Duffinator
Spin The Black Circle
Skip the Fatty and check these out: 2452H SL's on "not so Economy" PT-F95HF waveguides. Zilch has done a lot of work with the EWG and BMS and that's what got me going down this path. See the thread here.
Zilch's AK Design Collaborative - Econowave Speaker
- Thread starter Zilch
- Start date
Russellc
Addicted Member
Skip the Fatty and check these out: 2452H SL's on "not so Economy" PT-F95HF waveguides. Zilch has done a lot of work with the EWG and BMS and that's what got me going down this path. See the thread here.
That's the one! It is supposed to be the end all deal here.
I wish I could find a pair of those compression drivers cheap:thumbsdn:Russellc
Russell called it. I damped it with spray-on mineral-filled elastomeric yada yada, and it thumped duller. I found measureable differences, but could not hear any. It's also different when mounted on a baffle.
A vibraton expert suggested the way to make it truly inert is to back it up with a box filled with pool-filter silica -- sand. That thread's over in LHF. It does not appear that any of these measures are necessary or warranted at this level; just fire 'em up and enjoy for now. Get anal later, if you like.... :thmbsp:
http://audiokarma.org/forums/showthread.php?p=1886228#post1886228
Kegger
R.I.P. 1/12/1966 - 6/1/2017
Are there any online or free downloadable software that will let you build/test xovers
to where you can put in your own passive components?
One reason I ask is being able to use parts you have an compare what one set of say
cap an coil combos does (2nd order) compaired to another set where you may use the
same frequency you want but for example say your building a low pass section an need
a 1mh coil and a 15uf cap to get you 2k but you have a 1.2mh coil an would like to see
if by using a smaller cap you have on hand can you get the same frequency cutoff.
Or doing known compensation networks an tweaking there values to see how it responds.
Most of the calculators I find you just plug in the resistance and what freq you want and
it spits out what passive components they say to use, I want to do the opposite an pick
components then maybe have it spit out a response or at least see a change from previous.
to where you can put in your own passive components?
One reason I ask is being able to use parts you have an compare what one set of say
cap an coil combos does (2nd order) compaired to another set where you may use the
same frequency you want but for example say your building a low pass section an need
a 1mh coil and a 15uf cap to get you 2k but you have a 1.2mh coil an would like to see
if by using a smaller cap you have on hand can you get the same frequency cutoff.
Or doing known compensation networks an tweaking there values to see how it responds.
Most of the calculators I find you just plug in the resistance and what freq you want and
it spits out what passive components they say to use, I want to do the opposite an pick
components then maybe have it spit out a response or at least see a change from previous.
I believe what it boils down to is the absence of horn coloration. There are other factors as well, but, bottom line, you get compression driver clarity and dynamics with extended HF without horny sound. Horn theory has advanced considerably in recent years, and you're hearing the payoff in these waveguides. It's just silly how good they are.
Russell has called this one correctly, also. The 1" BMS drivers, 4550 or 4552nd will mount on the waveguides you have using a throat adapter. The primary difference is in smoothness, their diaphragm being Mylar, as opposed to titanium. An adaption of E'Wave's Pi-Speakers filter design could work well with that combination. 2226H is a good choice for the woofer, but you'll need a sub to provide extended bass. If you want that in one 15" package, use 2235H. Gordon earlier suggested that 2234H is also a logical choice, since 1+ kHz is a "stretch" for an extended-bass driver of this size.
Duff shows the premium version above. This $100 edition of the PT Waveguide has a 1.5" throat and is otherwise identical in size to E'Wave, but made of a different, thicker material. The BMS 4555 is a direct bolt-on with it. The JBL 2452H-SL he's using is a large-format (4" diameter) TOTL driver with a damped titanium diaphragm. We're still doing development with these, and Duff's presently enjoying the first-generation attempt. For a while, these drivers were available used, apparently from tour partouts, and Duff scored a pair. The secret's out now, however, and they've become scarce at bargain prices. Available new for $1429 MSRP. Yeah, that's apiece, alas....
ity:
dnewma04
The Healer
Kegger, speakerworkshop does that, but you'll need to import frequency response and impedance data. They can be imported from online sources or created if you have response graphs.
here are a number of useful tools.
http://www.pvconsultants.com/audio/frdgroup.htm
here are a number of useful tools.
http://www.pvconsultants.com/audio/frdgroup.htm
Yes, but you're moving into "Expert" territory here. Read in the Indignia thread how Dave, Ray, Gordon, and I did it for that project, and what tools we used.
