Looking for comment and insight: GA-EP45-UD3R

[BonzaiDuck]

I've seen this and the "UD3P" and "L" boards touted here, although people ask questions when they have problems.

I've had a bumpy road getting this sucker up for initial test and OS installation. It had the following initial configuration:

E6600 C2D
BFG GS-550 PSU [and wait just a minute . . . ]
2x(2x512MB) OCZ Platinum DDR2-800 5-5-5-15
1 Seagate 7200.10 320 GB HD SATA2
1 Samsung DVD/RW burner SATA
1 (old) Sony floppy-drive
eVGA 7600 GT graphics

After several reboots, including an initial entry into BIOS setup, and discovery that I'd mistakenly reversed the floppy signal-cable, I shut down, unplugged the system, grounded myself and corrected the floppy-cable problem. Re-booting resulted in a "site-wiring-fault" alarm on the APC battery-backup. I suspected the BFG PSU -- possibly a weak solder-joint that went south when I connected the power-cable a second time. Replaced the PSU with a Seasonic 550HT. Site-wiring-fault status: CORRECTED.

I had also sent e-mail to Gigabyte asking about fan-plug amperage limits. I had a spare DELTA 92mm fan available for my TR-Ultima-90 cooler. The DELTA was rated by Sidewinder Computers at 0.70A, but the label suggests that start-up amperage was more like 1.04A. Gigabyte's reply: "Don't use a fan rated at more than 1.00A. Three-pin fans will not take advantage of PWM control from the motherboard."

Even with the Seasonic PSU, "weird things" seemed to occur.

First, I noticed that the "6-Phase-LED" lights didn't light up after I replaced the PSU, even though the system posted and I got into BIOS just fine.

Second, I tried running v.1.70 MEMTEST86+ against the four OCZ modules. Keep in mind, that except for resetting the CPU multiplier to 9, I didn't change the BIOS settings -- all the memory settings were set to "Auto." [This also meant that "Enhanced Performance" was initially set to "Turbo."] After 21 minutes, MEMTEST86 was no longer updating its "Wall Time," -- the system was locked.

Third -- so I rebooted, and the system wouldn't post.

Fourth -- I decided to do three things: replace the DELTA fan with a 0.48A Tt "BLUE LED;" pull one of the two memory kits from their sockets (leaving 2x512MB or 1GB); and set the memory "Enhanced Performance" to "Standard."

Now -- I've gone through two iterations of MEMTEST86+ without a hitch. But I notice this: When I fired up initially and entered BIOS, the 6-phase-LED lights on the motherboard were on. But when I rebooted to MEMTEST86+, they're off.

Is this normal? After reading some of the remarks in another thread about "Gigabyte RMA," I hope I don't have to pursue such a thing. On the plus side, the MEMTEST86+ testing seems to be "going forward normally."

Also -- did I correctly recall some other posts suggesting that these motherboards "don't always play nice with OCZ memory kits?"

EDIT-UPDATE: MEMTEST86+ seems to freeze now after three successful iterations. Same symptoms upon reboot as mentioned: Doesn't post to monitor. Forgot to mention earlier (corrected, above) that system used eVGA 7600 GT PCI-E graphics.

 

[LOUISSSSS]

i wouldn't worry about the 6phase LEDS if all other things are working well. sorry you've had a crummy experience. my ud3r is coming this friday.

GBT RMA is a PITA. but if you gotta do it then do it. not much else you can do except: buy another board in the mean time and sell the MB they send back to you. So you can avoid the 1 month down time without a PC.

did you try to clear the CMOS, then load optimized defaults, then manually set all your voltages and timings in the BIOS? never leave them at default.

 

[BonzaiDuck]

Originally posted by: LOUISSSSS

did you try to clear the CMOS, then load optimized defaults, then manually set all your voltages and timings in the BIOS? never leave them at default.

Good point, because usually, and recently for the last few years, I built with the actual intention of over-clocking and I set the voltages manually right away.

This time, the objective is the replacement of my existing file-server, and I'm not inclined to over-clock the system. So I left the voltages at default settings.

The BIOS Monitor shows (what appears to be) a DIMM or RAM voltage at 1.9V. The VCORE for the E6600 shows to be about 1.30V -- which I think is adequate, but could be set higher. If the other voltages need adjustment, I need to do some background reading. This is my first board with an Intel chipset for a few years.

