My MOTU 828 MKII is set at 96Khz and is serving as a clock master to my Octa-Capture also set at 96 Khz. They are connected via SPDIF and they seem to play well together. The MOTU is on a firewire interface directly connected to my computer (specs below) and the Octa-Capture is connected to the same PC via USB 2.0. The mix bus 1 output of the Octa-Capture goes to the SPDIF port so it is sent to the MOTU. The PC has the latest (8/5/11) MOTU drivers installed as does the Octa-Capture.

My desire is to be able to use the interfaces on BOTH devices inside of Sonar to separately track instruments that are in channels on each of them. I made the mistake of using ASIO at first and found I could select one or the other device. Did my homework and realized I had to go with WDM/KS driver mode. Set that up in the record/playback preferences (ummm this was not obvious Cakewalk ;-) and now I could see inputs and outputs for both devices. Spent a few hours but it was learning and at this point all mistakes were user error. Fine.

I created a project that had a simple audio track associated with inputs associated with MOTU analog channels 3,4; a track with inputs associated with the Octa-Capture analog channels 5,6; and a z3ta softsynth just to make sure I eliminated external hardware issues if necessary. Recording timing set to MOTU mix bus 1 and playback timing was set to the MOTU main outs. Output of all audio tracks and the metronome was just sent to the Master. 

What I got when I did this was pure chaos. When starting to record the metronome click sound was just a disaster buzz and I was getting pure garbage on my other tracks. It immediately struck me that I probably had a clocking issue for my audio. Maybe I was wrong but it sounded like it. I went back into the preferences and tried to enable my RealTek Audio driver as a device for input and output hoping that maybe if I used the sound card for clocking everything from both devices I could get what I wanted. 

Enter chaos...Sonar complained that the 96Khz was not supported and proceeded to then claim that the MOTU device was no longer present and change audio to 44.1Khz/16 bits. I got a crash and the lovely SONARPRD process hang.

If I changed the input from the Octa-Capture analog channels to the MOTU SPDIF stereo channels all was well by the way. The metronome was fine as were all recorded tracks.

What I want to know is if what I am trying to do is impossible. Can I hook up both these devices to my PC with WDM and use the ports on each separately or do I really have to go through a single device? Is there some magic clocking issue? Am I just missing something totally obvious?

Thanks in advance to all you folks in the forum by the way. This is an incredible resource. I would also like to thank Albert Collins for cheering me up in the middle of this muck :-)

leave the realtek out of it completely.

I could be mistaken and maybe someone else will correct me if I'm wrong, but I thought you had to use 48k for SPDIF clocking?

you can use 44.1 also, my old Yamaha card was 44.1 and I had to downgrade the sb-live (as a sampler) I was using from 48 using a converter box. Although sometimes it don't work as should with sync'ing , not sure if other factors come into it. I had 2 cards that would loose sync by different makers.. maybe the bitrate comes into it or something like that also. I ended up wiring them 2 cards up via the audio .. 1 was used as midi only , so I could still use the asio driver.

FWIW, If you need 16 simultaneous analog inputs... I'd sell the MOTU and pickup a 2nd Octa-Capture.
All your issues will be gone...
The pair of Octa-Capture units will function (as a single unit) under a single driver.

Fog


you can use 44.1 also, my old Yamaha card was 44.1 and I had to downgrade the sb-live (as a sampler) I was using from 48 using a converter box. Although sometimes it don't work as should with sync'ing , not sure if other factors come into it. I had 2 cards that would loose sync by different makers.. maybe the bitrate comes into it or something like that also. I ended up wiring them 2 cards up via the audio .. 1 was used as midi only , so I could still use the asio driver.

ok - but NOT 96k, right?  only 44.1 and 48k for the clock?
 

S/PDIF is not limited to 48K. In fact, there is no specific limit on the transmission rate of the interface. It really comes down to whether the transmitter/receiver pair can stay in sync which is a function of the cable interconnect (because of signal integrity) and the quality of the hardware.  I have had no issues whatsoever getting the MOTU and the Octa-Capture to send clock/data over the S/PDIF coax links at 96K. They are short cables and are specifically made for S/PDIF. No digital noise discernible at all.

As far as getting another Octa-Capture and strapping the two together, I think I may go that route although perhaps with a Quad-Capture since I really don't need all that many interfaces. I know they are specifically designed to work well together and the Octa-Capture has been fine in my setup so far. It would really be nice to get the MOTU and the Octa-Capture to play on separate interfaces but maybe my time is just more valuable than all this screwing around trying to make this stuff work correctly.

