ECS nforce4a-939 locking up like crazy

[bboy]

Originally posted by: vck6v3
About the USB I didn't have any problems transferring several hundred megs of MP3s to my memory stick duo through a card reader plugged into one of my cases front usb ports.
I also have no problems flashing my mobile phone using the USB cable.

I just checked the ECS Site and they just updated the BIOS to 1.1d.

BIOS Name BIOS for NFORCE4-A939 (PCB:1.0)
Version 1.1d
Release Date 2005/12/26
-Support FX-60 VID for 1.35v CPU
-Support AMD FX-60 Dual Core processors
-Update BIOS for AMD new E4 CPU.
-Fix Patch PCI device just need option resource cannot work.
-Support Dual Core 3800+ CPU
-Dual Core CPU Support
-Support Rev.E3 CPU.

What brand of card reader is it? Do you leave the reader plugged in while it's not being used?

 

[Zap]

It may not be an overall USB issue, but perhaps may affect only certain devices. I say that because I can use a USB 2.0 flash drive or USB 2.0 card reader seemingly without problems, but my USB 2.0 HDD bay will give me lockups. Of course non-2.0 USB devices don't seem to cause a problem (mouse & scanner).

 

[bboy]

Originally posted by: Zap
It may not be an overall USB issue, but perhaps may affect only certain devices. I say that because I can use a USB 2.0 flash drive or USB 2.0 card reader seemingly without problems, but my USB 2.0 HDD bay will give me lockups. Of course non-2.0 USB devices don't seem to cause a problem (mouse & scanner).

What brand of card reader is it? Do you leave the reader plugged in while it's not being used?

 

[Zap]

It's a SanDisk external 6 in 1 or something... I normally don't leave it plugged in because all the extra drive letters bug me, but I'll leave it plugged in for a day or two to see. I had my USB 2.0 flash drive (Emprex 256MB) plugged in and accessed for a number of hours without problems.

 

[AkumaX]

Hmm.. i have some generic 10 in 1 usb2.0 reader that i use to transfer stuff from my MSPD to my psp, but no problems there.

i did have some inconsistency with using an ide->usb cable on a hdd (it would just die, after maybe about 5 min of sustained file transfer, but no freezing)

for those of you installing your main windows installation using SATA, are you guys using floppy drive + F6? i used nlite to slipstream my drivers, but i don't have a SATA drive to see if it worked



Also, do you guys trust your Vcore1, +5v, and +12v ratings?

 

[Zap]

Originally posted by: AkumaX
i did have some inconsistency with using an ide->usb cable on a hdd (it would just die, after maybe about 5 min of sustained file transfer, but no freezing)

My external drive bay is pretty much like that,especially since I never put the cover on. Drives overheat in that box with cover on really quick, and sometimes even uncovered. I noticed that my Seagates can handle it, but my one Samsung 160GB tends to overheat and then performance becomes erratic. If I shut it off and let it cool down, performance goes back up.

I haven't checked voltages and I don't have any SATA drives in this system. Too many of the hot deals have been on EIDE drives.

My external USB 2.0 SanDisk card reader has been plugged in since my previous post. No problems so far.

 

[Zap]

I ended up with my external drive bay plugged in for a week without problems, with occasional use. So far the only lockups have come from the USB 2.0 external drive bay. It seems as if it can be plugged in for a long time without problems, but only under constant use does the lockup problem rear it's ugly head. Actually have used the external drive about 3 times in the last week and it locked once, after about an hour of sustained usage.

 

[IlluminiX]

Hey all,

I actually own the Epox 9npa Ultra board and as most of you are aware, it has the same problems as the ECS board. Abit and Shuttle also both released boards with the freezing with USB 2.0 issues, but Shuttle managed to fix them up for the SN25P board (Info) I have tried to get Shuttle to tell me what they did in their Y Bios to fix thie issue, but they won't tell me, but maybe someone else can wiggle it out of them.

I hope most of you know by now that it is an issue with X2 processors ONLY in tandem with certain motherboards and USB 2.0 enabled along with high-speed devices like USB HDDs and network adapters. I have done a ton of research and much discussion is given to this topic at the Epox 9npa Anandtech forum, and the only solutions out there are to disable one of your cores by putting /onecpu in the boot.ini (but USB2 shouldn't be worth doing that!), buying a USB 2.0 PCI (or PCI-E if you have the dough), or only using USB 1.1.

I think right now the issue might be related to this:
Errata 123 - Link to BIOS info

Errata 123 refers to the 123rd bug identified in AMD Athlon X2 and Opteron processors. This bug affects the cache bypass feature in those processors.

These processors have an internal data path that allows the processor to bypass the L2 cache and initiate an early DRAM read for certain cache line fill requests, even before receiving the hit/miss status from the L2 cache.