By the way -- did anyone ever figure out why these GA-EP45-UD3x boards seem to set the CPU multiplier a notch lower than the stock value? Maybe someone explained that in another post, but I recall that this was a common experience -- for instance, an E6600 @ 2.1+ Ghz or 8 x 266, instead of a multiplier setting of 9. This . . . . was the one thing . . . that I changed in BIOS.

 

[BonzaiDuck]

Replying to myself and LOUISSSSS

This is what happens as you get older and overconfident with your "vast experience."

I have a lot of work to do with some web-searches on this motherboard.

FIRST: The silly "PHASE LED" lights: They ONLY light up according to CPU loading!! THIS . . . is freakin' OBVIOUS . . . . from the g**d*** "User's Manual!" What on EARTH am I THINKING -- to proceed without getting familiar with stuff like this?!

So today -- I'm going to CLR CMOS, reset to "Optimized Defaults," and start over. Apparently, this feature of the GA-EP45-UD3(x) motherboards that sets a default multiplier lower than the Intel CPU spec is -- or may be -- deliberate.

I can afford to take a month or two sorting this out. Reminds me of the the Meyer-Lansky prototype in one of the "Godfather" movies -- Hyman Roth:

" . . . because . . . DAT . . . is da BIDNIS . . . we're IN!!"

Which means -- "Accept the consequences of the new motherboard/chipset complexities."

HOW---EVER!! If anyone wants to continue this thread with suggested settings or observations relevant to any C2D or C2Q [I'm interested in Conroe E6600 for the moment] -- please chime in with your effusive expertise. Especially if you just LUV to type on your keyboard !! . . . .

 

[BonzaiDuck]

I don't want to start another thread if it's not necessary.

But I'll bet I'm in the ballpark, to suggest that the BIOS setup and features of all these motherboards is nearly the same:

GA-EP45-UD3L
GA-EP45-UD3P
GA-EP45-UD3R
GA-EP45-UD3LR

and possibly:

GA-EP45-EXTREME
GA-EP45-DS4P
GA-EP45-DQ6

I've read the thread on the "Over-Clocking" forum entitled "Is Your Processor Running Slower Than it Should Be? . . . . . "

I'm still mystified that "default" settings for my E6600 (stock: 9 x 0.266= 2.4 Ghz) show a multiplier of 8 in BIOS.

Troubles I've mentioned in posts above disappeared when I let the motherboard have its way for the time being, although I've set VCORE and other voltages to fixed values guided by either BIOS Monitor, User-Guide or other sources.

BIOS shows the processor running at 2.133 Ghz; MEMTEST86+ v.1.70 shows this as a "Core 2 Duo 2.4 Ghz." I've yet to install the OS, because I want to test the memory and other features at their stock settings (at minimum).

What's going on with this multiplier business?

 

[BonzaiDuck]

OK . . . . I seem to have a consistent problem getting the system to activate the monitor and post to the monitor. That doesn't mean it doesn't "post;" It means that shutting down the system -- even turning off the PSU and switching it back on -- doesn't re-activate the monitor through the graphics card and show post messages.

This is an eVGA 7600 GT PCI-E. It's only been used a few hours -- comparatively speaking -- and was purchased as a card for preliminary testing. It has been handled carefully when used on perhaps two or three different systems.

IS IT POSSIBLE . . . THAT THERE IS SOME SORT OF . . . INCOMPATIBILITY . . . . WITH THIS old, budget-model PCI-E card? I just don't get it.

The system ran through some 10 iteratiions of MEMTEST86+ without a hitch, before I come back to this . . . intractable situation. . . . .

 

[LOUISSSSS]

do you have another graphics card to test? or an old old PCI graphics card that can output to analog for you? that should be your backup, sometimes unpluggin the cables from your monitor helps to "clear" some crap. it happened once with my Viewsonic VX2025WM.

 

[BonzaiDuck]

Originally posted by: LOUISSSSS
do you have another graphics card to test? or an old old PCI graphics card that can output to analog for you? that should be your backup, sometimes unpluggin the cables from your monitor helps to "clear" some crap. it happened once with my Viewsonic VX2025WM.

On the video forum, I cleared up any questions about PCI-E 2.0 and older cards. There shouldn't be any problem.

