Shahinian obelisk rebuild project

[eb15]

Today I got a pair of used cabinets with crossovers of Shahinian Obelisk speakers that I would like to rebuild with new drivers, passive radiators, and crossovers.



I thought the obelisks were an interesting speaker back in the 80s but much too expensive for me back then. Shahinian has gone through several revs through the years with differing woofers/passive radiators, tweeters and mid-domes on top in 3 way versions of the obelisk.

This is a very old version which had 3 tweeters, an 8" woofer and a 10" passive radiator on the back.
Newer versions added a driver or two on the back face of the pyramid top which this doesn't have.

Internally there's a board diagonally across dividing the box into a transmission line that folds
at the pyramid top.

The internal cross over is very simple with a fuse, one 6.8uf film capacitor and an inductor which doesn't
seem to have any obvious markings on it. The three tweeters were connected in parallel.

I would be interested in opinions of what to stuff in these. I have no problems soldering, etc, but am a little woodworking challenged, though I can probably ask someone local for assistance if necessary.

[---k---]

Those are very interesting. They break many of the common rules of speaker building. Based on the crossover being just a fuse, it is seems that they built a pretty box and threw drivers in it and just adjusted the tweeter level to balance out the woofer.

Going from an 8" woofer to a tweeter would normally say you want a tweeter with a very low crossover and wideband woofer. You'd also want to minimize the woofer to tweeter distance - which you can't do. The three tweeters on top may cause some comb filtering and just weird effects. But, maybe in a omni-layout it won't be so bad.

So, I guess the questions are:

1) what are your expectations? Do you just want to reuse a cool looking box with some new drivers and a crossover that generally gets the levels correct and sounds good enough but not as good as many of the published designs? Or are you looking for a world class speaker with ultimate drivers and crossover?

2) what is your budget? Are you looking at Dayton RS level drivers or Scan Speak Drivers?

3) do you want to reuse the box exactly as is? do you want to re-use all three tweeters? or are you considering just one? are you going to re-veneer?

4) Do you need drop in replacement size drivers? If so, what are the opening sizes. That metric versus english thing might come in to play when talking generically about 8" woofers.

Seems like an interesting project.

[eb15]

I'm sorry if I didn't state it clearly The original crossover is a first order crossover.
one inductor to the woofer. one fuse and one capacitor to the tweeters.

From what I've read newer obelisks went with a three-way design and more complex crossover.
Here's a photo I found on the net..


I'll have to measure the holes to verify things will fit or that I'm able tweak the cabinets to fit things.
I'm haven't yet received the grills, so that's another part of the project.

I'm considering the 8" SB acoustics aluminum woofer that Jon mentioned in the '8" woofer crossed over to a 28mm tweeter' thread or siimilar drivers as used in past two ways. HiVi m8a, Dayton RS225. It looks like there are screws holding the pyramid baffles in place, so I might be able to try multiple configurations of upper drivers if I can fashion new ones. If not I'd probably use three 4-ohm tweeters connected in series, or cut a fourth hole for another tweeter and series/parallel connect 8ohm units to keep the impedance up.

If I go with a three-way, I'd probably get the Dayton RS-52ANs and try the Dueland crossover approach Jon has in 'Proposal for a compact three way' thread.

[eb15]

correction

I unscrewed the crossover and tweeter baffle off one to take a closer look.
I was mistaken, the crossover hard wired the woofer (except the fuse on the negative input terminal for all the drivers). It is a second order network to the tweeter.

The capacitor measured about 7uF
My multifunction meter meter couldn't measure the inductor's value.

I took some pictures after I put new batteries into my camera.

[eb15]

I measured the driver cut outs today.

tweeters -- 87mm inner cutout, 117mm outer diameter
woofer -- 180mm inner cutout, 223mm outer diameter
passive radiator -- 221mm inner cutout, 250mm outer diameter

It seems that I'll have to enlarge the passive radiator and woofer cut outs.

A quick look over available drivers has ScanSpeak passive radiator 23W/0-00 as closest fit but slightly too big
Dayton RS225-8 woofer is also slightly too big for the woofer hole.

[BClarkeVA]

EB15, as a Shahinian Arc owner I'd love to find out how you finished this project up. And specifically interested to see what 8 inch woofer you used. Recent research indicates the current OEM passive radiator is now the SEAS Prestige SP26R (H9946) 10" Passive Radiator.

[eb15]

Sorry, it remains an unfinished project. I never went further with the obelisks.
I can't believe its been so long, with me always thinking I'll get to them in a month or so.

[eb15]

Reviving this old project...
I got assorted drivers in to evaluate for this and other projects, and
I'm awaiting delivery of an inexpensive loudspeaker measurement tool that I ordered thru AliExpress.


I hope it will work well with REW measuring software and VituixCAD2 modeling software to help me develop this and other speaker projects.

For the obelisk cabinets...
It looks like I'll be using all Dayton drivers that fit the current driver cutouts without additional work beyond drilling new mounting screw holes.

Dayton DSA215-8 woofer (DA215-8 also fits, RSS210HF-4 is too big to fit cutout)
Dayton DSA270-PR passive radiator will fit mounted on cabinet exterior. (RSS265-PR is too big to fit existing cutout)
Box is approx 40 liters. I'm not sure how much weight I should add to the PR, and haven't modeled the passive radiator loaded
transmission line with hornresp either.

Dayton DC28F-8 or shielded DC28FS-8 tweeters
I ordered the non-shielded one to compare to some other tweeters on hand and found it sounded OK and fits perfectly, screw holes and all...
then poof it went out of stock til late November, so I have 6 DC28FS-8 tweeters on order, which are supposed to be equivalent plus magnetic shielding.

Studying the original crossover, I see it actually ran the (paper cone) woofer full range and had 12/dB high pass on the tweeters.
With the aluminum Dayton woofers I'll have to decide how I want to low pass the woofers.

[JonMarsh]

Software such as Unibox in Excel or VituixCAD (free, doesn’t require office) would permit modeling the performance and figuring out the best mass loading of the PR. If that is difficult for you, just let us know what the internal volume is, and I should be able to model the PR system

[eb15]

2023 update...
This project is coming along well.
One cabinet has drivers installed. grill covers re-installed over new woofers and tweeters.
Tweeters were installed parallel connected with meniscus 16ga twisted pair wires
and externally 4" felt rings from Madisound were applied around the dome tweeters.

These seem almost seemed tailor made for this application where they cover the tweeter mounting screws and the slightly recessed
to the baffle edge, and are quick peel and stick installation.
A further step will be trying more baffle covering felt/foam or thicker felt pads around the tweeters.

The cabinet volume, dsa215-8 and dsa270-pr was modeled in unibox, with no washers was close to a flat response.
I picked up a 30 pack of fender washers at homedepot to use as weights, divided between the two speakers models similar to an EBS6 type of alignment.
I can't properly seal up the cabinet until the crossovers are internal.

Shahinian moved to an 18dB/octave tweeter crossovers with their obelisk updates, and adding a low pass network with their woofer replacements,
so I decided to try modeling something similar for the Dayton drivers.

An initial test REW software and JBOX hardware measured the installed drivers from an approximate listening position and then exported impedance measurements
and 1/6 octave smoothed SPL into VituixCAD for crossover simulation with off the shelf parts values.

Note the tweeter measurements were for 3 paralleled units pointed at different angles due to the pyramid baffle but measured
as on axis from the plane of the box front at an approximately 39 inch height, and impedance as they're parallel connected to act as one for the crossover.

Its my first attempt measuring speakers, so I probably made some rookie mistakes, but maybe good enough for now.




The VituixCAD simulated crossover parts and some perf boards to prototype crossovers on were ordered from Madisound and Parts-Express.
Quickly clipped together It sounds reasonably smooth and detailed with Clarity Cap PUR capacitors, Jantzen litz 15gauge inductors, and Dayton 20W resistors.

I also ordered some other inductors to evaluate if I hear any obvious differences between their differing construction in this 2 way use case.
The litz inductors had pre-tinned ends so were quickest to connect up. I'll have to pick up some sandpaper to prep the coated wire inductor ends.

Another area Shahinian updates their obelisks is to reduce cabinet vibration with what they called a 'harmonic brace' and later units were built with
a high quality ply sheet in place of MDF.

I'm going to see if I can improve these cabinets by dampening the walls internally with some FatMat 70mil black butyl self adhesive sound deadener.

[eb15]

here's some pics of the dayton drivers installed

[theSven]

The restore of these is really coming along! I enjoy reading up on the progress you're making.

[eb15]

FatMat sound deadener arrived. It came as two sheets rolled up with a cutter and roller to help apply it.
It has no foil side, just the butyl and backer paper.

Its easy to cut with scissors, peel off the white 'paper backing' and apply to the interior walls and baffle.
I'm using one sheet per speaker cabinet, and I'll probably buy some more for some other projects.



This seems to have removed or at least reduced some 'ahhh' boxy coloration to the speaker's sound increasing overall clarity.
I do think the cabinet could use some more bracing for the lower frequency resonances, but this is good for now.

I wasn't careful to measure and cut beforehand so its a little messy.
It also would have been easier to cleanly fit and install before the dayton drivers were mounted.
I'll recommend the fatmat deadener for other projects.

Next steps here will be trying some other external tweeter baffle treatments with felt or foam products and
evaluating interior volume damping foam/fiberfill/etc. for the woofer chamber.

