Three-way tower with Dayton RS225's, Help needed.

[m_bathke]

So, I gather that the only real drawback to using multiple caps in parrallel is that the P.I.T.A. factor goes up because you have to measure each one and match it to others that will make the right value.

The hassle part for me really isn't that big of a deal, at least it will give me something to do for a few nights.

[JonMarsh]

That's a good price on a spool of that Monster cable. OTOH, how much do you actually need?



What the marketing BS means (flux tube, etc) is that there is a small plastic core for each conductor around which the wires strands for each wire are wrapped, so that proximity effect doesn't have as big an impact on the high frequency loss- were there wire in the center, it would tend to carry less high frequencies. This technique is used in a number of other speaker cables. The funny thing is that the net wire AWG for this cable is not mentioned, even on Monster's own page for this product.

OTOH, what I use for internal cabinet wiring that is moderately priced and works well is Cardas Crosslink- it also uses a hollow tube array for the wire, but instead of being twisted pair, as the Monster, it's a four conductor array, and uses Cardas's patented gold mean ratio of wire guage sizes, to distribute the signature of the conductors more uniformly. Each wire strand is individuially insulated (like Litz wire), the wire cable insulation is teflon (not LPE or PVC) and the net AWG for the cable is AWG10; this comes in handy for MTM connections, as I slit the insulation partially, and run one AWG12 pair to each midwoofer.


cardas_crosslinksp.jpg Image not available


crosslinksp1.jpg Image not available


crosslinksp2.jpg Image not available


This cable is available in bulk from DIYCABLE, Welborne Labs, and others, at about $3.75 per foot. I also use this for my HT speaker runs. (My main two channel system uses Cardas Golden Reference).


~Jon

[ThomasW]

So, I gather that the only real drawback to using multiple caps in parrallel is that the P.I.T.A. factor goes up because you have to measure each one and match it to others that will make the right value.

Actually it's much easier than that. Caps are almost always lower then the rated value (saves the mfgr money). So glue up the cluster and solder the caps. Leave one cap off the cluster. Now measure the cluster. 90% of the time you'll be adding a cap or two to get the needed value.

[m_bathke]

That's a good price on a spool of that Monster cable. OTOH, how much do you actually need?

Yeah, its a good price because I get it for cost. The wire is 16ga. they never say the gauge because there wire is supposed to have lower resistance then standard gauges, so you don't need to go as big. At least thats the BS they tell us.

Really, I will probably only need about 40 feet total. So, I guess now it is up for me to decide, thanks for the suggestions you guys are a great help.

And Thomas, thats a pretty good idea, I wasn't really looking forward to measuring 60+ caps. :P

[cjd]

It just now sank in that I managed to get the cap value requirements so they're all too handy to achieve using those excellent GE 10µF caps! Cool. Everything just seemed to come together for this one - pretty good phase tracking, nice reverse nulls, good (I think - never sure if I'm using the tools right) power response, reasonable impedance. . .

FWIW, Cardas is a good source for binding posts (spades and wire having been mentioned already ).

I'm trying to convince my dad he needs a trio of these (well, a pair plus CC version) to go with his Tumult. Next up, side and rear surrounds. (I just wanna build 'em!)

C

[m_bathke]

So I was looking at all of the cap values that are needed for this project and another question came to my mind. Would it be okay to use the less expensive GE1040L400 10uf caps ($1.15 each) for the values that are not in series with the drivers, I would still measure them to make sure the values are acurate? By doing this I would save about $40 which would offset the cost of the more expensive RS28 tweeter and still put me right around my budget.

I would still use the GE1041L600 10uf caps ($2.40 each) for the values in series and comparable caps for the smaller values.

This would go *really* well with a nice LM3886 chip-amp IMHO. Hmmm. heh.C

I looked into the chip amp thing and I must say that I am very interested in trying it out. I assume that you are suggesting a Gainclone varient. They looked extremely simple to build, that is as compaired to other DIY amplifiers out there, having only 9-12 parts for the amp sections. Do you have any idea how much a "nice" one would cost to build. I saw plans for one that uses two chips per channel and from what I gathered from other forums this would be the best for driving a 4ohm load.

