Hi everyone. Longtime lurker, first post here.
I’ve always loved the Tc9fd-18. 2 weeks ago on randomly on Facebook MP a guy is selling new old stock TC9 very cheap, so naturally I bought a box of 45pcs. 😎
I have in mind a straight array using 24 pieces each side. I plan to use dsp to attenuate/shade, add various amounts of delay, and low-pass the drivers based on the wavelength/line length ratio. I have a dsp box with 24 output channels, and am exploring amp boards from China or 12ch zone amps to have 24 channels of amplification.
My room has 7’ ceilings, so the array will cover 94% of the height.
Right now I’m trying to understand what issues might come from a delay-curved cbt array going floor to ceiling vs going infinite line/no delay.
the nice part is I’ll have the ability to adjust delay/attenuation/roll off pretty finely once it’s all assembled.
I’d love to hear any tips or cautionary tales anyone has or has seen for tc9 line arrays.
Unrelated but im putting danish oil on my current project: open baffle H frame with grs18-pt woofers. Then im going to get started on this array. hoping to do the CNC this summer and the wiring this fall and winter.
I’ve always loved the Tc9fd-18. 2 weeks ago on randomly on Facebook MP a guy is selling new old stock TC9 very cheap, so naturally I bought a box of 45pcs. 😎
I have in mind a straight array using 24 pieces each side. I plan to use dsp to attenuate/shade, add various amounts of delay, and low-pass the drivers based on the wavelength/line length ratio. I have a dsp box with 24 output channels, and am exploring amp boards from China or 12ch zone amps to have 24 channels of amplification.
My room has 7’ ceilings, so the array will cover 94% of the height.
Right now I’m trying to understand what issues might come from a delay-curved cbt array going floor to ceiling vs going infinite line/no delay.
the nice part is I’ll have the ability to adjust delay/attenuation/roll off pretty finely once it’s all assembled.
I’d love to hear any tips or cautionary tales anyone has or has seen for tc9 line arrays.
Unrelated but im putting danish oil on my current project: open baffle H frame with grs18-pt woofers. Then im going to get started on this array. hoping to do the CNC this summer and the wiring this fall and winter.
Looks like it:
https://www.facebook.com/marketplace/item/550101146568057/
I sent him a DM to arrange a larger order than FB allows.
Mine arrived in the original box from digikey, a little beaten up from ups but everything tested fine.
https://www.facebook.com/marketplace/item/550101146568057/
I sent him a DM to arrange a larger order than FB allows.
Mine arrived in the original box from digikey, a little beaten up from ups but everything tested fine.
Attachments
Hello @ericwestpheling and welcome to the line array enthusiasts😄.
I’m now working on a small 9 elements array of TC9FD18 but, as you may know, is not easy to obtain a reasonable vertical coverage with such a small array. Perhaps you provided me with a solution to scale up my project to the original 25 elements array! Thanks for the link!
My advice is to take a look to the @wesayso long thread (if you haven’t already checked it out) in which a lot of problems are already addressed and solved. A floor-to-ceiling straight array can just do very well without shading using only EQ so perhaps you can start experimenting without any shading and add them later if needed. One question: could you please share some info about your 24 channel DSP?
thanks,
Manuele
I’m now working on a small 9 elements array of TC9FD18 but, as you may know, is not easy to obtain a reasonable vertical coverage with such a small array. Perhaps you provided me with a solution to scale up my project to the original 25 elements array! Thanks for the link!
My advice is to take a look to the @wesayso long thread (if you haven’t already checked it out) in which a lot of problems are already addressed and solved. A floor-to-ceiling straight array can just do very well without shading using only EQ so perhaps you can start experimenting without any shading and add them later if needed. One question: could you please share some info about your 24 channel DSP?
thanks,
Manuele
Thanks for the welcome. I hope that seller has more. Good for him and for us!
I've read chunks of Wesayso's build thread but I should sit down and do it start to finish one night so I can get a real understanding of what hes done with all the JRiver software.
