• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Burning in Mushkin blues on Ultra-D

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

aka-DoC-

Registered
Joined
Mar 26, 2004
My specs are in sig. Purchased the Mushkin blues from outpost. Tried running on optimized settings and this ram just doesn't like 2-3-3-10. It will boot at 2-2-2-8 when it won't with 2-3-3-10. First time burning in ram, and read alot of the stuff around. Sort of fond of Uwackme's method cause the name rulz... anyway.

I am running at 2-2-8-2 with CPC on and getting errors in the thousands at just 220 FSB. This is with vdim at 3.3. Run it down to 215 and get none. This just seems way to low to me........ I am assuming it's necessary to run the mem divider 1:1 otherwise the burn in would be pointless? Should I just go at 216 or 217 as a starting point and go from there? Also, what is the best way to get these heatspreaders off? I know it voids warranty but they have to go.

Settings as of now:
FSB: 215
LDT/FSB: 4
CPU startup v: 1.425
CPU v: 1.475 110%
LDT v: 1.4
Chipset: 1.8v and DRAM at 3.3v

CPC on
Tcl 2
Trcd 2
Tras 8
Trp 2
Trc 7
Trfc 15
Trrd 5
Twr 2
Twtr 1
Trwt 2
Tref 3120
Twcl Auto?
 
215MHZ.. that is above spec. You will need to burn these in a bit. Check your BIOS also. Also try to test 1 stick at a time and see if one is giving you any probs.

Duonger
 
Yeah, I tried one at a time yesterday to make sure both sticks were good. They both ran at spec with no errors in memtest as long as I up the volts to 2.9 which I think is what they were rated with at 2-3-3-10.

What should I check in bios....
 
see if you have latest bios. you can post at the mushkin forums and one of our tech guys will reply.

Duonger
 
Nice, I gave those a shot. Last night was getting up to 300 errors a pass, this morning when woke up had dropped down to 150. I switched a few of those numbers, enabled blank sleeve interleave and dropped the tref and it's runnin about 10-20 a pass now @ 216. Which bios are you running ninjazx?

I feel like such a noob now with this new board, lol. I have to say this stuff doesn't have any problems on this board as compared to when I was tryin to get it to run on my NF7S v2.

I wish A64 tweaker worked with XP64 so I could screenshot :eh?:

I figured out the heatspreaders..... Here's a pic>>>

Blues.jpg
 
Last edited:
Nice pic...hope you don't mind, but I've saved it, highlighting the Winbond identifying marks, and will use it from now on when asked how to identify Winbond.

White = BP PCB identified
Blue = small hole in corner
Red = metal staples on edges of chip
Yellow = dimples on chip

What do you think?
 

Attachments

  • Indentify_UTT.jpg
    Indentify_UTT.jpg
    28.9 KB · Views: 271
LOL, I only took 8 pics to get a good shot. Glad I have finally contributed something! :attn:
 
If it only took 8 tries, you did good. I had to buy myself a tri-pod after trying out my newly purchased camera. It was that or quick drinking coffee.

Heck, I think I would have had to be dead in order to be still enough to take a clear picture without a tri-pod. So I figured that $13 for a cheap one at Big Lots was better than the other alternative.

Again, nice pic and a very good contribution. The rest of us will benefit from that picture of yours.
 
Well, I tested some of my limits initially with the 1-25 bios and had this:

Max HTT/FSB 280 @1.4V
Max CPU 270 @1.5v
Mem 210 Mhz @ 3.2v

Chipset @ 1.8 v

Setting everything back down at default my initial 3DMARK: 3532, 3DMARK03: 7988. Just using prime95 and memtest to check for stability.

With some of the bios tweaking help and memory burning the memory is hitting 216 now @ 1.5-2-2-5. Fun stuff.... I just wish I had another system set up just to burn in memory. Once I figure out my new limits I'll stick them in here. Any recommendations are welcomed. This is only my second oc, so you people know 100% more than I.
 
We only see a portion of the PCB, but I have to say that it must be a recent production by looking at how clean the board is (free of clutter and excess traces). If think you could only get better with BGA chips. I've never believed in "burn in" with semicons most E-Engineers would say you don't burn in chips. They may deteriorate overtime, but if you see changes in them overnight, then you'll have dead chips in days or weeks. Temperature does affect operation, when they warm up they work a little better due to voltages and timings. This means that every time you "cold-boot), you may see crashes for the first minutes until they reach warmer temps. Point us to an artical or white paper on "burn In", I would like to know more about it.
 
