https://audioxpress.com/article/tes...h-dome-tweeter-from-sb-acoustics#&gid=1&pid=3
Something unexpected is going on a 4KHz.
Unless you wanted this dome in smaller face plate, the standard one measures looks better…
They certainly do, especially given the pedigree of the people making them. I would not make any decision based on them as the only source.
Btw, also realize how insanely zoomed in that distortion graph is.
(graphs at VC are a mess in general btw).
Which can give the impression of small bumps being big.
It's basically the exact same thing as the SB29RDNC-C000-4, except for the dome.
That one is a really good performing tweeter.
Sorry for the hijacking business, but Google actually recommended me to read this thread 
Looking at post #299 these two chipamps, although I'd love to use them for their integrated features, are not very flattery for their performance. And, as always, they lack usb-input and iis2, but insist of always having BT. What's the point of not exposing a digital input when the board has dsp?!
I've been looking for dsp-amp board(s) for small desktop speakers. They would be for desktop use only, so usb would be the preferred input. Does anyone have suggestions? I would like to use an all-in-one solution, but going with separates is still on the table.
Looking at post #299 these two chipamps, although I'd love to use them for their integrated features, are not very flattery for their performance. And, as always, they lack usb-input and iis2, but insist of always having BT. What's the point of not exposing a digital input when the board has dsp?!
I've been looking for dsp-amp board(s) for small desktop speakers. They would be for desktop use only, so usb would be the preferred input. Does anyone have suggestions? I would like to use an all-in-one solution, but going with separates is still on the table.
@hifijim These Decware mini radials are kind of semi-DIY for woodworkers. Unusual in that they are omnidirectional.
Electronics parts list and wood working instructions given.
Much smaller than the 3-way you have in mind, so shared to possibly give you an idea for an approach.
DecWare Tiny Radials
Electronics parts list and wood working instructions given.
Much smaller than the 3-way you have in mind, so shared to possibly give you an idea for an approach.
DecWare Tiny Radials
Sorry for the hijacking business, but Google actually recommended me to read this thread
Looking at post #299 these two chipamps, although I'd love to use them for their integrated features, are not very flattery for their performance. And, as always, they lack usb-input and iis2, but insist of always having BT. What's the point of not exposing a digital input when the board has dsp?!
I've been looking for dsp-amp board(s) for small desktop speakers. They would be for desktop use only, so usb would be the preferred input. Does anyone have suggestions? I would like to use an all-in-one solution, but going with separates is still on the table.
I've looked at TinySine options too:
Have a look- you can skip the Bluetooth-they have a few options with SPDIF input.
https://www.tinyosshop.com/index.php?route=product/category&path=158_193
Then you use an USB-SPDIF convertor.
Tinysine use the TPA3221 in their 3 channel amp (sans DSP).
The new TI TPA3223 is a pin compatible drop in replacement, and provides a bit more power:
Output Power at 1% THD+N
– 170-W Stereo into 4 Ω in BTL Configuration
I've enqured at Tinysine. Will keep you posted.
Last edited:
I was thinking of getting MA12070, TPA3116 or TPA325x boards, a dsp unit and usb input board, but I'm afraid it grows too large for desktop use. That also rules out most plates. There is a discount going on Arylic units. Maybe I'll sacrifice some performance for usability, because it's not going to be for critical listening.
Way back in post 230, @stv provided a link which showed this technique for cutting tapered bevels, or what we sometimes call facets. I want to make sure this information does not get lost. https://www.diyaudio.com/community/threads/compact-low-cost-active-3-way-speaker.402812/post-7454720
The more I think about this technique, the more I like it. For anyone without access to a full-size table saw, this is probably the best way to go.
The more I think about this technique, the more I like it. For anyone without access to a full-size table saw, this is probably the best way to go.
Super-easy facets with a circular saw - no table saw. Very similar to the idea above except using a circular saw, so your arm's won't get tired!
- Mock up the design so that you know what angle you need to set the saw blade. I typically use 52 degrees as a compromise between 45 and 60 degrees. But I mock it up in SketchUp just to make sure.