I use a DSP box called Audia Flex made by Biamp. I have 3 now
They are 24ch DSP boxes with 12 card slots that can be configured with various input/output cards. They were set up for conference rooms or telephone/voip purposes. Occasionally you find one with a 12in/12out configuration, even rarer was the one I found a few weeks ago setup for 24out. Since I have another unit I can link them via ethernet to send inputs on one to the output on another- or I may just swap in one 2ch input card and live with 22ch output.
They are probably about 10+ years old at this point, so a lot of facilities are clearing them out for newer solutions, and they end up on electronics surplus sellers on ebay. I bought mine for between $50-$70/ea, usually best offer way under the asking.
They need to be configured over ethernet with software that is windows-only (this was super annoying since I'm a mac person). The software lets you assemble "DSP blocks" into a flowchart. Its easy enough to use. It has some quirks if you are serious about getting one I can explain them- they haven't been issues for me but might be for some uses.
If you have any questions about them lmk and i'll try to answer. I talk them up to anyone since they are such a great deal for what they are, they still are super cheap even if you had to buy 2-3 different ones to put together the i/o card combo you wanted.
I've read chunks of Wesayso's build thread but I should sit down and do it start to finish one night so I can get a real understanding of what hes done with all the JRiver software.
I use a DSP box called Audia Flex made by Biamp. I have 3 now
They are 24ch DSP boxes with 12 card slots that can be configured with various input/output cards. They were set up for conference rooms or telephone/voip purposes. Occasionally you find one with a 12in/12out configuration, even rarer was the one I found a few weeks ago setup for 24out. Since I have another unit I can link them via ethernet to send inputs on one to the output on another- or I may just swap in one 2ch input card and live with 22ch output.
They are probably about 10+ years old at this point, so a lot of facilities are clearing them out for newer solutions, and they end up on electronics surplus sellers on ebay. I bought mine for between $50-$70/ea, usually best offer way under the asking.
They need to be configured over ethernet with software that is windows-only (this was super annoying since I'm a mac person). The software lets you assemble "DSP blocks" into a flowchart. Its easy enough to use. It has some quirks if you are serious about getting one I can explain them- they haven't been issues for me but might be for some uses.
If you have any questions about them lmk and i'll try to answer. I talk them up to anyone since they are such a great deal for what they are, they still are super cheap even if you had to buy 2-3 different ones to put together the i/o card combo you wanted.
Found this link- this guy did exactly what I'm trying to do with full DSP control over the array. Great read:
http://www.audiodevelopers.com/10-case-study-3-a-line-array-with-dsp/
He sounds pretty happy with the results. I won't be doing the full wifi/bt app design like he has, but I might be able to make it switchable between 2-3 different beam widths using the external controls/logic box available for the Audia. Now im excited that this might work!
http://www.audiodevelopers.com/10-case-study-3-a-line-array-with-dsp/
He sounds pretty happy with the results. I won't be doing the full wifi/bt app design like he has, but I might be able to make it switchable between 2-3 different beam widths using the external controls/logic box available for the Audia. Now im excited that this might work!
Could you specify which part do you think is in need of simulation? I haven’t worked in vituixcad yet, will have to see how far I can get on building a model.I'd say try and simulate something like that in Vituixcad first! See what it does before committing to it
The build you link to has small enough drivers for the theory to work, I'm not so sure about the 3.5" drivers...
My thought was by using all channels dsp /active it would be able to be tuned/curved/shaded/low pass to whatever degree needed. Not quite “figure it out later” but kind of.
Shading and delaying these bigger drivers is what I'd check out prior to building it.
There will be combing in a straight array, but I'm not convinced making a CBT out of drivers this size is going to fix the combing or lobes.
We ran an awful lot of sims in the thread by nc535: https://www.diyaudio.com/community/...-corner-placement.337956/page-21#post-6164403
(not the start of simulations in the thread, but I think it's where the CBT was compared to the straight array)
At first I thought that @nc535 was going bonkers. Making all those simulations. That is until I started recognizing my measurements in his "in room" predictions. That's where I decided to jump in and start simming too. @nc535 kindly provided a starter model and shared it in the thread.
