Development of open baffle 3-way plus subs

Lewinskih01
Posts: 57
Joined: 19 Oct 2019 22:12

Re: Development of open baffle 3-way plus subs

Post by Lewinskih01 » 27 Sep 2020 14:34

UliBru wrote:
27 Sep 2020 11:21
Horacio,

the 10G40 seems to be quite inefficient below 200 Hz. Can the 12-inch subs run up to 200 Hz?
Yes, it is. The main reason is the open baffle.
The subs are 2 DIY sealed boxes for 12" Rythmiks servo, which aren't good up to 200Hz. Plus people people talk about the bass definition provided by OB, so want to try it. Maybe my drivers aren't suited for OB and I will need to get new ones. I guess it's part of the development. I want to try the brute force approach, taking advantage of their power handling capacity, significant X-max, and the 400W amp driving them, and that I listen at 85-90dB SPL. Question of course is how they will sound.

I'm running NT2 at 70/500/1500Hz to be able to listen, as a next step. Running the 10G40 amps off the phones out from the Hilo wasn't a good idea because of the low gain available. Now I'm running the 10G40 off the line out, and the tweeters off the phones out (which isn't ideal either because this output is probably not the best quality output...tradeoffs). Let's see how it sounds.

What do you think of using NT2 for this?
User avatar
UliBru
Posts: 468
Joined: 19 Oct 2019 13:58

Re: Development of open baffle 3-way plus subs

Post by UliBru » 27 Sep 2020 15:25

First try to get a decent result. Later you can test the UB filters instead of NT
Stay well tuned
Uli

Moderator
Acourate system: JRiver/Roon -> AcourateConvolver -> miniDSP U-DIO8 -> TacT M/S2150 amps -> DIY horn speakers
Lewinskih01
Posts: 57
Joined: 19 Oct 2019 22:12

Re: Development of open baffle 3-way plus subs

Post by Lewinskih01 » 27 Sep 2020 15:30

UliBru wrote:
27 Sep 2020 15:25
First try to get a decent result. Later you can test the UB filters instead of NT
Makes sense.

How do I generate a mono signal? Since I will be running 1 speaker for the test, I should blend together L & R into mono. I thought there would be a mono option for XOs, but I don't see it in the Generate Crossover window. Maybe XOs are stereo and I mix them into mono later?
User avatar
UliBru
Posts: 468
Joined: 19 Oct 2019 13:58

Re: Development of open baffle 3-way plus subs

Post by UliBru » 27 Sep 2020 16:15

Horacio,

please check if the Hilo mixer allows to route both playback channels (left/right) to a single output channel (RME Totalmix can do this).

Another solution is to measure over AcourateConvolver.
Stay well tuned
Uli

Moderator
Acourate system: JRiver/Roon -> AcourateConvolver -> miniDSP U-DIO8 -> TacT M/S2150 amps -> DIY horn speakers
Lewinskih01
Posts: 57
Joined: 19 Oct 2019 22:12

Re: Development of open baffle 3-way plus subs

Post by Lewinskih01 » 28 Sep 2020 14:34

UliBru wrote:
27 Sep 2020 16:15
Horacio,

please check if the Hilo mixer allows to route both playback channels (left/right) to a single output channel (RME Totalmix can do this).

Another solution is to measure over AcourateConvolver.
Thank you. Indeed, Hilo mixer was good for this.

Things didn't fully turn out as expected: the corrected sound was very lean and mids/treble at higher level (SPL) than the target curve defined. Hilo mixer allowed to make level adjustments by channel to where I can live with. So I'm getting familiar with the sound and will do another correction iteration next weekend. Also need to allow the drivers to break in.

Bass is definitely different. More defined and maybe leaner. That's what I need to get used to and understand if I need a steeper downward slope for bass than with boxed speakers. Or maybe I need larger woofer for OB midbass.

BTW, interestingly I turned the speaker 90° so now the back wall is 3m away and I still get the dip centered around 100Hz and 1 octave wide (this is likely affecting my perception of lean bass mentioned above). Maybe rather than back wall interference it was related to floor or ceiling.
User avatar
UliBru
Posts: 468
Joined: 19 Oct 2019 13:58

Re: Development of open baffle 3-way plus subs

Post by UliBru » 28 Sep 2020 14:42

Keep in mind: you can also create a multiway wav from Cor filters (insted of XO filters) and then re-measure your setup.
The result will show if the intended target is reached.

PS: select a different workspace for this measurement to prevent overwriting previous recordings
Stay well tuned
Uli

Moderator
Acourate system: JRiver/Roon -> AcourateConvolver -> miniDSP U-DIO8 -> TacT M/S2150 amps -> DIY horn speakers
Lewinskih01
Posts: 57
Joined: 19 Oct 2019 22:12

Re: Development of open baffle 3-way plus subs

Post by Lewinskih01 » 04 Nov 2020 11:24

I've been listening to my one OB speaker in mono, made some changes, opened the tweeter's back so it was dipole above 1.5kHz, and I'm really liking it. Adjusting the bass level gave me more of the sound I was seeking while keeping the bass definition the OB delivers. This is a path I want to explore further!