I agree with keeping the RealTek out of the mix. That was not a healthy choice based on what a disaster it caused for Sonar. I was just trying to see if a clock source local to the PC would be the way to go. If it is, I definitely was not doing it right.

If I went with a pure Octa/Quad capture solution I know I could use the ASIO drivers which seem to be rock solid. Who knows? Maybe I can actually get back to playing music again instead of doing hardware engineering :-)

If I went with a pure Octa/Quad capture solution I know I could use the ASIO drivers which seem to be rock solid. Who knows? Maybe I can actually get back to playing music again instead of doing hardware engineering :-)

 
Working with two separate audio interfaces (not units that act as a single unit under a single driver) is always a bit dicey.
 
Two Octa-Capture units will be solid from the get-go... using ASIO  
 

Ok based on the replies I am getting I think I am going to just get another Roland Octa-Capture and strap them together to get 16 coherent channels that can use a single ASIO driver. I give up the ADAT port of the MOTU which I use for my MPC 5000 but that is just not worth worrying about. It would be nice to not spend the money but oh well. At least I was not just being a dummy about putting the stuff together and it really is a problem to do it.

I am going to take my MOTU and use it for mobile recording my piano in place of the Tascam US-144 unit I had been using. If I do that then I am probably going to save my sanity and I can mic the piano and also the bedroom to capture the music and also my wife yelling at me to stop playing so late at night. It will give a truly live feel at 96/24.

inaheartbeat
My MOTU 828 MKII is set at 96Khz and is serving as a clock master to my Octa-Capture also set at 96 Khz. They are connected via SPDIF and they seem to play well together. The MOTU is on a firewire interface directly connected to my computer (specs below) and the Octa-Capture is connected to the same PC via USB 2.0. The mix bus 1 output of the Octa-Capture goes to the SPDIF port so it is sent to the MOTU. The PC has the latest (8/5/11) MOTU drivers installed as does the Octa-Capture.

My desire is to be able to use the interfaces on BOTH devices inside of Sonar to separately track instruments that are in channels on each of them. I made the mistake of using ASIO at first and found I could select one or the other device. Did my homework and realized I had to go with WDM/KS driver mode. Set that up in the record/playback preferences (ummm this was not obvious Cakewalk ;-) and now I could see inputs and outputs for both devices. Spent a few hours but it was learning and at this point all mistakes were user error. Fine.

I created a project that had a simple audio track associated with inputs associated with MOTU analog channels 3,4; a track with inputs associated with the Octa-Capture analog channels 5,6; and a z3ta softsynth just to make sure I eliminated external hardware issues if necessary. Recording timing set to MOTU mix bus 1 and playback timing was set to the MOTU main outs. Output of all audio tracks and the metronome was just sent to the Master. 

What I got when I did this was pure chaos. When starting to record the metronome click sound was just a disaster buzz and I was getting pure garbage on my other tracks. It immediately struck me that I probably had a clocking issue for my audio. Maybe I was wrong but it sounded like it. I went back into the preferences and tried to enable my RealTek Audio driver as a device for input and output hoping that maybe if I used the sound card for clocking everything from both devices I could get what I wanted. 

Enter chaos...Sonar complained that the 96Khz was not supported and proceeded to then claim that the MOTU device was no longer present and change audio to 44.1Khz/16 bits. I got a crash and the lovely SONARPRD process hang.

If I changed the input from the Octa-Capture analog channels to the MOTU SPDIF stereo channels all was well by the way. The metronome was fine as were all recorded tracks.

What I want to know is if what I am trying to do is impossible. Can I hook up both these devices to my PC with WDM and use the ports on each separately or do I really have to go through a single device? Is there some magic clocking issue? Am I just missing something totally obvious?

Thanks in advance to all you folks in the forum by the way. This is an incredible resource. I would also like to thank Albert Collins for cheering me up in the middle of this muck :-)

I know this thread is old but will Platinum handle the two devices together better? I have the Motu MK3 and Roland Octa-Capture and if I can save money and use the two together that would be great.  I have no clue how this works, if come out of the Motu Spdif and in to the Roland Spdif which unit is in control?  If I plan to use the analog channels only is there a need for the clocks to match?  I would use 48k for both if that matters, thanks in advance for shedding any light on this for me

Using SPDIF, yes, you can do that.  SPDIF out from one, into the other.  the one "in control" will be the one you choose as the one with the SPDIF input.
 
using analog only need for clocks to match?  yes. but SPDIF will sync with its clock.  if you don't use SPDIF, however, and try to use them both as inputs at the same time going into the computer, one of them will have to be the master clock and the other will slip out of sync since you are not clocking using a word clock or other syncing mechanism.