Avid Pro Audio Community

Avid Pro Audio Community

How to Join & Post  •  Community Terms of Use  •  Help Us Help You

Knowledge Base Search  •  Community Search  •  Learn & Support


Avid Home Page

Go Back   Avid Pro Audio Community > Pro Tools Software > Windows

Reply
 
Thread Tools Search this Thread Display Modes
  #1  
Old 05-17-2012, 09:59 AM
Bob Nagy Bob Nagy is offline
Member
 
Join Date: Jan 1999
Posts: 111
Default An empowering adventure

The following story is a bit long, but I tell it here in the hope that it may be of use to others who take the road of build-your-own.

For those who want the short version, it is this : Check your voltages.

It all began a couple of years ago when I built a new machine, which was to be used for my general work plus also Pro Tools HD. The machine worked fine, or seemed to at first. Pro Tools ran fine. Everything ran fine. But once every one or two months, it would blue screen, or lock up, or reboot, or something. It was rare, but it kept happening, and of course I wanted to fix it.

The first thing I found, eventually, was the CPU cooler. I had used the stock one that comes from Intel. The one with those silly plastic push pins. The pins had worked loose, and the cooler was not in firm contact with the CPU. This lead to high temps. Ok, I replaced it with a decent cooler, the CPU now sits slightly above room temp, and I thought that was the end of it. But the blue screen parade continued at the same uni-or-bi-monthly intervals.

The next thing I found, eventually, was the voltages. I had known that the power supply I had used was old, and that some of its voltages looked a bit low. But it worked. So I forgot about it. I decided to check it out again. I did some looking around on the internet, and found the recommended voltages for the CPU and memory, plus memory timings, etc. I plugged all of this into the BIOS settings, which had all been set to Auto. After doing this all problems stopped. No more blue screens, no more crashes. Rock solid operation.

All was well until a couple of days ago. I had not used PT for some time. I set it up again now, and of course it was not working right.

DigiTest showed no errors. I ran it many times, and it always completed with success. But when I’d load Pro Tools and open a session, most of the time it would throw a 9161, 9162, or 9031 error. Pro Tools would also become very sluggish, as if something was having trouble and timing out. To my understanding, these DAE errors indicate trouble communicating with the DSPs.

Sometimes PT would run fine on one session, but throw errors on another session. Sometimes it would throw errors even before opening any session. It was not a 100% failure, it was semi-intermittent, but it was failing a lot.
My system contains two HD cards, one HD PCI-X core, and one PCI-X Accel process.

I tried to troubleshoot it, and did the following :

1) Tried PT versions 10, 9, and 8. All with the same results.
2) Tried a different flex cable. Same result.
3) Swapped the cards’ PCI slots. My motherboard has only two PCI slots. I traded slots for the two cards. Same result.
4) Ran Windows 7 (the usual OS) and also OSX on the same system. Same result.
5) Since the PCI slots on my motherboard are not contiguous, I am using one of my TDM extenders. Of course the thought occurred to me that this may be the culprit. I made another one and tried it. Same result. (I should also note here that I and many other people have been using these extenders for a long time with no problems.)

If I pull the Accel card and run with only the core, everything works fine. I loaded up the session that always fails (when the Accel card is in), and ran it overnight in loop playback mode with only the core card. Worked perfectly.

Ok, so now it starts looking like the Accel card is somehow failing. Even though DigiTest says it’s fine. This seems somewhat odd, but also seems possible. No test app is ever going to be as thorough as the real app, and thus the possibility exists that Pro Tools is exercising the card in some way that DigiTest does not, and which is failing.

I don’t have another card here to swap, which makes troubleshooting harder… I find a guy near me who I have never met but who graciously allows me to bring my Accel card to his studio, and who then spends an hour of his time to put my card in his PT rig and test it for me. And of course, the card works fine in his system.

As I drive back home after doing this, I spend half the time in disbelief, still thinking that the Accel card is really bad, but somehow is not being exercised in the same way in his machine. And then it hits me.

Check your voltages.

I arrive back home, put the Accel card back in the machine, and fire up my computer, going into the BIOS screen that shows the power supply voltages. I notice that the 3.3 volt buss is at 3.07. Hmm. That seems rather low. I put all of the BIOS settings back to Auto, removing my prior adjustments of CPU and memory voltage, etc. The 3.3 volt buss is now up to 3.10. Still low, but higher. I fire up PT, Accel card and all. And it works.

There is a well known saying “If it ain’t broke, don’t fix it.” Many people are likely nodding in agreement right now, having proven the wisdom in that sentence. I think the companion sentence to that is : “If you change anything, you WILL break something.”

I started with a power supply that had a 3.3 volt buss that was under voltage. But it was high enough that everything worked. My changes in the BIOS to set the CPU and memory voltages (which I think is not related to the 3.3 volt buss being low) to stop the blue screen song and dance fixed one problem but created another. It caused additional strain on the power supply, and caused the 3.3 volt buss to dip just enough to cross the borderline from working to not working. With two HD cards in the machine, the power draw was too much. The HD cards would not be reliable and the PT software would have trouble communicating with the hardware. With only one HD card in the machine, the load would be less, the 3.3 volt buss would raise slightly and cross the borderline in the other direction, and PT would work.

I borrowed a power supply from another friend, and the 3.3 volt buss is now at 3.29. And of course everything is working fine. I put my CPU and memory voltage changes back into the BIOS settings, and machine and Pro Tools are both happily running and adequately powered.

Yet another friend commented to me a long time ago, and then again recently during this adventure that he had a similar but different problem. He thought he had a bad HD card too. After much drama he eventually found out that it was the CPU memory. He swapped the memory for another, and the problems with PT went away. He concluded that the memory was bad. Well, I am starting to suspect that it was his power supply voltages.

