3way XO help greatly appreciated!

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Grant, You seem to be bangin on all cylinders. Like you , I was not aware of the idea that ferrite cores were different than iron core inductors. If memory serves, I thought I read an article discussing this. As you may be able to relate, that when you read TONS of articles pulled from the internet, after a while, it's difficult to remember sources of information, and exactly who wrote what, when, etc.............the info can start to become garbled in ones mind, unless we print EVERY document we ever read, and create an organized file system.........And I certainly don't do that.. But it's also weird that particular details seem to stick in ones mind, available for recall, at any point in time, when needed..........here's an example.............I thought I read somewhere that poly switches had a 5 second delay, which was reported to be too slow for effective tweeter protection. I am not certain here and will have to re-research it. Anybody got any input on this?.........At any rate, currently I am dealing with problems in my baffle finishing which is causing delays, between paint coat drying times and repainting sessions, and surface problems, moving to the next step of finalizing the cabinets and mounting the drivers for more impedance measurements, is currently on the back burner, not to mention severe cost overruns......... I guess I will use this time to study those links Sreten provided in his latest posts............I hear quite often that there is a reason for everything?................Regards........Omni
 
Hi,

Polyswitch time to trip depends on its thermal time constant and the
amount of overdrive beyond the notional constant tripping current.

Similarly whether a drivers coil (adhesive) melts depends on pretty
much the same parameters, thermal time constant and rms rating.

I cannot see why you cannot build one to protect a tweeter.

I've seen 5s as the maximum time to trip, which is meaningless.

:)/sreten.
 
Hi,

Yup, the D75 doesn't give you that much bandwidth used properly.
The off axis is a phase dip that should not be taken literally.

Simple question - how much it it attenuated in the current c/o ?
If its not by around 6dB it never going to sound that good.
Also means probably no BSC on the bass driver.

but if they are damaged.......


:)/sreten.
 
Hi Sreten,
The D75 c/o has a 3.3 Ohm resistor 'before' the network, ie closest to amp. I think that attenuation should be placed between the network and driver (or Zobeled driver, which these are not).
I haven't worked it out but it would maybe be around -3dB?
Thanks for confirming its limited bandwidth. Yes, I'm sure they are damaged and nearly 20 years old anyway.

I've been playing with Basta, a terrific concept. I need to reread the User Guide. Might also try Speaker Workshop. thanks , grant
 
Hi again,
I've been playing in Basta, it really is very good, thanks Sreten.
I've managed to get a plot (attached) for the woofer and CA15 mid in their respective boxes.
Took a while to get the parameters right though!
Does the unlocked version give a summed response? I ticked both driver boxes. The baffle step is at the bottom I think. It would add in nicely, maybe to the plotted curves.
Also not sure why the CA15 goes out to 5Khz? (Lalena ~290/~2900 LR2). Also does Basta include the 2dB band pass gain?

My tweeters are actually D25TG-55-06! (I unscrewed one) but not sure how to specify in Basta as I don't have Rms, Cms,Qes,Qts (or Mms&BL). Not in Tymphany's archive. But I do know: Fs=1500, Sd=7.1, Re=4.6 and that Mms may be about 0.27?, and BL
possibly about 3.3? Do these guesses for Mms and BL seem reasonable?

I tried SpeakerWorkshop (no manual 'coz no Excel) but I couldn't get inductors to 'place-vertically'!

Omni did you get a chance to use it? If so, am wondering if you got a good result?

best regards, and thanks, grant
 
Basta plot:
 

Attachments

  • basta_p25_ca15.jpg
    basta_p25_ca15.jpg
    96 KB · Views: 206
Grant, I have not done anything with Basta yet, however I read the introduction on the link. I could only see that it deals with high pass and lowpass filters, couldn't see an application with it dealing with the band pass. Maybe I wasn't thorough enough in my initial research? I downloaded it once, a few months ago, and couldn't make heads or tails of it. Again,, maybe I wasn't thorough enough. I have checked out the other link Sreten posted and that simulation shows me a bit more promise. I will be reviewing that program again tonight, will print out the introduction and use instructions so I can read and digest the concepts. I will keep you posted when I get a solid foundation of its' use. At any rate....What parameters are you plugging into the program for the CA15RLY? You mentioned that you were not sure why the CA plot goes out to 5000. I am wondering if the plot you produced with that driver is in the high pass section of the program...............Keep me posted................Respectfully..........Omni
 