[cjd]

On the caps, I'm probably not the person to ask. If it's not an electrolytic cap, I think it will be at least acceptable. I was under the impression that the GE caps commonly used here go for $.90 when purchased in batches of 100.

On the chip-amps - I have a couple based on the LM3875 chip, purchased off chipamp.com

Were I purchasing again, I would go for the LM3886 kit or the LM4780 kit and I would use the "big cap" power supply. Pair the former with a 25-0-25 toroid (Parts Express carries one at 330VA that's perfect) and it'll love a 4ohm load. Or go dual mono on a pair of the same trafo's (though you could drop to the 250VA) and the LM4780 for more headroom. The LM3886 version should cost you maybe $150 once you've all the parts in hand, box built, heat-sinks, etc.

However, I have other amps based on these chips somewhere in the process between design and component purchase that, while more complex, should also vault these into, perhaps, the same class occupied by stuff like the Pass X series. Of course, Mr. Pass contributed the initial concept I'm messing with, so I can't really take credit for more than PCB design and probably component value selection. It should be no more complex (though more time consuming) to assemble though, once completed.

C

[m_bathke]

So I cut up the wood for the boxes for these beasts this past weekend and I am starting to get an idea just how heavy they are going to be(really HEAVY). In the past I have done quite a bit of wood working, but I have never used MDF which is alot more absorbant than wood. So it got me thinking of what kind of glue to use to get a good bond and an air tight seal for the encloser? Any suggestions?

[cjd]

Standard PVA works great. I usually use TiteBond 1 (I often find I'm working in conditions where TiteBond 2 doesn't adhere properly - too cold). You can also use a PVA/water mix to seal the raw edges of MDF. May take a couple-few coats. Slather it on, clamp it real good, and let it dry.

I also tend to stay away from MDF. Nasty stuff. :P 13 ply birch is lighter, to boot. Oh yeah - more expensive too.

Aaaand, I have a network now for the RS28A tweeter :B :B :B :B :B instead of the 27TDFC. Personally, it's what I would go with. There's a good version and a great version. The latter takes an extra fifteen bucks per crossover though (moves the -50dB point on the tweeter from ~450Hz to ~750Hz or so). :T :B

Oh yeah. Also a CC version (with both shielded and un-shielded tweeter used).

[JonMarsh]

YOU have been BUSY! :T

I'm jealous- work conitnues in the panic- rearrange the deck chairs on the Titanic mode, with all kinds of extra clueless people trying to get involved who should know what's going on in regards to this one issue, but DON'T, so I'm going to be doing some product marketing definition for a key customer in the coming week, because at least I have a couple of clues (i.e., I talk with them on a regular basis and visit them quite frequently, unlike the assigned Sales people and account manager....)

So, ya gonna share some of the new work with us? Inquiring minds are curious.... :B

~Jon

[cjd]

Share?! Yeah. I want to tweak a bit more, but sharing means getting things posted to my website and I just didn't have time for that yesterday after mowing the lawn and all. :P I did get things tweaked to a comfortable spot, but I think I may be able to improve the phase relationships still.

And, I need to do a "minimal BSC" version still. Though it likely won't be much (if any) different on the tweeter side of things. I'll probably actually toss this stuff to a web-page at this point, just to collect all my thoughts. . .

Busy? Yeah. I like staying busy. If you happened to glance away from my work-bench in the basement you might notice the 120+ tomato plants we started in early March (plus the peppers, the onions, and the rest). And, of course, the sauce from last years crop. :lol:

But then, methinks you know what busy means! At least I get paid hourly! (well, for the "work" I do - this speaker stuff is "play" and isn't paid, except in the satisfaction of a job [hopefully] well done...)

C

[m_bathke]

CJD

Earlier in this thread you stated the relationship between parts in the woofer crossover which gave you various BSC's. According to your statements I calculated that for 3dB of BCS would require a 2.7mH ~.9ohms, 121uF for the cap, and .62mH, I think that is about right. The BCS seems to be pretty sensitive to the DCR of the first inductor and the cap value, but does the DCR of the .62mH effect it very much? I ask because on the original the DCR of the .68mH was .24ohms and the only .62mH air core that I can find is DCR= .16 ohms. Will it make any difference or am I just thinking too hard again?

