Build Intention:
A “simple” two way speaker with the advantages of the Econowave, but deeper bass than my current woofers, and more directivity by using a compression driver/waveguide that extends down to a lower frequency.
Eventually building the speakers with “modular” cabinets, to make changing parts easier.

I’ve been enjoying my Econowave speakers but I know the Selenium D220ti is not as good as it gets with compression drivers – plus I’ve been quite impressed with the Crites speakers “Cornscala type D” speakers, a DIY two way variant on Klipsch speakers that use the Faital Pro LTH142 Tractrix horn as one of the variants.
These horns even have published directivity heatmaps (although only to 12.5Khz), quite unusual for info for the manufacturer to publish.

Although the Cornscala uses the HF140 compression driver, I’d read that the HF144 has a similar frequency response (usable from 500hz – 18/20Khz) but smoother sound as it uses a “ketone polymer” (whatever that is ) diaphragm material instead of titanium.

So a pair of HF144 compression drivers and LTH142 Tractrix horns purchased from usspeaker.com and eventually arrived..

The compression drivers installed with 66uF protection caps in series, to stop an inadvertant blast of LF signal frying them.

also after getting a bit of hiss from the HF amp I’m using (a hifimediy T2), I added some resistors in an approx 9dB LPAD (using 5.1 and 3.9 ohm resistors)to attenuate the output as well.

initial dissapointment – notch in the HF response

left and right compression drivers, no EQ on miniDSP or crossover:

I was disappointed to see on measuring these they had quite a notch in the HF response.. although above the range I can easily hear (about 15K) it was still disapointing when the compression drivers mention a range up to 18K and the manufacturers graph indicates smooth response in that region (but what level of smoothing have they applied?). In fact I stopped using the drivers and waited until Col came over with his WooferTester3 setup so we could measure the impedance curve so we could get a handle on wether the units were faulty – but they tested electrically as fine anyway. Also one of the units has a bit more of a HF rolloff than the other.
After the shock of my first measurements (with a fairly “zoomed in” and unsmoothed graph, using the same scale as the manufacturers graph, with a bit of smoothing, which Faital appears to have used in their graph, I get something closer to the “factory graph” taking into account low frequency room artifacts and rolloff due to the protection caps, but the HF notch is still there where it isnt on the “factory” graph:

As I know my hearing does not go up to the level where the HF notches are any more the notches are maybe not worth worrying about, but still it’s annoying when something doesn’t literally measure up to it’s promise.

On the other hand the units do sound very good, once I set them up with MiniDSP to flatten the response a bit.

Presently thanks to my MiniDSP setup it was fairly easy to add these with my current Econowave speaker cabinets to get started quickly.
I have tilted the cabinets on their side so I can have the tractrix horn as close to the Delta 12LFA woofer as possible.. it looks a bit awkward, but the sound is very good.
I’ve got it sounding clear and pleasant good with a range of crossover points from 500hz to 900hz, although one problem is room effects kick in below 700hz so below there it becomes guesstimation – until I get an opportunity to take the speaker cabinets outside to measure, which I may not bother with as they are only “temporary” until….

total response of Delta12LFA on its side, plus HF144/LTH141 combo, crossed at 900hz, with miniDSP eq. Gated. Measured @ 1M from WG :

total response of Delta12LFA on its side, plus HF144/LTH141 combo, crossed at 900hz. Smoothed to 1/6 octave. Measured @ 1M from WG – so this shows the room effects as well.

THE NEXT STEP: CHOOSING A NEW WOOFER
I’ve been looking at a range of 12″ and 15″ low frequency drivers to match to the new High Frequency part of the system. I started looking at 12″ drivers such as the Beyma 12B100 but have been convinced (especially as how good the HF144 sounds even crossing over at 500hz) that 15″ drivers are worth seriously looking at.
There are a vast number to check out, and I’m working on a shortlist of candidate woofers, and I’m in no hurry as the current rig with the Faitals sitting on top of the sideways econowave cabinets sound pretty good.

Well I had made a shortlist of all sorts of great drivers to work through the cabinet simulations with, including the B&C 15TBX100 (Italy), BEYMA 15B100/R (Spain), RCF LF15N401 (Italy), Selenium 15WS800 (Brazil), and even some from Lorantz (Dandenong, Australia), though the Lorantz ones were out of my budget, despite being locally made.

