Choosing a lil' project

[technodanvan]

While I have several large speakers projects in various states of completion, house remodeling stuff to do, and a platform bed (and other furniture) to wrap up, I figured I should embark on yet another speaker project, this time something small (or, at least, small-ish) for a friends computer. I have several sets of parts for some bookshelf two ways, but many of those use 6-7" drivers which I believe would be too large for this application. I'm currently looking at the parts I already have in an attempt to keep costs down and remove some excess inventory. I'm leaning towards doing a traditional rectangular box with no facets or anything to keep it simple and fast, but I could see the argument for going 'all out' with some of the below.

I will also be purchasing an amplifier for use with these. Maybe a Wiim or SMSL model of some variety? Whatever it is will be 4 ohm capable. These will have a passive crossover that needs to fit in a reasonably small box. Placement will probably be on a desk resting on a thick foam pad. The desk is not currently up against a wall. Usage will be for gaming and music, with the occasional movie stream. I am trying to avoid the need for a small subwoofer, but expectations should be reasonable.

In no particular order, these are the options I have come up with:

Purifi PTT4.0X08-NFC-01 + the "Baby" Raal 64-10: Could be a fun project, probably with a PR or two.

Kartesian "Kourtesane" 4.2 Kit w/ Sub120 + Twt28_vHP: I like the size a lot, but wolf_teeth's troubles with the Sub120 kind of give me pause if I'm to gift these to someone.

Sica 5,5 C 1,5 CP coaxial + Dayton Epique E150HE-44: This would be a bit larger but, I think, doable for desktop duty. Probably with a PR, I've seen several use the SB Racetrack one with success, or the 7" Epique one maybe.

Visaton W130-X + Visaton KE 25 SC: Not sure what to think about these parts as I haven't seen them used much, which makes them interesting to me. I do have a pair of waveguides for the tweeter as well, though I believe they are wider than the woofer is so that might look a bit odd.

SB Acoustics SB12CACS25-04 + SB Acoustics SB29BNC-C000-4: I could be talked into purchasing a 5" variant of the CAC woofer as well. Unsure a beryllium tweeter is a great option here, but I currently have no other plans for them. This might be the smallest option, which would make it enticing.

Dayton Audio RS150-4 + Dayton Audio "Classic" RS28A: I don't technically have any RS150s right now, but do have some on backorder until May, I think - the 4 ohm version. Could be a nice speaker and I bet there is an established design somewhere. In fact, I think someone just reminded me of their project here at the forum a few months back. Having an established design would make this a top contender...if I had the woofers in hand.

Anarchy 554 + Peerless DA25TX00-08: Speaking of backordering until May, models from Peerless are expected to be here around that time as well and I have a few sets on order. That said, I already have three DA25TX tweeters in hand. I'm unsure of what woofer I would pair with them, but I do have some of the smaller Anarchies that currently have no purpose in life.

I'm definitely open to some input on the above, and if other combinations of these drivers might work better.

[JonMarsh]

Why be selective? Just try all of them, and see which you like best!

Note, I have quite a few pairs of various SB Be tweeters- they seem more suited to 3 ways, due to relatively low Xmax. And like many Scanspeak designs, they have much higher HD2 than HD3. Doesn't help the IM.

But I would vote for something... maybe more like this?

[technodanvan]

I have gone through my other parts and discovered I do have a pair of DXT tweeters in addition to small Wavecor woofers and tweeters. Your suggestion of doing them all isn't taken lightly - that is generally the plan for these unless something else comes up, but I do have a bit of a time crunch given my other projects.

I'm leaning towards using the little Anarchy for this project as I have several sets, both 4 and 8 ohm, if it becomes necessary to switch. And with the newly discovered DXT, it may be a toss-up between that and the Peerless. I need to find some data on the Anarchy to see how high it'll play though - I'm pretty sure you've looked at this part in the past so I'll do some searches here.

Edit: Annnnnd I found one set of data from a while back. https://www.htguide.com/forum/forum/...076#post829076

[JonMarsh]

I'd say a 2kHz crossover point would look pretty reasonable for those Anarchy's... keep in mind with the right ultrasonic breakup filter, you can suppress a lot of the distortion at and above 2kHz.

[technodanvan]

I've been doing a little research on the DXT, you said back in 2019 (ish) with a Purifi 6.5 (Modula PSD thread) that it was a bit quirky - then showed a heckuva crossover:



Would this sort of network (i.e. one that has well over a dozen components) be recommended for the DXT? I only ask as I'd like the crossovers to fit in some pretty small cabinets! I understand it would not be this exact crossover, but if there are several bits that need 'tamed' I would expect they should be dealt with regardless of the woofer selection. The DA25TX appears a bit easier to work with (subjectively speaking) from other design implementations.