- Build a jig for the circular saw: Basically a DIY "track saw". Use 1/4" or 1/2" material. The guide piece on top needs to be straight. If you don't have a table saw, buy a piece of straight board - but it may not be as simple as a straight 1" x 2" or 1" x 4" because if it is too thick it will hit the motor of the saw. The back overhang needs to be wide enough to clamp. Make the front portion where the saw will glide wider than the base of the saw and blade. Set the circular saw blade to the angle determined in step 1 and run the saw down the guide, trimming off the excess. You now have a guide that is EXACTLY where the saw will cut, because in fact it was made by cutting it with the saw.
- Mark the facets: Use an accurate ruler and fine tip pencil. Place the guide on the pencil line and clamp it down. The guide is going to cut the baffle exactly where you place it, so it will only be as accurate as your pencil line. I use a 0.5mm mechanical pencil or carpenter pencil.
- Cut: Run the saw down the guide being careful to keep it nice and straight against the guide piece.
I have constructed a prototype box from XPS foam. I am evaluating the two tweeter options to see how they integrate with the SB15NBAC mid driver.
The face frame for the SB26STAC is 100 mm, while the SB29SDAC is 104 mm. As a result, I had to rout a recess of 104mm. When I mounted the smaller SB26STAC, I had a gap that had to be covered in tape.
The 50 mm bevels are not tapered, but I believe they will be representative of tapered bevels around the tweeter. I don't expect these straight bevels to be representative around the mid driver... in reality the mid driver will "see" about 25 mm bevel.
I hope to post some measurements in a few days.
The face frame for the SB26STAC is 100 mm, while the SB29SDAC is 104 mm. As a result, I had to rout a recess of 104mm. When I mounted the smaller SB26STAC, I had a gap that had to be covered in tape.
The 50 mm bevels are not tapered, but I believe they will be representative of tapered bevels around the tweeter. I don't expect these straight bevels to be representative around the mid driver... in reality the mid driver will "see" about 25 mm bevel.
I hope to post some measurements in a few days.
SB26STAC vs SB29SDAC
I ran a full comparison of these two tweeters, as installed in my test box. This post is the first six plots.
The FR plots show the diffraction impact of the box, and we see a similar on-axis shape to both drivers. This is the effect of the baffle… a broad hump at 1k, and another hump at 2k. I like that the SB29SDAC is 3 dB higher in sensitivity. In this plot we get the first hint that something is going on at 2k with the SB29SDAC driver.
Next we have the horizontal polar plots. In general, I like that the SB26STAC has a gently narrowing directivity, while the SB29SDAC has a more abrupt transition from 4k-7k. We also see the 2k wobble on the SB29SDAC.
Next we have the Power&DI plots. The DI curve for the SB26STAC looks like it would be easy to deal with, and would integrate well with the mid. The SB29SDAC actually has less directivity from 1k – 4k than does the SB26STAC, which is unexpected. This shows the limitation of simulating drivers with an ideal flat-disk piston model, because the real world behavior differs from prediction. The baffle diffraction simulation gets us remarkably close, but to match reality to within a dB or so would require a more advanced kind of simulation… or building and measuring a prototype.
Again we see the wobble in the SB29SDAC DI curve at 2k. Since the on-axis shows a peak here, and the power curve does not, it shows up as a bump in the DI curve.
I ran a full comparison of these two tweeters, as installed in my test box. This post is the first six plots.
The FR plots show the diffraction impact of the box, and we see a similar on-axis shape to both drivers. This is the effect of the baffle… a broad hump at 1k, and another hump at 2k. I like that the SB29SDAC is 3 dB higher in sensitivity. In this plot we get the first hint that something is going on at 2k with the SB29SDAC driver.
Next we have the horizontal polar plots. In general, I like that the SB26STAC has a gently narrowing directivity, while the SB29SDAC has a more abrupt transition from 4k-7k. We also see the 2k wobble on the SB29SDAC.
Next we have the Power&DI plots. The DI curve for the SB26STAC looks like it would be easy to deal with, and would integrate well with the mid. The SB29SDAC actually has less directivity from 1k – 4k than does the SB26STAC, which is unexpected. This shows the limitation of simulating drivers with an ideal flat-disk piston model, because the real world behavior differs from prediction. The baffle diffraction simulation gets us remarkably close, but to match reality to within a dB or so would require a more advanced kind of simulation… or building and measuring a prototype.
