AOpen AK77-8XN Motherboard ReviewFormer staff writer -
The AWARD bios is colourful, full featured, and easy to use. Right off the bat, I updated the bios to the latest R1.06 from the shipping R1.00 version. It recognized the default settings for the CPU (11X133 fsb) and, on auto, set the RAM for DDR333. CPU core voltage is available up to 1.85v, which isn't going to do it for the phase change/water chiller crowd, but on air I maxed out at 1.75v. Vdimm (memory) only goes up to 2.65v, which is quite low and I feel I could have got more clock cycles out of the board with 2.85v or so - again, the extreme/volt mod crowd go as high as 3.3-3.4v on RAM, AOpen needs to allow a little more latitude if they want to see more of the enthusiast crowd. AGP voltage goes to 1.8v, and wasn't moved from default 1.50v during testing.
When I implied a moment ago that AOpen may want to see more of the enthusiasts' business, here's an example of what I'm talking about. Repeatedly, in the manual, the installation guide, and on their website, they refer to overclocking - not that they support it, of course, but the implication is there. Speaking of support, it's on the weak side for sure. Their website is almost constantly flooded, in fact during weekday business hours, I could not access their downloads section or their forums, although I was able to send an email to tech support. Which I am still waiting for an answer on.
And here's where the biggest problem in my experience with this board is. Remeber I said the basic fsb levels of 100, 133, 166 & 200 had to be set by a jumper? I could NEVER get the system to post with the jumper in the 166 position. Here are some, but not all, of the things I tried: RAM on Auto, RAM on 333, tight, medium, relaxed, default and default turbo RAM timings, max RAM voltage (well, I used it at max all the time anyways, 2.65 isn't going to hurt anything), multiplier on default, multiplier at each .5 step from 7 to 12, (7 X 166 would have run it at 1162, I'm sitting here writing this at 13 X 136 = 1768), and probably about 4-5 hours of messing with other bios settings. When I posted in the AOpen forums (which I was able to access late Sunday night, EST), I got a reply the following day, although not from an AOpen tech. This is what it said:
- "I had the same issue with wanting to run an XP2400 at 12 X 166 instead of 15 X 133. I could go into the bios and lower the multiplier to 12X. Save and boot into Windows....then shut down but not power off. Change the jumper to 166 and it would boot as 12 X 166. Great until you power off, then you have to set the jumper back to 133."
Tell you what, ladies and gentlemen. I have 3 kids to feed and I'm not sticking my fingers in there with power on the board. Actually, I have a guess that the board might be OK at 166 with a 166 fsb native cpu (2700+, 2800+) but I don't have one handy to test with - so it's going to remain a mystery to me, at least for now.
So where does that leave me? Well, I did manage to get an OK but not spectacular 20.5% overclock, from 1467MHz to 1768 (no, I'm not going to tell you AMD's "+ rating", you'll figure it out if you want to know!) very stable, with fast 2-3-2-5-1T RAM timings, on air cooling and eventually settling in at 1.725v - by bumping the multiplier from 11 to 13 and slightly overclocking the fsb to 136. Oh yes, you can still manipulate the fsb in 1 MHz increments, with the jumper in the 133 position the range available in the bios is 130-160. The problem is, there's no PCI/AGP lock and the 1/5 divider isn't enabled until you get the #&@%#$@ 166 fsb and so, no bandwith for YOU, keto. According to AOpen's hardware monitor, I'm idling along at 35C, under load the most I've seen is 38C. On the subject of monitoring, I'll mention another beef - neither AOpen's utility nor MotherBoardMonitor5 show either a) Vdimm or b) chipset fan rotation - and yes, the connection is good and there is a monitoring wire on the chipset fan.
I'm surprised I got any memory increase at all, given I've only upped the system bus by 3 MHz. I will also note that I DID get better results running memory at DDR333, as opposed to in sync at DDR266, but the differences were slight. The better results are what you see above.
In line with what would be expected - significant cpu increase due to higher clock speed, insignificant memory increase, and slight degradation of hard drive scores, either due to statistical variation or possibly because of slightly out of spec system bus when overclocked.
The extra 300MHz equals 869 points increased yield, or 11.3%. Woot, first time I've had this video card over 8k.
4% in Elephant Atrium and < 1% in Citadel. Repeatable results. Not sure, video card limitations? Nahhh. I repeat, repeatable results. Didn't try OpenGL this time around, will next time.
16.7% increase, closest scaling to MHz increase, not a surprise given the age of the technology in the game - 1999 product is ANCIENT by computer hardware standards and though I have a somewhat aged video card, it's still a generation or 2 newer than Quake3. Heh, for a laugh, I went to www.idsoftware.com and checked out the system req's for it, there's not many of you reading this running a PII @ 300 MHz or K2 @ 350, 64 MB RAM and a 16 MB video card, are there. Oh, and DirectX 3.
Overall, it's a stable and well built board, with the added plus that the shiny "pearl black" PCB makes it gorgeous to look at. Other than in my overclocking attempts, I haven't had any issues with stability that couldn't be directly pointed to operator error. (I loaded their system restore utility without enabling it in bios, resulting in a Windows reload. RTFM, keto.) Within the bios, you can find a "credits" menu with the names of the individuals on the development team, I think this points to pride in their work, which in any endeavour will tend to show through.
And I DID get a reasonable overclock, given my components - and hey, that's why I wrote this and you're reading it, most likely. HOWEVER, I cannot get past the fact that I could not get the 166fsb and 1/5 divider to work. The voltage maximums for the vcore and especially the vdimm were much lower than what I would like to see - given the above mentioned quality, there's no reason this motherboard shouldn't have more headroom built in.
I would still give the AK77-8XN a high recommendation to anyone who isn't interested in exploring the extreme limitations of their hardware - providing they are fairly self sufficient with regards to support. It's priced anywhere from even with (MSI) to $75(Abit) CAD less than competing boards with onboard LAN and sound. For you overclockers on a tight budget, this will probably satisfy - provided you have an unlocked processor and can bump the multiplier. If you have a few more bucks to spend, or are going for the extremes, look elsewhere.
One final note, if I do hear back from AOpen on the 166 fsb jumper issue, I will update with their response.