Tired of trying to make NCCH-DL/Magma/UAD-1/Sonar work!

[TweakFreq]

Through many trials and great expense, I finally have a semi-stable system (only crashes occasionally now with BSD) that has incredible horsepower (never exceeds 55% CPU, 10% disk I/O) but can only manage to successfully mix about 32 simutaneous tracks and 88% UAD-1 CPU without crashing. I won't get into all of the nitty gritty details of the hardware configurations, problems and solutions that I have tried here because I could write a novel if I did. At this point, I'm ready to scrap it all and start over. Basically, I'm interested in finding out if anyone has personal experience with a 100% stable DAW running Sonar 4 Producer with 4 UAD-1's. I would like to be able to mix at least 48 simultaneous tracks with the full dsp power of the 4 UAD-1's. I already have a Magma chassis which I found to be an absolute requirement for the UAD-1's to work at all with the NCCH-DL. Any suggestions? Thanks...

My current DAW hardware configuration: Asus NCCH-DL; Dual Xeon Noconna 3.0; 2GB Corsair 2-2-2-5 DDR400; Ati Radeon 9500 Dual Monitor; 3 WD 74GB Raptors; Firewire push/pull drive bay w/ multiple backup drives; Magma/Avid 7 slot PCI expansion chassis; 4 UAD-1's; MOTU PCI 424 w/ 2408mk3 & 24i/o; 660W Enermax PSU.

 

[billybk1]

Have you tried contacting UA support directly? I would give them a call and hear what they have to say. In the Support Report section of the October 2005 issue, of the UA webzine, it was reported that UA in it's beta testing of the 4.0 update found excellent UAD-1 stability with minimal host CPU loading using an Intel dual-Xeon system. This system was built by UA around a PCI-Express-based motherboard utilizing the Intel E7520 chipset: A SuperMicro X6DHE-XG2 with dual 2.8GHz Xeon processors, five PCI-X slots, 1GB DDR2-400 RAM, and the nVidia GeForce PCX 5300 PCI-Express graphics adapter. If all else fails, worst case scenerio would be to swap out your current MB with the SuperMirco X6DHE-XG2 using the Intel E7520 chipset. With it's (5) PCI-X slots (which work with the UAD-1), you may not even need the Magma anymore.
That way you can get back to making music again and get the full use out of all your UAD-1's. Anyway, just thinking out loud here.

http://www.uaudio.com/webzine/2005/october/index.html

I would still give UA a call and see what they have to say.

 

[TweakFreq]

In the Support Report section of the October 2005 issue, of the UA webzine, it was reported that UA in it's beta testing of the 4.0 update found excellent UAD-1 stability with minimal host CPU loading using an Intel dual-Xeon system.

Thanks for the information. I have spoken with tech support directly on many occasions and we have successfully worked out solutions for many of the issues that I had with the NCCH-DL. Although my DAW is functional in its current state it still crashes sometimes and it just is not at the level of performance that it should be considering the machine that it is. By your post you certainly have insight into my frustration. I'm really tired of playing the role of computer tech. I just want to get back to producing music. I will look into the SuperMicro X6DHE-XG2 option. Thanks again.

 

[winnie_pooh]

... for many of the issues that I had with the NCCH-DL. Although my DAW is functional in its current state it still crashes sometimes and it just is not at the level of performance that it should be considering the machine that it is ...

Hi TweakFreq,

I’m absolutely with you and the frustration resulting from an unstable system. I had a very similar problem with my Sonar box which I built around the MSI K8N Neo2 Plat, RME HDSP and 4 UAD-1 recently. I got BSOD with driver errors, IRQ_NOT_EQUAL etc. ... BS messages especially under heavy load and excessive memory usage without being able to identify the problem. All drivers like Matrox, RME, UAD-1, Midisport8x8s have been crashing somehow randomly. I had tried a lot of testing without results. As a last resort I came across MemmtestV1.60 through some OC forums to find out that the fucking highly expensive 2x512MB OZC 2-2-2-5 Plat something memory kit was faulty on both modules and always failed in test #4! Now the crazy thing: After adjusting the memory timing to 2-3-2-5 latency these two suckers worked for hours without a glitch!