[technodanvan]

I think I'd like to measure the Visaton KE 25 SC as well (both with and without the waveguide), so I guess that's my first step. It's an interesting little guy...

From the Visaton website:

"Magnetically shielded 25 mm (1") high end dome tweeter made of ceramic for crystal clear, perfectly reproduced sound even at high volume settings and with no distortion. The pole piece perforation coupled with the flow-optimised rounded shape, the damping in the cavity and the ventilated voice coil carrier guarantee a very low resonance frequency. The impedance is linearised by means of an integrated LRC element. This means that the dome can be easily integrated in a network of crossovers. The protective grille is fitted on the inside with an acoustic low-pass filter in the shape of a transparent ring, which does not begin to operate until 25 kHz are achieved, when it damps any material-borne resonance. The grille has the added advantage of improving omnidirectional sound distribution at higher frequencies."




Presumably this is on axis, but I didn't see anything describing the measurement technique from their website.



Edit: It seems augerpro tested these back in 2007 or so, but I don't think his website showing the results is up anymore. He did describe some odd distortion artifacts though.

R-Carpenter also tested these and presented the results on PE. https://techtalk.parts-express.com/f...e-25-sc-design

[JonMarsh]

I've got those both, have measurements somewhere around here... It is an interesting little part, but with some small issues as I recall... I may still have the tweeter and waveguide on a test fixture.

[JonMarsh]

I've been doing a little research on the DXT, you said back in 2019 (ish) with a Purifi 6.5 (Modula PSD thread) that it was a bit quirky - then showed a heckuva crossover:



Would this sort of network (i.e. one that has well over a dozen components) be recommended for the DXT? I only ask as I'd like the crossovers to fit in some pretty small cabinets! I understand it would not be this exact crossover, but if there are several bits that need 'tamed' I would expect they should be dealt with regardless of the woofer selection. The DA25TX appears a bit easier to work with (subjectively speaking) from other design implementations.

The DXT has a relatively low and fairly high level breakup mode. It can be addressed, as well as the quirks in the overall response... I'm expecting that the Kii Audio systems using this use all active compensation networks, but if they don't use a physical ultrasonic filter, then you can't fix the resonant amplification of distortion products in the tweeter.


To just "normalize" the DXT and provide a first order high pass to get it ready for a planned crossover requires a network like this:





Then you can wind up with an on axis response like this:



This is the raw half plane response of a DXT tweeter- you can take the route many have done in the past, and try to do a crossover with a single cap.



I don't recommend that, but then, no one made me Speaker Dictator for Life!

In the Kii inspired three way design, the high pass looks like this:





The two way system design you reproduced above includes elements for impedance flattening the amplifier crossover load, which also optimizes the current phase behavior. A lot of folks don't pay attention to those things (most, from what I can tell), but unless you've got a Halcro or Cambridge Edge W, amplifiers are happier with relatively smooth impedance loads and not too serious a current phase shift relative to the voltage output. Just sayin'. Food for thought.

[ergo]

Some time back (around 2016 apparently) I did a small bookshelf with DXT tweeter and Peerless HDS134.

Just posting the Xover for reference - these are still playing in the cottage "tiny house" and I'm still happy how they sound. So perhaps an option to to try and simulate

[JonMarsh]

I've been doing a little research on the DXT, you said back in 2019 (ish) with a Purifi 6.5 (Modula PSD thread) that it was a bit quirky - then showed a heckuva crossover:



Would this sort of network (i.e. one that has well over a dozen components) be recommended for the DXT? I only ask as I'd like the crossovers to fit in some pretty small cabinets! I understand it would not be this exact crossover, but if there are several bits that need 'tamed' I would expect they should be dealt with regardless of the woofer selection. The DA25TX appears a bit easier to work with (subjectively speaking) from other design implementations.

Just to be certain you "grok" some of what may seem superfluous in crossover components, R12, C12, and L9 form a series shunt zobel network, to stabilize the input impedance to the speaker at a more reasonable range. Most passive crossovers for the DXT I've seen online, like many other waveguide based solutions, have an impedance peak in the mid woofer to tweeter that like Ergo's, substantially exceeds 40-50 ohms. There is usually significant input current phase shift that goes along with this. A lot of amplifiers don't respond well to these conditions... especially the more low budget amps that might logically be used with a budget friendly speaker design.

