Series crossover project?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Does anyone have any input on series topologies for fairly simple two-ways? Links to projects with clear explanations of why series vs. parallel crossovers were implemented would be great... I've read the AR page, but don't really understand why none of the crossover component values are driver-specific.
 
You want a better reference book, like Vance Dickason's Loudspeaker Cookbook.

The xover effects everything - including the impedance looking at the speaker from the amp, the freq response, the response in the xover region, etc etc...

Series xovers are lower parts count.
Parallel xovers are higher parts count (usually) and easier to implement different "splits" in frequency and filter shape (type), almost no inter-driver dependency on the electrical part of the results.

That's a simple answer...

_-_-bear
 
Does anyone have any input on series topologies for fairly simple two-ways? Links to projects with clear explanations of why series vs. parallel crossovers were implemented would be great... I've read the AR page, but don't really understand why none of the crossover component values are driver-specific.

Hello,

On the basis of the impedance curves provided by a friend I used to design for him a series crossover for a 2 ways system using a B&C DE25 compression driver on a RCF H100 horn and a Beyma SM112/N boomer in a 50 liters bass-reflex enclosure (Fc = 50Hz)


The schematic is here :
http://forums.melaudia.net/attachment.php?aid=2276

The simulations are here :
http://forums.melaudia.net/attachment.php?aid=2275

It is very important that for a better operation of the filter to equalize the impedance curves of the 2 loudspeakers inside an interval of frequency of +/- 1 octave around the common cut-off frequency.

Best regards from Paris, France

Jean-Michel Le Cléac'h
 
I was a long time parts kit customer of Bud Fried . When he switched from parallel to series circuit on later C2 and O2 he used same exact values. In a 3 way he deleted Zoebel on sub crossover. He used 6db values close to textbook. I switched to series on my Model C1 and couldn't believe the difference. Series circuits are more efficent then parallel. Current is not wasted like parallel circuits. The current that the inductor prevents from the woofer seeing goes back to the tweeter. The same thing happens with the capacitor in parallel of the tweeter circuit. You can also use a RC circuit to shelve tweeters low end response. You need high quality drivers or you will hear intermodulation distortion. Ashley,Kaminsky Small and others AES papers cover this. I only use series now. The resistance in the circuit being balanced is what makes the crossover sum flat to the tweeter and woofer on 2 ways. Way to much hype on this subject. Many people have argued with me over this. I have the plans to prove it.
 
but don't really understand why none of the crossover component values are driver-specific.

As far as I can remember, it is a fast track to design a filter, without simulation and measurement. For proper filter, of course each components are driver specific (tweeter Fs and also tweeter sensitivity that will collide with system total impedance).

The nature of the series filter is that the drivers are automatically "matched" (not necessarily perfect). This makes the result will (if the drivers are smooth enough) never sound bad. Just like a tube amp. No tube amp sounds bad enough. But with a proper parallel filter, I always prefer a parallel.

I see pro and contra is related with whether you can build a good parallel filter or not. I always try both and the parallel always win.
 
Try a few of these links as may help.

Tony Gee's series information and also has some projects.
Humble Homemade Hifi

Andy G's series page.
series cross overs

Some examples from Andy G.
Andy G's home page

Troels has a lot of projects using series and parallel and some explain why he used series / parallel xo.
DIY Loudspeaker Projects Troels Gravesen

The diyAudio 2-way ref speaker uses a series xo.... classic series as well as AR series.
diyaudio-reference-speaker-project-summary

The Esquire is a similar speaker to the above but with a parallel xo which you can compare the design.
Esquire by Scott C. Blaier

The AR site information is only a starting point guide as each speaker has to be individually designed the same as other crossovers even though final real world values usually fall within their range in the guide. Small changes can make a huge difference.

The most important thing with series is driver choice and usually requires well behaved drivers with smooth acoustic slopes and no peaks. For example from my experience:
1. Vifa P13-WH-00-08 works well with classic 1st order series.
2. Scan Speak 18W8531G00 works well with an AR series but not a 1st order classic series.
3. SEAS Nextel W15LY001 needed a parallel crossover and wouldn't work with either series.

In other words use the right tool for the right job but a well implemented series design can be special and preferred by some builders (me included).
 
With flat impedance of drivers such as the one obtained through application of zobel, first order series is absolutely the same as first order parallel.
Same phase shifts and same output.
Even... to be exactly correct, with simple loads (flat impedance) all orders series and parallel are equal.

But, with real drivers, first order series has the advantage of self adjusting.
Second order and higher series does not have this advantage.

First order series crossovers are a very good way to add a super tweeter to a multi way speaker. Or for connecting two separate crossovers, for instance sub bass and midbass part that have their own crossover and a multi way top consisting of 2 or more midranges and tweeters that have their own crossovers.

Recently I tried a two way (and still experimenting on it) consisting of a 12 inch woofer and a small fullrange with first order crossover at ~1250-1500 Hz.

Without a Zobel on the woofer the series variant sounded better, with Zobel it was the paralel crossover that was better ;)
Around crossover point the bass has ~14 ohms and the full range has ~10 ohms inmpedance. I used 0.82 to 1 mh inductor and 10-12 mf capacitor.
Zobel was applied only to the woofer - 10 ohm and 8.2 mf.

In first place first order series is the network that allows self adjustment to impedance variations, but according to my humble experience (which is really humble!) the sound can be really jammed around crossover if values are not very exactly correct!
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.