I-Appliance BBS
The Official Source for Internet Appliance Upgrades and Mods

Click Here!
BBS Main List | Sign In | Sign Up | Search | Help | Linux-Hacker.netReply to Thread | Printer |

Home / I-Opener Areas / I-Opener Technical Stuff
Stability above 450 Mhz

New MessageStability above 450 Mhz (modified 0 times) Ragnar
Profile
Gang,
I know have my V4b, 128ram, 2.01voltage amd k6-II+ processor stable (for the most part) at 450mhz. The only problem know is when the video is trying to load a picture above ~2mg it will back screen. Everything else continues to work. My question is this a current draw problem or simply a problem caused by the limitted 2mg video? I have ran at 500mhz, but have not determined if the problem that I am having at 450 is the same one(s) at 500. On my V1, 64ram, 2.1voltage, amd k6-III 450mhz, it runs rock solid, but at 500 almost always does the continual rebooting thing. I can live very happily with my V1 at 450, but would like to have my V4 running at 500 or above. Would dropping the voltage on my v4 help? Any suggestions? Thanks.
07-06-2001 09:43:56

New MessageRE:Stability above 450 Mhz (modified 0 times) Ragnar
Profile
Please replace "know" with "now" and it should read a little better. Ouch!
07-06-2001 09:46:42

New MessageRE:Stability above 450 Mhz (modified 0 times) Ragnar
Profile
OK,
After a little research, according to Turbo3 in the thread "V1/V2 mod for 2.2Vcore" he says 500 mhz is stable at Vcore = 1.8V using an amd k6-2+. Since I have the famous S4 switch I didn't have to bother doing a hack to change the Vcore. Now, it's just a matter of figuring out the easiest hack to get Vcore down to 1.8V on a V4b. I quess that doing a resistor divider would be the best way to go. Any ideas?
07-06-2001 22:28:45

New MessageRE:Stability above 450 Mhz (modified 0 times) BadFlash
Profile | Email
Easiest way is to adjust R343. I'm not sure what values to use, but that is the base reference point. On a V3 10K sets it to 2.0 volts, so that would be a starting point. Take out the CPU of course, but you can measure the core voltage with the CPU out just fine. If you can find a SMD pot that would be the best.
07-07-2001 14:33:09

New MessageRE:Stability above 450 Mhz (modified 0 times) Ragnar
Profile
Ok,
After a little more research, in Badflash's thread "Hacked V3 boost to 350 Mhz, low risk mod" Tackhead points out specifically that when the screen blanks out you have a current ( i.e. R302 ) problem. So I opened up my machine, got out the magnifying glass and inspected my first R302 soldering job. Let's just say that it was a little sloppy at best. You REALLY need to stack the resistors up nice and neat. Redid it with a little more care. Played around with R343 set a slightly lower vcore. Also changed the resistor on my hd light. Closed back up. Powered up. Let Sandra run her tests. Over the last four hours I have done everything I can to crash. No go. And this is at running 500 Mhz. The video screen is brighter. I shoulda figured it out when the colors had started to fade. And the hd light is like a search light. Does anyone else get a good feeling when you finally get something like this to go your way?
07-08-2001 04:55:40

New MessageRE:Stability above 450 Mhz (modified 0 times) BadFlash
Profile | Email
Yea, it just don't happen that often. Butonce in a while is all I need. Sorta like a gambling addiction. I guess it really is.
07-09-2001 10:08:33

New MessageRE:Stability above 450 Mhz (modified 0 times) Ragnar
Profile
Post script

I have been running rock solid at 550 mhz for over a week now. This is with an AMD k6-II+ 450 mhz booting at 300 mhz then warping to 550 with k6clk.exe in the autoexec.bat file. The heat problem is solved with Jack's dropping resistor and a socket7 12V heatsink/fan combo but externally powered using 9V. This is to reduce the fan noise generated at 12V. A second 5V fan is blowing into the case at the ram door. I cut out slots at the top of the metal radio shield above the cpu and the hard drive to allow hot air to exit the top. Max cpu temp is 100F. Max motherboard temp is 120F. Vcore is 1.97V. Instead of the stock power supply, I am using a converted laptop power supply that supplies 18.5V @ 3.5 amps, which is about twice what the original power supply does. Everything is so stable I just may push my luck and try for 600 mhz in the near future. Has anyone else experimented with 600mhz?

07-22-2001 22:00:40

New MessageRE:Stability above 450 Mhz (modified 0 times) GWIZAH
Profile | Email
We are in awe of your overclocking prowess ragnar.....:)
07-23-2001 10:08:54

New MessageRE:Stability above 450 Mhz (modified 0 times) Ragnar
Profile
I have been running at 600 mhz for several hours this morning. No problems so far. My processor info as reported by Sandra is: AMD k6-III model info K6-2+/III+ (Mobile) 450-600 2.0V revision D. Notice the 450-600 mhz. This is the only reason that I decided to test at greater than 500 mhz. I bought this processor at SoftwareandStuff several months ago. I do think that I lucked up with this processor. The AMD K6-2(no +) 500 mhz that I have in my V1 will not go above 450 mhz.
07-24-2001 08:45:26

New MessageRE:Stability above 450 Mhz (modified 0 times) Kcmjr
Profile
Ragnar,
What OS are you running? I have a similar setup using a K62+ and Windows 2000 Pro. My core is at 1.97v and I boot to 300 Mhz cool and stable with BadFlash's mods. I'd like to bump it to 400 or 450 via software but I can't find any references to K6CLK working with Win2K.