Hey Kegger,
If you have an electronics simulator like PSPICE or Micro-cap then you can do it there. You can go simplistic and use a resistor as the load or you can poke around the 'net and find a fairly reasonable passive imitation of a speaker driver. I believe Ken Kantor did something like that once. You put the XO together in the simulator and change whatever you like. I do this in Micro-cap 8 (free version from years ago) and it will "sweep" through values. It doesn't give you acoustic output but it gives you an idea of what the changes do in the electronic domain. Thought that might be something you have on hand being the builder you are.
Ray
Kegger
R.I.P. 1/12/1966 - 6/1/2017
Funny you mention that Ray, as I just fired up my spice to see what I could do there.
I've used spice for many things, an while there is a learning curve once you get it down
it is quite powerful, I've built/tested many circuits with it and find it to be quite accurate.
At the moment I'm not trying to build an actual complete xover but I want to sim known
circuits with there predicted outcomes then change components to see how they react.
I've got some things in mind and I want to test an see what they do.
-----------
Thanks for the other suggestions guy's, and I'm wading through them, just looking for a
simple program that will let me change values.
I've used spice for many things, an while there is a learning curve once you get it down
it is quite powerful, I've built/tested many circuits with it and find it to be quite accurate.
At the moment I'm not trying to build an actual complete xover but I want to sim known
circuits with there predicted outcomes then change components to see how they react.
I've got some things in mind and I want to test an see what they do.
-----------
Thanks for the other suggestions guy's, and I'm wading through them, just looking for a
simple program that will let me change values.
Wayne Parham
Active Member
Groovy Spice Stuff
Here are a bunch of posts with Spice stuff. Some have links to download Spice, others have links to models, others have good information about using Spice.
If you want to really get fancy, use the ICD in WTPro. It lets you import your Spice models, but instead of simulating load impedance (Le, Re and tank circuits to simulate diaphragm and horn resonances), you can define the load using an impedance measurement ZMA file.
Here are a bunch of posts with Spice stuff. Some have links to download Spice, others have links to models, others have good information about using Spice.
If you want to really get fancy, use the ICD in WTPro. It lets you import your Spice models, but instead of simulating load impedance (Le, Re and tank circuits to simulate diaphragm and horn resonances), you can define the load using an impedance measurement ZMA file.
Wayne Parham
Active Member
A few useful documents
Here are a few useful documents to help you get started using Spice to design loudspeaker crossovers:
The documents at the following link go beyond the basic crossover filter design, into the acoustic realm where phase angles and baffle spacing affect summing at various angles. This is what sets the positions of lobes and nulls. The idea is to get a useful wide forward lobe with a large angle between nulls.
Here are a few useful documents to help you get started using Spice to design loudspeaker crossovers:
The documents at the following link go beyond the basic crossover filter design, into the acoustic realm where phase angles and baffle spacing affect summing at various angles. This is what sets the positions of lobes and nulls. The idea is to get a useful wide forward lobe with a large angle between nulls.
Use Bode plot of the scanned audio spectrum, the voltage drive. The voltmeter and frequency counter aren't telling you what you want to know. Measurement point is your resistive load....While it's not simple but I do get along with spice decently, looks like it has promise. (long way to go)
Kegger
R.I.P. 1/12/1966 - 6/1/2017
Yah I'm still fidling around with what I can do with this spice an what measurement tools
plus sources I do have available, an yes my measurement point is to the top side of my
8.2 ohm load resistor in the pic.
I've used spice quite a bit, mostly on designing tube audio gear.
What I'm trying to find in this spice at the moment is a pink noise gen and an analyzer.
plus sources I do have available, an yes my measurement point is to the top side of my
8.2 ohm load resistor in the pic.
I've used spice quite a bit, mostly on designing tube audio gear.
What I'm trying to find in this spice at the moment is a pink noise gen and an analyzer.
Attachments
It's just an AC signal generator and measuring the voltage. Set the vertical scale for dB magnitude and the horizontal for log frequency.
You can see I merely duplicated the circuit and used two test points to compare the performance with a different value C4.
I could adjust L2 now and match the frequency with different damping at the crossover point. With more tweaking, I could get the slopes to superimpose, if that were the objective.
