Tweet
Okay all you who got drivers from Pmazz, let's discuss what to build with these things. I am getting two, and I want to build either an MTM center or two MT bookshelves. Any other ideas are welcome. Since I haven't found any speaker designs using this driver, except a couple with multiple drivers (Two in each tower, etc), we need to start from scratch.
-
-
I like the MT bookshelf idea. Both because my current ones are weak & because I believe these are unshielded & my CC has to be on my tv. On the PE specs sheet, it looks like the xover would be best to get us off the woofer by ~1600 hz or so. They suggest a .76 ft^3 enclosure has an F3 of 35 hz. Zaph Audio has some tests on this driver, look at the 6.5 & 7" driver comparison, there are a few graphs that will come in handy (thanks, Mr. Krutke). Is there any other testing that should be done on the speakers before proceeding? I'd be willing to send mine out to someone with equipment. -
Nathan,
If you want to go with a bookshelf version of this driver then you are looking at probably going sealed. If you do a sealed 12L enclosure (roughly .42ft^3) you get a Qtc of .707 with no stuffing and an F3 of 59Hz. If you want lower extension than that then you might as well build a tower or a really big monitor because the smallest enclosure I'd do ported is .75ft^3 tuned to 35Hz. Even then, bass would be a bit boomy in room unless you overstuff it a bit to bring the response down around the tuning frequency.Comment
-
That .42 ft^3 does plot well. Vented would mean a bigger cabinet, but after my 1st post, I got to thinking..maybe little mini-towers would be cool. Something similar to the Northcreek Okara II TM minis. I'm not sure how well they could fit on a bookshelf, that would obviously depend on final dimensions, but it's something that could fit into what I will need soon..not sure if anyone else is interested in that route, though.
Edit: After a little playing around with my mini-tower idea, 8"w x 24"h x 11"d is .8ft^3, if it can be taken down to .7ft^3 after bracing & such, a tuning of 37 hz gives pretty flat response to 50 hz & an f3 of 36 hz. Eh? ehhh? Anyone else want to make cute little TM towers with me?
Comment
-
Another thought....
Being one to think out of the box a lot of the times. I just recently did a TMM 2.5 way with woofers wired in series since it was a 4ohm woofer. Came out darn nice and with the woofers being in series you don't need that large inductor which is very expensive. A cap and resistor did the trick on mine. I've not posted the final yet as I still have two more crossovers to try and then tweek the final one I like best. Why would one put them in series instead of parallel? After all in parallel you get a 6db boost in efficiency. Well, first these were 4ohms so if I had decided to go this route I would have been battling the impedance. In series, you don't get the 6db boost but you do get more power handling (theory of 2x) and lower distortion.
I've never seen a 2.5 way series TMM before I tried it and I'm amazed at the sound quality. This will give you a mini-tower to boot :-) and it looks cool to me. Way cooler than an MTM IMHO. RomanB came up with a shunt circuit on the lower woofer to lower the response which should eliminate any lobing issues. Overall, I'm impressed by this unique design. Once I get it done, I will post the results. Cheers, Paul.Comment
-
I gave him my address, and i think im one of the people who get them...I did however ask if he had a paypal account to pay for shipping but he never replied? Guess that was a dumb question lol...
My goal is to make a pair of bookshelf speakers with these drivers. I thought about a DIY center, but I couldn't imagine them blending well with my NHT SB2's. If its possible id like to build something that will outperform my Infinity primus 150s.Comment
-
These are 8 ohm drivers, do doing a TMM wouldn't be neccesary to get the correct impedancel, it's a cool idea, but I was planning on using each of these in an MT so that I wouldn't have to buy as many drivers, as buying the tweeters is about all I can afford. I'm also not opposed to doing some smaller towers, let the majority rule, do we want small towers or bookshelves?Comment
-
Nathan, bookshelfs for me. you?Comment
-
I'm for bookshelves myself, but hey, I wouldn't mind a set of smallish towers, so I won't be brokenhearted either way, but for now I say bookshelves.Comment
-
Id like these to fit on my computer desk. so towers just isnt an option for me. If you do go the towers route, i bet they'll make a good driver for this application.Comment
-
I'll help however I can & follow a bookshelves thread...but if mine go the mini-tower route...don't be angry! I have some bookshelves already, they work well for me (though I admit they are not of any great quality, my 1st speaker building foray). Regradless, I may build bookshelves from these after seeing this thread progress. Or maybe my towers, either/or, I'm excited to read over the coming weeks / months!