KCM

08-13-2001 14:51:28

New MessageRE:Stability above 450 Mhz (modified 0 times) Ragnar
Profile
I'm using win98se. In the k6clk readme file it say " This is probably only good for dos, win3.1, win95, win98, and win98se". It doesn't say that it won't work in win2000. If you want to try it and need the file I can e-mail you the k6clk file. The only specific directive is to make sure emm386.exe is not loaded into your config.sys file. I don't know how similiar 98se and 2000 are so the only way to find out is to test it. If it causes problems the easy way to solve it is on boot up hit the f8 key, choose dos mode delete the file or the lines that you put in autoexec.bat.
08-13-2001 18:41:12

New MessageRE:Stability above 450 Mhz (modified 0 times) Kcmjr
Profile
98 and 2000 are very different animals. 95, 98 and ME (I think) are all DOS based. they're just fancy shells riding on a 16 bit DOS subsystem. 2000 and NT have done away with DOS and are true 32 bit operating systems. That's one reason that if an app crashed in 95 or 98 then the whole system tends to go belly up. I'm not sure about ME, I've avoided it since it's just a big patch for 98. In NT and 2000 the 16 bit subsystem is in it's own protected memory space so if it crashes it usually doesn't tend to affect the rest of the system, just all the current 16 bit apps which normally share the same memory space.

Because of the fact that DOS is not really there there's a good chance that K6CLK won't work. I already have a copy of the file, I just haven't tried it yet. I guess I'll have to give it a try and see what does or doesn't flame out.

08-14-2001 14:40:40

New MessageRE:Stability above 450 Mhz (modified 0 times) Ragnar
Profile
I use the dos command k6clk.exe in autoexec.bat. That may or may not work in 2000. The only way to avoid using dos at all is to manually run k6clkini.com first and then k6clkwin.exe right after that. This way you would have to manually load it every time that you boot.
08-14-2001 15:24:46

New MessageRE:Stability above 450 Mhz (modified 0 times) Programmer
Profile
The reason so many things that affect the hardware don't work in NT is because of protections built into the operating system to prevent malicious programs from damaging the system. Win9X has these protections, but disables them as soon as a program tries to access the hardware. WinNT (including 2K) keep the restrictions in place and they can only be circumvented by a driver that has made it known to the OS that it needs access to specific hardware (which can be rather general). It is possible that the k6clk programs will work as they don't actually change anything (that I know of) in the hardware. They just reconfigure the processer. That said, They may be restricted because of the instructions they execute, I really don't know, and would like to be informed if anyone has success running this on Win2K
08-15-2001 06:22:13

New MessageRE:Stability above 450 Mhz (modified 0 times) Kcmjr
Profile
Well here's your answer.......
On Win2K Pro the k6clkwin program runs. But, at startup and each time you press "go" a window pops up stating "priviledged instruction" and Sandra doesn't show and speed change. Seems that the programs are not NT/2000 compatible. I didn't try the autoexec version. I figure that since the kernel is already loaded and the HAL is in place it still won't work. Looks like I'll be flipping switches....

KCM

08-15-2001 22:27:02

New MessageRE:Stability above 450 Mhz (modified 0 times) ckbone
Profile
The shareware program cpucool does the same kind of tweaking as k6clk. It can adjust the FSB and the multiplier, also can do it automatically upon a reboot. Works great on 98se....It also claims to work on 2k. Try the free version. The registered version has more tweaks.
08-16-2001 04:12:25

New MessageRE:Stability above 450 Mhz (modified 0 times) Ragnar
Profile
I haven't yet found a software program that reports the true speed of the k6 + processors that use k6clk to boost to higher speeds. Only by having a good sense of timing when programs load and using a stop watch to time certain programs do I know that I am truly operating at higher speeds than 300.
08-16-2001 10:42:40

New MessageRE:Stability above 450 Mhz (modified 0 times) Programmer
Profile
The cpuinfo.exe tool from AMD works fine.
08-16-2001 11:36:51

New MessageRE:Stability above 450 Mhz (modified 0 times) Ragnar
Profile
Gang,
I am no longer able to stay stable at 600 mhz. Looking around for an answer I noticed in my bios that my vcore had gone from 2.01V to 2.5V. This is on my V4 that only has the dropping resistor mod along with the sw1,sw2 mod. I am at a lost to explain this jump in voltage. Do I have a potential short somewhere? Has Q16 failed somehow? Testing the various settings of sw4 I get higher voltage from 2.5V to 3.5V. This is perplexing. Does anyone have a clue about what to do? I sure would appreciate any help on this one....
08-16-2001 21:36:31

New MessageRE:Stability above 450 Mhz (modified 0 times) Ragnar
Profile
I think that I have figured it out. Somehow I have disabled R343 or R342, the two resistors that make up the final voltage division before going to the processor. On my V4b, R342 is 25k ohms and R343 is 100k ohms. In the equation to drop voltage from 2.5V to 2.0V these are the right values. k= 100/(100 + 25)= 4/5 so that (4/5)*2.5 = 2.0. I had earlier in the quest for 600mhz soldered a resistor across R343 to drop the Vcore below 2.0V. When I discovered that was unnecessary I remove part of it. In cleaning up the rest of it yesterday I must of cut a trace or something. Now it's just a problem of repairing the trace or setting up another voltabe divider. Now is it the north end or the south end or R343 that goes to ground?
08-17-2001 08:05:47

Reply to Thread | Printer |
All times are PSTPowered by UltraBoard v1.62



Copyright © 2000, Netmake Inc. All Rights Reserved.
See Terms and Conditions for more information.




i-opener opener laptop notebook computer help drivers dll free windows dos repair fix linux mac macintosh 2000 95 98 nt pc configure hardware software sound video netscape explorer network networking lan wan software cmos fat bios printer card mouse modem ide scsi cd rom controllers scanner tape hard drive cgi scripts source code mp3