Thanks for all the help, and I am really looking forward to getting these done, its kinda like waiting for christmas....

[cjd]

The less BSC you go for, the less the series resistance matters in the circuit - indeed, it can help to have more. Rather simply stated, more series resistance decreases the output on the bass end of things thus reducing BSC. This also goes for box tuning BTW - the higher the total series impedance, the larger the box needed to maintain the same Q. Or, as you increase impedance without altering the box, the response rolls off a bit faster.

As long as it's in the ballpark you'll probably be close enough to be happy. It's still an estimation on my part.

Given what you normally run into as far as room gains, nulls, reflections, and all the other mess, I think close enough is going to be at a point where response has much more significant issues from things not having to do with the circuit, but with the room. Room can take a good setup and make it sound horrible.

C

[m_bathke]

I have the boxes assembled and I applied a thin coat of bondo to them(accept for the front baffle) which I then sanded smooth. For the front baffle I was thinking of using a Cork veneer partially just because I like the way it looked and also because it was much cheaper then using anything else that was kind of exotic.

So here are the questions that I have.
-If you were going for a high gloss finnish in black, what kind of paint and primer would you use, would you use laquer, and how would you apply the paint, I would asume a spray gun with a series of sandings but I am not really sure? Just kind of looking for direction on this one, I would assume that some of you have done a paint job like this before and would know what works best and what doesn't.

-Secondly, is there any big drawback to using cork for a veneer on the baffle besides that it is softer and may be damaged a little easier than hardwood. I have seen it use only on one other speaker and I like the look of it I just don't know if helps or hurts the sound of the speaker. Any insight into it would be appreciated as I wanted to avoid a mistake before it happens.

CJD, I was thinking of calling the speakers the "Jam Sessions", or has that already been taken? I was just wondered if you were cool with the name because you designed them so I guess the naming thing is up to you. I am going to CNC badges/name plates for these out of 1/4" alluminum with raised lettering in a few weeks for a project for a cad class.

[PMazz]

If you were going for a high gloss finnish in black, what kind of paint and primer would you use, would you use laquer, and how would you apply the paint, I would asume a spray gun with a series of sandings but I am not really sure? Just kind of looking for direction on this one, I would assume that some of you have done a paint job like this before and would know what works best and what doesn't.

Use lacquer based black primer/sealer followed by 4 or more coats of gloss lacquer. If you don't have access to a spray setup, rent an HVLP (High Volume Low Pressure) setup from a rental company. Prepping the surface is most important. Fill any problem spots and flat sand. The better the surface the better the finish. Next apply a couple coats of the black primer. Check the surface for defects as you go and fix it now. After priming, apply a couple coats of gloss, let dry some and flat sand with ~320 grit. If the paper clogs quickly, let dry some more. Apply as many more coats as you want, sanding as needed with finer grit paper, until the depth of finish you want is achieved.

BTW, if you've never sprayed before, practice on something disposable first . Another alternative, although non-DIY, is take them to a local body shop and have them sprayed up.

Pete

[cjd]

You could just call 'em "J.S." and let those of us with some taste think "Bach"

I need to put a disclaimer out here at this point, unfortunately. I have discovered that the results I am getting in Speaker Workshop may not be accurate. I'm trying to work through this with someone that is doing a center channel version of this speaker (He's modeling with LSPCad). This is not the first time I have questioned the results I get in SW.

It's not huge differences specifically, but I'm not happy with the differences we are seeing.

C

[JonMarsh]

My experience is that I got different results in some areas even between early versions of LspCAD 5, and the final 5.25 version.

If you have driver data files and a netwrok schematic, I'd enjoy playing with them, too, if you'll let me into your sandbox...

Give me something to do on the airplane trip to Houston next week....

~Jon

[cjd]

Now that I've got a server again, I can toss this stuff out there for folks in a zipfile or something. I'm not so much concerned that I can't fix it as I am that Speaker Workshop doesn't do the job correctly. I'll post it up sometime today.

C

[JonMarsh]

Just let us know when we can grab it...