However, at low core frequencies, the DRAM data read may reach the processor core before it is ready. This causes data corruption and/or the processor to hang.

That is currently my only lead, but I feel as though I exhausted Google's info on the subject, so the only other option is to call different motherboard manufacturers and find out how they dealt with the issue. Epox tells me they are working on it, but it's been 6 months since the X2's came out so I am not really hopeful anytime soon.

Just thought I'd post, letting y'all know you are not alone!

IlluminiX

 

[Zap]

Thank you IlluminiX. After reading a couple of other posts like this one you've certainly shed more light on this issue. I think with the ECS board since they were bundled with the X2 CPU (as you've said the issue only pops up with the dual cores) the issue popped up very frequently. Also, since they are "ECS" people tend to scrutinize them more. These two things combined in a one day sale where untold number of people purchased this exact combo resulting in this long thread that assumed it was an "ECS problem."

 

[tcsenter]

"Yes, there were some very low end OEM boards were manufactured by ECS but they were never sold in the retail channel. Up to this point, NONE of the ABIT boards are manufactured by ECS, period." -- dated 07-01-2003

Note the date of this statement.

Now compare these side-by-side photos:

ECS 661FX-M v. ABIT SG-72

ECS KM400A-M2 v. ABIT VA-20

Both ABIT retail models.

Abit, however, said this is simply another step in a long term plan. "In fact, we began to outsource our very low end production, like MicroATX, to ECS [Elitegroup Computer Systems] two or three years ago," said the representative.

"Now, we're moving low end ATX products to other manufacturers, once we're sure about their quality. ECS is one of our production partners, but we're also looking for other qualified EMS [Electronics Manufacturing Services]."

lol! How could ECS be a 'production partner' for a few years now, but never produce any boards for ABIT? DUH!

 

[Zap]

Originally posted by: tcsenter
ECS 661FX-M v. ABIT SG-72

These look similar, but there are numerous layout differences including obvious differences in traces visible on these two pictures. These boards are definately NOT clones of each other. The similarity may be due to most manufacturers copying "close to" reference layouts. This doesn't mean to say that the Abit board in the picture wasn't made by ECS, just that it isn't the "identical" PCB.

A few of the differences:

-Between the IDE ports and ATX power connector and close to the edge of the PCB the Abit board has what looks like jumper blocks. You can see the traces go around the bottom of these blocks. On the ECS board they go straight through the area.

-The solder pads for the missing SATA ports on the Abit board are closer to the Southbridge than on the ECS board.

-The layout of the CPU power circuitry is different on the Abit board, even down to the silkscreening of the missing capacitors. Also the mosfets have a different layout and differ in number.

-The battery is in a totally different spot.

-The AMR/CNR (whatever it is) slot on the ECS board is not present on the Abit board unlike in the next example (as solder pads).

-In the corner of the PCB by the "front" of the third PCI slot, the Abit board is pretty busy with what looks like the audio chip and lots of supporting circuitry. The ECS board actually has almost nothing in that corner and the audio chip is located where the network chip is located on the Abit board.

-The area around the top of the AGP slot in the picture (between the AGP slot and the rear audio/USB ports) has obviously different trace layout between the boards.

Originally posted by: tcsenter
ECS KM400A-M2 v. ABIT VA-20

These do look completely identical and most likely rolled off the same production line.

 

[tcsenter]

These look similar, but there are numerous layout differences including obvious differences in traces visible on these two pictures. These boards are definately NOT clones of each other. The similarity may be due to most manufacturers copying "close to" reference layouts.

No other SIS661FX mATX board from any other manufacturer even comes close in similarity to these two boards. I've looked. So much for the 'reference' design explanation.

The boards are similar enough in layout and even PCB markings that a learned eye reasonably can assume they are different major revisions or contract options of the same board that came from the same manufacturer. ECS owns a huge contract PCB and electronics design firm and designs at least as many boards for ABIT than it manufactures.

 

[Zap]

Originally posted by: tcsenter

These look similar, but there are numerous layout differences including obvious differences in traces visible on these two pictures. These boards are definately NOT clones of each other. The similarity may be due to most manufacturers copying "close to" reference layouts.

No other SIS661FX mATX board from any other manufacturer even comes close in similarity to these two boards. I've looked. So much for the 'reference' design explanation.

AOpen s661FXm-US looks quite similar to the Abit board but same "differences" I saw between the Abit and ECS. So, perhaps Abit and AOpen use closer to the reference design while ECS strayed a bit more? Perhaps AOpen made the board for Abit? Perhaps ECS makes boards for both Abit and AOpen?

