I have a Soyo Ultra dragon kt600 there is a hard ware monitor on the cd are you says at 1.82 vcore those are your temps? Get the best thermalright

0

Well Jam, I didn't have any dreams about it. lol! But I do know this about the MBM5. If you change any settings, which is basically what you did (the hardware way), the MBM5 recognizes it and will change in the way it responds. It also changes if you update your bios as well. I noticed that the other night when I was fooling around with mine. YO

0

I know my nf7-s undervolts also.I have it set at 1.6v and it also shows 1.568v(cpu-z). Mostly in my oc'ing exploits I've found that the increase in 'multi' usually effects temps more than the increase in 'fsb'. THAT is low temps for over 1.8v though. Knowing you don't like wasting money you must have the standard heatsink? I bet your smiling while scratching your head lol. Tt Lanfire nf7-s v2.0 XP-m 2500@209x11 SP 97 512mb pc3200 Jou Jye 550w psu FX5600 WDCaviar 160gb sata WDCaviar 160gb sata ;~}

0

My board is the Soyo SY-KT600 Dragon Plus v2.0 (aka K7VKPE). I updated to the latest BIOS the day I installed it. This is probably the most basic, stripped down BIOS I've ever seen. It does support 200MHz (400FSB) & PC3200 RAM, but all I have is PC2700. It's a 2ndary machine anyway & will be used mostly for the internet & maybe some light gaming. I checked the CD that came with the board & there's no monitoring software included. The Soyo website has monitoring software for download but for some reason, this board isn't supported. I tried the software anyway but the readings didn't jive with the BIOS readings. The heatsink itself is bearly warm to the touch, so I'm guessing that 30-32C at idle is correct. I didn't have much time last nite but I did run Toast for about 10 min just to get an idea...the temp maxed at 44C. I'm happy with the temps but I'm not happy with the voltage. I plan on pulling the heatsink later today & changing the wires to see if I can get between 1.70-1.75v Asus A7N8X-X 1800+ @ 8 x 210MHz 512MB PC3200 Asus Ti4200 128MB WinME/WinXP Pro

0

The Vcore formula for AMD is something like 1.85 - a*.025 - b*.05 - c*.1 - d*.2 - e*.4 = Vcore where a-e are either 1's or 0's. a-e are the static voltages of certain pins. A 1 is produced when the pin is connected to a Vcc pin, and a 0 is produced when connected to a Vss pin. All this makes sense only if you have the sempron tech doc (from amd.com) in front of you. In there you will find the Vcore pins (which I am calling a-e) and the nearby Vcc and Vss pins. So in your case, "d" in the formula is being pulled low instead of high, producing 1.825v instead 1.725v. Which means one of the socket pins is connected to Vss and it needs to be connected to Vcc.

0

Well, I just pulled the HSF & CPU, removed one of the wires & put it back together. I don't know if that site is incorrect or if it has something to do with the Sempron, but when I had the wires set for 1.725v, I got 1.825v & when set for 1.675v, I got 1.775v - consistently 0.10v off. I think I can live with 1.775v. The idle temp was 32C & system temp 26C. Right now, I'm running Toast. It's been about 25 min & the temp is holding at 41C. I'm gonna let it run a while longer, then move on to Prime95. Asus A7N8X-X 1800+ @ 8 x 210MHz 512MB PC3200 Asus Ti4200 128MB WinME/WinXP Pro

0

post 5 below has more detail about how pin-mods adjust vcore and multipliers. http://computing.net/cpus/wwwboard/forum/6204.html

0

Hey johnoh, Thanks for the tip. I'll go to AMD & see if I can find the tech papers on the Sempron. Everything I've read about the Sempron says that it's essentially just a Athlon XP reprogrammed for 333FSB, so I assumed the pin layout would be the same as the Tbred. Maybe it is & that site is just incorrect? I'll have a look & see what I can find out. In the meantime, I think it'll be fine at 1.775v. I just checked & after 45 min of Toast, the CPU temp maxed at 42C. I've just moved on to Prime95. Thanks again. Asus A7N8X-X 1800+ @ 8 x 210MHz 512MB PC3200 Asus Ti4200 128MB WinME/WinXP Pro

0

I do not think that site is correct. If you are only going to connect three pin pairs, cpus of varying voltages would need different connections, and the tbred came in a 1.5v, 1.6v, and 1.65v default vcore. To use that pin-out they have on the site on each of those three tbreds would produce three different vcores.

0

"So in your case, d in the formula is being pulled low instead of high" I meant c, not d. my bad.

0

johnoh, I think you're on to something there. I didn't consider that there are different default voltages for the different CPUs, & as I wrote in the original post, "The site doesn't show a pin mod for the Sempron, but it does have the Athlon XP, T-bred, & Barton and the mod is the same for all of them" I just downloaded the tech paper for the Model 8 Sempron. I'll look it over & see if I can decypher the pins. Like I said, I'll probably leave it at 1.775v. That way, I shouldn't have to change anything if I decide to upgrade to PC3200 & then switch over the CPU bus to 200MHz. I may try running async to see if the board/CPU/voltage can run stably. 1st things 1st though...lol. Thanks for you help. Asus A7N8X-X 1800+ @ 8 x 210MHz 512MB PC3200 Asus Ti4200 128MB WinME/WinXP Pro

0

Here's an update. I tried running Prime95 but got kicked out of it almost immediately with a FATAL ERROR. Since Prime tests the RAM as well as the CPU, & since I'd never tested this RAM before, I decided to run memtest86. I ran one series at 185MHz but unfortunately, I got a lot of errors & it appears that both sticks are bad. I tried again at 166MHz & still got the errors. Later today, I'll try one stick at a time, 1st at 166MHz, then 185MHz. I got the RAM on sale at CompUSA (Centon & PNY) just before xmas so I may not be able to exchange them, but they both have a lifetime warranty. I may have to fire off a couple of e-mails & do the exchange thru the mail :( Asus A7N8X-X 1800+ @ 8 x 210MHz 512MB PC3200 Asus Ti4200 128MB WinME/WinXP Pro

0

If anyone's still following this thread, I tested the RAM individually. Luckily the Centon 512MB PC2700 tested good, the PNY 256MB PC2700 stick had the errors. I'm currently running Prime95 with the 512MB stick installed & running at 185MHz. So far, so good..... Asus A7N8X-X 1800+ @ 8 x 210MHz 512MB PC3200 Asus Ti4200 128MB WinME/WinXP Pro

0

To finish this thread....I ran Prime95 for over 3 hrs with just the 512MB stick installed & CPU/RAM both running at 185MHz FSB. There were no errors & the temp maxed at 39C. Idle temp was 32C, system was 26C prior to testing. I know that seems low, but the heatsink was only slighly warm to the touch, so I assume it's correct. I have an 80mm front intake fan, 80mm side panel intake (don't like it but case came with it), & 80mm rear exhaust (plus PSU exhaust). The last time I installed the CPU, I tried a different paste than my usual Radio Shack stuff. I had picked up a freebie syringe of "Stars-700" Silver paste several months ago & finally decided to give it a shot. So far, the temps seem comparable. I already sumbitted my info to PNY using their online RMA form. Hopefully the 256MB stick will be replaced with minimal hassle. Asus A7N8X-X 1800+ @ 8 x 210MHz 512MB PC3200 Asus Ti4200 128MB WinME/WinXP Pro

0

Jam, I got a question since you brought up RMA. When you submit an RMA, do they ask for the receipts? Or only the info on the sticker that's on the ram? The reason I ask is the receipts that I have are so faded you can't read anything. Some of that BestBuy quality paper. lol! YO

0

Hi YO, I had copies of the receipts because the RAM was an "after rebate" deal. I always hang on to them until the checks come in. Different manufacturer's do it differently. I've exchanged PC133 RAM with both Kingston & Spectek before & didn't need a receipt. I just went to their website, found the RMA form, & filled out what they asked for. In a few days, I got an RMA number w/instructions in my e-mail, & I just packaged up the RAM & shipped it back. Cost a few dollars for shipping & I got the replacment RAM very quickly. I wasn't sure which stick of RAM was bad at 1st, so I checked both websites. Centon actually asks that you test the RAM with memtest-86 v3.2 & that you supply them with the error results...I don't recall any mention of a receipt. PNY asked for the the date & place of purchase, the UPC number, & a detailed description of the problem. Then, if the RMA is approved, they want a copy of the receipt too. I'm still waiting for a reply from them.... If you've got some bad RAM & it has a lifetime warranty, I suggest you check the manufacturer's website & see what their policy is. Asus A7N8X-X 1800+ @ 8 x 210MHz 512MB PC3200 Asus Ti4200 128MB WinME/WinXP Pro

0

Thanks Jam! I hope they like blank receipts! lol! YO

0

jam did you get the vcore worked out?