But I got very lucky with ebay and scored a pair of Lorantz 15″ C390X B1 for a great price. I’m told these are asbolutely amazing quality speakers.

Conveniently the auction even included some PA cabinets – though not great for loungeroom use long term at least I can get some instant gratification and get started with measurements and integrating the speakers before I start the big task of building some nice permament enclosures. When placed upside down they even position the speaker quite close to the waveguide, which is how I want the new cabinet to have it.

Not measurements yet. I’ve just quickly hooked up the bass drivers with some temporary banana sockets connected via the bass reflex port.

from the spec sheet of the manufacturer:
The C390X/B1 is An Australian made professional low frequency 15″ bass loudspeaker with a useful upper limit of 2.5 KHz. This model offers superb bass performance, high program power
handling, high efficiency, and therefore capable of producing extreme levels. The smooth response, wide frequency range, wide dynamics is attributed to features not often found in one model.
The C 390X range features rigid diecast aluminum frame, CNC precision components. The massive FE optimised ferrite magnet-assembly permit large linear voice coil excursion with efficiency.
Improved BL linearity, lower distortion , less wind noise is achieved with an undercut and flared vented polepiece. The stiff damped ribbed cone is product of our OFP technology and is molded in -house from a blend of premium air dried wood pulp and Kevlar fibres resulting in smooth controlled mid response.
The spider is made of Aramid material chosen for its high rigidity
and long term stability in demanding applicat ions. The accordion cloth cone surround assures extreme levels with minimal distortion.

I wanted to try some different waveguide options besides the FaitalPro LTH142 before I start making enclosures – and noticed the Pyle PH916 was a quite decent looking, inexpensive, candidate – but very little information about people using it at all on the web, unlike lots of other waveguides. But it is sold as a constant directivity waveguide, and in fact the waveguide I use in my Econowave speakers is a smaller model from Pyle so I gave it a try.


The shipping costs when ordered alone from Parts Express to Australia are quite high, but I noticed it was available on the Amazon website with much cheaper shipping, so I ordered it from there. I was suprised they offered shipping to Australia, but apparently they do now.

The waveguide does not seem as well made as the FaitalPro one, and is all plastic where the FaitalPro has metal reinforcement at the base where the compression driver bolts on, but it is made well enough for it’s purpose. It does seem a bit rough around the “mouth” and I think I’ll smoothen some spiky bits that seem artifacts of the moulding.

But it does seem to sound better, and measure better. In fact, for “head on” (no off axis directivity measurements yet), it doesnt have the 15khz dip that the FaitalPro LTH142 has. This is quite interesting.

For starters here is a measurement of the LTH142 horn + HF144, from 1M away, smoothed to 20/octave. Unless stated otherwise all other measurements will be done this way. Room effects take over at about 700hz so disregard the frequency response below there.
Crossover is at 500hz, using LR24db. One plot is with the miniDSP parametric EQ to flatten the response as much as possible, the other with no EQ. The 15Khz “notch” is quite visible – and would be even deeper with no smoothing applied:

Next is the same measurement, but with the Pyle PH916 (with and without EQ) – no 15Khz notch, in fact smoother abobe 10K. The Pyle tails off lower at the higher frequencies so I had to add about 3 or 4dB of extra EQ to compensate, but barring the notch I got the frequency response almost identical when measured from directly in front:

Lorantz woofer and Pyle waveguide: this is a BIG waveguide!

Just having a listen to music on this and I think it sounds substantially better. Now I need to do some measurements from different angles to get a handle on the directivity..

Also one of the waveguides seems to have a gap or space where the waveguide meets the body of the compression driver, maybe bolting it on differently will help here?

The Pyle specs state this is usable down to 1000hz but I am crossing over at 500hz, so it will be very interesting to see what it is doing in the 500-100hz region where I’m running it outside of the manufacturers recommendations.

Frequency response with holmimpulse gating in place instead of simple smoothing:

The MiniDSP EQ parameters I’m currently using to try and keep this flat – this is a rough first try:

closeup down the throat showing where it is not quite lined up on the left waveguide with the compression driver opening, it seems off centre. I’ll try rebolting that on from the other way to see if it helps?