I decided to replace the twin 10" by a single 18" per side for midbass duties. That's a BIG midbass! My starting point is to run it nude (no baffle) and on a swing. I will likely test using smallish baffles made of light, non-resonant materials attached to the basket, keeping the assembly on a swing - but that's for later.
Here is a first iteration test frame, which I took last weekend for outside measurements:
20201101_155648.jpg
20201101_155648.jpg (1.33 MiB) Viewed 3976 times
20201101_154608.jpg
20201101_154608.jpg (2.73 MiB) Viewed 3976 times
Measurements look as expected. Here's the front measurements taken 0 to 90° every 15°:
18HP1010 0 to 90 degrees horiz.jpg
18HP1010 0 to 90 degrees horiz.jpg (144.83 KiB) Viewed 3976 times
The dip around 380Hz might be a reflection from the edge of the rotation table. Will have to remeasure with a different setup to confirm this. Dipole peak showing around 600Hz, like the modelling tools predicted (The Edge).
Edge simulation on 18 nude OB on axis.jpg
Edge simulation on 18 nude OB on axis.jpg (56.05 KiB) Viewed 3976 times
I am yet to run distortion measurements, but my guess is xo will be around 70Hz and run up to somewhere between 300 or 400Hz. This of course will imply compensating for about 12dB. What's the best practice to do this on Acourate? Should I run a "regular" process where the target curve level is set at 70Hz and let Acourate bring the rest of the rising response down to the target level (with the consequent loss in gain), or would it be better to apply a 12dB boost at 70Hz gradually reducing to 0dB at 300Hz and apply the target level to this? Or an alternative approach?

cheers!
User avatar
UliBru
Posts: 468
Joined: 19 Oct 2019 13:58

Re: Development of open baffle 3-way plus subs

Post by UliBru » 04 Nov 2020 15:14

Horacio,

setting the target level at 70 Hz or boosting at 70 Hz by 12 dB (plus gradual reduction) is basically the same stuff. If you use the target then Acourate will do the rest for you. But keep in mind that in both cases the filter will reduce the level at higher frequencies and thus loose some resolution.
The only other solution is to apply such a boost by an according circuit in the analog domain.
Stay well tuned
Uli

Moderator
Acourate system: JRiver/Roon -> AcourateConvolver -> miniDSP U-DIO8 -> TacT M/S2150 amps -> DIY horn speakers
Lewinskih01
Posts: 57
Joined: 19 Oct 2019 22:12

Re: Development of open baffle 3-way plus subs

Post by Lewinskih01 » 22 Mar 2021 11:37

UliBru wrote:
04 Nov 2020 15:14
Horacio,

setting the target level at 70 Hz or boosting at 70 Hz by 12 dB (plus gradual reduction) is basically the same stuff. If you use the target then Acourate will do the rest for you. But keep in mind that in both cases the filter will reduce the level at higher frequencies and thus loose some resolution.
The only other solution is to apply such a boost by an according circuit in the analog domain.
Hello Uli.

I continue (sporadically) working on this design, and I was reviewing this thread and your input, and came to realize I don't really understand the implications of the sentence in blue above. What does resolution mean here?

How does this affect the contiguous filter? Let's say my midbass on open baffle is 80dB/2.83V at 80Hz and 92 dB/2.83V at 250Hz (xo), the midrange is 85 dB/2.83V at 250 and 97 dB/2.83V at 1600 Hz, and the tweeter is 102 dB/2.83, let's assume the drivers display smooth slope between the noted low and high xo for each one, and I use a flat target curve to simplify the example. I would set the target curve at 80dB and bring everything down: 12 dB down for the midbass at 250Hz, and 5dB for the midrange at 250Hz all the way to 17dB at 1600Hz, and the tweeter by 22dB. I guess the midrange and treble will be impacted by lower resolution as well, and that sounds to me like lower "sound quality", but trying to understand it better.

Regards,
Horacio
User avatar
UliBru
Posts: 468
Joined: 19 Oct 2019 13:58

Re: Development of open baffle 3-way plus subs

Post by UliBru » 22 Mar 2021 11:55

Horacio,

your calculation is correct.

Indeed the summary is: if there are speaker drivers of different efficieny it would be possible to use a digital filter with a target curve at/below the driver with the lowest efficiency. It is not possible to play to this driver with a digital signal > 0 dBFS. This approach will result in gain loss of the digital filter.
The other solution is to apply a different analog gain to the drivers in such a way that the measurement shows all drivers at the same level or following the intended target. Thus the digital filter will keep its max. resolution.

But we can also discuss a bit about the digital gain loss. -12 dB corresponds to loosing two bits of resolution. Today the convolver programs typically feed a DAC with 24 bits. And it is known that the DACs will use 20 to 21 bits in best case (see typical SNR values). So the loss of two bits is neglegible.

At the end it depends on given circumstances. If the amplifers for the driver have a fixed gain the job can be done by digital filters. If they allow a variable gain the filter can be optimized for best resolution.
Stay well tuned
Uli

Moderator
Acourate system: JRiver/Roon -> AcourateConvolver -> miniDSP U-DIO8 -> TacT M/S2150 amps -> DIY horn speakers
Post Reply