Hi,

I'd say with something like Basta and the FRD tools :

first sim something like a subwoofer that you can cross check that
the thing is outputting reasonable numbers, then try adding a low
pass passive filter to it at various frequencies, soon learn why "cookbook"
c/o values can be meaningless when used in incorrect circumstances.

Then move on to a two way (again ideally play with something you
can cross check) so now you have BSC and diffraction ripples and
the mid/ treble c/o to deal with.

The above is obviously the top half of a 3-way. To add the bottom
half of a 3-way the main issue is BSC and how it is achieved.
Things get complicated depending on driver sensitivities and where
the target c/o frequency is. BSC needs to be included and this usually
means the actual c/o is way off "ccokbook" c/o values.

Also note in real life c/o's do not have to be, and seldom are
symmetrical. If they are this should be acoustic symmetry as
apparent electrical symmetry is pointless. Asymmetrical c/o's
can be extremely useful in certain circumstances.

My trial of Basta has expired, so I can't check what works / doesn't.
A bandpass should be OK, its a low pass + a high pass filter.

:)/sreten.
 
Hi Omni,
Basta will do 2&3ways 1st to 4th orders. The locked version doesn't give summed response. The trial however, still does seem worthwhile. Paremeters areinput in two groups of 5, ie: group 1) Fs Vas Sd Mms and Cms; group 2) Re Qes Qts BL and Rms - But you only input 3 of each group, Basta calculates the others.
Tick checkboxes to specify parameter, but in practice this is tricky. It graphs immediately, just toggle to the response tab.

CA15 data used was Zaph's, see attachment. I double checked the HP/LP sections, a circuit diagram appears in the lower 'passive filter' Tab. Sreten has some good ideas re a logical method to verify/get acqainted with it.

BTW, the user guide warns not to use the 'table-lookup' Tabs for xo frequencies to determine the component values - I tried it and it gives weird results. best regards, grant
 

Attachments

  • ca15driver.jpg
    ca15driver.jpg
    69.4 KB · Views: 178
Grant, I printed and read the introduction sheets to FRD consortium tools late last nite. Seems to be do-able. I imagine Basta has similar functions to FRD ? At any rate, it seems the approach I will take is to experiment with both, and compare results....From my initial study of the FRD consortium tools, the parameters are obtained, in part with a few of its features, along with some of my information from manufacturer or possibly Zaphs measurements, then plugged into a crossover simulator, something like Speaker Workshop to derive the crossover components...........Is this your experience with Basta? Also, I noticed that you used Zaphs measurements for the CA15RLY. Do you think his measurements are more accurate than the manufacturer published specs? Do you plan to model with Zaphs measurements exclusively or will you also be experimenting with manufacturer parameters ?...........Sreten, it appears you have extensive experience here............question for you: Are the manufacturer published parameters reasonably accurate? Or would it be advisable to model utilizing say, for example, with Zaphs measurements? Zaph has published his measurements on 2 of the drivers I am using in my system. Or try both? Or am I simply complicating everything.............Your input here will be appreciated.........At any rate, I look forward to comparing notes.....Sreten, your links have added a new dimension to this pursuit...............Respectfully........Omni
 
Hi,

Seas are generally pretty accurate. Note that some of the curves
shown are measured in boxes and show BSC, others in a baffle.

Re Zaph's measurements : you can't argue with them, if they are
similar to spec's shows consistency, but if they are very different
you have to go along with what Zaph measured, e.g. I'd take
Zaph's measurements for a silver flute driver over specs anyday.

Always cross check as much as possible, e.g. the data here :

http://www.rjbaudio.com/Calypso/calypso.html

:)/sreten.
 