~Jon

[dawaro]

Chris do you want me to just send the files to Jon? When I did it in SW the other night the biggest difference I had was in the midrange response. If I get time tonight I will play with it some more. Today we are cleaning the woodshop in hopes of building this speaker in the next week or two.

[cjd]

File is here:

http://www.eldamar.net/audio/rswwmt/cjd-RS-WWMT.zip

Contains: impedance (.zma), corrected frequency response (.frd) and a screenshot of Speaker Workshop with the circuit and response. Also a text file with the inductor impedances (and where I got that number). FRD files are minimum phase (Hilbert was run after summing in baffle response).

Impedance on this particular version dips a tad low (to 3ohm) but not too bad. It's with full baffle step.

I've already discovered one annoying habit of Speaker Workshop. You can't have multiple sources and expect them to behave properly. Makes for much more miserable network readability.

It's almost exclusively the mid/tweeter relationship that seems to be all whacked.

And, I'm sure this network can be further optimized.

Jon: I really appreciate you taking the time with this.

Anyone else with Speaker Workshop, SoundEasy, LSPCad, or anything else, is welcome to take a crack at this. At this point, I'm curious to see what the summing results are for these networks with the provided driver data.

C

[m_bathke]

Well I haven't ordered any of the crossover parts or the drivers yet because I am moving in a week, I really didn't want to have $500 of stuff delivered to the wrong place. So if there are any changes made to the network thats ok, I was still planning on using the 27TDFC.

I am still wondering if cork would be ok to use on the front baffle for the veneer, and how much it will affect the sound? Any insight into this would be appreciated.

[cjd]

Chris do you want me to just send the files to Jon? When I did it in SW the other night the biggest difference I had was in the midrange response. If I get time tonight I will play with it some more. Today we are cleaning the woodshop in hopes of building this speaker in the next week or two.

I want to start out with the vertical version, which is what's in the zipfile attached (also, since it's related to the initial topic and I was tweaking it a bit more to work with the RS28A today).

I can also put a snapshot of the CC network I put together with response in there though (using your source files David). In fact, I'll do that in a separate zipfile. That way they're up for anyone to DL and mess with if they choose.

My guess is we're going to discover a bug in Speaker Workshop. Yes, it *is* the midrange that's different. Both predicted response and resulting phase.

C

[dawaro]

BTW my results are from LspCAD 5.25 2003-11-04.
Chris I will pull the verticle files tonight and run them also and send you the results.

[cjd]

I was still planning on using the 27TDFC.

OK. The results of this experiment will pose the same trouble whatever tweeter is used. I am glad to hear you don't have crossover parts yet though!

Cork should be fine I think. It's wood. Bark really. If it's particularly thick, I suppose it might alter diffraction a little (perhaps like felt might) though I doubt it will do much. Certainly, I doubt anything significant.

C

[cjd]

Alright. Center Channel data posted.

http://www.eldamar.net/audio/rswwmt/cjd-rs-wtmwCC.zip

Center channel stuff. Again, impedance and response files (with baffle summed into frd and minimum phase run). Text file with inductor impedances. Graphic with network and the response I get in SW.

I think the polarity thing has been figured out at the very least, thankfully. David, this network is tweaked a bit from what you last saw from me.

C

[m_bathke]

CJD
So, I have noticed that all of the new networks that you are posting are for the new Dayton RS28. I am was wondering how much the tweeter and mid networks would change if you used the 27TDFC because that is what I am planning on using?

[Brian Bunge]

This isn't meant to sound like a smartass question (I'm hoping Jon will add in his $.02), but is there a real reason to go with the 27TDFC over the RS28 other than the minimal, IMHO, cost savings? I've heard the 27TDFC and it is a mighty fine tweeter, but what you get with the RS28 for about $15 more sounds almost like a steal.

P.S. Jon, I'm working on my tower cabinets this week so feel free to get busy on your XO sims!

[JonMarsh]

The biggest advatage the 27TDFC had over the RS28A up until a couple of weeks ago was that it was actually available! And it was available back last Christmas, too, when I started some of these projects.

For those needing to make a speaker with a "kind" impedance curve (more like 6-8 ohms) the 27TDFC would be a better choice.