There are many things out there that can be observed, but not adequately explained. Burning in seems to currently be one of them. Search the net for burning in RAM, you should find a bunch of anecdotal evidence.

CJ
 
Well all I can say is that these would barely cold boot at 210 to start with at 2-3-3-10. I had to put them at 3.2 v to get them there stable. I stuck them down to 1.5-2-2-5 and I had to clear cmos to even boot, lol. If I went to 212 it BSOD'd me to death no matter what volts. After 20 hours or so of burnin they cold boot at 216 with no problem at 1.5-2-2-5 @ 3.1 v. It's slow process just because I can only burn them at night, but it's working so far.

More of the stick for Alabama.... http://www.adobe.kodakgallery.com/BrowsePhotos.jsp?&collid=36630283108&page=1&sort_order=0
 
Last edited:
aka-DoC- said:
My specs are in sig. Purchased the Mushkin blues from outpost. Tried running on optimized settings and this ram just doesn't like 2-3-3-10. It will boot at 2-2-2-8 when it won't with 2-3-3-10. First time burning in ram, and read alot of the stuff around. Sort of fond of Uwackme's method cause the name rulz... anyway.

I am running at 2-2-8-2 with CPC on and getting errors in the thousands at just 220 FSB. This is with vdim at 3.3. Run it down to 215 and get none. This just seems way to low to me........ I am assuming it's necessary to run the mem divider 1:1 otherwise the burn in would be pointless? Should I just go at 216 or 217 as a starting point and go from there? Also, what is the best way to get these heatspreaders off? I know it voids warranty but they have to go.

Settings as of now:
FSB: 215
LDT/FSB: 4
CPU startup v: 1.425
CPU v: 1.475 110%
LDT v: 1.4
Chipset: 1.8v and DRAM at 3.3v

CPC on
Tcl 2
Trcd 2
Tras 8
Trp 2
Trc 7
Trfc 15
Trrd 5
Twr 2
Twtr 1
Trwt 2
Tref 3120
Twcl Auto?

I find that I cannot do TRTW so low. It has to be 4 or else thousands of memtest errors. I am stable at 230,1:1 1T,2,2,10,2,9,13,01,2,2,4,3120,1 Bank Interleave=on, 3.1v

You might want to try TRTW=4 and see what happens.
 
I am an Elecrical Engineer and have had BH-5 and UTT that, after "burning-in" have improved stability at higher frequencies and more voltage. I have not, however, really tried developing an explaination of this process and how it works. I do know that it has helped. To answer the OP's questions, I had Mushkin Blues that initially did not like the voltage coming from the 5v rail, so I reset the jumper to the 3.3 volt rail and used 3.2v and noticed better results.
 
What sort of bandwidth should you shoot for on these sticks? I really don't know much about it, but I know if it's too low your OC really doesn't mean anything.
 
Over the years weve heard a lot of good ones,
Frogs cause warts, cell phones blowup gas stations and Elvis is alive (in Vegas he is).
Meant as a constructive argument, keep this thread alive, we might all learn something!

Ram burn-in is suspicious but there may be a bit of quantum physics at play here. When you stuff a gazillion capacitors into a few milimeters of silicon and it may be possible that these suckers change slightly by altering the metal used in the mfg process. (Theory:G:) This would mean that the atomic structure of the semicon changes to an improved chemistry allowing the caps to charge and dump faster. I would doubt that any change would occur in the cas/ras driver or the data registers due to the nature of logic gates though some may argue that point. Still if all this is true, I would like to know what the effect is over the duration of time. Does it improve in ever smaller fractions over time or does it eventually deteriorate the component over time till it fails. Google was useless in seaching this matter as every forum out there has discussions of this phenominon. This might explain also why after a failed OC and a BSOD or 2 mysterious events occurr making you thing you fragged a chip or finally toasted ole celly, only to find it boots the next day.
 
Trying out the 414-2 Bios. New timings I have been running and burning at for a few days.
220 1:1 @3.3v
CPC Enabled
1.5, 2, 5, 2, 10, 16, 2, 2, 1, 4, 1168, 1
Blank Interleave enabled
+ Skew, 0, 8, 4, 7ns, 256, 16X, 7X
Bandwidth 2531
 
Back