More measurements to come: but so far I consider this an improvement to the sound.

OK, we know that simple on axis, straight into the speaker frequency measurements aren’t the whole story – getting a range of measurements at different angles gives a better idea for speakers getting general use, not the whole glued to the magic sweetspot that a lot of hifi has.

So it was time to do some directivity measurements. I don’t have the a licensed copy of ARTA so I can’t do the heatmaps ala Gainphile’s site – but at least Room EQ Wizard is a nice free package that makes it fairly easy to overlay lots of frequency response graphs in different colours.

To do the directivity plots, I took the entire speaker (box + waveguide) outside and sat them on a turntable made (with help from Col) from 2 bits of marine ply attached together by a lazy susan bearing purchased from Ebay. My backyard isnt huge so I could still have some environmental effects – but there is no ceiling and it should be bigger and less obstructed than the loungeroom.

The below measurements were done with the mic @ 2 metres distance, pointing direct into the compression driver. Then the turntable was rotated 7.5 degrees and another measurement taken, up until 60 degrees off axis. I then smoothed the response graphs to 12/octave to make them more readable.

I turned off the crossover, so lower frequency was only being attenuated by the 66uF protection capacitor – which gives a single order high pass filter @ about 400hz

I did leave in the miniDSP correction I was using for the PH916.. this was set up to give as flat as possible on axis response:

First I measured the PH916, with the following response:

Then I measured the LTH142, with the following response:

Now one gotcha here is I used the same EQ settings, and the LTH142 measures a bit brighter above 7khz than the PH916, so its a bit harder to eyeball and compare directly.
But the results are interesting, the PH916 has a closer spread below 10khz, so better directivity. But more chaotic dips above 10Khz. The EQ I am using has a big notch at 11.5khz – looking at the graphs this only helps on axis, and maybe makes things worse with the PH916 off axis.
For starters I’m now crossing over at 700hz to avoid the first dip at 600hz.

Also as a comparison I took a measurement of the Lorantz woofer the same way:

though this is using a 2khz LR24 cutoff. as expected, no directivity issues. Not sure what the huge dip at 280hz is??

Also to measure the angles properly I purchased a CRAFTRIGHT 300MM DIGITAL ANGLE FINDER (i.e. protractor) from Bunnings – this is a handy but of gear, I’ve used it to toe in the speakers in the loungeroom so they are 22.5 degrees off axis pointed at the couch (about 3M away) – this translates to a very slight toe in. For now I’ll use the 22.5 degree toe in as the basis for measurements.

I’ve been running through ideas for some cabinets for the Lorantz woofers.

The current fairly rough PA cabinets that came with them OK, but I’m not getting the full bass the speakers are capable of.


driver measurement:

First step was some measurement with WooferTester3 to get the Thiele Small parameters. I have the values from the Lorantz Data sheet for this driver, but as I could borrow Col’s WT3 rig I made the measurements myself.

My measured values were close, but not quite the same as the ones in the Lorantz data sheet. For one thing the F(s) measured was 37.68 Hz instead of 40 (this is actually a bonus as I’d prefer lower), and the Vas was 221.2L instead of 170L. But, I used the added weight method to measure this (piling a known weight of coins onto the speaker cone), which may not be the most accurate way.

Regardless, the cabinet size I’m looking at work equally well with either “factory” or “measured” data so regardless of which numbers are more accurate, I should be OK.

Measured values for the left speaker in WT3:

Measured values for the right speaker in WT3 (named “splash” as it has some cosmetic splash marks on it):

sanity test with 110L enclosure in the Lorantz specs
One encouraging result is if I run the measured numbers in WinISD I get a curve very close to that in the Lorantz data sheet:

comparing with running these same parameters in WinISD with the same scales chosen:

note: the strong line is with the measured T/S parameters, the weaker line with the ones from the data sheet. For the transfer function they do not vary more than 1dB.