Overall, I agree; the extra $15 or so is money well worth spent. But if one was under real budget contraints, I wouldn't feel bad at all about winding up with the 27TDFC in order to make a project affordable.

~Jon

[dawaro]

I ran the data from Chris for the WWMT in LspCAD 5.25 and the FR results are almost identicle between the two. I took the results from both Speaker Workshop and LspCAD and traced them with SPL Trace and then loaded them into FRC so I could overlay them and I saw no major differences. Where there does appear to be a difference is the phase, if I am reading the response chart Chris posted correctly. But everything sums the same and the polarity for all the drivers is unchanged in both instances.

[cjd]

m_bathke: The 27TDFC needs a different network. I still have it around so you'll not be left wishing you had the RS28A. The reason these recent posts have the RS28A is because that's what I have been most recently working with and tweaking, and it was the simplest to put out there for these comparisons. Not only that, these have full baffle step. I believe you were looking at a minimal baffle step requirement - these recent posts are all full baffle step.

dawaro: Phase appears to be rotated 180 degrees between Speaker Workshop and LSPCad - that could easily be a "source" that's wired the opposite way. Probably should verify the CC works out also. I'm wondering if x or y-axis offset is perhaps the source of some of my troubles - Speaker Workshop either has no place to enter this (just z-axis offset) OR I'm daft and can not find it.

C

[m_bathke]

The real reasons I want to stay with the 27TDFC are that it will provide a slightly easier load to the amp and also I was kind of discourgage by the frequency dip on it. I was hoping to have something that could really sing up high, and I guess if I am wrong about the RS28A please set me strait.

CJD
Full baffle step is fine, the receiver I will be using will adjust for to much bass out. I realize that that is not optimal, but I will probably only have limited space for two more years and after that I would need full baffle step anyway.
I guess what ever the general consensous is on this one I will go with, and I really appreciate all the effort that you guys are putting into this.

[cjd]

Uh, the tweeter isn't going to dictate how easy a load this is - it'll be the woofer circuit. Which will take a good 4ohm stable amp to dish out at high volumes.

But, there is nothing wrong with the TDFC at all. It's a great tweeter. I have no plans to replace the TDFC's in my MTM's.

Don't second guess yourself.

C

[Brian Bunge]

Well, the easier load definitely makes sense to me. As far as the dip in the response, I played around with a test disk a few years ago and found that once I got up above 16KHz or so I couldn't hear anything anyway. Whether or not you can I couldn't say. Also, whether or not it's an issue would be more of a question for Jon and others that have more experience than I.

For me, the potential for lower distortion at higher SPL's makes the RS28A worth the extra money for me. At the same time, if you've heard the 27TDFC and like it, that would make a good arguement for using it. (I love playing devil's advocate)

[m_bathke]

Well the reason I am second guessing myself on the tweeter is because this project will be my biggest investment in my home theater to date and I don't want to suffer from buyers remorse. 8O I haven't heard any of the speakers in the project, but what I have heard from speakers that are mass produced in the price(at cost) range of these I know I don't like at all. In the future I plan to build a center and surrounds to match, so this will be the base for my system as far as the speakers are concerned. I want to build a really good base .

Now I have no experience with building home audio speakers and am putting full trust into the knowledge and experience that you guys have, which seems to be extremely extensive. I really appreciate your willingness to share it. I guess I just want these to be the best speakers I can get for my money, and I think I am getting closer to that everyday.

[Brian Bunge]

I highly doubt you'd ever be disappointed in the 27TDFC. I've heard it in several DIY designs and think it's one of the best I've heard at any price. Not that I've got extensive experience with many tweeters, but that one made a big impact on me.

For me personally, the minimal cost difference and the fact that my surround receiver can handle 4 ohm loads made the RS28A a no-brainer. That and the fact that I'll probably build some of those chip amps in the near future.

[dawaro]

Hey Jon you still coming to Houston this week? I am interested in seeing what kind of results you get when you run Chris's files.

[Brian Bunge]

Duh, just saw Chris' response saying the tweeter won't dictate the resultant load, the woofers will (I knew that, I guess I can claim suffering from undue stress!). I guess it would help to read other people's responses before responding on my own.

[m_bathke]

So I ordered some parts for the boxes from Parts Express in an effort to get the boxes finished so I could just drop in the Crossover and Drivers.
I got 2- 4" Percision port tube Kits, planned on making the ports 14"(not cut yet) which I figured should make the boxes of 5.25cu.ft. tuned to about 20hz.(please correct me if I am wrong)
Also ordered:
2 sets of Dayton Audio gold binding posts
Kester Silver Solder
and 10 pairs of Gold Female 1/4" disconnects
All arrived on time but the disconnects were wrong, they gave me Male ones instead, I guess there entire stock was mislabled and they are out of the female disconnects until the end of July or early August. Any ideas on what I should do, I really don't want to have to wait until late July to get them.

I am curious to see what Jon comes up with for the networks

I still plan on using the 27TDFC, but after doing a little reading and thinking it makes alot of sense why the tweeter would have little to no effect on the load of the speaker. The tweeter gets padded increases resistance, and it takes up less then 20% of the load, where the woofers will make up just over 60% of the load, I think.

[cjd]

I use solder to connect to the speakers.

You could go with whatever you find at your local home improvement store or radio-shack as well. Or check to see if someone else has what you need (Madisound, whomever).

When considering impedance, understand also where most of the energy in music exists - it's solidly in the mid-woofer range. Above 5kHz is fairly light use by comparison - you could probably get away with unusually low impedance on the tweeter with less impact on the ability of your amp to drive it without running into trouble, compared to the mid-woof range.

Also, it is ultimately an issue of how hard you drive these. I use an amp that's not recommended for 4ohm use (if I used a lower voltage transformer it would do better) yet it doesn't complain at all with my MTM's 4ohm impedance.

Did I post up a ported design? I don't remember! :P (I also use PVC for my ports, and a round-over bit at the exit).

C

[m_bathke]

Actually I was going to use PVC for the ports at first but I really like the way that boxes looked that had used the percision ports, and I decided to go with the 4" over your suggested 3" because it offered a much lower air speed, not that it would make much of a difference in the actual function of the speakers but it makes me feel better.

[Jim Holtz]

Monte K's distortion tests...

I found Monte K's tests ( http://www.mfk-projects.com/tdfc_mill_rs.htm ) very interesting in that there was agreement with Jon's and Mark K's tweeter tests but differences too. Monte's tests point to the TDFC still being a very slight leader at higher volumes in the 1.5K region but both trailing the Millineum, as they should considering the price differences. Darn close for inexpensive tweets, though. Perhaps Jon can add his thoughts to Monte's results.

It seems to me that it'd be hard to go wrong with either the TDFC or the RS. :T

Jim


[/QUOTE]
For me, the potential for lower distortion at higher SPL's makes the RS28A worth the extra money for me. At the same time, if you've heard the 27TDFC and like it, that would make a good arguement for using it. (I love playing devil's advocate)[/QUOTE]

[m_bathke]

Jon
I really curious to see what you came up with for any changes/improvements to the networks, if any.

Also I was reading one of your posts in another thread that discussed room issues with bass and ported enclosers. It brought to my attention that the actual port location might influence the bass response to some extent. SO what would would be the best place for the port tube on a speaker that was going to be placed closer to the wall, say with in 6"-12"? I have seen in some of your designs that you have put the port on the bottom and on other you have put them on the back, but never on the front, so I would assume on the front is NOT optimal.

Hope you enjoyed your trip

[JonMarsh]

I'm taking this data on my current trip to review, 'cause I've been working like a whipped dog and have had essentially NO time for DIY the last few weeks- actually since a little before Easter. Hopefully I should have a little time to tweak and play on some flights this week.

Just got another magazine article accepted (for my day job), too, so the whipped dog thing is going to continue for a while.

~Jon

[m_bathke]

Just wondering if anything new has come up for the crossover. I have been pretty busy the past couple weeks with working full time and taking a full load of college classes for the summer but I would like to get the parts and install them in the boxes in the next couple of weeks. Anyway if anyone has anything new I would apreciate seeing it, plus it will give me something to play with in the hour or two a day that I have free.

Thanks