After replacement against a 2GB OCZ 2-3-2-5 memory kit everything went back to normal. I have roasted these modules at least 10 hours without one error. Based on that good experience I have ordered a second kit to get 4GB installed. You won’t imagine what happened: After installation of the additional 2GB they have been recognised by the BIOS as 4GB but after a couple of cold starts only 3GB were detected by the BIOS! After testing of all modules individually it turned out that one of two new ones must have died after a couple of minutes of operation. The defective one wasn’t detected by the BIOS POST routine as defective during operation. Only after testing each module individually the defective one was detected as such. Another disturbing side effect was that after inserting the defective module and replacing it against a good one the MOBO was always dead! The LED on the USB backplane bracket and the POST beep codes indicated a memory/VGA error. It was required to set the 'Clear CMOS jumper' all the time for at least 10 minutes to get the MOBO back to live. Well that could have been a BIOS (ESCD) problem and seems to be better with version 1.C3 now. But I’m still waiting on the 2GB replacement. :roll:

To make this long story short: I will never trust any memory module without excessive MemtestV1.60esting before use especially these low latency modules. That’s why I would recommend that you should test your memory modules before you take any further cost intensive action. That’s what I learned from the OC guys! :wink:

 

[TweakFreq]

[/quote]I will never trust any memory module without excessive MemtestV1.60esting before use especially these low latency modules. That’s why I would recommend that you should test your memory modules before you take any further cost intensive action. That’s what I learned from the OC guys![/quote]

Well, it appears that bad memory is something that we share in common. I've already been down that same road. In my situation, one stick of my Corsair memory failed the Memtest v1.60 test. I tried to return it to the store where I purchased it and wouldn't you know it...the module tested perfectly fine on their test equipment. They insisted that my problem was with the chipset on my mobo. So I purchased a replacement mobo. They also managed to mix up my TWINX matched pairs. Arggghh! I was originally running 2 sets of 2x512 TWINX DDR400. So, I went on ebay and bought 2 TWINX sets of 2x512 Corsair XMS DDR400 2-2-2-5 of which one stick ended up being bad as well. The guy that I bought it from replaced it thank goodness. I installed the replacement and all of the RAM tests passed with flying colors! Now for the weirdness. Next I re-installed the original pair of 2x512 TWINX sets into the original mobo and now that mobo and original memory has passed all memtest v1.60 tests as well. Did Memtest v1.60 give me false readings the first time or were the planets not aligned on that particular day? Odd thing is that after replacing and shuffling all of this RAM around, my specific problems that were related to bad memory went away. Go figure...

 

[billybk1]

Odd thing is that after replacing and shuffling all of this RAM around, my specific problems that were related to bad memory went away. Go figure...

I had a similar problem, about 14 months ago, that coincidentally started to rear it's ulgy head right after I installed my 2nd UAD-1 and the v3.5 update. I would get a BSOD, in SONAR, while using my UAD-1 plugins after a short while. This went on for weeks and I really thought it had something to do with either my 2nd UAD-1 card or the 3.5 driver update.
To make a long story short, it had nothing to do with the UAD-1, but faulty RAM. After running the memtest diagnostic I was able to pin it down to a particular RAM slot & module. One of the first suggestions was to simply pop the RAM moudule out, clean the contacts with any anti-static cloth and firmly re-seat it. Sure enough that worked and I have been problem free ever since and I took back everything bad I said about my 2nd UAD-1 card. So sometimes simply re-seating your RAM modules will do the trick. They say you should re-seat all of your cards/modules on your MB, once a year, as part of a computer preventitive maintenance program.

 

[TweakFreq]

[/quote]They say you should re-seat all of your cards/modules on your MB, once a year, as part of a computer preventitive maintenance program.[/quote]

That makes perfect sense...consider how you have to clean the contact points in all of your audio gear periodically as routine maintenance. I wonder if a good cleaner like DeoxIT D5 or ProGold G5 would be beneficial or detrimental to the computer components. Or maybe reseating is all that is necessary...Just a thought.

 

[winnie_pooh]

... to clean the contact points in all of your audio gear periodically ... reseating is all that is necessary ...

Just received the 2GB replacement so I’m back on testing. :roll: I will check the single modules first without the two good ones. I will keep the cleaning idea in mind just in case. :idea:

Reseating is more or less the nature of the beast I fear. :wink: