Able to use multiple Topping DACs of same or different types simultaneously with PC?

[mike7877]

I'm in the planning stage of converting one of my passive 2 ways to active.

Originally, I was planning to use my RME Babyface Pro's 4 channel output. But now that I have an LA90D, DX1 and G5, and I'm probably getting an E30 II to power the [very low distortion] tweeter (power is enough if anyone's doubting - the tweeter is just 4% shy of the ideal 80% woofer / 20% tweeter power ratio (the tweeter being 6.5dB/W more efficient than the woofer helps, along with the LA90D only putting 41W into 8 ohms...). Amp choice isn't the point of this thread though.

I'm wondering if I can I use my DX1 for the lows and G5 for the highs, or buy another DX1 to add to my first one so I can use them both for the 4 channels my actives will need, and keep my G5 for headphones only.

I've noticed the DX1 actually shows up as "G5" in Windows' Control Panel -> Sound. I'm hoping having two cards with the same name isn't going to cause some issue. In the past I've actually had issues (granted it was a while ago with Windows 7 lol, but it is M$, they could've been just sitting on their hands with their thumbs up or something). Will this naming cause issues? The same XMOS chip twice cause issues? Can anything I'm not considering cause issues?

 

[gvl]

ESS chips are running in async mode in Topping DACs, so technically there can be non-constant phase differences on analog outputs between DACs. Whether or not it’s a big issue I don’t know but if I wanted to build an active crossover I’d probably use a single multichannel DAC just to make sure all channels are in sync.

 

[ppataki]

It will not work (or at least I could not make it work)
See here my thread about it and how I solved it: https://www.audiosciencereview.com/...ichannel-full-digital-amp-fda-solution.40232/

TLDR version: you will need a hardware device to solve this; either a multichannel DAC as mentioned above by @gvl or a multichannel DDC (that is the way I went about it)

 

[Sokel]

How would you time-align them in such a configuration?
Also,have you considered your power needs for peaks (not max) at the tweeter area?

They can be pretty high up there (even momentarily) with the right material and clipping is the No1 tweeter enemy.

 

[mike7877]

It will not work (or at least I could not make it work)
See here my thread about it and how I solved it: https://www.audiosciencereview.com/...ichannel-full-digital-amp-fda-solution.40232/

TLDR version: you will need a hardware device to solve this; either a multichannel DAC as mentioned above by @gvl or a multichannel DDC (that is the way I went about it)

So when you feed two identical DACs with SPDIF, the delay is the same, so everything works out?

I wonder if I could set up my RME Babyface Pro for its XLR outs to do the woofers through the LA90D, and sync the RME Babyface Pro SPDIF out through the DX1 on it with my oscilloscope...

If I couldn't do that, does ADI-4 DD look like it'd work? Out of production but so is my Babyface Pro - they should be very compatible:

Babyface Pro has optical out, can be either ADAT or SPDIF. Plug ADAT into ADI4-DD's ADAT IN 1, and finally, plug in dual DX1s to SPDIF OUT 1 and SPDIF OUT 2?

Used they seem to be a little less than half the price of what you bought, which is why I'm exploring other options. Hopefully I can even sync with o-scope

 

[ppataki]

So when you feed two identical DACs with SPDIF, the delay is the same, so everything works out?

Yes, correct
Provided that the DDC you use can sync the outputs (Digiface can, MiniDSP UDIO-8 can, MOTU LP32 can, and I guess all such devices can)

What you descibe above could work if the delays introduced at various outputs (like the XLR and ADAT of the Babyface and the SPDIF outs of the ADI-4) are constant
Then you can compensate that with additional delay
But in real-life how that would look....I have no idea - maybe somebody here with more pro audio experience could comment on this

I would definitely use just one device for simplcity and for less hassle (+ it will guarantee that it will work fine)

 

[mike7877]

How would you time-align them in such a configuration?
Also,have you considered your power needs for peaks (not max) at the tweeter area?

They can be pretty high up there (even momentarily) with the right material and clipping is the No1 tweeter enemy.

I think that RME allows you to add delay to outputs. I don't know how exact the adjustment is yet, but I'd wager at least down to 1ms. Individual samples, though, I think crossover software would allow to make up the difference in delay if RME doesn't do samples.

I'm thinking what I said in post 5 now - Babyface Pro XLR analog outs to LA90D, Babyface Pro SPDIF out to DX1.
I think I'd play a sine wave and send it to both the Babyface Pro XLR 1 and 2, and the DX1 L and R. I'd then hook up my favourite o-scope

Send 90kHz, level match them, then time align them with delay

Yes, correct
Provided that the DDC you use can sync the outputs (Digiface can, MiniDSP UDIO-8 can, MOTU LP32 can, and I guess all such devices can)

What you descibe above could work if the delays introduced at various outputs (like the XLR and ADAT of the Babyface and the SPDIF outs of the ADI-4) are constant
Then you can compensate that with additional delay
But in real-life how that would look....I have no idea - maybe somebody here with more pro audio experience could comment on this

I would definitely use just one device for simplcity and for less hassle (+ it will guarantee that it will work fine)

Hmm... I'm just getting this idea here - could I not use the DX1 for the tweeter, Babyface for the woofer, and just add a delay to the analog XLR out of my Babyface Pro to match the latency of the DX1?

