistance would. The formula is ([Re + AmpOutputImpedance + AnyOtherSeriesResistance]/Re) x Qes = effective Qes. Sorry I don't know some of the proper names for these terms!
Anyway for a given amplifier output impedance, higher values of Re (driver DC resistance) result in less change in the effective Qes. So 16 ohms speakers are less sensitive to variations in amplifier damping factor than 8 ohm speakers are. Also we can see that low amplifier output impedances have very little impact on effective Qes, in which case the series resistance of the inductor(s) in the lowpass filter usually becomes the dominant modifier of Qes.
Duke
Me being a dealer makes you leery?? It gets worse... I'm a manufacturer too.
Edits: 10/06/14
Drive series-connected woofers from an OPT with grounded center tap.
The voltage at the speaker mid-point will also be zero. The mid-point could be shorted to the CT, making the speakers electrically independent, with no change in response.
True, but no matter how you slice it, it's better to have electrically identical (matched) drivers whether using series or parallel connections. Not just left to right, but within the same box.
Scan Speak, for example, is known for drivers with *very* good consistency from drive unit to drive unit.
If a guy is using off-the-shelf drivers as an OEM in one of his designs, this is very important.
No matter what you do with connection, differences between drivers (if large enough) will show up. Series, parallel, whatever.
You can't get the acoustic output to match with different drivers unless you use driver specific impedance or output equalization.
Cheers,
Presto
As I see it the back EMF's of matched woofers in series will cancel and the damping & (free-air) resonant peak will be the same as for a single woofer.
I think you'll find that fairly common when large numbers of identical drivers are used. The Bose 901 uses 1 ohm speakers in series to provide a nominal 8 ohm impedance. I think Roger Russell uses a series-parallel approach with his IDS-25 speakers which use twenty-five identical drivers.
For all the time/bandwidth/energy that has been expended slagging Bose in general, and the 901 in specific, that is probably a little known fact.
You won't find too many other cone & coil drivers that are 1 Ohm.
Doesn't it seem like an outstanding challenge to rewire a pair of 901s in parallel & hook them up to those few amplifiers that are advertised as being stable into a "dead short??"
If I had the time, I'd sure be inclined to try it, just for grins! :)
The Bose car stuff often used the 4", 1 ohm driver from a 901 with an amp built in to a plastic box that housed both and was of a proper volume for the driver to perform "well". It also had an EQ circuit based on the environment and placement. I've often called them one of the least offensive pieces of car audio. We repaired them in the 80's and 90's. The amps would go bad, usually caps like 90% of the time, not transistors. The drivers almost never went bad.
ET
Edits: 09/30/14
Good description!
I rather liked some Bose car systems. The old "No highs, no lows, must be Bose," is pretty much ideal for the high background noise environment of an automobile & leads to pretty fair sounding mids. :)
As to the amps? Still see tons of posts on NSX boards about Bose amp failures.
Well, sure, they're probably more than forty years old!
I did car audio for many years and we had good stuff, tube amps and lots of good speakers as most are trash. We had a calibrated analyzer too. Implementing a (mid)woofer/tweeter in a car is not easy, well its easy, just making it sound good and have any image isn't. The typical 6" mounted low on the door panel and tweeter on the A pillar usually sounds horrible. So the Bose with a full range driver EQ'd at the amp worked pretty good.
I found just adding a sub and amp to those Bose systems which then allowed you to turn down the bass from flat at 12 o'clock to say about -10dB at 8 o'clock(AM) gave much better sound from the Bose setup by not asking it to produce all the bass. That job went to the sub that did it much better. It was that way in my 92 Maxima. I just added a single sealed 12" in a very small box and A $200 Kenwood mono bass amp with speaker level inputs (needed since the stock radio had no pre outs)and it did quite well.
ET
Well... a couple things come to mind.
If there are two mids in an MTM design, for example, and they are super sensitive, then you surely could put two 4 ohm mids in series for an 8 ohm load. Not all drivers are 8 ohm despite that being the most common.
8 ohm is still the vast majority, with a lot of 4 ohm designs being two 8 ohm drivers in parallel. 16 ohm is not that common but it's around. 12 ohm is rare but exists as well, as 12 in parallel is 6 ohm.
It's not a problem to have drivers in series.
So long as the acoustic output of the speaker is good (meets design criterion) then series connections are valid.
Now... define "good"! ;)
Cheers,
Presto
Please check the Speaker Art Clef as its drivers are connected in series. There are some other interesting features of this speaker as well.
. . . such as that used in the linked speakers, is a whole different situation from the OP's speakers, which apparently have two midrange drivers wired in series and connected to the bandpass section of a more conventional parallel crossover.
The series crossover, in its simplest form, has the tweeter and midbass driver wired in series with each other, and the reactive components in parallel with the individual drivers: the inductor shunts across the tweeter, and the capacitor across the midbass. This yields a first-order electrical transfer function, similar to the more common situation of a cap in series with the tweeter, a coil in series with the woofer, and both legs of the circuit in parallel. Steeper orders may be achieved by adding appropriate additional components in series and parallel with this arrangement. The linked article provides a good comparison.
Series crossovers have been around for a very long time -- Bud Fried among others favored them, and Jeff Bagby used one in his Fried-inspired "Tributes" design -- but because of the increased interaction between highpass and lowpass components, and the way everything affects everything else, they are much more difficult to get right. Hence, perhaps, the small but enthusiastic cult following among some speaker designers.
The x.7 line of Magnepans use a series crossover as does the old SMGa. Rod Elliott has a detailed discussion of the pro's and con's of series and parallel crossovers.
Edits: 10/01/14
Thanks - that's very interesting! Would be great to know more about their approach.
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: