Dipoles sorta like bob/arvo - long

[Paul H]

My property is just like Sea Ranch, although it's on the Atlantic rather than Pacific, and I have to drive 20 minutes to get to the ocean, and of course our place is covered in snow now ....

I've been making minor progress here and there - some quick measurements with the 10" speakers to confirm that my understanding of baffle width, bass response and speaker measurement were okay - it checked out :wink:

I've more or less settled on a 18-20" wide flat front face with wings angling back at some angle and some depth to be determined, but 6-12" wide and 60-70 degrees back to start. The transition may be rounded - details, details :wink:

I've attached an autocad sketch and a picture of my current test baffle. I didn't start out thinking about sloping the lower baffle like the arvo, but it just makes sense, for bass response and dispersion.



Paul

[Paul H]

I'm familiar with baffle step, diffraction, near and far field measurements, mls length, windowing of measurements ... and I still have to ask, how do you measure the lower frequency response (say under 250Hz) from dipoles?

Most of the methods I've seen suggest nearfield measurement, followed by a calculated adjustment for baffle step - that's how LspCAD is set up (you're not forced to use a baffle step but it's there). This assumes you know what the baffle step is, which is fine for a rectangular box, but not for dipoles if the whole purpose of the measurement is to check the baffle response.

Is using a very long time window on the measurement (in spite of room interactions) the answer?

Suggestions and comments are appreciated.

Paul

[Paul H]

Bump the last question on measuring dipole bass, anyone?


Also, here's a snapshot of where I am so far - attached crossover layout and frequency response - as always your comments and feedback are appreciated.

Paul

Edit: Revised crossover layout to more legible picture

[JonMarsh]

I'm familiar with baffle step, diffraction, near and far field measurements, mls length, windowing of measurements ... and I still have to ask, how do you measure the lower frequency response (say under 250Hz) from dipoles?

Most of the methods I've seen suggest nearfield measurement, followed by a calculated adjustment for baffle step - that's how LspCAD is set up (you're not forced to use a baffle step but it's there). This assumes you know what the baffle step is, which is fine for a rectangular box, but not for dipoles if the whole purpose of the measurement is to check the baffle response.

Is using a very long time window on the measurement (in spite of room interactions) the answer?

Suggestions and comments are appreciated.

Paul

Hello Paul,

With a dipole system, it might be confusing to call the panel effect "baffle ste" since the usual way that term is used is in reference to the change from 2pi to 4pi space radiation in a closed box system, depending on the relationship of the radiating wavelength to the baffle size.

For dipoles, below the cut-off point of the baffle, the response falls at 6 dB/octave. PLUS whatever the natural response of the driver is. So, for example, if you have a driver with a Qts of 0.4 at Fs, the response will be down by over six db at the driver resonance, plus the cancellation effect from sound going around the panel. This is why low Qts drivers really get penalized in dipoles, and one often prefers to work with a Qts around 0.5 or higher, if possible.

Now, am I correct in thinking that right now you're just evaluating your midbass/midrange driver?

What I've done is to do MLS mesaurements with long gating windows, and also I've done swept gated sine measurements. (CLIO supports this). The latter shows you everything going on from room effects; its a good idea to pull the system out into the room away from walls. Do a high resolution measurements, warts and all, then apply smoothing to the data if you want to see the more general picture and minimize room reflections. I do that at 1 meter. I also do some semi-nearfield stuff, about 6", to check the latter against what the driver itself is doing.

Note that I've found it helpful to use BDS in developing front panel layouts for the midrange/high frequency portion of a diople- I don't find that a symmetric layout like Linkwitz's Phoenix works best either in theory or practice.

BTW, can you adjust your Acrobat setting so that the PDF's aren't downsampling the graphics? I can almost read them, but not quite... old eyes, you know.

[Paul H]

Thanks Jon,

I revised the crossover layout - anyone's eyes would have had a problem reading the previous file. :wink:

Right now I've got 4th order crossovers between the tweeter and upper mid (6 1/2"), and 2nd order from the mid to a pair of upper bass (10"). The 10" also have a notch filter for the resonance. The 6 1/2" doesn't as it has the benefit of the 4th order crossover.

Paul

[Saurav]

I've had similar questions with in-room bass measurements. Increasing the gating time on the MLS measurements always makes the higher frequency portions of the graph look more ragged. If I set the gating window based on the impulse response so I keep the first reflection out, I get a graph that's usable down to about 500Hz. Below that I ended up doing a swept sine wave measurement, and then performing an FFT on it to get the frequency response. I did this from pretty far away, about 2m. My woofers are < 12" above the floor, so I wasn't sure what I would be missing by being close to the drivers.

What's a swept gated sine measurement? Are you gating on the same window that you set up from the impulse response? Do you then set up the sine sweep to cover the frequency range you want within the gating time?

[sfdoddsy]

For the bass measurements in my system I used a combination of the sub tests in ETF (long gate MLS) and TrueRTA. Both produced pretty similar results as it happens. I didn't even bother trying for quasi-anechoic, but set the speakers up where they would normally go, and then did the measurements from the listening position. At the same time I took the opportunity to correct a couple of room modes using the digital xover. It takes a while doing it this way, but it seems to work out.

Steve

[Paul H]

I wired up a speaker this morning - nothing permanent, just a test setup using a behringer CX2310 as an active crossover for the tweeter, 6 1/2" and pair of 10's.

The behringer crossover is great for the current purpose. It's dirt cheap, and because crossover frequency adjustments can be made with the turn of a dial it's easy to experiment with different crossover frequencies.

I bought the crossover for this purpose, and possibly to use as an adjustable sub crossover later.

My initial impressions are that these drivers are detailed and crystal clear - which should be no surprise given their reputation and cost.

However, making them fully cooperate with each other is not going to happen without some work.

I'm still debating crossover types - the xover to the 15" drivers will be active, and the tweeter to mid will be passive, but I'm not sure about the mid to woofer (it'll be around 300Hz).

Paul

[Paul H]

No matter which way I go with the crossovers for this dipole, I'm going to end up with a maximum 4-5 ohm nominal speaker, and many of the configurations I've looked at are lower impedance.

What minimum impedance is generally recommmended, assuming use of quality but not exotic amps (ie Bryston 3B)?

Thanks,

Paul

[JonMarsh]

It depends on both the impedance AND the phase angle - if the latter is benign, the impedance is not generally a big deal for a good solid state separate or even many integrateds. If the impedance is ~ 4 ohms and the phase angle is much greater than 40 degrees, things get ugly in the output stage (the current lags or leads the voltage, so you can have combinations of high current and high voltage simultaneously across the output transistors.

Brystrons, the better Aragons, Krells, etc, don't much give a rat's patootie down to about 2 ohms and normal phase shift. For most amps, it's best to keep them at 3 ohms + if rated for 4 ohm loads. Heck, some of the tweeters I see which are rated at 4 ohms are really more like 3 ohms. That's what I call bending the rules. Not a lot of power up there normally, though.

I'd suggest trying to keep the impedance no lower than 3 ohms with low phase shift, and 4 ohms if the phase is ugly. Higher is better, of course. That's been an issue with the Arvo, and is part of why the midwoofers are wired in series, aside from the fact that you don't need higher sensitivity for them, anyway.

[Paul H]

Thanks Jon,

I've been playing with two 10" woofers in parallel for the 100-400 Hz range. The extra efficiency helps to compensate for the baffle roll-off and match with the mid.

That most excellent theory runs smack bang into problems with impedance.

Back to the virtual drawing board..

Paul

[JonMarsh]

How low an impedance are you seeing? In that range the phase angle shouldn't be too bad; I can see why you'd want to parallel, depending on what you need to pull them into range with the mid. Are you planning an all active crossover, or hybrid?

[Paul H]

How low an impedance are you seeing? In that range the phase angle shouldn't be too bad; I can see why you'd want to parallel, depending on what you need to pull them into range with the mid. Are you planning an all active crossover, or hybrid?

I did some more modelling today, and got a +4 ohm reasonably smooth impedance, except a dip below 100 Hz, which shouldn't matter, as I'll be cutting off below ~100Hz with an active filter to go to separately amplified 15" woofers.

I'm heading towards a hybrid, with a passive speaker containing 2-10", 6.5" and the tweeter. It appears this speaker could cross over to a sub at a typical 80Hz setting, but I'll be crossing higher as noted above.

I've got 2nd order crossovers at ~400Hz and a cauer elliptic at ~2000Hz. I started with a 4th order for the higher crossover, but added in the extra cauer components to beat the "Excel breakup mode ringing" into submission. :wink:

Since my initial speaker measurements I've modified the baffle (added segmented 4" radius curves each side) and given the speakers some run-in time. I'll need to do the measurements over complete and them I can start to finalize the crossovers.

Paul

[JonMarsh]

Sounds like you're making great progress, Paul!

I think using the cauer elliptic to "beat the Excel breakup mode into submission" is a good idea; I didn't like the way the Thor used a relatively slow roll off with 2.1 kHz crossover and just notched down the level; it doesn't do anything to deal with the energy storage, even if the slope of the amplitude looks nice.

I'm back to working on the Modula MTM's, laying out the crossover boards. May also get to do some RS28 measurements tomorrow, if I'm productive enough!

~Jon

[Paul H]

Microphone Calibration

My tweeter (SS9700) measurements are all showing a decided (~4 db) hump in the 10K to 20K range. I know from other's measurements that this shouldn't be the case, and I'm nervous about proceeding further without understanding this (what if it's not a hump above 10K, but a trough below 10k? ).

I'm using a Behringer UB802 pre with ECM8000 mic, without the benefit of a calibration file.

I've checked preamp settings, in case something a setting was accidentally wrong. I've tried different measurement distances and configurations with the same results. I've tried a second tweeter.

Is there something I could be doing measurement-wise that would cause this 'bump'?

Or do I simply need a calibrated mic? If so, where/how would you recommend I get a calibrated mic (or a mic calibration), and does the pre-amp need to be calibrated also?

Thanks,

Paul

[JonMarsh]

My tweeter (SS9700) measurements are all showing a decided (~4 db) hump in the 10K to 20K range. I know from other's measurements that this shouldn't be the case, and I'm nervous about proceeding further without understanding this (what if it's not a hump above 10K, but a trough below 10k? ).

I'm using a Behringer UB802 pre with ECM8000 mic, without the benefit of a calibration file.

I've checked preamp settings, in case something a setting was accidentally wrong. I've tried different measurement distances and configurations with the same results. I've tried a second tweeter.

Is there something I could be doing measurement-wise that would cause this 'bump'?

Or do I simply need a calibrated mic? If so, where/how would you recommend I get a calibrated mic (or a mic calibration), and does the pre-amp need to be calibrated also?

Thanks,

Paul

Can you post what you're getting, and the measurement conditions? IME, the 9700 does have a 3-4 dB hump centered between 10 and 20 kHz.

With my Behringer ECM8000, the highs are rolled off a bit above abut 13K; you can compensate this with the Beheringer mic cal file from True Audio that comes with the True RTA product (free down load for level 1). Still, the ideal way to go is a true calibrated mic; the calibration costs more than the mic, as in the case of Girardin's.

[Paul H]

Jon,

One of my Scan Speak 9700 measurements is as follows:

SS9700 Freq Response.pdf

This was taken on-axis at 75cm with a 96,000 sampling rate, total mls length 4096 and sample length 3ms.

Various other measurements at 48,000 sampling rate, distances from 50-100cm, and samples from 2-6ms produce similar results. I checked my preamp (using a rs meter alone and then through the preamp with a before and after comparison) and it seems pretty flat or unobtrusive.

Is this what you would expect?

Paul

Edit in - PS: If anyone can tell me how to get the #$%^&* picture to display within the post I'd appreciate it.

[JonMarsh]

Attached files will only display if they're graphic files that are browser compatible, like .jpg, .png, .gif. But reading your PDF in a seperate window is no problemo.

Was this measured on a test baffle with a round over, or on your actual dipole panel?

It "looks" to me like their could be diffraction effects at 4-5 kHz (peaking) and 8-9 kHz (dip). What I do to seperate the tweeter performance from the current baffle is a nearfield measurement, say, 2-3"; that can even be swept sine with smoothing if you want oodles of detail; then compare this to what you're seeing at your current measuring distances and windows. You can do that nearfield measurement with MLS, too; you'll probably need to, if you're just using "Just MLS". Some day I'll have to try JustMLS, but the PC I'm finishing build which could do that with (just got in an Audiophile 192 card for it) will probably be equipped with Praxis shortly (hopefully in April), and I'm just going to keep my CLIOWIN setup for backup and correlation if needed.

Do both 9700's you have measure identically? Another option if you're patient is some correlation measurements- exchange drivers and measure. You've got XT25's in your MTM's, right? THey're pretty dang flat on axis from 2 kHz up to 20 kHz- (assuming large diffraction free panel, or nearfield measurements- you could pull one of those and check for comparison.

Mark K has a nice tweeter panel for doing farfield measurements- I should build something like that- pretty good sized, and then rather large roundovers at the edge, with the center panel replacable with a few screws in T Nuts for moutning different tweeters. It's a good setup for somewhat more farfield measurements in driver testing. OTOH, you can get the same infor with true nearfield. Just harder to do polar response- his fixture is good for that.

~Jon

[Paul H]

Was this measured on a test baffle with a round over, or on your actual dipole panel?
~Jon

Yes ... It was measured on a test panel made up with scrap material in the approximate shape of the dipole panel I'm considering, which is approximately 20" wide plus 3" radius each side. The radius 'curves' are done with segments. The tweeter is mounted off-centre, at a location suggested by baffle simulation software. So, this isn't the actual dipole panel, but is a close resemblance to what I was planning for it to be.

Here's the just-now-measured nearfield response:



This was measured at 96K, 4092 mls length, 4ms sample time, 2cm distance, 1/12th octave smoothing - the previous file had no smoothing.

Paul

[Dennis H]

I think Girardin charges about $40 to calibrate a mic and about $125 for one of his mics. Plus, you'd probably need a new preamp to use one of his (low phantom voltage, unbalanced RCA connector) so getting the Behringer calibrated looks like a better deal.

[RonS]

FYI I've bought a few calibrated Panasonic mic capsules from Liberty Instruments, Kim G. calibrates them to 40kHz. Cost was $34 & $3 for shipping, you provide the brass tube and connector.

[Paul H]

Thanks Ron and Dennis,

Jon, I forgot to mention one item on measurements above - the tweeter is surface mounted, not flush.

Paul

[Paul H]

I flush-mounted my tweeter and re-measured (okay, I surrounded the tweeter in cardboard) with much smoother, flatter results except above 10kHz. Then I did some modelling. This is what I have so far:

Crossover Schematic:




Frequency Response:



Frequency response mid inverted to check phase (also indicates crossover frequencies):




Impedance:




The low end (100 Hz) drops off a little early with a slight hump at 350Hz, but those were trade-offs to get in-phase response at the crossover.

There are no impedance compensation adjustments. The resonance modes from the seas drivers are down 60dB, as is the response from the tweeter at its 500Hz resonance peak.

I have a small cap in parallel with the woofer inductor which seems to work wonders on the woofers breakup resonance at 4-5kHz. This type of filter isn't from a book, but is sort of an adopted cauer elliptic type thing, which seems to work - is there any problem with using this cap here that I don't know about?

Any comments, suggestions, etc on either my overall approach or details are much appreciated.

Thanks,

Paul

[JonMarsh]

I have a small cap in parallel with the woofer inductor which seems to work wonders on the woofers breakup resonance at 4-5kHz. This type of filter isn't from a book, but is sort of an adopted cauer elliptic type thing, which seems to work - is there any problem with using this cap here that I don't know about?


Thanks,

Paul

It looks like you're making a lot of progress, Paul. That cap is in the right place for what you want to do- just like acrosso the second inductor in the Arvo midrange crossover, for example.

I think your crossover frequency for the W18's should be about perfect from a power handling viewpoint- the W18 is one of the few drivers with great ETC plots even up to 1.8 kHz.

Do you have the Pro version of LspCAD? Can you try emulating this through the PC sound card to give it a listen? Am I correct in assuming that the plots you show are for TF1 bypassed (not in circuit, effectively?) Would be interesting to see how the off axis stuff looks, say, out to 30 or 45 degrees at least.

If you keep steaming along at this rate, you'll get this finished before I do my current version of Arvos! Of course, it would help if I would stop changing drivers or configuration, wouldn't it?

Keep us posted!

~Jon

[Paul H]

Thanks Jon.

The reponses shown above are with the active filter bypassed - when the filter is on at 4th order 120Hz it looks like this:



The off-axis response looks like so (per the LspCAD simple simulation):


I would like to try the simulator, but unfortunately I don't have the professional version of LspCAD, and the regular version only has two channel simulations.

Paul

[Paul H]

Finally received my order for 4 of the AE IB15's yesterday. :T

Last night I did some checking on Linkwitz's spreadsheet for volumes from these drivers. I found that one of these, with an 800mm 'path', will hit 105dB at 42Hz when it hits x-max, but 2 of them will only go to 34Hz for the same volume at xmax.

Given that I'm crossing over to a sealed/ported sub (in some sort of box) for the bottom octave or so, is there any real benefit in using two rather than one of these 15" drivers in the dipole speaker?

Paul

[JonMarsh]

Rember Xmax is defined at an arbitrary droop in BL product, it sort of like soft clipping in an amplifier- while I can understand the advantages (size wise) of only using one, since you've already got them, consider your linearity wil be better with two, and you'll have three dB of headroom.

If you have any capability of sine testing and FFT, (can be done on the relatively cheap- TrueRTA in high res mode without bars) you might want to try some basic distortion testing at a few frequencies in the range you're interested in- near the bottom of the planned range for the driver, in the middle, and in the top.

Then again, you might not. Knowledge isn't always pleasant, I've found.

[Paul H]

Knowledge is always good, and if, after testing and listening, I end up consigning these 4 speakers to an new IB subwoofer while I hunt for other products for the main speakers, so be it

My initial impressions (from cobbling together some amps and and a behringer crossover so that I could listen to the IB15's) is that the dipole bass is phenominal, and these speakers seem pretty good for their intended purpose.

I'll have a look at TrueRTA - I've got the free version now but was considering the "real" version for in-room response testing.

Thanks,

Paul

[Paul H]

I've been making a little progress here and there. I've decided to hold off on ordering components until I can re-measure the drivers in the finished baffle.

The photo here is the partially completed baffles (2 top, 2 bottoms) with the drivers holes cut:



The next major piece of work on the baffles is rounding over the edges at each side of the panel to an approximately 2.75" radius. This is being done by cutting away multiple sections for the curve on the table saw, followed by hand planing close to a penciled-in curve on the end of the panel, followed by sanding using a custom-made wooden sanding block shaped to the curve's radius.

The almost-finished curved edge, along with the plane and sanding block are in this photo:





Just a little dusty

For anyone who's noticed the overall width of these baffles and are wondering, yes I do have full wife approval to build speakers as high and wide as I want, as long as they're going in the basement. The rest of the house is subject to the usual rules of invisible sound systems

Paul

[JonMarsh]

Oooh, invisible sound systems... you could make a mint selling those... is it Klingon cloaking technology, or are they nano speakers?

Some tasty looking sawdust you've got there, Paul!

[mante]

Pluto and Evil Twin

Does anyone have additional information regarding Linkwitz's new Pluto? Any thoughts about the speaker's design? For 20% of the price of the Orion's, you get 80% of the sound of the Orion's. A very interesting kind of plumbing look to them.




Mante

[ThomasW]

Does anyone have additional information regarding Linkwitz's new Pluto?

No just rumor, but speculation would be a good topic for a new thread.

[Paul H]

Jon posted some interesting comments re active/passive filters in the Modula MTM sticky thread, and I thought I'd ask for clarification/comments here rather than derail the other thread.

As a quick summary, this thread was started to document a 4-way with 2-15" crossing at ~120Hz to 2-10" crossing at ~400Hz to 1-6.5" crossing at ~2000Hz to the tweeter.

What say the pros and semi-pros about active vs passive for the various crossovers for this 4-way beast I'm looking at? I keep thinking that the higher crossover can (should?) be passive, that the lower one must be active, and the middle one is .. not sure.

Thanks,

Paul

[JonMarsh]

I think you've summarized it fairly well, Paul..., though at 400 Hz, I might be leaning more towards the "Let's go passive" ~ depends on how many amps you have around and want to run. How are you thinking about configuring the system?

ThomasW's Arvos are passive on the midwoofer to tweeter, and active on the 12's to midwoofers, and IB to twelves. Flexible, but it's a lot of stuff hooked up. He's got it racked up down in the basement.

To show what a bunch of Luddites team "Star Wars" is, Evil Twin plans only a single electronics crossover in the complete Saint-Saens system (between the dipole midwoofers on the panels, and dipole "sub" enclosures)- that's about as bad as doing the Arvo's fully passive, excepting a buffered inline EQ box. (BUF03 + passive network).

For certain, going passive as low as I am requires some honking inductors- I've posted pictures of the NorthCreek AWG12 parts.... they're beautiful, but a big investment. In a enclosure style system, I'd be worried about messing up the Q- not so in this case.

Also, these systems were kind of planned around the fact I have two PAIRS of Palladium monblocks, which put out a lot of voltage, and a lot of current, so sensitvity issues or crossover insertion losses don't much concern me (they clip at ~ 600 watts @ 8 ohms... might as well be infinity for home speakers like these). So having a net sensitivity of 83-84 dB is no big deal. Would be like having 150 watts with 90-92 dB/watt.

[JonMarsh]

Does anyone have additional information regarding Linkwitz's new Pluto? Any thoughts about the speaker's design? For 20% of the price of the Orion's, you get 80% of the sound of the Orion's. A very interesting kind of plumbing look to them.


Mante

Paradigm would endorse these, or sue him for infringement if their recent patents are worth more than warm spit. ;^)

~Jon

[Dennis H]

Seems to me, in for a penny, in for a pound. Once you accept an active XO anywhere in the signal path, there's no real penalty (other than the extra amps) for going active all the way.

[JonMarsh]

Well, ask ThomasW about the complexity and the impact on reliability... more electronics, more cables, unfortunately, = more things to go wrong. With the right gear, I think that's surmountable, but you pays your money and you takes your chances.

Does make things easy to tweak, but conversely, easy to mess up.

[mante]

100% Active Arvo

Dear Evil Twin,

Speaking of active crossovers, I realize this isn't Jon's intend of the Arvo Project, but will we be missing a lot if we go with the Behringer dcx and add, say, Davey's Marantz SR4500 (7 channel receiver). If we are trying to obtain the best drivers verses cost, wouldn’t $350 for six amps and $250 for the crossover get us pretty darn good sound?

Your Young Padawan

[JonMarsh]

Well, your proposal is a very creative way to get the required number of channels of amplification and a flexible crossover. I actually plan to add a DCX to my repetoire of investigatory tools, but I don't know if I want one always in the system. Still, compared to what speakers typically do wrong, it's impact isn't probably that great. OTOH, I cringe at the idea of running my DAC1 or SCD777ES SACD player into a DCS for listening... but who knows, might work ok? Not perfect, but OK?