simulation of current trap cabinet

I have guestimated using some trapezoid volume calculator web page that the rough volume of these current cabinets are about 90L, and they have two 8cm diameter bass reflex ports that are 6cm deep.
Playing around with WinISD it was pretty easy to come up with the transfer function of this cabinet, which shows it is tuned to about 54hz, with a very flat transfer function. So for what they are, these cabinets are quite OK for getting a lot of power out of the Lorantz speakers without making use of the lower bass frequencies they are capable of.
transfer function:

max power:

from the above I’m guessing I can drive these speakers with up to 340W each in these cabinets and they will still cope with it, at least if I stick to frequencies above 46hz?

idea for cabinet to build

Current idea is to have wide, not so deep cabinets something along the lines of the “Altec 19″ older style speakers boxes. Right now 60cm wide, 75cm high and 39.6cm deep. roughly including bracing and the volume of the driver this gives me an internal volume of about 138.9 litres.
I’ve already got a sample of “BB Pine” plywood from Allboard distributors which is the current proposed material for the build, using 18mm ply for the sides and 25mm for the front baffle. The woofer will be mounted as high up the front of the baffle as possible to be as close to the waveguide as possible.

Using this volume, and tuning to 33Hz I get the following transfer function:

Not very linear, but then I’m intending to cross over at 500-700hz anyway, and the resulting “bump” at 300hz may already cancel out the measured dip at this frequency (baffle step??).

The other price for tuning the enclosure this low is loss of power handling, but according to the WinISP graph I still have plenty of headroom for the 130W amps I am using.. and thats in the “never gonna happen” case of me running them at full power:

My current setup is with a LR24 crossover low pass filter at 500hz for the crossover, handily I can add this filter in to WinISD to get an idea of the “crossover + cabinet” transfer function.. although I’ll very likely be runnig with quite a bit more miniDSP correction anyway:

I’m still a bit of a neophyte at WinISD, UNIBOX (I cranked the numbers into Unibox as well and got the same results), so any advice appreciated, it’s all just an idea on paper subject to refinement at this stage.

Other “design axioms” I am working from here, that on consideration I actually need to read up on rather then let them remain as rules of thumb clattering around in my mind are:

• make the drivers (compression driver and woofer) as close together as possible. insert reasoning here.
• roundover of sides of cabinet.. the bigger the roundover the better, but I’ve got 25mm of ply to work with.
• baffle dimensions – wider is better – justification.. not just because I like how it looks.
• plywood better than MDF (insert link to justify this)

..and a crappy sketch to round things off with:

There aren’t many constant directivity waveguides for 1.4″ compression drivers, but one I had not been aware of until Paul Spencer and SvenR on SNA forums brought it to my attention was the 18sound Xt1464.

After reading up on it I found it online for sale (inc shipping) not too expensive at www.thomann.de. It did take ages to show up though, with very detailed notifications of the first and last 2 days of shipping, and nothing for the 3 weeks inbetween.

I’ve decided to read up on doing the crossover “properly” so also have ordered a copy of “Sound Reproduction: The Acoustics and Psychoacoustics of Loudspeakers and Rooms” by Floyd Toole, as I have the speakers in a working state already I can take my time and learn some stuff on the way.

One odd thing with the XT1464 waveguides – despite the claims of super hiogh tech the tolerance on the holes for bolting them to the waveguides seemed a bit off, I ended up having to file them a bit to get them to fit. Other than that they appear very well made and strong plastic though.

Now I’m building up quite a collection of 1.4″ throat entry bolt on waveguides:

Even with a “quick and dirty” adaption of the previous MiniDSP used with the Pyle PH916 they sounded not bad at all, and doing a “from scratch” config, choosing a 600hz crossover point I got them sounding very nice with fairly rudimentary measurements and MiniDSP / Room EQ Wizard autoconfig.

The current 139 litre volume of the cabinets I’m planning need a two 100mm diameter 24 cm bass reflex ports to tune to my proposed frequency of 33Hz.. Any of the bass reflex tube products cost quite a few bucks to send over to Australia unfortunately. Thanks to the useful suggestions of Stereo.net.au forum members, I got the idea of making my own from PVC water piping, using a heat gun to form the flares. But I’m not so handy with being able to fashion them presentably well, so I went with a combined approach of using some less expensive fixed lenth 200mm long ports (the same kind from Altronics I’ve already used), with some 90mm PVC piping to make short extensions with flaring on the end.