I wanted to try a frequency shaded array, as the CBT shading means not all drivers help out on the bottom end, loosing that potential. In the end, that's what I have build. With one exception, swapping the TC9 driver for the Scan Speak 10F 8414G10.
You can get that basic model (that needs some adjustment for your own situation) here: https://www.diyaudio.com/community/...-corner-placement.337956/page-28#post-6205129
Just build up the new array you are thinking of, using the TC9 driver data in that model. I can't remember if this model had the mirror drivers beneath the floor or not, you don't need those. But it's good to see what floor and ceiling add to the "in room" prediction.
The "0" level is the mic height, from there on the drivers are placed (so my floor is set at -1000, as a sample). You can play with delay and attenuation.
The hard work, making a model of the TC9, is done already. Based on a model created in Vituixcad's diffraction tool I believe combined with actual data, not actual measurements of a driver in the enclosure (which would be even better or more accurate).
A mimic of floor and/or ceiling reflections with actual attenuated drivers is possible too, and gave me more insight to compare to my measured in-room IR's. In other words, it was useful to me.
There will be combing in a straight array, but I'm not convinced making a CBT out of drivers this size is going to fix the combing or lobes.
We ran an awful lot of sims in the thread by nc535: https://www.diyaudio.com/community/...-corner-placement.337956/page-21#post-6164403
(not the start of simulations in the thread, but I think it's where the CBT was compared to the straight array)
At first I thought that @nc535 was going bonkers. Making all those simulations. That is until I started recognizing my measurements in his "in room" predictions. That's where I decided to jump in and start simming too. @nc535 kindly provided a starter model and shared it in the thread.
I wanted to try a frequency shaded array, as the CBT shading means not all drivers help out on the bottom end, loosing that potential. In the end, that's what I have build. With one exception, swapping the TC9 driver for the Scan Speak 10F 8414G10.
You can get that basic model (that needs some adjustment for your own situation) here: https://www.diyaudio.com/community/...-corner-placement.337956/page-28#post-6205129
Just build up the new array you are thinking of, using the TC9 driver data in that model. I can't remember if this model had the mirror drivers beneath the floor or not, you don't need those. But it's good to see what floor and ceiling add to the "in room" prediction.
The "0" level is the mic height, from there on the drivers are placed (so my floor is set at -1000, as a sample). You can play with delay and attenuation.
The hard work, making a model of the TC9, is done already. Based on a model created in Vituixcad's diffraction tool I believe combined with actual data, not actual measurements of a driver in the enclosure (which would be even better or more accurate).
A mimic of floor and/or ceiling reflections with actual attenuated drivers is possible too, and gave me more insight to compare to my measured in-room IR's. In other words, it was useful to me.
Wow thanks so much for this. A lot to digest in that thread of testing. What a great resource.
One quick thing was the low-pass shading I was thinking of seems different than what you’re describing (Iirc it’s discussed in the near field line array paper). I believe I also saw it in another thread you participated in about the “eXpanding” array. My understanding was that low pass was applied so that the wavelength was proportional to the length of the drivers in the array.
Regardless I’ll get familiar with the cad software and see what can be accomplished.
for your array you aren’t using any shading or dsp? Just eq in jriver? I’m still reading through your build thread but it’s a lot!
Well, I started with an unshaded 25 driver straight array. With EQ + FIR correction within JRiver (and then some ).
But about 1.5/2 years ago, when we were running sims like crazy I decided to try to make it a frequency shaded array that I had come up with in Vituixcad, inspired by the prior work done by @nc535. So that's what I have today. I still need to redo all DSP, which means: new FIR correction of both main array speakers, new FIR correction on my 2 ambience back-channels and new FIR correction on both subwoofers. I'm running on old versions of everything but due to a room renovation 2 years back it is all up for a refreshment. The arrays are running FIR correction made for the TC9 FD18-08 version of the (frequency shaded) arrays. That too needs a refresh as I now run the Scan Speak 10F drivers.