More times than I can count on one hand over the last ten years, I've had components -- either the motherboard, or the memory -- go south such that the computer exhibits the same symptom per the graphics card and raising the monitor out of its sleep-state.

I'm pretty certain the video card has not gone bad.

Tomorrow, I'm going to try a different type of memory module that has a lower voltage requirement. I've already made the preliminary contact with Gigabyte tech-support for the RMA process, so unless I can resolve this as a problem with another component, RMA may be the terminus for all this effort. If it's really the motherboard, then this is the second Gigabyte board I've had that was defective. The first board's defect was minor, and circumvented with a PCI LAN card. But I'm beginning to develop a more clarified opinion of the manufacturer at this point.

 

[elconejito]

Sorry to hear about your troubles, I'd definitely suggest an RMA. That's a whole host of really oddball problems. I'm in the process of overclocking my quad with the same board and love it so far (about 2-3 months).

Just out of curiosity, was the prior defect the gigabit port on a P35-DS3L? I had heard that was a common problem (after I bought one of course) and sure enough about 6 months into it, it died.

 

[BonzaiDuck]

Originally posted by: elconejito
Sorry to hear about your troubles, I'd definitely suggest an RMA. That's a whole host of really oddball problems. I'm in the process of overclocking my quad with the same board and love it so far (about 2-3 months).

Just out of curiosity, was the prior defect the gigabit port on a P35-DS3L? I had heard that was a common problem (after I bought one of course) and sure enough about 6 months into it, it died.

UPDATE TO ALL INTERESTED PARTIES: THIS IS TENTATIVE, but interesting.

Customer-reviews a the reseller site [NewEgg], when filtered on RAM-specific strings, suggest that some models of OCZ DDR2 RAM do not play nice with this motherboard. Other users seem to have no problem with Reaper kits.

I spent two days or so watching TV and doing choirs while ruminating over options for my next step. Gigabyte's tech-support has yet to respond to my questions, posted through their messaging services.

I finally decided today to just pull the (2x512) 1GB OCZ DDR2-800 kit and replace it with a G.SKILL 2x2GB 4GB kit of DDR2-1000's. I know for a fact that these G.SKILLs, when defaulted to native DDR2-800, will run at a low "Auto" setting that yields 1.8V RAM voltage on other motherboards without fail.

System booted back up -- albeit, with the E6600 multiplier set down a notch to 8 -- and is now running MEMTEST86+ v.1.70.

elconejito :

The Gigabyte board I mentioned was a budget-model GA-73VM-S2 mATX board. I think we got two of 'em for about $65 each -- while they were available. One of them seemed to have a defective PCI-E x1 slot (which I'd hoped to use for a gigabit-LAN-PCI-E-x1 card), and the other one seemed to have a defective 10/100 LAN. I was going to use the PCI-E-LAN on the former, and just went to my parts locker and found a PCI-gigabit LAN, so no problem. Then second board was no problem with its defect, since I wanted gigabit-LAN anyway -- and its x1 slot seems to work fine.

For the price of these boards, I never bothered to RMA them for these defects -- which I circumvented with expansion cards -- by necessity or desire.

 

[BonzaiDuck]

MORE "UPDATE" ON GA-EP45-UD3R EXPERIENCE:

I'd be interested for anyone with the following motherboards to chime in:

GA-EP45-UD3L
GA-EP45-UD3P
GA-EP45-UD3R
GA-EP45-UD3LR

Here's what I've done:

1) Replaced the OCZ Platinum XTC DDR2-800's with a set of G.SKILL DDR2-1000 2x2GB I have in the parts-locker.

NOTE: I really want to replace these DDR2-1000's with a 2GB kit, because I think it's overkill for a Windows-Home-Server file-server system. I'm looking at these G.SKILL (2x1GB) 2GB DDR2-800 " . . NQ" modules because 2,500 customer-reviews can't be wrong, one of them specifically mentions the UD3R motherboard with great success, they're CHEAP, they run at 1.8V, and they're one of the few 2GB kits still offered.

2) I've turned off all power-management features on this motherboard. It's now clipping through MEMTEST86+ with no trouble.

HERE IS WHAT I SUSPECT. Most of you running these GA-EP45-UD3(xx) boards are configured with Penryn-based Wolfdale or Yorkfield cores. I strongly suspect that the boards are currently shipping with a BIOS version optimized for these Penryn cores.