Remember, I'm a certifiable nut case... (go price out what my main speaker cables cost- 2-1/2 meter Cardas Golden Reference, or my Ayre balanced signature intercconects).

Your concept here is the fast way to get 90% of the way there, I expect. Is that 80W in the all discrete amplifiers with all channels driven simultaneously? At 33 lb it's a little light for a 560 watt amp, much less for the 4 ohm draw, but if you offloaded the bass to a larger external amp, that would probably be fine.

As digital crossover's go, the DCX is widely regarded as an amazing low cost alterntive to more expensive solutions like the DEQX. But from reviewers I trust, the DEQX is not the last word in transperency and imaging, either. But it's a fast way to evalute some interesting options. Just don't get over sold on the ultra steep slope stuff with no phase delay- there are side effects of the digitial filters with very steep slopes which manifest in the time domain- witness the pre and post ringing on most CD players on 1 kHz square waves. The digital filter xovers, when driven to "extremes", will sum OK on axis, but will be a bit wonky anywhere else- well, that's what SL says, I haven't tried it yet.

What I will do is provide transfer function curves for the passive crossovers- if implemented in a digital, then you'll get the same net balance and phase reponse. But no 96 dB/octave crossovers, please!

[mante]

Digital Crossover

Jon,

Remember, I'm the Young Padawan who doesn't know how to read a schematic, I need color pictures with big part numbers as well as a map.

So we bump up our power to the Marantz SR5500 90 watts x 7ch. Discrete Amplification X (All 7ch). Regarding all active, isn't it possible to combine additional passive crossover components after the dxc and amp if needed? For example, "if taming the beast is called for", just add your Cauer Elliptical thing before the speakers.

Finally (well, for now), would this get 90% of your system, or more like 99% of my system which includes a mid level cd player, an older 100 watt Luxman amp., and a passive pre.

Many thanks,

Young Padawan

[Dennis H]

wouldn’t $350 for six amps and $250 for the crossover get us pretty darn good sound?

I'd add another $300 for a pro amp to drive the woofers. Those cheap receivers don't like low impedance so each midrange driver, as well as the tweeter, would like its own amp for a total of 6 channels. Used that way, the receiver should have plenty of oomph for the top section of the Arvo. The woofers need more power so that's where the pro amp comes in. Hook the preamp outputs of the two unused receiver channels to the pro amp so the receiver's volume control still controls everything.

[cjd]

I would be seriously considering some LM4780 based amps in here. If you build these such that you do not need binding posts, you could comfortably do ~100+W@4ohm (and 2ohm stable amps) for ~$85/ch give or take. Depends a whole lot on what you do as far as transformers, and whether you go with existing kits or some that don't exist yet (i.e. do your own or, uh, some that I have somewhere in that vague nebulous "in the works" state).

Of course, you can also be a little pickier about what amp goes to what drivers, so you may not need that much power on the tweeter, etc.

C

[Dennis H]

Trouble is, a 6-channel volume control (with remote) for the chip amps will probably cost more than the Marantz receiver.

[cjd]

I think I missed it being a receiver. I would be using it with a proper pre-amp, you know? Otherwise, you would still put a simple stereo volume control somewhere in front. Seems like the power here is the low point in this system proposal?

Besides that, I'm surprised volume comes *after* the XO? I thought a pure power amp was the norm there.

C

[JonMarsh]

Jon,

Remember, I'm the Young Padawan who doesn't know how to read a schematic, I need color pictures with big part numbers as well as a map.

So we bump up our power to the Marantz SR5500 90 watts x 7ch. Discrete Amplification X (All 7ch). Regarding all active, isn't it possible to combine additional passive crossover components after the dxc and amp if needed? For example, "if taming the beast is called for", just add your Cauer Elliptical thing before the speakers.

Finally (well, for now), would this get 90% of your system, or more like 99% of my system which includes a mid level cd player, an older 100 watt Luxman amp., and a passive pre.