0

johnoh, I've decided to leave it as-is for now. Vcore is at 1.775v. The temps are good, system is stable. I reran Prime95 with CPU & RAM at 185MHz for approx 3.5 hrs with no errors. Temp maxed at 44C. I'm sure you understand these pins better than me. I have the Sempron doc from AMD. To get were I am now (1.775v), I have pins VID[3] & VID[4] both jumpered to the same VSS. I'm kinda confused by this because there is a table that lists the various voltages along with the "zeros & ones" under the heading VID[4:0]. At 1.775v it's 00011. If those numbers run from VID[4] to VID[0], I would think it would be reversed (11000). Can you explain? For the hell of it, I tried CPU & RAM at 200MHz. The system booted up, I got into Windows & everything seemed OK, so I tried Prime again...it stopped immediately. I suspected the RAM, so I went back to the BIOS set the RAM at 166MHZ & kept the CPU at 200MHz, but Prime still wouldn't run. My guess is 1.775v isn't quite enough for 200MHz. I needed 1.8v to get the Sempron to run stably at 200MHz on my Asus board. There's no way to lock the PCI/AGP. At 185MHz, the PCI speed should be 37MHz. I don't see any reason to push it any higher than that just to gain another 20-60MHz for the CPU. Asus A7N8X-X 1800+ @ 8 x 210MHz 512MB PC3200 Asus Ti4200 128MB WinME/WinXP Pro

0

"I have pins VID[3] & VID[4] both jumpered to the same VSS. I'm kinda confused by this because there is a table that lists the various voltages along with the "zeros & ones" under the heading VID[4:0]. At 1.775v it's 00011. If those numbers run from VID[4] to VID[0], I would think it would be reversed (11000)." jam, Here's how it works. first, post 11 isn't helping the situation. It says: 1.85 - a*.025 - b*.05 - c*.1 - d*.2 - e*.4 = Vcore but it would be clearer to have written: 1.85 - a*.4 - b*.2 - c*.1 - d*.05 - e*.025 = Vcore where a=VID[4], b=VID[3], etc your default vcore of 1.60 is: 01010=1.600, so a=0, b=1, c=0, d=1, and e=0 by default. i.e. vcore=1.85 - 0*.4 - 1*.2 - 0*.1 - 1*.05 - 0*.025, which does in fact equal 1.60 you are connecting vid4 (which is a) and vid3 (which is b) to Vss, making each of them 0. Therefore your vcore is now represented by: 00010, which is 1.80v. You mentioned your mobo is undervolting a little (or your mobo is underreporting a little - no way to be sure which it is without a probe), so perhaps its showing up as about 1.775v instead of 1.80v note that you could accomplish the same thing buy leaving vid4 as it is, since its 0 by default so to get to 1.725v (ignoring for now that it may be reported as 1.70v by your mobo), since you are starting at 1.60v, here is what you need to do 01010=1.600v 00101=1.725v which means you'll need to change 4 of the 5 values. vid3 and vid1 are currently 1's and to make them 0's connect each to a Vss. vid0 and vid2 are currently 0's and to make them 0's connect each to a Vcc. Anytime you are messing with vcore pins and you end up not being able to boot its probably because you've accidentally set the vcore too low, so if that happens just remove the pins, reboot and start over.

0

"vid0 and vid2 are currently 0's and to make them 0's connect each to a Vcc." oops. should read: "vid0 and vid2 are currently 0's and to make them 1's connect each to a Vcc."

0

Thanks johnoh, A bit confusing, but I think I've got a grasp on it. I didn't take into consideration that I'm already at 1.60v by default & the fact that the board apparently undervolts (or under reports) kinda threw me off as well. So just to test my understanding, let's say I wanna shoot for 1.75v. The AMD doc lists it as 00100. Since 1.60V is 01010, I'd have to connect VID3 to VSS, VID2 to VCC, & VID1 to VSS. Is that correct? Got another question for ya. In the case of VID4 which is already at zero - does the wire to VSS (which I placed following the diagram from the website in my original post) do anything? Asus A7N8X-X 1800+ @ 8 x 210MHz 512MB PC3200 Asus Ti4200 128MB WinME/WinXP Pro

0

"Is that correct?" yes "does the wire... do anything? no

0

Thanks... Asus A7N8X-X 1800+ @ 8 x 210MHz 512MB PC3200 Asus Ti4200 128MB WinME/WinXP Pro

0

no problem jam

0