His system is large, with (I think) four HD cards in the computer. Each HD card is a significant load on the power supply, with the nine DSPs on the card pulling lots of juice, pretty close to the maximum that a PCI slot can deliver (IIRC). Memory is also a large power draw. It could be that his power supply was also being taxed heavily by all of this, and that the new memory he put in somehow drew slightly less power. There is no way to prove this now, but it does seem possible that this is what happened in his case.

Now, as for DigiTest. Why would it show everything is OK, yet Pro Tools would crash and burn ? My guess is that DigiTest runs one DSP at a time, testing them individually. Pro Tools puts the pedal to the metal, turns everything on at once, and hopes it all works. PT would thus cause the HD cards to draw more power than DigiTest does. This of course suggests (Hello Avid !) that a new test mode in DigiTest that turns on all DSPs full blast would be a great addition. It could be the final test phase. This would help uncover potential power supply problems like the one described here.

In the PC universe, with the number of power supplies available, we really have a high probability of things like this biting us. Certainly we can and should check power supply voltages. I did, saw mine was a bit low when I built the machine two years ago, and foolishly did nothing about it. That was my mistake. But having DigiTest be more thorough in this regard would certainly be helpful.

In any event, the lessons for me in all of this are :

1) Throw the push pin CPU cooler in the trash and buy a decent cooler.
2) Never trust the Auto settings on your motherboard BIOS. At the minimum, set the CPU voltage, DRAM bus voltage, DRAM frequency, and DRAM timings if you can. This holds true even if you are not overclocking anything. I am running everything at supposedly stock timings. The Auto settings just don’t do the right thing all the time, and in my case caused intermittent failures. The CPU voltage was too low, etc.
3) Check your power supply voltages. The HD cards (IIRC) draw a considerable amount of 3.3 volts. If this buss is low, you’re going to have problems.

My thanks to friends Scott (aka Sunburst), Tim (aka x9), and Ben for all of their help during this adventure.
Reply With Quote
  #2  
Old 05-17-2012, 04:36 PM
Trainwrecker Trainwrecker is offline
Member
 
Join Date: Sep 2008
Posts: 124
Default Re: An empowering adventure

Thanks for the helpful post Bob.

I always spend on the power supply on all builds, You need to rule out when building a machine. Current build has a CORSAIR AX850. Worth the money for a good DAW machine.

There is a good choice of quality PS these days, with a coupon or a sale, you can rule out the voltage issues...
Reply With Quote
  #3  
Old 05-17-2012, 05:00 PM
Dism Dism is offline
Member
 
Join Date: Apr 2008
Location: Toronto, Ontario, Canada
Posts: 6,154
Default Re: An empowering adventure

I want to highlight these points, because they are very important, yet often overlooked factors when building a computer. Especially one being used for professional work or high end gaming.

Quote:
Originally Posted by Bob Nagy View Post
1) Throw the push pin CPU cooler in the trash and buy a decent cooler.
A thousand times, yes. If you intend to really push your CPU, a stock cooler will not cut it, even without overclocking. There's a reason why there are fancy cases with a million vent holes for fans, and expensive liquid-cooling options. Heat dissipation is of uncanny importance when it comes to both performance and lifespan of your system.

Quote:
Originally Posted by Bob Nagy View Post
2) Never trust the Auto settings on your motherboard BIOS. At the minimum, set the CPU voltage, DRAM bus voltage, DRAM frequency, and DRAM timings if you can. This holds true even if you are not overclocking anything. I am running everything at supposedly stock timings. The Auto settings just don’t do the right thing all the time, and in my case caused intermittent failures. The CPU voltage was too low, etc.
Another huge, important point. Default/Auto settings are only default for the motherboard, and may not have any use in the real world, especially when you are using an unusual (for the 99% of PC users out there) addon like a TDM card. A small voltage change can be the difference between constant blue screens, and a system running smooth as silk.

Quote:
Originally Posted by Bob Nagy View Post
3) Check your power supply voltages. The HD cards (IIRC) draw a considerable amount of 3.3 volts. If this buss is low, you’re going to have problems.
In line with the last point, also extremely important. Something a lot of people do (who build their own computers) is keep using the same old power supply because "it works," without realizing it may not be supplying enough power to keep everything on all the time. Plus, like everything else, power supplies fail and lose efficiency over time. A 650w power supply from 10 years ago isn't going to be as efficient as a brand new one.
__________________

D
a n t h e I n c r e d i b l e S o u n d M a n

"Svetlana" v1 - 4.2GHz i7, 16GB RAM, OSX 10.7.4
Liquid Saffire 56 - PT10.2 - BFD2, VCC, Duende Native, Play 3.0

_C U R R E N T-D V E R B-S C O R E:515
Reply With Quote
Reply

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
A little iLok adventure... audiofp Pro Tools 9 41 05-14-2013 07:47 AM
Losfer Words - A DC Modern adventure adamqlw Eleven Rack 9 02-04-2011 12:50 PM
The Adventure Continues....... AngelEarth 003, Mbox 2, Digi 002, original Mbox, Digi 001 (Win) 3 04-16-2006 10:48 AM
Karel and Baz's Excellent Adventure Bazzle 003, Mbox 2, Digi 002, original Mbox, Digi 001 (Win) 81 09-22-2004 10:29 AM


All times are GMT -7. The time now is 01:41 AM.


Powered by: vBulletin, Copyright ©2000 - 2008, Jelsoft Enterprises Limited. Forum Hosted By: URLJet.com