[eb15]

Some felt pads from the dollar store to help reduce diffraction off the woofer cabinet lip and to damp the metal grill.
clean new acoustistuff similar to original damping material in woofer side chamber loose fill.

Earlier I tried some 1-inch high density foam from the sewing store, but it was a sonic failure. too much? overdamped the low end and sucked the life out.

[theSven]

It's good to do testing like that to see how performance is impacted. The dollar store is super helpful for little things like that. I like to get foam brushes from there for all the clear coating I have been doing on wood at this time. It's to cold to spray so I have to resort to foam for now.

[technodanvan]

Maybe it wasn't open-cell foam?

This is a fun rebuild!

[eb15]

inductor evaluation this morning. first on the high pass network for the tweeter array.

A jantzen litz or 18AWG seemed to have a slight mid/treble coloration (15awg litz) or compression (18awg)
not present with the Solen 0.47 mH Perfect Lay Inductors 14 AWG.
Then the low pass circuit for the woofer.

jantzen 15awg litz, 15awg solid, vs Solen 2.0 mH Perfect Lay Inductors 14 AWG. Harder to hear and describe audible differences,
but I'm gonna call it for the Solens for both uses in this speaker.

[eb15]

I decided to order some alternative crossover inductors, capacitors, and sonic barrier 1-inch foam available from parts-express, and today the order came in.

Testing a couple cored inductors requiring unrolling them a bit to hit the correct nominal value, but in measuring inductance
values on an old sencore Z meter I found the previously preferred solen .47mH inductor was under at .417mH and its twin ,419.mH
but the Solen 2.0mH in the bass section was 2.1mH and the Jantzen litz inductor was .472mH and the Jantzen solid wire .457mH. a Dayton foil inductor .445mH

I can't be sure if the preference was due to the value change or the different construction. hmmm.

OK, I unwound the 18ga Jantzen .47mH and the two cored inductors to match the measured inductance of the Solen plus or minus .002mH. and the 14AWG Solen still sounds better, less 'congestion' more dynamic, blends more cohesively like the midrange of a full range electrostat and some vocals can. sound spookily real even with the large spacing between woofer and tweeter array. The litz and foil inductors are a little more work to unwind than I want to do right now and I'm guessing they'll still have 'issues' if I did.

I wanted also listen and see if low cost bipolar electrolytics, Dayton dmpc, or Jantzen crosscaps could make acceptable budget substitutes.
I needed 6.8uF, 20uF, and 27uF values so that limited what is affordable.
The bipolar were mostly inoffensive, but lacked the clarity to draw me into the music I've heard hundreds of times.
The Dayton film caps were different but not significantly better to recommend them.
The Jantzen crosscaps started to have the desired affects so they're my budget choice in caps here.
The more costly clarity cap PUR remain superior even in this budget 'audiophile' system I've put together.
Macbook air or PC laptop with old CDs ripped into lossless format
usb input of geshelli jnog2 DAC (akm chip version)
Ayima A07 power amp (mfr upgraded opamps LM4562NA US plug version thru aliexpress)
hosa interconnect and speaker cables (from amazon) with changed out bananas since the amp puts the connectors too close for standard double plugs.

[eb15]

I've been evaluating more components for use in crossovers.
first up inductors.
The 16ga solen perfect lay inductors seem to have slightly better "open" midrange and trebles while being comparable to the thicker gauge siblings. smaller, cheaper, win/win.
Of the inductors I ordered the Solens seemed to sound the best. The bigger ones might be more 'solid' sounding but the 16ga seemed to be better in this use case.

second up resistors.
The Mundorf 20W m-resist seem to remove a slight midrange fog with the Dayton 20W wire wounds.
Some expensive 10W cast reststors were tried, and sounded perhaps a little laid back and 'faster' than the Mundorf, but less dynamically interesting and the value measured higher than specified. 2.2ohms from 1.8ohms specified as measured on my fluke DMM where the Mundorfs were right on the value.
I'd say the mundorfs were worth the upgrade cost in this case (series resistance element in tweeter crossover), the cast resistors different and questionable.

third up capacitors. ( note: no small value bypass caps were used in these evaluations 6.8uF, 20uF values (10uF in parallel with 10uF where necessary)
The Jantzen Z-standard seems to be a nice clarity upgrade from the crosscap so that's probably the budget pick now with the updated inductors and resistors.

Jantzen Z-superiors and Mundorf supreme are better differing in which area they seem to have better performance. midrange to Mundorf, trebles to z-superior or is it the recording?
having to parallel 2 10uF caps to reach the 20uF value makes these caps a pain to work with size wise.
Clarity Cap PUR somewhere near those two in quality but smaller and easier to manage. I would be happy with these except for...

Auricap XO capacitors sounded the 'fastest' almost like I'd expect from using a better tweeter. Maybe not as 'full bodied' sounding as the above two, but the 'speed' gets my attention in a good way.
Auricap XO is what I'm going to use here. They're also a little smaller than PUR caps, though the weird twisted pair leads will mean a little more work in installation.

