TOPIC:

Hi there!

Is there a recommended way to hook up an active stereo 3-way (triamp) configuration with the SHD series?
I see how an active 2-way configuration would be configured, but is there a simple trick for 3-way? I'm mostly interested in the internal routing options of the processor...I've read thru the manuals, but not sure it's going to be elegant?

It seems I would need to use two SHD units to pull this off?

I believe the first unit would be the master where DIRAC and all that happens....then I would send the HF and MF signals thru the analog outputs into those amplifiers. Then I would send a full range signal from the digital output into the second SHD digital input. I would do the LF processing and send it from the analog output into that final set of amps?

Can the digital outputs be different than the analog outputs?
Am I able to do both a high-pass and low-pass on every analog output?
What is the latency thru the box so I can offset the extra DSP latency? (Digital IN to Analog OUT)
Can I bypass DIRAC on the second unit? Does it get the DIRAC output from the first unit?
How would the volume control be handled? Is this applied to the input source?

I'm currently running three stereo Crown XTi amplifiers (w/ built in DSP) and I want to upgrade the quality of my setup. I'm planning to use Nord Acoustics nCore multichannel amplifiers (www.nordacoustics.co.uk/product-page/nor...nnel-amplifier-black)
...and want to use the SHD for the streaming music, EQ, and DAC functions.

Will MiniDSP be releasing a 6 channel version of the SHD so we can do stereo 3-way? I'm using home-brew horn loaded speakers that require constant directivity frequency contouring, so a passive xover is out of the question (I also think direct from amplifier sounds better).


I'm also considering the Cambridge CX Series 2 with DEQX, but that's a lot more money, and I hear the DACs in this new SHD series are pretty good?
www.cambridgeaudio.com/usa/en/products/hi-fi/cx-series-2/cxn
www.deqx.com/products/hdp-4/
This seems overpriced compared to the SHD and the good reviews and measurements I've seen thus far. I also like the idea of having the units combined into one box...I'd love to have the SHD Power offered with nCore amps and the option to build it out the way Nord Acoustics (and others doing this nCore "kits") are doing, but splitting into two boxes is better than 3 total units (I hate cables).

I would love to get everyone's thoughts. How would you tackle this?

Thanks!

The SHD only has 4 channels of output so for an active 3 way you need an additional SHD.

Another way to do this would be to use a SHD Studio which only has digital outputs and another product with digital inputs and enough analog outputs for a 3 way (OpenDRC-DA8, 4X10HD, etc), although it sounds like part of the reason you are interested in the SHD is for its DAC so that may not be an attractive option for you.

The analog and digital outputs are the same but you can still achieve what you want with two SHDs. You would use analog outputs 1-3 to handle the woofer / mid / tweeter for one side (left) and then use the routing matrix to send a full range signal for the other side (right) to output 4. You would route the channel 3/4 digital output from the first SHD to the digital input of the second SHD and handle the right woofer / mid / tweeter in the second SHD.

You can both high pass and low pass every output.

I believe the latency is 25 ms.

You can use the SHD without using Dirac (as you would on your second SHD). As the SHD only does stereo Dirac it has no understanding of individual output channels, just left/right. As a result if use Dirac on your first SHD it would apply to both left (first SHD) and right (second SHD).

I assume you would use the volume control on the first SHD although you could also use individual source volume control.

Hope this helps.

Michael

I’m sure this has come up before with folks looking to build a 3-way active system (or 2-way with sub)

Can minidsp respond?

Daisy chaining products is not an elegant solution. Is there a 6 or 8 output product on the roadmap?

minidsp never comment about future products, only released products

fair enough... but note to minidsp... it seems like a natural product to release

Thanks Michael!

One challenge with your approach is that there is only 30ms of delay available per channel. If the latency is 25ms, then it seems only 5ms is left over for speaker processing? I think I need about 18ms with the bass horns in my system...

I'm also worried about symmetry for the stereo image. If the first unit is doing volume control, then there will be different dynamic range between the L and R channel. There can be differences in jitter too, and different power supply noises, etc...I think it would really harm the mono center image - even if the artifacts are super small. It's just how our brains are wired.


Totally agree with the others about having an easy multi-way expansion option in the future. I like the direction this system is going, and feel like it's a great price point. I don't mind the idea of purchasing two units, but it'd be nice if there was a simple "digital daisy chain" where you can choose the tap point, and then have a clean way to maintain stereo symmetry and have a shorter latency (tap it before the delay buffers).


Thinking about the delay limitations further.....


Can I send the LF signal to Channels 1/2 on the first unit, and then a full range on Channels 3/4 to the digital outputs?
Then on the second unit, Channels 1/2 are MF, and 3/4 are HF?

This way, the 18ms of delay from the subwoofer ends up being 7ms early....so I do the time-alignment on the second unit.


Does anyone know if these units have PLL's to do clock recovery, or is the jitter locked to the signal source? I imagine there's gotta be some level of clock recovery if doing network streaming...I'd rather have the HF/MF on the first processor, but having it on the second should be fine I guess.

Am I missing anything here?

All miniDSPs have an asynchronous sample rate convertor on the digital inputs. In the case of the SHD it converts to 96kHz.

I think your plan would work but I'd suggest drawing it out in detail on paper