Alright SRETEN............Rock and Roll!!!!!!!!!!!!! I logged onto the link you posted in your post..........Serendipity, man...............Cool. Thank you very much for that link and the time you are spending with us here..... I downloaded the crossover simulation for the Calypso, and 2 of the drivers which I am using in my 3 way are those in the Calypso................Grant, check it out. So I will need to read the entire Calypso project and check out the downloaded file. If you can bear with me, I will try to digest it and learn as much as I can before coming back with more educated questions, but I will throw a few at you now, thoughts that entered my mind as I checked out the site: For my crossover simulation in SW for a three way will I be using Nearfield, or Farfield measurements, or both, ungaited, etc? Since the Calypso is a 2 way, and I'm doing 3 ways, are there any different things I need to be aware of? So, as of now, information is starting to come together for the Midbass, and Tweeter, Next in line will be my Peerless SLS 830669 woofer. Grant............you are working with a Vifa woofer aren't you?........ Also, if memory serves, you mentioned about switching gears into a 12dB per octave network at 300 and 3000 Hz? Is that still your plan?...................Gotta go to work, Keep in touch..............Respectfully.........Omni
 
Grant, Its late here in the US............Seems I post when you are asleep and you post when I am asleep, Sreten seems to be somewhere in between.....timewise, that is...........lol. At any rate I got a few more questions for Sreten, and Grant , I hope I am Not stealing your thread, since we are both on similar paths, it's my hope we can both learn here. If I am infringing, please let me know.........Sreten, I downloaded SPL Trace. Are the frequency and impedance graphs that are to be imported into SPL Trace to come from manufacturers website or Zaphs website? I can't open other parts of Trace for some reason; it's a 3 part download and part 2 and 3 will not open because Windows doesn't know what program created it, and I am not sure if the other parts of the program have a driver database or not. Also I am having problems with Speaker Workshop to get to a crossover designer, however I saw in the FRD Consortium a Program written by Jeff Bagby which uses these parameters etc. to design crossovers......What do you think ? Is Bagbys program viable. He has 2 programs: One which is described as textbook crossover designer, and the other one which uses an Excel spreadsheet format which uses what seems to be the Frequency Response Data input...................Respectfully............Omni
 
We would like to thank the Madisound site for their wonderful
audio discussion board, where most of these programs were
first discussed, and eventually beta tested.

Hi,

from FRD consortium, so the place to go for questions.


But from setup.txt (read this !)

Full Spl Tool Setup Procedure
------------------------------------
1) Create a folder on your system for the Spl Tool installation
files (for example: C:\SplTools\Setup).
2) Download all three installation files: SplPart1.exe, SplPart2.cab
and SplPart3.cab. Download the files into the folder created in step 1.
The size of each file is approximately 700kb.
3) In "Windows Explorer", or "My Computer", navigate to the setup
folder created in step 1, and double-click the first installation file,
SplPart1.exe. The program will start, and will extract the individual
setup files.
4) Run the Setup.exe program extracted in step 3, and follow the
dialog instructions.


so download all 3 without opening them.

:)/sreten.
 
Hi,

OK some pointers :

in both your cases you have bass units with higher sensitivity than the mids.

You therefore are essentially building a two-way with no BSC and
then turning it into a 3-way with BSC - this gives the right balance.

(If the mid was say only 3dB lower than the the bass unit you would
build a 2-way with some (but not all) BSC, e.g. 2dB +3dB from bass
unit = total 5dB BSC.)

An example speaks a thousand words see :

http://users.nac.net/markowitzgd/david/david.htm

An externally hosted image should be here but it was not working when we last tested it.


Here the bass units as a pair are 96dB/2.83V, the midrange 91dB/2.83V.

Generally speaking midband sensitivity is 4dB to 6dB lower than the bass unit(s).

The Calypso has all its BSC built in so you cannot use the c/o,
unless you add the bass unit as a powered subwoofer.
(Dennis Murphy routinely uses full BSC in his designs and does
not discuss the issue, so you can bet BSC is in the CAOW1)

:)/sreten.
 
