New Open Baffle Project

[Ali Tait]

Good result Simon, enjoy.

[Docfoster]

UPDATE
I have now added the Monacor SPH-210 bass-mid drivers to the mix. The mid/tweeter is running from 5000khz with a 1st order XO. The 15 inch SP-382PA crosses over to the bass/mid @ 160hz with a 4th order LR XO implemented on the miniDSP. All drivers are being run flat with no EQ or level adjustments for any driver.

The whole thing seems to integrate really well. Soundstage is very 3D and instrument position is pretty good. I feel that it's good enough to warrant a nice baffle now.

On a side note, the Monacor SPH-210 is a fine driver and could be a great alternative to a doc-modded 8 inch Skytronic / QTX type driver. It's flat to 4khz with a nice smooth roll-off, especially with the blutak worm around the dust cap join. Efficiency is very decent at 90db, and a low Fs of 28hz seems to give a nice bottom end.

Especially pleasing is the performance of the TPA3116 Class D amps. Unless I have a radical change of mind, the valve amp will be heading for the World's favourite auction site.

Loads of useful information there. Thanks. Must bookmark!
So you have no attenuation at all on any of the drivers? Not even the tweeter?

[Simon Hickie]

Yes, no attenuation on anything. The sensitivities just seem to match. If I was to attenuate anything, it might be just up to 2db on top section (8 inch & FR mid/tweeter).

Output on the bass section is pretty solid at 40hz with some output even at 30hz. The overall sound seems less 'forced' than with the typical box speaker.

[Ali Tait]

Yes, the bass just seems to breathe better with an OB.

[Simon Hickie]

One thing that has become very obvious with the miniDSP is how very EQ small changes of as little as 0.5db can have quite significant changes on the voicing of the speakers. For example, I've applied a mild 'S curve' to the response zeroed at around 1khz with about 1.5db of lift towards 100hz, and 1.5db of cut towards 10khz. Adjusting either or both of these by just 0.5db in either direction is easily audible.

I've resisted the temptation to do any measuring with anything other that my ears!

[karatestu]

Enjoying your thread Simon. MiniDSP is something i have never played with. My purist beliefs probably wouldn't let me Silly isn't it. I won't even allow myself a smartphone.

[Simon Hickie]

Enjoying your thread Simon. MiniDSP is something i have never played with. My purist beliefs probably wouldn't let me Silly isn't it. I won't even allow myself a smartphone.

Yes, I don't have a smartphone either - hateful things! A month ago I would have agree with you re. the miniDSP. However, the clearly audible differences are quite remarkable. I need to dial in another -0.3 db on upper-mids - what chance would there be with traditional caps, coils & resistors!

[CN211276]

My smart phone is for playing music only via four apps, Qobuz, My Cloud, BubbleUp and Audio Player Pro. Would not dream of using it for anything else. I have an old style pay as you go phone for calls.

[Docfoster]

One thing that has become very obvious with the miniDSP is how very EQ small changes of as little as 0.5db can have quite significant changes on the voicing of the speakers. For example, I've applied a mild 'S curve' to the response zeroed at around 1khz with about 1.5db of lift towards 100hz, and 1.5db of cut towards 10khz. Adjusting either or both of these by just 0.5db in either direction is easily audible.

I've resisted the temptation to do any measuring with anything other that my ears!

Couldn’t agree more.
And i’m not saying I have golden ears or the world’s most revealing system. I’m confident most people here would hear the difference between a mid or tweeter attenuated to 3.4dB and one to 3.5 or 3.6dB.
When I am using to JRiver to subjectively assess levels in order to calculate L-pad values, I absolutely go to tenths of a dB.

[Simon Hickie]

The project continues! The latest step is to replace the Monacor FR driver that was doing tweeter duties with a pair of SEAS 27TDFC fabric dome tweeters that I already had. Needless to say, they are a distinct improvement with a sweeter top end. Efficiency @ 90db matches the bass/mid driver that is operating from 150hz.

As things bed in and many familiar test tracks played, I've applied a cut of 3db to the non bass driver section. I've also added a high 'Q' notch filter at 3600hz where the bass/mid driver is a bit peaky. Finally, there's a modest cut with fairly low 'Q' around the 4000hz to tame a bit of excessive brightness to vocals. The miniDSP values for the various bits need tweaking as there's a fine balance between reducing listener fatigue and making the sound too lifeless. As Docfoster mentioned above, small changes can have significant effects.

The next step will be to replace some 4mm2 densely stranded speaker cable with some solid core stuff. I have generally had the best results with several runs of solid core magnet type wire per terminal as opposed to the stranded in PVC stuff. I'll put together enough runs to give around 4mm2 cross section over around 11 feet per cable. From what I can make out, the LS series adopts a similar approach, but with a lot more runs of thinner, better quality wire. Stripping magnet wire ends is a pain, so I've gone for fewer strands than is optimal.

Once I've done as good a job as I can with the experimental baffles, I will then go for something better looking.

A parallel project sees more work on the amplification side. The TPA3116 class D amps are currently being fed with laptop power supplies. A Mean Well 24v 300w PSU will be replacing these when I return from holiday. They seem quite well regarded in the DIY chip amp community.

Finally, the digi player side of things is there or thereabouts. My digital music is on a hard drive attached to a Raspberry Pi with HifiBerry Digi+ Pro S/PDIF HAT which in turn feeds a Khadas Tone Board DAC. The network player I'm using is Volumio.

The whole thing is a mass of wires and bare boards, but things should be much more presentable in a couple of weeks.