Sins of a scan speak user..

[Beau]

I know use of scanspeak is strictly forbidden on this foum, but... i am in the middle of constructing a 3 way using ss25w8565-01 for bass, Possibly Hiquophon ow1 for tweeter and either 15w8530k01 or18w8531g00 for mid duties. I alreadyhave both mids andwould really like to avoid spending any more money. Any pointers?? (ps anyone want to design me axover?? )

[ThomasW]

Any pointers??

Yep make measurements of the drivers mounted in the baffle where you intend to use them. Otherwise there's no way for you or anyone else to accurately design a crossover

[Beau]

Thanks,I would but I lack the proper gear to do so, in any case coulsd you say whether you would use one driver over the other?

[cjd]

Given the cost of the SS drivers, a mic and pre-amp will not significantly impact the total project cost IMHO - especially if you pick up one of the "diy" kits. I think ESP may be in your neck of the woods and able to help at a good cost in this regard (I *think* he still has the mic/pre kits). All the rest you need is a sound card on your computer (though not all work) and Speaker Workshop (free).

ThomasW: don't say that too loud. I'm not sure you require accurate measurements on your final baffle, but you certainly need accurate measurements taken in a very controlled environment. Certainly, measurements as they'll end up is the safe way to go and preferred.

[ThomasW]

Chris,

Mfgr's specs are usually anechoic, or on an industry standard baffle. That's a far cry from the baffle most people end up using. In addition mfgr's specs often have a ton of smoothing.

All this means that their specs aren't that reliable for a XO designer to use.

If one can't make their own measurements then the only viable option is to have a company like Madisound design the XO for a fee. In the long run that's cheaper and more accurate, than buying test equipment and software

[JonMarsh]

I know use of scanspeak is strictly forbidden on this foum, but... i am in the middle of constructing a 3 way using ss25w8565-01 for bass, Possibly Hiquophon ow1 for tweeter and either 15w8530k01 or18w8531g00 for mid duties. I alreadyhave both mids andwould really like to avoid spending any more money. Any pointers?? (ps anyone want to design me axover?? )

With no measurement data of the drivers in your built or planned cabinets, there is a way to proceed, but it's a bit time consuming. You've painted yourself into a corner on this, so you wont' really have much choice.

You can solve your design problems using free tools from the FRD consortium (google is your friend), and possibly Speaker Workshop also (free, too).

First, post a few forums and check the FRD site to see if anyone has FRD and ZMA files for the drivers you're using. These are measurements normalized to an IEC baffle context (i.e., no cabinet boundary or diffraction effects- the "raw" driver response.). If they don't already exist, you can use the FRD tool for tracing the response graphic and doing a Hilbert transform on the amplitude data to get the phase data. A little time and patience, and a copy of a recent version of Excel are all that are needed. (forget Excel 97).

Then you'll need to simulate your planned cabinet design using Baffle diffraction simulator, and use the resulting profiles to process the FRD files to develop a "simulation" of of your drivers mounted in the cabinet.

Then, you can use this data to develop your crossover, keeping in mind whatever strategy you had in selecting these speakers for the range you want them to operate in and how you'll deal with baffle step compensation. The crossover could be developed in Speaker Workshop using the optimizer if you find that helps, or using the passive crossover program from the FRD consortium- the latter, IIRC, doesn't have an optimizer and has limited flexibilty with regards to network configuration (it's not a true simulation, more a ladder filter calc).

BTW, I wouldn't typically recommned a three way system for a first project- if this is your first, try doing a couple of two way project simulations first to get warmed up.




18W8531G00 - note impedance curve glitch at 800 Hz- either a surround resonance or first cone mode.




15W8535K01 Better dispersion to 2 kHz, but has characteristic SS bend in the response at 1 kHz- optimized for BSC on midwoofer with single pole transfer function.


We have nothing "religious" against ScanSpeak drivers here- I use some of their tweeters from time to time- and in the 90's did some designs using SS midwoofers. Why, I just threw out some new/old stock SS midwoofers back in February, some 8" kevlar modesl like used in the Audio Artistry and Phoenix designs.

The basic motor distortion is quite low in the SS midwoofers, but some of their design decisions with regards to cones and surrounds make them less desirable to work with, IMO. YMMV.

Take a look at CJD's work on three ways with the Dayton RS drivers; may give you some useful ideas. He also just published a nice design with the RS150, dual RS225, and 27TDFC. Maybe you could adapt that to your project.

Good luck, and have fund with your project.

~Jon

[TacoD]

I have in my own system a SS 15W8530K01, it is a nice driver. In a closed box of 12 liters, it hits 60Hz -3 dB. There is something annoying to this unit, the unit has some cone mode / resonance in such way you do not want to listen anymore to the loudspeaker after some hours (tried a lot of x-overs, Cauer, Series, Parallel 1st, 2nd, 3th, 4th order). The 18W is the better choice, not as resolving as the 15W, but more listenable and enjoyable. But I doubt if you can cross it to a OW1, I would recommend a bigger dome, 9900 or 9800 from Scanspeak. I have crossed the 15W at 1500Hz, 1800Hz, 2000 Hz, 3000 Hz, 3500 Hz, 4000 Hz, try to notch the shelving and stay around 2kHz.

[cjd]

Mfgr's specs are usually anechoic, or on an industry standard baffle. That's a far cry from the baffle most people end up using. In addition mfgr's specs often have a ton of smoothing.

I agree, manufacturer specs are usually not worth a whole lot. Though I've had quite good luck with PE's published data and they're technically the manufacturer of the Dayton line. But finding measurements taken in controlled conditions (that is, either IEC or IB) where the raw data is available combined with the FRD Consortium tools can yield results that surprised even me with their accuracy.

Jon (and whomever): The project I just did was RS225/RS150/27TDFC (not RS180).

C

[JonMarsh]

I meant RS150, was thinking RS150, but look what my rebellious fingers typed out...
I'll fix that.

[Beau]

Thanks to all who replied, I have a company who will design me a xover, my main concern was matching the mid sensitivity. Would the 18w sens be a waste when matched to the 25w?
Beau.