Lucky I still had the offcuts from the ones I had shortened for the Econowave speakers so I could use that to check the dimensions were OK:

It was hard to get them consistantly identical, but I think they will be close enough when it comes to frequency measurements:

Unfortunately, although it is still mentioned on their web page, Altronics no longer stock this model of Bass reflex port. But I was able to find it (model number HP100) on a German web site with reasonable shipping costs – in fact including shipping it came to only slightly mroe than it would have cost for Altronics. They are taking quite a while to arrive here though.

Heating and forming the PVC plastic took a while to get right:

though I was careful to do it all outside in a breeze as I’ve been told the fumes from heated PVC are not healthy at all.

Lucky the jar and milkshake glas were just the right dimensions for forming the shape I was after.

After a very long delay I’ve finally gotten around to working on better cabinets for the Lorantz woofers that will unleash their full potential. Having adequate enclosures so I could be lazy and use the speakers as is no doubt contributed to the delay, but also some non audio related stuff got in the way as well, as happens.

boards assembled to check but not glued yet

I ended up using a wider design, 60cm wide, 75cm high and 40cm deep. Using the PVC port extenders described in a previous post I calculate I can tune to about 33Hz, or if I remove them it will be at about 37Hz.. this gives me some tuning options.

The cabinet will use 25mm pine plywood for the front baffle and 18mm for the other sides and bracing.

With 3 internal braces, plus the estimated volume of the driver and BR port, I get a volume of 131.3 litres. This should give me the bass extension to make use of the 30Hz rated range of the driver (well down to 33 Hz), at the expense of throwing away power. As I’m using these in a domestic situation they are never going to need the 400W they are rated at, the 150W these cabinets can give should be more than enough with 99dB sensitivity, especially since I’m driving them with an approximately 110W Tripath amplifier.

I had the plywood cut at Allboard Distributors who have done a good job.. one minor error on MY part was I wrote down a slightly incorrect dimension for one of the internal braces, but luckily this should be easily fixed by adding a bit of offcut to close the gap, as by dumb luck my error was out by 18mm.

Although I’m a computer person I couldn’t be bothered learning CAD just to spec out some wood so ended up doing it the old fashioned analogue way:

Next stage.. clamping and glueing, and adding in the holes for the driver screws and binding posts, and the stuffing, after reading up the ESP pages on the matter in more detail.

At long last the enclosures for the Lorantz 15″ woofers are (mostly) done:

Big thanks to Col for helping with the build – his expertise gleaned from the multitude of enclosures he has created was a big help in streamlining the assembly process.

Assembly:

Most importantly the specifications I had given for the rebated hole for the woofer was spot on: the driver fits in the hole perfectly:

The driver was attached using M6 “t-nuts” and Allen bolts purchased from ebay, these worked out really well. The T-nuts were stapled on so they dont lift off the baffle while bing unscrewed:

Then the tricky bit of actually glueing – both cabinets used almost an entire 500ML bottle of Selleys PVA wood glue. Everything has mostly lined up, not absolutely perfect but nothing that a little bit of sanding wont fix, more importantly the enclosures ended up sound and airtight. The 4 corner strap clamp turned out to be a major lifesaver here, and the 5 boxes of spare floor tiles turned out handy as well:

A picture of the dried enclosure the day after – this shows a detail of the bracing, and where we fixed the mistake in the upper brace dimension by adding a chunk of 18mm plywood offcut into the gap. Maybe this is actually a unique “Audiophile” design feature and not a mistake?

First layer of stuffing added:

Binding posts added, these are nice chunky Dayton Audio BPA-38G purchased from local DIY vendor theloudspeakerkit:

I wanted to move away from the plastic plate style used before and go for these ultra chunky, simple ones. It does mean I have to be careful about bashing them as they jut out the rear of the enclosures.

A short run of chunky speaker wire (can’t remember the type but it’s the heavy duty stuff from Altronics or Jaycar) added, as well as an additional rear layer of Clark Rubber convoluted foam stuffing against the back panel:

Bass reflex ports with PVC pipe flared DIY extenders added:

..with some gaffa tape to keep the extenders attached. Excessively documented here. The specified cutout hole size of 138mm turned out to be spot on, these just tapped into place with a mallet and fit perfectly.