Comment
-
By bookshelf speaker you mean that you intend to place the speakers either on a shelf or near a wall?
I did a quick and dirty crossover design for the DA175-8 and the Seas 27TDFC that has just 3 dB of baffle step compensation. It still needs a little tweaking, but its a 4th order acoustic XO at 1600. It is based on PE's CLIO data and tweeter data I found on the web. Baffle step assumes a 10" wide baffle. I'll post it tomorrow sometime.
Nathan, the Seas tweeters are around $30, so hopefully they won't break the bank.
If you want to do MTM's the 27TDFC models well in the PE kit crossover that was mentioned in the other thread. That's probably not the best choice for a center channel because of its relatively high crossover. The catch with that XO is the woofer breakup is not as well suppressed as a 4th order electrical filter. For just one more cap in the XO a 4th order filter drops the breakup nastiness 15 dB.Comment
-
That's not too bad, but it's a little more than I was hoping for, but I'm cool with it. How about we have a parallel design thread or two design threads with common drivers etc, just different enclosures. I would think that the xover would be the same, but I'm not sure on that.Comment
-
So I'm wondering, when designing an enclosure, you design it to the correct size etc for the woofer right? I thought I read somewhere that enclosure size had less effect on the highs than on the lows.Comment
-
Bob, I'm new to most of this, but the info I see lists the 27tdfc @ 90 db efficiency & the DA175 @ 86. My question is your 3 db of compensation. Is that 1 db difference inaudible, or was the 3 db only for the baffle, since you assumed they will be placed very near a shelf or wall? That may be standard for that type of placement, I don't know..Comment
-
That is correct. The width of the baffle usually affects only the frequencies in the lower crossover region. Also, the internal volume & port tuning is calculated on the woofer's specs; a tweeter isn't really affected by box volume (someone correct me if I'm wrong somewhere).Originally posted by Nathan PComment
-
okay, so in that case, it shouldn't be too hard to get the best volume and size box for both a bookshelf and a tower speaker. We could do a ported tower, and a sealed bookshelf or something like that.Comment
-
The problem lies in the width of the front baffle and how far the speaker is intended to be placed from a rear boundary; that's where variables in baffle step compensation come in, and thusly would not have identical crossovers. It may be possible to design a sealed bookshelf speaker with the same baffle width as a ported mini-tower, then "guesstimate" at tweeter attenuation in relation to room placement...someone more advanced than I would have to comment on how accurate this could be. Don't worry about the little towers just for me
.
Comment
-
Okay, so are we going to concentrate on the bookshelves for now, I know mine will be probably within 2' of the wall, as they will go next to my dresser or cabinet in my room. I won't shove them up against the wall, but they can't be like 4' from the wall, as my room just isn't that big.Comment
-
Assuming a monitor and a tower would be placed the same within the room, you could use the same crossover. So if Nathan's monitors will be next to his dresser (assuming sitting on some sort of stand) then he could also just build them as towers with the same baffle width and place them in the same location and not have to change the crossover.Comment
-
That was what I was thinking. Actually, if given the option, I may build towers because I don't have a sub to go with them, and it might be kind of cool. So if we can do both, I would be really happy! So let's say that both will use the same baffle width, and design both together for the same crossover.Comment
-
Sounds sweet! I had previously stated 8" wide baffles, Bob mentioned 10", I don't know if anyone really has any preference. I think the depth & height are somewhat adjustable by the end builder, as long as internal volume & baffle width are kept consistent, correct? Personally, as the room they're going in is small, I will most likely build towers <30" tall. Also, Nathan mentioned 2' in from the rear boundary; this is good for me, anyone else have placement grossly different?Comment
-
I think 10" would give a little more leeway for driver cutout for the novice builder, of course, lets just see which one models up the best.Comment
-
Oh, exorcer, I don't think these are shielded, so be careful with your monitor, unless there is a way to shield the boxes or something.Comment
-
nathan, you bring up a good point. would lining the walls with steel help do the trick? is there an easier way?Comment
-
I think I've read about that method being used successfully, but just a thin layer may not do it. Bucking magnets can be used also, try a search for that & you may come across the use of steel also.Comment
-
Are there any free programs I can use to model this? Also, is that Seas tweeter the best one for sure? Let's find the best possible match, of course, I guess you wouldn't have picked it if it wasn't good.Comment
-
I had thought about the Seas 27TBFC, very similar to that tweeter, but a metal dome. Some tests show it has lower distortion, but not as well behaved in higher regions..especially given the ultrasonic breakup mode..I dunno, metal cone / metal dome? Meh, not too big of a deal, some things I've read say it is almost a drop-in replacement between those two & the Dayton RS tweet.Comment
-
So which of those two Seas would be the best? What's the real world difference between the two?Comment
-
I have the 27TDFC and like it a lot. I haven't heard the other.