Many thanks,

Young Padawan

The "best" way to use the DCX is with SPDIF input, and a volume control after it, before the amps; that way, A/D is bypassed, and the D/A operates using lots of bit depth. That poses problems for a mult-way system, as level control tracking is an issue.

Real studio gear pushes the signal levels on the analog side, and so gets closer to 22 bits of resolution- this is why the Benchmark DAC1 is setup for such hot output levels, so that the analog noise floor is less of a limiting factor. Considering how we usually listen to music, at much lower SPLs, the dynamic range and resolution is compromised a bit at 80-90 dB playback levels if it's analog in and no post attenuation. Does it really matter in practice? Well, good question. But I wouldnt' want to be re-digitizing my SACD's and truncating the effective resolution.


Everyone's perceptions are colored by their past experience - me, I've been a bit anal about DAC's and stuff, but I'm coming from the position of someone who's helped build a recording studio, done live classical recording, yadda yadda yadda... sometimes (maybe often?) I'm a little restrictive in my outlook.

With that many channels to work with, going active all around might make a fair amount of sense- it partly depends on how capable the DCX is - what kind of filter options are possible (I wonder if one can download a PDF manual - ought to try google tonight for that). If it can do tunable elliptical filters, cool. (I'll be surprised, though ) If it can't, I'd probably recommend a passive upper crossover (midwoofer to tweeter), and the rest active.

With the genearl line up you're describing, I think you could put something together as a hybrid crossover and be very pleased with how your system sounds. Remember, I made a conscious decision to go all passive partly becuase I wanted this system as a test bed for listening/testing active electronics (mostly power amps).

Now, when you consider that even mediocre power amps measure much better than most speakers, you'd think I was stark raving bonkers to be saying something like that- but then, you wouldn't believe the number of power amps I've hooked into a system (by name brand manufacturers- Crown, Bose, Proton, even Luxman) and gone "eewwwwwh, take it out". (there are some Luman amps I really liked, their vacuum tube, and their low TIM one in the late 70's, but forget the 4000 and 6000, for example. Which Luxman do you have? (I used to sell them- a LONG time ago)). All I can say is that active electronics, especially ones with loop feedback around multiple gain stages, often do some rather un-nice things to the music (not always, but often). IMO. YMMV. Usual disclaimers apply. Flame suits not needed here, we're among friends.

With ANY system it's a case of working to get the best combination of sound and performance for the money you're willing to spend. I think you're making some good choices, as regards bang for the buck, and where to put your effort. Speakers and room interface being #1 issue.

Discrete amplifiers are also a good thing, if designed and built well, and Marantz has gotten some good reviews on a number of their relatively "budget" components, including the CD5400 player, the integrated amp, etc. There's more focus on the music quality at Marantz these days than is common at that price point from many other lines (Sony comes to mind).

It comes to mind, how we first tested the set of Arvo's at ThomasW's- a couple of Aragon amps, but just a Liteon 2001 DVD player for the disk source, nothing at all special in cables, etc. Still, the magic was definitely there.

When you refine the source and the rest of the chain, you just get more of the magic. A key would be not spending disproportionately on any one element- I think you understand that. Even with speakers, you wouldn't want to hook up a set of Maggie 20.1 with an old Sansui reciever and a cheap DVD player... somethings are long term investments, like my Ayre and Cardas cables, for example.

As to your system, without knowing what your CD player is, cables, etc, I can't say, but I don't think the basic configuration you're proposing would be the limiting factor. Probably next step would be a player upgrade, or a DAC.

You're being innovative about controlling amplifier costs (imagine the difference between what you propose and even a multi-channel Aragon HT amp!) Probably it's like the difference between a 27TDFC and a Millennium Excel- it's there, but doesn't scale with the financial cost.

Would you hear the difference if you build this speaker, then at a later date got an Ayre preamp, true seperate power amps, etc? I imagine so, especially at more demanding playback levels, but the improvement in what you hear might not scale for the investment required. That's a very personal decision.

Boy, have I ever rambled on too long! Must have been the sugar rush from the Oreo's I had at lunch....


~Jon