BUT, as always, this is DIYW- Do I Your Way, and evaluate the tradeoffs for yourself.

[ergo]

Noted, this is indeed an issue with the Xover I've proposed as well.... Just for education, have you Jon ever tried an Xover with the big impedance variation and then a version that will tame it. How does it present itself in the sound character. I would guess that while it does stress the amp this would probably be a kind of sound influence that is hard to "put your finger on". What I'm thinking is that if I envision myself building a new pair of speakers with woofer/tweeter I don't know before one could easily also put this down to the driver "character". So a very good point and learning.

I don't remember too well anymore, but that pair o speakers did get quite a few xover alterations, so I did hunt for something, but can't remember what. Eventually was happy with this, but maybe I did change the amp along the way 🤔

I did need to rerun VituixCAD v1 again to open the project - takes one back and gives perspective how much good work Kimmo has done on VituixCAD

[technodanvan]

Thank you for the additional input gents! Every post is a learning experience for me.

[Jonasz]

Jon, what does distortion look like in the DXT with the ultrasonic notch?

[JonMarsh]

Noted, this is indeed an issue with the Xover I've proposed as well.... Just for education, have you Jon ever tried an Xover with the big impedance variation and then a version that will tame it. How does it present itself in the sound character. I would guess that while it does stress the amp this would probably be a kind of sound influence that is hard to "put your finger on". What I'm thinking is that if I envision myself building a new pair of speakers with woofer/tweeter I don't know before one could easily also put this down to the driver "character". So a very good point and learning.

I don't remember too well anymore, but that pair o speakers did get quite a few xover alterations, so I did hunt for something, but can't remember what. Eventually was happy with this, but maybe I did change the amp along the way 🤔

I did need to rerun VituixCAD v1 again to open the project - takes one back and gives perspective how much good work Kimmo has done on VituixCAD

Kimmo has done huge amounts of key work and detail improvements in VituixCAD, and that's why I have a standing donation to him every month.

Now, the obvious extreme comparison amplifier wise would be a Dynamo ST-70 or the MkIII's I have, versus my Halcro DM68's. BUT, a less obvious but also relevant comparison would be an low or no feedback design like the Ayre AX-5 versus Cambridge Edge W. Ayre specializes in zero loop feedback designs, solid state. There older models, before the diamond feedback circuit (something I was using in preamps decades ago) may have a nominal output impedance of 1/2 ohm, rising to higher levels at high frequencies. This is in vacuum tube territory.

(Note: Disclaimer of sorts- Charlie Hansen, the founder of Ayre, used to work for me as an electronic tech when he was studying physics at CU. He founded Avalon acoustics, and their first speaker had some similarities to prototypes I gave him after my collaborator William Keith Kennedy passed away in 1980 from an aortic aneurysm while I was at CES. He sold Avalon and later founded Ayre Acoustics. I did some design work for him including PCBs on the first Ayre preamp, and he did offer me a job, but couldn't afford me compared with what Siemens Semiconductors was paying. I still have a couple of pieces of Ayre gear, including a K5-x preamp that we came up with some more tweaks over the original design, based on a dynamic analysis of the FET cascode biasing impedance desirable (fancy way of saying adding some capacitance to the AC grounded gate to reduce signal related fluctuations- I.E., distortion.

The holy grail for the Ayre gear was total avoidance of TIMD (Transient InterModulation Distortion, which is a function of high feedback designs that enter into interstage limiting or clipping due to high internal signal swings when the amplifier is "stressed" externally. Ayre did this by abandoning any loop feedback connection. The Leach Amplifier, developed by William Marshall Leach, and designs from Dr. Matti Otala for Electrocompaniet took a somewhat more conventional approach with linearization of internal stages and using a high bandwidth reduced gain approach to the input and error amplifier design. )

And since nothing is real these days on the Internet without a picture (and now, with AI, even with a picture!), from my assembly work/bedroom, soon to be A/V Room:

[technodanvan]

Jon, what does distortion look like in the DXT with the ultrasonic notch?

Hey JonMarsh, just highlighting this for your visibility. I am curious what you have seen as well.

[letiennam]

I would like to contribute to your thread with my project. The speaker is in the process of voiceing. My project is progressing but a little slowly.

[JonMarsh]

Jon, what does distortion look like in the DXT with the ultrasonic notch?

Two different samples, and somewhat different notch/normalization filter designs- note HD3 levels. All at 2.83VRMS.

Plus a raw direct plot measured under same conditions, same test fixture mounting.

[Jonasz]

Thanks Jon, do you think the 3rd order reduction is actually audible?