New Open Baffle Project

[Docfoster]

The development baffles are now made. They have adjustable wings to help determine whether a standard baffle or a U frame, or something in between works best. Drivers installed. The miniDSP has been configured with a 4th order LR XO @ 160hz on both drivers. I have applied a 5db cut to the full range driver, but another db or two may be needed to get the sensitivity matching right. MJK's rule of thumb is that the bass driver needs to be about 6-10db more efficient than the full range driver. The Monacor SP-382PA is around 94db, the Monacor SPX-32M full range is around 88db, so I'm there or thereabouts (I tend to be rather treble sensitive).

Overall first impressions are very positive. Soundstage is wide and high. Bass is solid and sufficiently extended. Room modes are not excited in the same way as with a box speaker, so one can go louder with less boominess. More fine tuning is obviously needed, but I suspect this is going to be the way forward from now on. Once I have settled on the right baffle configuration, the next step will be to make something that actually looks decent.

You’ve made goof progress!
Love the building in of experiment and on-the-fly tuning.
My ears agree with the “6-10dB rule of thumb” you mention.

[Simon Hickie]

After a rather unsatisfactory listening session yesterday evening, I began to suspect the Sony home cinema amp I was using for testing (I needed 4 channel input from the miniDSP).

Therefore, this morning I have rigged up a pair of Sure TPA3116 class D amp modules in a monoblock type configuration. Voila! Bass boominess has gone and treble distortion has disappeared. The Sony really was messing things up.

The challenge now is to get the balance right between bass and full range drivers. The Monacor SPX-32M FR drivers are now at -3db, giving about a nominal 9db difference between the drivers - very much in the MJK ballpark. There's still some glare in the sound, and a brief look at the frequency response chart for the FR driver does indeed show a 5db lift from around 3khz. Therefore, the next step is to try a high shelf filter to flatten things a bit. The beauty of the miniDSP is that one can make changes very quickly and there's no messing with lossy inductors, caps & resistors,

[Docfoster]

Well done for identifying the cause of the dissatisfaction. And for the having to hand the necessary gear.
I agree that DSP is a great way to quickly establish ways forward. Though once I've found the values I like with the shelf filters in JRivers DSP, I can't resist building physicla L-pads to meet those values.
The results I get to when subjectively balancing driver levels often baffle me. I think the effective sensitivity of my tweeters are about 5dB lower than the mids and 13dB lower than the woofers. "Trust your ears!" as they say. (Mine must be very trebble sensitive! Other than my ears I don't have the necessary measuring equipment to assess respective driver levels, as you clearly have.)
Enjoying reading your updates on this!

[Simon Hickie]

Well done for identifying the cause of the dissatisfaction. And for the having to hand the necessary gear.
I agree that DSP is a great way to quickly establish ways forward. Though once I've found the values I like with the shelf filters in JRivers DSP, I can't resist building physicla L-pads to meet those values.
The results I get to when subjectively balancing driver levels often baffle me. I think the effective sensitivity of my tweeters are about 5dB lower than the mids and 13dB lower than the woofers. "Trust your ears!" as they say. (Mine must be very trebble sensitive! Other than my ears I don't have the necessary measuring equipment to assess respective driver levels, as you clearly have.)
Enjoying reading your updates on this!

Thanks for the kind words! I'm not using any measuring equipment apart from my ears and some music tracks that I know very well. The published Monacor frequency response charts seem pretty reliable. I also have a very basic understanding of mixing and which frequency areas have what effect and what I'm hearing is consistent with both the fr chart and mixing guidelines.

The high shelf filter had the desired effect of reducing the glassiness and glare I was hearing. I'm now running the full ranger with no attenuation, along with the high shelf filter. It's still running a fraction hot. It only seems to need very small adjustments to alter the overall sound quite significantly. I have no plans to fiddle around with much more in the way of EQ. However, I might try something different from a blanket high shelf filter. The nice thing about the miniDSP software is that I can see in chart form how the response curve changes.

I too am very treble sensitive. A friend's ATC SCM40 speakers tend to make my ears bleed (metaphorically speaking), and there's no soundstage worth talking about.

My overall impression is that the OB speakers are delivering a huge soundstage that is high, wide & deep. It's streets ahead of my semi-omnis in soundstage and openness. The room is loaded in a completely different way and probably the best word to describe the presentation is 'effortless'. Bass is cleaner and more extended than with any box speaker I've had. They are pig ugly and warrant something better in the way of baffle and supports, but that should be relatively easy to achieve once the final configuration is settled. MJK U frame seems the best to date.

On the amp front, the TPA3116 is a remarkable little class D unit. The boards were about £10 each if I remember correctly. The TPA3116 is certainly well regarded over at diyaudio. I do not have the knowledge to mod the boards, but some versions lend themselves very well to upgraded caps etc.

[karatestu]

Glad its working out for you Simon.

I am treble sensitive as well. Sibilance really does get on my tits. It is what led me to discover nva.

[Ali Tait]

Sounds like you are making good progress Simon. Agree on OB bass, makes most box speakers sound coloured.

[Jammy Dodger]

But Ob's are flawed........

Where can one place an ornament or vase. ?
Where is the port fitted. ?
No Easy stand placement. ?
Etc etc etc. ?

These failures would need to be addressed before folk actually took them seriously.

[Docfoster]

Well done for identifying the cause of the dissatisfaction. And for the having to hand the necessary gear.
I agree that DSP is a great way to quickly establish ways forward. Though once I've found the values I like with the shelf filters in JRivers DSP, I can't resist building physicla L-pads to meet those values.
The results I get to when subjectively balancing driver levels often baffle me. I think the effective sensitivity of my tweeters are about 5dB lower than the mids and 13dB lower than the woofers. "Trust your ears!" as they say. (Mine must be very trebble sensitive! Other than my ears I don't have the necessary measuring equipment to assess respective driver levels, as you clearly have.)
Enjoying reading your updates on this!

Thanks for the kind words! I'm not using any measuring equipment apart from my ears and some music tracks that I know very well. The published Monacor frequency response charts seem pretty reliable. I also have a very basic understanding of mixing and which frequency areas have what effect and what I'm hearing is consistent with both the fr chart and mixing guidelines.

The high shelf filter had the desired effect of reducing the glassiness and glare I was hearing. I'm now running the full ranger with no attenuation, along with the high shelf filter. It's still running a fraction hot. It only seems to need very small adjustments to alter the overall sound quite significantly. I have no plans to fiddle around with much more in the way of EQ. However, I might try something different from a blanket high shelf filter. The nice thing about the miniDSP software is that I can see in chart form how the response curve changes.

I too am very treble sensitive. A friend's ATC SCM40 speakers tend to make my ears bleed (metaphorically speaking), and there's no soundstage worth talking about.

My overall impression is that the OB speakers are delivering a huge soundstage that is high, wide & deep. It's streets ahead of my semi-omnis in soundstage and openness. The room is loaded in a completely different way and probably the best word to describe the presentation is 'effortless'. Bass is cleaner and more extended than with any box speaker I've had. They are pig ugly and warrant something better in the way of baffle and supports, but that should be relatively easy to achieve once the final configuration is settled. MJK U frame seems the best to date.

On the amp front, the TPA3116 is a remarkable little class D unit. The boards were about £10 each if I remember correctly. The TPA3116 is certainly well regarded over at diyaudio. I do not have the knowledge to mod the boards, but some versions lend themselves very well to upgraded caps etc.

You sound chuffed and excited by how things are progressing.
Any chance of some photos?
Apologies for misunderstanding your previous words on frequency response. All ears then. Great.

[Simon Hickie]

I know, Jammy. These are serious issues for mass market box shifters. Even semi-omnis fail in the vase department.

Yes DocF, I'm delighted with the progress so far. I'll sort out some pictures when I've tidied the listening room. At the moment it's a spaghetti of cables, boxes, bags of bits etc. It's also development setup, so not pretty! I also need to get the class D amps in a box - that's tomorrow's job. I'm also an interconnect short of a full set, so to speak, so I need to nail another one together. It will be another low capacitance special, loosely based on DNM solid core jobs with no twisting or shielding, 0.4mm copper in teflon tubing with plastic phonos. I've finished some new solid core speaker cables too. Again, DNM inspired with single solid cores. I've gone for low inductance over low capacitance.

[Jammy Dodger]

Cheers Simon.........Nice of you to acknowledge a wee bit off light hearted banter.

Some folks heads are so far up their own arseholes, they may well miss Xmas.

I think OBs are ugly.....a very lazy option, they may well sing like an angel, but look like DIY cack.

Ali call back.....