Tweeter level matching is taken care of with a padding resistor. The baffle step relates to the width of the baffle. (I think cobbpa had these mixed up)
Here is the response of my XO for 1 DA175-8 and a 27TDFC modeled in Speaker Workshop. (free) Note that I haven't spliced in low end response predictions. The PE CLIO data look like it was infinite baffle.
With Tweeter reversed - you can see I need to work on it a bit more. I had a beautiful notch when it had a 2 db response bump at the XO frequency
Current filter configuration:
Comment
-
From personal experience, I like the look you get with the woofer on a 9" wide baffle and 3/4" rounded vertical edges.Comment
-
You're right, I mixed these up, my mistake! I'm fine with the 27TDFC, I have never heard either; just from reading I understand they can work with similar crossovers. I may try to play around with SpeakerWorkshop some; I never have before. Is there much difficulty, Bob, in making the compensation in line with the baffle Brian suggested? Although, I guess it is only an inch, and it seems that you have come up with a simple tweeter network with rather round values, so maybe it isn't worth changing? Or would someone only change R2?Originally posted by BobEllis
This is definitely a learning experience.
Thanks to the more experienced guys jumping in here, helping out us budget-minded, lesser experienced enthusiasts!Comment
-
If you want to change the tweeter level, you can simply change R2 without affecting much else.
If I understand Brian correctly, his baffle has 9" of flat space with round-overs applied beyond that, making the cabinet 10.5" wide. That's close enough - I usually count half the round-over as part of the baffle width when making a SWAG at BSC. It seems to work out pretty well.
Increasing L1 will increase baffle step, but also affect the crossover region. I tried to use stock values to get in the ballpark. Unwinding inductors and paralleling caps isn't hard, but if occ the shelf values work, why not? Note that the crossover isn't optimized since the reverse null is ralatively shallow. It assumes .75" offset between the driver acoustic centers. Probably fairly close for a flat baffle. You could get fancy and throw in some compensation for the tweeter's response, but then you're runnig up the parts count.Comment
-
Bob,
No, the actual cabinet/baffle width is 9" wide. You can see the finished product here:
Image not availableComment
-
Actually, now that I think about it, I believe the actual cabinet is 10" wide. We just were able to wrap a 48" wide piece of veneer around the cabinets and place a single seam in the back. The 3/4" roundovers removed enough material to allow the veneer to overlap and allow for trimming.Comment
-
That last pic was figured cherry that I did for a customer a year or so ago. Here are my dad's in African (Ribbon) Mahogany. My dad has been completely amazed at the sound of these. Bass response is pretty awesome for low cost, dual 7" drivers. Crossover work was courtesy of Curt Campbell.
Image not availableComment
-
Nice work Brian. Those cabinets sure are pretty.Originally posted by Brian Bunge
Hey, we're on the same page for baffle step close enough for a start.
Remember, though, that this is still a prediction based on someone else's measurements. You really ought to measure your drivers in your cabinets and then work out the details of the crossover in SW, perhaps with my design as a starting point. You can see a bit of the woofer irregularity impacting the response around 3 KHz, If I could afford it I'd do something similar to Jon's Modula MTM crossover with very steep initial cutoffs. SW models them pretty well, as far as I can tell. I haven't built one yet, but SW models seem to match Jon's predicted response.Last edited by BobEllis; 21 March 2006, 17:26 Tuesday.Comment
-
This is progressing very well! So what do we need to do with the crossover to get it perfected? I'm not entirely sure how crossovers work, so any explanation would be great.Comment
-
The simple version of how crossovers work:
Start with the basics: (simplified)
Impedance is a measure of how much a device resists the flow of current. It generally has real (pure resistive) and imaginary (reactive) components -
resistors provide pure resistance that is constant with varying frequency
Capacitors have reactive impedance that decreases with frequency
Inductors have reactive impedance that increases with frequency
real components have bits of all three types of impedance built in, but ignore that for now.