Other boards using this chipset do differ a bit, but most of the major things (chipset location, drive connector location, etc.) are in the same locations. One thing I did notice was that some of the boards do not have SATA ports, not like the Abit where the solder pads are there but the ports are physically missing, but that the boards don't even have the solder pads. I'm thinking there may be some differences in the Southbridge used. Indeed all the boards I could find that have SATA (ports or just solder pads) look almost the same.

BACK TO THE TOPIC...

I had another thought regarding the USB problem. I'm thinking that it definately is some bad juju going on between certain USB 2.0 devices and the motherboard. I continue to have problems with the external drive bay, but continue to not have problems with my flash drive and card reader. One thing came to my attention the other day as I was disconnecting my drive was that oftentimes Windows will say that there is still something accessing the device, so it cannot disconnect it. This is the only device I own with this problem and coincidentally is the only device I own that causes lockups on my ECS board. I've used this for a few years now and this problem of Windows saying something's still accessing device has been present no matter what computer/motherboard it is hooked up to. It doesn't do it all the time, but enough (at least 1/3 of the time or more) that it is annoying.

I don't know what good this information does for our cause, but it just seems an eerie coincidence.

 

[yukit]

Originally posted by: Zap

BACK TO THE TOPIC...

I had another thought regarding the USB problem. I'm thinking that it definately is some bad juju going on between certain USB 2.0 devices and the motherboard. I continue to have problems with the external drive bay, but continue to not have problems with my flash drive and card reader. One thing came to my attention the other day as I was disconnecting my drive was that oftentimes Windows will say that there is still something accessing the device, so it cannot disconnect it. This is the only device I own with this problem and coincidentally is the only device I own that causes lockups on my ECS board. I've used this for a few years now and this problem of Windows saying something's still accessing device has been present no matter what computer/motherboard it is hooked up to. It doesn't do it all the time, but enough (at least 1/3 of the time or more) that it is annoying.

I don't know what good this information does for our cause, but it just seems an eerie coincidence.

I use firewire for the external drive, therefore, avoiding the USB issue all together.
I also get the "still in use" error when I disconnect the firewire drive, but clicking the "stop" button the second time usually disconnects the device.

 

[bboy]

Been running for a day with no lockups on BIOS 1.1e. Keeping my fingers crossed.

b

 

[Zap]

Here's to hoping. If it cures all your USB related lockups, then I'm soooooo there! I usually don't bother updating BIOS "just for the heck of it" because my thinking is that if it ain't broke... but the USB is giving me the blues on just one of my devices. Use the heck out of your USB devices for the next few days and keep us updated. :thumbsup:

 

[bboy]

3 days with no lockups on 1.1e. I only have the flash reader plugged in though. Tonight I'm going to try my external hdd.

b

 

[bboy]

Originally posted by: Zap
Here's to hoping. If it cures all your USB related lockups, then I'm soooooo there! I usually don't bother updating BIOS "just for the heck of it" because my thinking is that if it ain't broke... but the USB is giving me the blues on just one of my devices. Use the heck out of your USB devices for the next few days and keep us updated. :thumbsup:

Just for curiosity, which USB port are you using for the flash reader?

b

 

[Zap]

-------N
U2---U4
U1---U3

There's my ASCII art depicting how the ports would look on the back if the machine were lying down on its side.

N = network
U1 = mouse
U2 = HDD
U3 = card reader
U4 = scanner

 

[Zap]

Just had an anomoly. Was using my external flash card reader and it disappeared from Windows. Had to pull the plug and re-plug it in, and then it was fine. No lockup or BSOD.

 

[yukit]

Originally posted by: bboy
Been running for a day with no lockups on BIOS 1.1e. Keeping my fingers crossed.

b

Did you need to clear the CMOS to boot after the flash?

 

[bboy]

Did you need to clear the CMOS to boot after the flash?

For the last couple I had to do it from a boot floppy because the WinFlash would fail. Are you seeing the same thing?

b

 

[bboy]

Something else I just noticed through device manager. Looks like the 4 ports on the back of the board are divided between 2 controllers actually. One is Standard Enhanced PCI to USB Host Controller and one is Standard OpenHCD USB Host Controller. Wonder if that has something to do with it.

b

 

[yukit]

Originally posted by: bboy

Did you need to clear the CMOS to boot after the flash?

For the last couple I had to do it from a boot floppy because the WinFlash would fail. Are you seeing the same thing?

b

I always flash BIOS from a boot floppy.
The last two times, I had to explicitly set the jumper on the MB to clear the CMOS settings before the system would boot to even get into the BIOS settings.

Maybe it was because I had my system overclocked.
I would not want to upgrade the BIOS until I have a reason to open my case to reset the CMOS. It maybe sooner than later though, I maybe swapping the stock HSF with something else.

 

[carstea1]

it's not the PSU, it's something driver/bios related. i have almost the same thing with an asrock dual 939. i've tried anything you can name and no result. it just happens randomly.