The filters that ended up in the arrays:
All done passively, as I want to keep using my favorite amplifier. Not pictured are 2 notch circuits per side.
So, yeah, I did use DSP in recent years and still do now. But also make use of passive means to get what I wanted.
The new drivers:
Making up the renewed arrays. Only thing left to do is spend the time to redo literary everything.
I know my thread is just about impossible to wade trough. But the first post should contain a few useful links.
Beside the EQ and FIR correction, and running ambience channels I also use some mid/side EQ as a cross talk solution.
So I would definitely say I do use DSP. I think I didn't mention all of it yet. But no stand alone DSP unit, it's all done
within the PC. Linear phase crossovers and a home brew ambience mix and phase correction... not for the faint of heart.
).But about 1.5/2 years ago, when we were running sims like crazy I decided to try to make it a frequency shaded array that I had come up with in Vituixcad, inspired by the prior work done by @nc535. So that's what I have today. I still need to redo all DSP, which means: new FIR correction of both main array speakers, new FIR correction on my 2 ambience back-channels and new FIR correction on both subwoofers. I'm running on old versions of everything but due to a room renovation 2 years back it is all up for a refreshment. The arrays are running FIR correction made for the TC9 FD18-08 version of the (frequency shaded) arrays. That too needs a refresh as I now run the Scan Speak 10F drivers.
The filters that ended up in the arrays:
All done passively, as I want to keep using my favorite amplifier. Not pictured are 2 notch circuits per side.
So, yeah, I did use DSP in recent years and still do now. But also make use of passive means to get what I wanted.
The new drivers:
Making up the renewed arrays. Only thing left to do is spend the time to redo literary everything.
I know my thread is just about impossible to wade trough. But the first post should contain a few useful links.
Beside the EQ and FIR correction, and running ambience channels I also use some mid/side EQ as a cross talk solution.
So I would definitely say I do use DSP
. I think I didn't mention all of it yet. But no stand alone DSP unit, it's all donewithin the PC. Linear phase crossovers and a home brew ambience mix and phase correction... not for the faint of heart.
Here’s the thread I’m reading right now about the shading/low pass:Well, I started with an unshaded 25 driver straight array. With EQ + FIR correction within JRiver (and then some
But about 1.5/2 years ago, when we were running sims like crazy I decided to try to make it a frequency shaded array that I had come up with in Vituixcad, inspired by the prior work done by @nc535. So that's what I have today. I still need to redo all DSP, which means: new FIR correction of both main array speakers, new FIR correction on my 2 ambience back-channels and new FIR correction on both subwoofers. I'm running on old versions of everything but due to a room renovation 2 years back it is all up for a refreshment. The arrays are running FIR correction made for the TC9 FD18-08 version of the (frequency shaded) arrays. That too needs a refresh as I now run the Scan Speak 10F drivers.
The filters that ended up in the arrays:
View attachment 1071063
All done passively, as I want to keep using my favorite amplifier. Not pictured are 2 notch circuits per side.
So, yeah, I did use DSP in recent years and still do now. But also make use of passive means to get what I wanted.
The new drivers:
View attachment 1071064
Making up the renewed arrays. Only thing left to do is spend the time to redo literary everything.
I know my thread is just about impossible to wade trough. But the first post should contain a few useful links.
Beside the EQ and FIR correction, and running ambience channels I also use some mid/side EQ as a cross talk solution.
So I would definitely say I do use DSP
within the PC. Linear phase crossovers and a home brew ambience mix and phase correction... not for the faint of heart.
https://www.diyaudio.com/community/...-wall-or-corner-placement.337956/post-6193737
it does add up dsp and amp channels quickly but I’ve managed to come across a dsp box that can do 24ch. I don’t think it’s the more advanced type of filtering like FIR but it has all the usual filters and parametric eq you could ever want. TBD. Journey just beginning. 👍
Hey just a thought after reading all those sims (only on page 50-something really)
If I know I’m going to be using a 6’w x 12” listening position (couch) and won’t be doing any standing listening - would it be worth exploring doing the reverse of a cbt? a concave curve- ie using delay to have the central driver to match the distance of the outer drivers, tapering the rest, etc.