Hope that helps a little.
Two impedances in series form a voltage divider. the voltage across each component is directly proportional to its impedance.
The crossover forms a voltage divider with the speaker. As frequency increases the inductors' impedance rises and the capacitors' falls.
Looking at the woofer section you see the first two components are an inductor and a capacitor. With rising frequency, the voltage at the junction of these parts drops, eventually reaching a rate of 12 dB/octave. This voltage is fed into another LC, causing another 12 db/octave drop. This is what makes it a 4th order filter.
There's a lot of math that goes into figuring out the component values, but fortunately for us, there are also a lot of programs available to do it for us.Comment
-
So what now? Wait until we have our speakers to test on-baffle? Or just further xover experimentation by those with more knowledge?Comment
-
So, how does a crossover cut out the frequencies that are too low on the tweeter, and cut out the frequencies that are too high on the woofer? Do the resistors, capacitors and Inductors of different values strip different frequencies from the signal?Comment
-
That's a way of looking at it. lets go back to the simplest crossover for a tweeter, a capacitor. Now we have a voltage divider consisting of a capacitor and the tweeter. At DC, our ideal capacitor has infinite impedance. No current flows, so the tweeter sees nothing. As the frequency increases, the impedance of the capacitor decreases (Z=1/(2piFC) where F is in Hertz and C is in Farads).Originally posted by Nathan P
Use an 8 ohm tweeter impedance and a 4.7 uf (microfarad = 1x10^-6 farad) cap for our model. Using the impedance formula we find that at 4232 Hz our capacitor has a Z of 8 ohms. This means that current can flow and the voltage across the tweeter is half the applied voltage (for simplicity, assume 1V). By 10KHZ the capacitor's impedance has dropped to 3.3 ohms. The voltage across the tweeter is now Vin*8/(8+3.3)=0.707 V. At 20Khz, the tweeter voltage is up to .82V.
If you plot the voltage against frequency on a log scale you'd wonder why you'll see that it sags terribly. The rolloff is apparent well into the pass band and the stop band rolls off at 6 db per octave.
Higher order filters allow a sharper corner and steeper cutoff slopes. The math gets a little difficult to show in this form, I suggest a good electronics text if you want to dig deeper into the theory.
As for do you wait for your drivers to measure (you can build the cabinets while the drivers are in transit, since you know the dimensions) or just build build what is modeled here, the choice is up to you. Do you want design that is optimised for your application or a generic, but pretty good simulation based design? The sim is just the first step in the process.
When you have built as many speakers as Jon Marsh ;x( and have his set of tools, you can be reasonably sure that you'll have something pretty darned good the first try based on sims and response measured in your cabinets. But you see in his threads that even he goes through a few revisions before calling a design complete. It's a trade off, like most of life. How much effort are you willing to put into it?Comment
-
So do we have some dimensions for the bookshelves and/or towers? Or are we waiting to get them to start with all that.Comment
-
For the bookshelf version, assuming a 10" baffle/cabinet, I'd go with external dimensions of 14"H x 10"W x 10"D. For a tower version, you could keep the 10"W and 10"D and make the cabinet 25.5"H. This is a bit short for a "tower" so I'd just add in another 12" or so in height with the bottom section sealed off. All that really matters is that the internal top section be 24"H to achieve the appropriate internal volume.
I'd use a 2" flared port 6.75" long, which tunes the cabinet to 30Hz and gives an F3 around 28Hz. If you stuff the cabinet some it'll raise the F3 but gives a more gradual rolloff, which should be pretty flat in room. Without the stuffing, I'd worry about boominess.
Oh, and both designs assume 3/4" material will be used.Comment
-
Brian, what program did you use to figure out that volume & tuning? I'm getting those dimensions (talking about the tower one) to be about 1 cube, but the 2 different sources I use are saying the port needs to be longer for that tuning. I'm guessing I'm doing something wrong or forgetting to consider something, but I'm curious about what it is.Comment
Comment