Then the drivers themselves were mounted in place, and the system set up with the compression drivers and horns – already looking a lot better in the loungeroom than the old PA cabs:

The two steps left to be done are:
a) some finish on the wood to make it look nicer – the unfinished plywood is still a step up from before, but some sanding, varnish and wood stain should really make these look like long term furnture items and not some nerd construction project, I hope.

b) miniDSP adjustment of the crossover. Now the cabinet is complete this can be done properly. I’ve already got some ideas and have done a “quick and dirty” miniDSP setup that is getting reasonable results but it’s worth hitting the books and reading up and doing it properly.

how does it sound
Great! Well it’s early days yet, but the bass definately extends lower, and is much less distorted in the lower frequencies. These cabinets are really getting the full bass capabilities of these speakers, at the expense of not being able to utilise their maximum power. But even with the power I can use I can get it to earsplitting levels in a domestic listening situation so thats OK. I expect this to only get better once the measurement and crossover refinement is done.

further technical stuff
Here is the winISD modelling of the transfer function of the 131.3 litre cabinet size eventually used (approx, given the guestimation of the volume of the driver itself and bass reflex ports, but probably close enough). Tuning is @ 34Hz, which seems to match the approximately 24cm length of the extended bass reflex ports.

and here is the winISD modelling of the max power:

as can be seen from the transfer function the extended bass does have some gradual 7db rolloff below 600Hz (remember this system currently crossing over at 600hz), but the miniDSP EQ functions can be used to correct for this.

It took me a while to get to adding some finish to the speakers, but I finally did using some “Cabots Satin Jarrah Water Based Stain and Varnish“.



Although it looks OK at a distance I’m not completely satisfied, its a bit blotchy and has some drips that show up when you look more closely, especially at the sides and back.

I followed the instructions and sanded between coats but it was quite hard to judge how much paint to apply so I ended up with some drips unfortunately.
The solution I guess is yet another sand and coat of varnish, but I’ll wait until warmer weather and do outside instead of in the garage so I can see better, and use a roller instead of a brush so I apply the varnish more evenly.

Both before and after I went overseas I’ve been experimenting with the most excellent Active Crossover Designer spreadsheets by Charlie Laub, with a lot of input and information from the Stereo.net.Au forum users, especially henry218. It’s a lot of fun to upload some measurements and get some crossover configs and ideas from helpful people and find they work! Later I’ve gotten to grips myself with cutting and pasting miniDSP biquad settings. There is quite a bit of info posted in the relevant thread, but this remains a work in progress while I improve the measurements needed to get the best results.

Also I’m quite curious to see the improved ACD that Charli Laub has in development that will include heatmap graph capability for evaluating directivity.

With the woofer enclosures now looking pretty good it was time to tidy up how the compression driver is supported, and tidy away the LPAD and protection circuitry.
I got some small pine boxes of the right size (from ETSY seller HIPROGRESS. I stained this the same colour as the speaker boxes, and got some self adhesive 4.5mm neoprene sheet from Clark Rubber to act as vibration damping, and protect the finish. This acts as a nice firm bed for the compression driver.

I got some 12AWG speaker cable and banana plugs from Selby Acoustics to redo the LPAD and protection cap wiring, mounting them inside the box with the cables threaded out through a small hole. Now the whole rig loooks a lot tidier. I was originally thinking of making a wooden enclosure for the waveguide, but I quite like having the compression driver and horn exposed, and they look OK now that the wiring is tucked away and the bricks are gone.

before:

after:

mounting the parts in the box:

some dabs of hot glue added after to keep it in place and stop the components rattling about.

Not strictly related to the speakers themselves, but as the miniDSP is a core component in driving the speakers I’ve noted this in the “beyond-econowave” posts:
For a long time the miniDSP was sitting bolted to a lump of wood, the “temporary” test chassis, with a cardboard box sitting on top of it to cover it. But this test chassis was quite noisy.

I’ve made a simple pine enclosure that makes up in solidity what it lacks in sophistication. It still containes the miniDSP and 2 simple OPA2134 based pre-amps to give me a bit of gain over the quiet miniDSP output, but I’ve got a 4 channel stepped attenuator (actually two stereo stepped attenuators ganged together) as a master volume, plus the software volume control as a fine volume pot.