Hi! Warmest regards to you both and thanks very much for staying in this thread and offering your continued and extremely helpful support! I feel very priveleged to be in such esteemed company!

Sreten, Re: WinISDPro - auto function, a big thank you!

Re: Yes, Seas specs 'from what I've read' are fairly accurate (some manufacturers are much > +- 20%?), It seems Zaph is very accurate, so I also would tend to believe him more too, but the differences for CA15's Fs, Vas, Qes and Qts are in some cases also much greater than 20%. This could be because of different box/baffle testing as you point out.

I have D/L'd SPLTrace also, with a view to creating *.frd and *.zma files for my ancient woofer and tweeters to use in SpeakerWorkshop. (I finally got past the 'code conversion' problem for Jay Buttermans? SW tutorial and saved it as html.) Learning SW, for me, will be more difficult than when I learnt Photoshop! SW will probably? be more useful than Basta because it takes driver specs into account and is incredibly powerful. Both have dis/advantages. Basta is easy by comparison.

SPLTrace and SPLView open OK here on my XP PC, I just haven't used them yet. (And thanks for the info. of course!) I went into Thiele/Small website, but alas no archives for my twt&woofs. I might email Tymphany.

Re: BSC and David Markowitz link, many thanks, I'll read tomorrow. I presume, as a newbie, that BPG would also need to be considered? Many thanks Sreten, as usual 'great stuff'! grant
 
Hi Omni!
Unfortunately most of FRD tools use Excel, I think?, which I don't have yet. I have to be careful of 'cost-over-runs' too! But I might buy Excel. 'SW' seems the best, most comprehensive, free 'modelling
choice'. Shame its such a steep learning-curve though.

So far, I've only delved superficially into Basta, AFAIK it doesn't include driver response, but I could be wrong. I know 'SW' does as it allows the import of *.frd (F.R.) and *.zma (IMP.) files.
Re: Zaph's testing, IMHO he's very thorough, but I am a beginner. Sreten would have a much more informed opinion on that than I would.

At the moment I'll be considering driver parameters from wherever(if) I can get them! Lol. Its fairly rough-and-ready over here. I might only be able to get 'ball-park' stuff. But there is 'method in my madness' hehe. If my project ultimately doesn't work, for whatever reason (and there's lots of those!); and as you know I plan on eventually getting the 27TDFC's, then, hey presto, Calypso's! In new boxes on top of P25's (in slightly baffle-modified boxes working as subs!) So, I haven't lost much! The only things that originally slightly 'put me off' the Calypso's were the 1400Hz dip and the homemade port(s) which superficially to me, didn't look good! (Even if 'my original project' doesn't work, I'll learn a lot, have fun and hopefully spend the least dollars!).

Re: (quote) "Sreten, it appears you have extensive experience" - he certainly does! I guess one thing that encouraged me is that he has no test facilities and yet, has still gained amazing insight and
expertise.

Yes Omni my woof is a Vifa P25 10" poly (better unscrew it to be sure after my error with tweeters!).
I must admit I've been vacillating between XO topologies, but I'm now seriously leaning to LR2 for each driver including bandpass at 300/3000. I thought that I had understood things from the 'Cookbook' but am still pondering Tinitus's enlightening reply. It raised even more questions re mixed order slopes etc. I might put an involved fundamentals type question here, but only if you have ~20mins to spare. So, at the moment a complete LR2 seems for me, ultimately easier , better documented and more likely to provide an 'acceptable' result. I'll just have to buy a good meter and unwind coils! The tweeter I think is rated at 100W (2ndO. XO) so 3000 should be OK (resonance 1500).

Re: (quote) "Grant , I hope I am Not stealing your thread" ...Of course not! No problem at all. I learn things! and so glad to read your thoughts, Omni. Besides we both benefit.

Re: SPLTools install, I think Jay B. in the SW tutorial mentioned an XP install problem, but I followed the install text message as Sreten kindly mentioned (I have XP SP2) and all OK. I must look at the
J. Bagby XO designer too, but maybe not the spreadsheet one just yet. cheers, grant
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.