[eb15]

I finished some listening tests of a couple more 'audiophile' resistors in a common value (1.5 ohm) in the tweeter series position the 10W Mundorf MResist Copper Nickel wirewounds were almost as good as the 20W, and preferable to Mills 12W wirewound. I'd recommend the 20W mundorfs where practical, 10W MResist wirewound where the cost savings and lower power rating is ok. The 1.8ohm value is preferred in this application for the best crossover 'blend' to the woofer.

[eb15]

I tried flipping the woofer cabinet upside down to get the woofer up off the floor a bit like a stand mounted bookshelf speaker.

This would be the way I'd go if I were designing new cabinets for a shahinian-esque 2-way speaker.
You get closer to a proper blend with the closer spaced drivers and woofer off the floor is better for a two way.

Back in the 80s I remember an audio shop doing demos of bookshelf speakers on stands vs directly on the floor,
then the sales person would lie down on the floor saying 'you don't normally expect the sound of instruments or a singers
voice to originate at floor level'

In family room settings, having the woofer up higher like this it puts the most of the woofer above coffee table height
so its not blocked and the angled tweeters aim up and away from the coffee table top surface so they're not reflecting off it.

[eb15]

I decided to try implementing more baffle step compensation...
Playing music quieter has a better balance now.

Increased woofer series inductor to 2.5mH to add another 2.5-3dB BSC
and increased tweeter series resistor to 3.3ohm to lower level of tweeter array accordingly.

[theSven]

I decided to try implementing more baffle step compensation...
Playing music quieter has a better balance now.

Increased woofer series inductor to 2.5mH to add another 2.5-3dB BSC
and increased tweeter series resistor to 3.3ohm to lower level of tweeter array accordingly.

Does it still sound good when you turn it too? Or is your preference to listen at lower levels only?

[eb15]

I think it sounds good for as loud as I play it.
I thought It was a little light sounding before.
The speaker is out in the room quite a bit from the wall.
I've turned it up, but It has obvious driver limits and/or the little a07 amp starts clipping,
This redesign has not been optimized for playing really loud.

[Frankreich]

Thanks for this very informative post. I have recently acquired a used pair Obelisks which I wanted to update. However I‘ve got stuck at the very beginning, as I was not able to remove to pyramidical grill. As you obviously removed them successfully I would like to ask how you managed to do so. Thanks in advance for your feedback.
Cheers
Frank

[eb15]

The seller had parted out all drivers before I saw the cabinets for sale on eBay so the perforated metal pyramid grills
and the round cloth on a plastic circle frame grill for the woofer were already loose.

It was just a friction fit as far as I know, but there are multiple versions of the obelisk so things probably have changed on newer editions.

[CH999]

Reviving this old project...
I got assorted drivers in to evaluate for this and other projects, and
I'm awaiting delivery of an inexpensive loudspeaker measurement tool that I ordered thru AliExpress.


I hope it will work well with REW measuring software and VituixCAD2 modeling software to help me develop this and other speaker projects.

For the obelisk cabinets...
It looks like I'll be using all Dayton drivers that fit the current driver cutouts without additional work beyond drilling new mounting screw holes.

Dayton DSA215-8 woofer (DA215-8 also fits, RSS210HF-4 is too big to fit cutout)
Dayton DSA270-PR passive radiator will fit mounted on cabinet exterior. (RSS265-PR is too big to fit existing cutout)
Box is approx 40 liters. I'm not sure how much weight I should add to the PR, and haven't modeled the passive radiator loaded
transmission line with hornresp either.

Dayton DC28F-8 or shielded DC28FS-8 tweeters
I ordered the non-shielded one to compare to some other tweeters on hand and found it sounded OK and fits perfectly, screw holes and all...
then poof it went out of stock til late November, so I have 6 DC28FS-8 tweeters on order, which are supposed to be equivalent plus magnetic shielding.

Studying the original crossover, I see it actually ran the (paper cone) woofer full range and had 12/dB high pass on the tweeters.
With the aluminum Dayton woofers I'll have to decide how I want to low pass the woofers.

Thank you for identifying a suitable drop-in passive radiator. I got the Daytons and replaced my worn original radiators. I wasn‘t able to remove them without damage. Maybe the weight of the original cone is useful to know as I used this as a reference point for the weight added to the Daytons. What can I say, it works very well. See attached photos.

[JonMarsh]

Great that it seems to be working out for you. When one has enough info on the woofer and PF specs, it's possible to do an analysis in tools like Unibox for Excel and VituixCAD, among others, and model the alignment. You've taken a very pragmatic approach, given resources, and it seems to be working for you!

Congrats, and

Happy New Year!