I was having a brain fart earlier, too, and the DX1 doesn't have optical in - only USB! There is a DAC I could exchange it for that costs about the same so that's not a huge issue. But what do you think about Babyface XLR out and DX1 getting 24/192, playing back 90kHz, level matching it using o-scope, and then adding delay to Babyface XLR out 1 & 2 until it matches RCA outs of DX1?

 

[mike7877]

Also,have you considered your power needs for peaks (not max) at the tweeter area?

They can be pretty high up there (even momentarily) with the right material and clipping is the No1 tweeter enemy.

So the speakers are going to be convertible - 2 way passive and 2 way active. Passive for louder listening, active for more critical listening.
The speakers actually come in an active version as well, but because the passive version costs ~$5,000, and the active version almost doubles that price, after looking at the design and coming across the -81dB THD+n @2.83V amp spec in a promotional video (it unintentionally included in a clip - only present a few frames and I had to zoom in reaaaally far), I decided that I was going to use my Kingergetics KBA-280 (140WPC x2 Class A MOSFET with 1000W toroidal transformer) to drive them. Their 3rd order butterworth made with 1% massive thick air core inductors and 1% custom capacitors spread as far apart and on 90 degree angles (when appropriate) seemed a better bet with my very powerful, clean amp for the time being - and if I wanted better sound in the future, $5000 is a lot of money!

-81dB THD+n at 1W is about the same as my Denon x3700H. I got 9 channels of it, they're capable of 135W each at -98dB THD+n, they came with a bunch of HDMI, switching circuitry, DACs, preamps, op-amps, etc. etc etc., and for just 1/3rd the price of the increase...

Anyway... I got a little off track. The active version uses a 12dB/oct crossover at 2.1kHz, and the tweeter gets 50W, while the woofer gets 200W

Since I'll be using an 18dB/oct crossover at 2.1kHz, the same performance should be able to be had with 45W

If we divide both 45W and 200W by 4, we get 11W and 50W.

The L30 II can effectively drive the tweeter with 7.9W (with it being 6.5dB/oct more efficient), and the LA90D is 41W

41/50 = 0.82

0.82 * 11 = 9.02

The ratio should be 9.02W to 41W, but it's 7.9W to 41W. Not ideal, but not far off either.

After implementation, if the tweeter seems to be running out of power before the woofer, I think I would try increasing the crossover frequency from 2.1kHz to 2.5kHz or 2.7kHz. I think that would for sure even things out.
I get your point that clipping is (especially) bad for the tweeter, and I really like these tweeters so I don't want anything bad to happen to them!
For louder listening, I'd flip a couple switches and change and input, and then BAM bi-amped, passive crossover!

 

[Philbo King]

I'm in the planning stage of converting one of my passive 2 ways to active.

Originally, I was planning to use my RME Babyface Pro's 4 channel output. But now that I have an LA90D, DX1 and G5, and I'm probably getting an E30 II to power the [very low distortion] tweeter (power is enough if anyone's doubting - the tweeter is just 4% shy of the ideal 80% woofer / 20% tweeter power ratio (the tweeter being 6.5dB/W more efficient than the woofer helps, along with the LA90D only putting 41W into 8 ohms...). Amp choice isn't the point of this thread though.

I'm wondering if I can I use my DX1 for the lows and G5 for the highs, or buy another DX1 to add to my first one so I can use them both for the 4 channels my actives will need, and keep my G5 for headphones only.

I've noticed the DX1 actually shows up as "G5" in Windows' Control Panel -> Sound. I'm hoping having two cards with the same name isn't going to cause some issue. In the past I've actually had issues (granted it was a while ago with Windows 7 lol, but it is M$, they could've been just sitting on their hands with their thumbs up or something). Will this naming cause issues? The same XMOS chip twice cause issues? Can anything I'm not considering cause issues?

I see this come up a lot on home recording forums. The basic problem is that the word clocks are not locked together, so they drift apart over time.
It is a bigger problem for recording than for playback; a 7 minute recording could have a significant time drift between 2 (or more) A/Ds running at once by the end. But in playback, you'd likely see at least variable phase shifts between the various outputs during playback.

A couple of 'cheats':
- If one interface has an SP/DIF input, and the other an SP/DIF output you can use that to lock the clocks by using one as the clock source and the other as a clock slaved to that source.
- Ditto as above for optical toslink DAT inputs & outputs.

If you want to try it with unlocked clocks, get the (free) program ASIO4ALL. It allows you to map multiple DACs and ADCs on a PC into a single ASIO software port on the PC.

Note: On Macs or Linux you *can* do this. Just not on Micro$oft...