Have you seen anyone test something like that?
Yep, I've seen it done with an actual bend line array. And he even had an anti cross talk devise (wearing it, actually).
Let me look for it...
https://www.diyaudio.com/community/...range-line-array.242171/page-128#post-4845884
Let me look for it...
https://www.diyaudio.com/community/...range-line-array.242171/page-128#post-4845884
I've seen several concave line array projects over the years on various speaker forums. Not many of those projects have been duplicated by others as the spatial coverage is just too limiting. Concave line array would limit spatial coverage for my listening habits. But you are welcome to try them. Creating an array that 'focus' the drivers is one way to overlap beam patterns without resorting to electronic time and phase delays. These tend to be one listener in the room approaches which is waste of coverage from either straight and CBT line arrays for more dynamic sound. Another thought is that you consider headphones instead of building a concave line array speaker.
If you wish to build a straight array or a CBT array, you will have a speaker that will be enjoyed with others for years. Either a near field straight line array or a CBT would help mitigate room reflection effects. My Modified CBT24 project is still my daily listening system, while my past straight line arrays were awesome as well.
https://www.diyaudio.com/community/threads/my-new-line-array-its-a-modified-cbt24.313352/
If you wish to build a straight array or a CBT array, you will have a speaker that will be enjoyed with others for years. Either a near field straight line array or a CBT would help mitigate room reflection effects. My Modified CBT24 project is still my daily listening system, while my past straight line arrays were awesome as well.
https://www.diyaudio.com/community/threads/my-new-line-array-its-a-modified-cbt24.313352/
I'd have to agree on that! The arrays are for home entertainment and movies in this house and are meant to be family entertainment.
If that were not the case, just maybe I would have build something else. Come to think of it, if I had a room all to myself... I don't think
that would be a good thing. I believe having some restrictions made me work harder to get great results.
If that were not the case, just maybe I would have build something else. Come to think of it, if I had a room all to myself... I don't think
that would be a good thing
. I believe having some restrictions made me work harder to get great results.Apologies- I dont think I was clear in the initial posts- this will be a straight line array, 24 drivers per side, floor to 7' ceiling. I have 50pc of the tc9, and the hardware for up to 36 DSP channels (hoping to keep it to 24 in practice). Amp channels coming shortly once I finalize the arrangement and figure out the power requirements per channel.
I think its most productive to leave off the questions for exploring on my own once I have it built. Not much changes about the cabinet build regardless.
I agree its a super limited use case and wouldn't merit building a curved cabinet. However, the DSP system allows having various presets- so for solo listening (most of what happens here tbh) having this available sounds intriguing. Could a concave/focused array could be created using tapered delay? It would create a very small sweet spot presumably, but would it have any benefits like reduced comb filtering, room reflections, etc?
That leaves me with a final question for now- trying to determine power requirements needed per channel. My initial thought is to group the 24 drivers into 6 groups of 2P, for a total of 12 channels per side. TC9 with 2 drivers in parallel has a sensitivity of 90.8db.
Until I can get more competent at using virtuixcad I won't have an exact answer, but if I were able to keep it below 1.5W per channel I think I have a shot at using the inexpensive PAM8406 in AB mode. Anyone have thoughts about the power requirements I can expect? Listening distance is 9'.
Do you plan on getting any bass out of it? As that would determine the limits I suppose,
I run a 400 watt amplifier. However the maximum I'd want to feed to the arrays is more like 200 watt.
But having more on tab usually means you have clean (non clipping) watts available for peaks.
A driver is more likely to be broken with an undersized amp that's clipping than an oversized amp delivering a few watts more
for a limited time.
I run a 400 watt amplifier
. However the maximum I'd want to feed to the arrays is more like 200 watt.But having more on tab usually means you have clean (non clipping) watts available for peaks.
A driver is more likely to be broken with an undersized amp that's clipping than an oversized amp delivering a few watts more
for a limited time.