There’s also a switch to switch over to the analogue input for “guest devices” .e.g. mates dropping over with DJ mixers or synths – so the sockets are on the front for easy access.
Made with bits of pine from the local hardware warehouse, and painted with leftover stain/varnish from my big speakers so it matches.

The tricky bit was mounting the two preamp PCBs (which didnt come with mounting holes) in a way that they were braced down firmly to cope with the stepped attenuators being turned.. solution some clamps fashioned from “polymorph plastic” and brass bolts to hold the PCBs firmly in place.

As the USB input on the miniDSP I find is super sensitive to USB cables quality, so a USB extender would be unreliable, and I couldnt think of an easy way to provide an SP/DIF optical socket on the outside, the left side has a simple “porthole” in the side to access the optical SPDIF (primary input) and USB (for making adjustments to the crossover).

Jaycar linear PSU is mounted inside the steel box from a PC power supply at the rear of the box, keeping all the mains/power stuff isolated from the rest of the circuitry.

Just to recap what is in the box:

These speakers have been seeing daily use in the ZOG lounge-room, are essentially complete, and continue to sound great.

three small upgrades done since the completion have been:

• addition of JBL/Selenium ST200 Supertweeters.
• low frequency amp upgrade to TPA3255
• replace the miniDSP “classic” 2×4 and op-amp boost with a miniDSP 2×4 HD.

ST200 Supertweeter

One small bit of experimentation done was to add some Selenium/JBL ST200 supertweeters. This was to get a bit more high end above 10khz without relying on extreme miniDSP frequency boost to try and push the higher end. They have been added to sit next to the main waveguide on rough and ready L-shaped stands. A simple capacitor has been added to bring in the ST200 above 10-12Khz using a basic first order crossover. These are not particularly expensive or “high end” devices but the specs are a good fit and they have been left in the system now for 4 years as they sound OK.

TPA3255 amplifier

This was a quick build using another “found object” enclosure – from an old busted home theatre amplifier kindly donated by a StereoNet.AU forum member.

At the time the Texas Instruments TPA3255 Evaluation board was on special so I ordered one and made a quick, but functional amp build using a Meanwell LRS-350-48 switch mode power supply. This has turned out to be a great amplifier to drive the Lorantz bass drivers, so has been left in the system.
I haven’t bothered documenting this build too much, and building the amp from the evaluation board is fairly straight forward. There is masses more information in threads like this in the DIYAUDIO.com Class D amplifier forum.

miniDSP 2×4 HD upgrade

The miniDSP “classic” crossover has been working great for years but always had a few annoyances. The more recently released miniDSP 2×4 HD looks like it addresses a lot of these issues so I got one and upgraded to it recently.

One initial issue is I could not copy over exactly my old config. They are quite different devices. Although most of the settings can be copied over by putting the two configs side by side and working through them, the biquad filter definitions do not copy over directly as the miniDSP HD has a 96Khz sample rate, while the older miniDSP has a 48Khz sample rate. But it turns out just “eyeballing” the resultant EQ curves and making ones that matches as closely as possible using basic filters gave me a miniDSP 2x4HD config that measures almost exactly the same as the old one, and sounds pretty good.

eyeballed config and previous config, back to back measurements at about 2M on axis:

A later project if I feel like it will be making a new miniDSP HD config from scratch that takes advantages of the additional features provided by the better software and CPU in the new device.

Also I don’t need to use intermediate preamps any more to boost levels – the miniDSP is plently loud enough to drive the amps and loud volumes. But it is a bit noisier than the previous setup with the TDA7297 15W “lunchbox” amp – preamp “hiss” audible within 1M of the waveguide, though not at the listening position. The noise levels however are quieter with the JLH 1969 Class A amplifier (different gain settings). So still usable, but not as quiet as the old one, but the two volume controls on the old one caused usability problems.

The remote control volume/mute button and 4 selectable saved configs is also handy, so I can switch between configs suitable for different amplifiers without getting the laptop out.
I got a small 12v linear PSU to drive the miniDSP just to power it with something a bit cleaner than the wallwart it came with, though this PSU made no audible difference to the hiss levels,