Again we see the wobble in the SB29SDAC DI curve at 2k. Since the on-axis shows a peak here, and the power curve does not, it shows up as a bump in the DI curve.
Now I show some distortion and burst decay plots for the two tweeters.
First is the harmonic distortion. I used STEPS to measure this, in 1/24 octave intervals, at an SPL of 92 dB at 36 inch (915 mm). I adjusted the gain on the tweeters so that both were playing 92 dB at 4k to 5k.
I think the advantage here has to go to the SB29SDAC. From 2k to 5k, it has lower distortion, particularly in the important H3 and H5. Above 5k, distortion becomes less audible (so I have heard)… However, the distortion differences between these two drivers are pretty small, and both tweeters are acceptable.
Next we see the burst decay plots. I have two for each driver, one plot is a frequency focused, and the other is time focused burst.
The decay is not as clean overall with the SB29SDAC as with the SB26STAC. We also see the 2k issue with the SB29SDAC. It seems odd to me that a tweeter would have such a prominent resonance at 2k. I wanted to be sure that my data was not corrupted by test technique or installation issues. So removed the SB29SDAC and carefully reinstalled it, then I applied tape around the perimeter to cover the small recess gap. This put it on “equal footing” with the SB26STAC. The data I collected was nearly identical to the first set of data, so I am confident the 2k issue is with the tweeter itself.
At this point I have enough data to decide that the SB26STAC is the tweeter for this project.
j.
First is the harmonic distortion. I used STEPS to measure this, in 1/24 octave intervals, at an SPL of 92 dB at 36 inch (915 mm). I adjusted the gain on the tweeters so that both were playing 92 dB at 4k to 5k.
I think the advantage here has to go to the SB29SDAC. From 2k to 5k, it has lower distortion, particularly in the important H3 and H5. Above 5k, distortion becomes less audible (so I have heard)… However, the distortion differences between these two drivers are pretty small, and both tweeters are acceptable.
Next we see the burst decay plots. I have two for each driver, one plot is a frequency focused, and the other is time focused burst.
The decay is not as clean overall with the SB29SDAC as with the SB26STAC. We also see the 2k issue with the SB29SDAC. It seems odd to me that a tweeter would have such a prominent resonance at 2k. I wanted to be sure that my data was not corrupted by test technique or installation issues. So removed the SB29SDAC and carefully reinstalled it, then I applied tape around the perimeter to cover the small recess gap. This put it on “equal footing” with the SB26STAC. The data I collected was nearly identical to the first set of data, so I am confident the 2k issue is with the tweeter itself.
At this point I have enough data to decide that the SB26STAC is the tweeter for this project.
j.
Having heard the 26STAC plenty, and used the 29SDAC' sibling the 29RDAC; I would concur with your choice. The STAC sounds wonderful every time I hear it, just clean and open sounding. The RDAC/RDC/SDAC brethren really increase in HD as it approaches 2k. In fact, I had to kill the Fs to get the last offender out of there, with a 75uF/0.4mH series notch across the tweeter. It had a nasal quality without it. Something of info, the RDC/RDAC did not exhibit a 2k peak in the conventional baffles I used.
@HiFiCompass tested the SB29SDAC, and the waterfall CSD plot does not show a 2k resonance. Comparing a burst decay plot to a CSD plot is not exactly apples to apples, but if the driver has an inherent resonance at 2k, I would think it would show up in the hificompass plot. So it may actually be some kind of interaction with my chosen baffle shape. In any case, from a practical standpoint, it does not matter if the driver or the baffle is to blame. With this baffle shape, the SB26STAC performs better.
I appreciate your thoughts on the two tweeters. I have never worked with either one. I love the performance of the SB26CDC and 26ADC, and I believe the 26STAC has the same motor.
j.
I appreciate your thoughts on the two tweeters. I have never worked with either one. I love the performance of the SB26CDC and 26ADC, and I believe the 26STAC has the same motor.
j.