Then, I put the revised values in the prototype, measure the new performance in comparison to the original, and see if works worth a whit accomplishing what I intended.
Is there a problem here? Yeah, two major ones:
1) The driver's not a resistor; you've got to use an accurate model of the specific one you've got for the transfer function to really be correct, and,
2) It's the acoustic response that counts. That's what you're controlling with the filter voltage drive, but it's very likely not the same. More sophisticated tools than SPICE are required to sim that, but measurement works, too.
I call this "The Hard Way," but have created many successful filter designs using it....
You can see I merely duplicated the circuit and used two test points to compare the performance with a different value C4.
I could adjust L2 now and match the frequency with different damping at the crossover point. With more tweaking, I could get the slopes to superimpose, if that were the objective.
Then, I put the revised values in the prototype, measure the new performance in comparison to the original, and see if works worth a whit accomplishing what I intended.
Is there a problem here? Yeah, two major ones:
1) The driver's not a resistor; you've got to use an accurate model of the specific one you've got for the transfer function to really be correct, and,
2) It's the acoustic response that counts. That's what you're controlling with the filter voltage drive, but it's very likely not the same. More sophisticated tools than SPICE are required to sim that, but measurement works, too.
I call this "The Hard Way," but have created many successful filter designs using it....
Kegger
R.I.P. 1/12/1966 - 6/1/2017
Understand what your saying completely Zilch, just like when I build my tube equipment.
If I don't have a good model then the sims are for nothing, but building the circuits then
changing things helps one to learn how changing parts makes the circuit "react", and if
I build a curcuit then tweak a tube model to show as I measure in the real world for the
given tube then I've built a new working tube model.
Again this is just an exercise to try different parts an see how they react, once I get a
feel for what I'd want to try an do and it seems like it's working the next step is to do
some building then actually test the changes/circuits out, take notes an go at it again.
But it's much easier and quicker to try a new idea or tweak in the sim first to see if it's
got potential before you go buy parts and build it let alone trying to use what you have
on hand that may do the job just as good as the "exact" part values.
(Once you learn how to build up the sim)
There's probably better software and or a better way to go about this then what I've been
trying today but spice is pretty powerful and I find it very informative, I may find something
a bit different in the end that is specific to speaker xover building that works better. Spice
is a kinda build whatever you want and we'll test it type of deal that can be very handy.
I've tried to follow your AC analysis and bode plot setup but it's just not working here.
Something I'm doing for sure, time to step away from the computer for a while.
If I don't have a good model then the sims are for nothing, but building the circuits then
changing things helps one to learn how changing parts makes the circuit "react", and if
I build a curcuit then tweak a tube model to show as I measure in the real world for the
given tube then I've built a new working tube model.
Again this is just an exercise to try different parts an see how they react, once I get a
feel for what I'd want to try an do and it seems like it's working the next step is to do
some building then actually test the changes/circuits out, take notes an go at it again.
But it's much easier and quicker to try a new idea or tweak in the sim first to see if it's
got potential before you go buy parts and build it let alone trying to use what you have
on hand that may do the job just as good as the "exact" part values.
(Once you learn how to build up the sim)
There's probably better software and or a better way to go about this then what I've been
trying today but spice is pretty powerful and I find it very informative, I may find something
a bit different in the end that is specific to speaker xover building that works better. Spice
is a kinda build whatever you want and we'll test it type of deal that can be very handy.
I've tried to follow your AC analysis and bode plot setup but it's just not working here.
Something I'm doing for sure, time to step away from the computer for a while.
A side note on "sim" design...
Keg handed off a design to me, I built a prototype and handed it back to him. Once all the bugs were ironed out, it was found that the model's approximation of some of the tubes we had used were a little off, but in the ball park so they didn't blow up. He tweaked the circuit until it performed the best by changing some values of some passive components. The sim allowed us to get close to what we wanted, and the experiment let us get the "perfect" values. The real world is the best sim, by far, but the virtual world can help to not make your first experiment a costly one!
Keg handed off a design to me, I built a prototype and handed it back to him. Once all the bugs were ironed out, it was found that the model's approximation of some of the tubes we had used were a little off, but in the ball park so they didn't blow up. He tweaked the circuit until it performed the best by changing some values of some passive components. The sim allowed us to get close to what we wanted, and the experiment let us get the "perfect" values. The real world is the best sim, by far, but the virtual world can help to not make your first experiment a costly one!