I'd be interested in hearing from anyone who is running their board with a Conroe/Kentsfield spec'd for a 1066 FSB. I'd be VERY interested if someone has identified a BIOS revision for this version 1.1 motherboard that works best with Conroe-based cores of the 1066-FSB flavor.

 

[imported_spong]

I'm using the EP45-UD3LR, and I have had this video issue you're referring to. I just updated my system to Ubuntu 9.04 (coincidental), causing the system to reset. It reset but did not display video. I could tell it was booting by noise and it was requiring the use of the 5s override to reset the system, but no video.

I have a 8800 GTX that I transferred from my XPS, it has 2 video ports. I tried the other, also no luck. I was using a DVI cable, but my monitor apparently only listens to the digital interface on that cable, so I put a DVI->VGA adapter on and voila, I got video. So how on earth does that happen? This weekend I will have to do some playing around, and try the latest BIOS (F5 I believe). I have no idea if digital video can be disabled via BIOS (it seems like video BIOS, which would be controlled by my video card, not the mobo)...I'm curious if you have done the same experiment.

I'll see if I can dig up the specs for things, I know I have a corsair 750W psu, G.Skill memory and a high end core 2 duo proc. I believe I am using 1066 FSB. I believe my phase shedding LEDs are all off after boot, and all light up during boot. Honestly I never took them seriously, it seems like a gimmick.

I've never run memtest86 on this machine and haven't had reason to, once up it's pretty solid. But I have had some mysterious problems with it on restart. The reset button for example doesn't seem to work, I usually have to flip the PSU switch. The power button usually works ok, but once in a while it doesn't boot at all. Very strange. I've thought about refunding but Fry's has a short refund period on motherboards and mfg RMA is such a pain...

 

[imported_spong]

PS When I say no video, I mean no post screen, no BIOS splash...nada. That's how I ruled out the OS update

 

[PCTC2]

Hey BD.

I have the EP45-UD3R and a Q6600, 2x2GB of Patriot DDR2-800, and an 8800GTS 512MB.
I have the Q6600 OC'd to 3.3GHz without a hitch. I haven't noticed any issues with the board, but maybe I'm just lucky. I have whatever BIOS shipped with my board, revision 1.1. I bought it a few weeks ago. I don't have a Windows License to install on it, and it's running Mac OS X right now, so I can't help you much on system monitoring in OS, but if you have any Q's about my BIOS settings, toss them my way. I'm at work right now so I don't know off the top of my head. It did default my multiplier to 8x, and I just manually changed it to 9x.

 

[BonzaiDuck]

Originally posted by: PCTC2
Hey BD.

I have the EP45-UD3R and a Q6600, 2x2GB of Patriot DDR2-800, and an 8800GTS 512MB.
I have the Q6600 OC'd to 3.3GHz without a hitch. I haven't noticed any issues with the board, but maybe I'm just lucky. I have whatever BIOS shipped with my board, revision 1.1. I bought it a few weeks ago. I don't have a Windows License to install on it, and it's running Mac OS X right now, so I can't help you much on system monitoring in OS, but if you have any Q's about my BIOS settings, toss them my way. I'm at work right now so I don't know off the top of my head. It did default my multiplier to 8x, and I just manually changed it to 9x.

Thanks to spong as well.

I experimented with both the E6600 and Q6600 (b3 steppings) quite a bit before. I cannot think of any reason why the E6600 might be "damaged," but even that is possible -- anything is possible. I sure handle these CPUs with TLC, though.

If you want to take the time, post your settings for the Q6600 and EP45-UD3R so I can have a looksee.

I'm trying to retire old equipment and still keep a "file-server" in the house. The appropriate equipment would be the 6-year-old P4/533/RAMBUS system. But it's "six-plus--years-old." Usually, I'll keep PC's for 8. However, it was an opportunity to jump on the Win Home Server bandwagon. I want high-capacity SATA drives, so I need a mobo that can handle them. But people say my C2D E8400 is over-kill, and the same E6600 that gave me troubles with the UD3R-board gave me trouble with a 680i board. The 680i board had the latest BIOS that made the E8400 work, and some of us have experienced troubles with different BIOS updates in conjunct with certain processors before.