Awww... maybe it's pining for the fjords.
Page 58 of this says you weren't all that near the max yet (4.6V) https://en.wikichip.org/w/images/f/f0/Am5x86_ … %2C_1996%29.pdf
Unicorn herding operations are proceeding, but all the totes of hens teeth and barrels of rocking horse poop give them plenty of hiding spots.
Yes, what voltage were you running? I’d expect 4.0v to be fairly safe?
BitWrangler wrote on 2022-02-13, 02:44:Awww... maybe it's pining for the fjords.
Page 58 of this says you weren't all that near the max yet (4.6V) https://en.wikichip.org/w/images/f/f0/Am5x86_ … %2C_1996%29.pdf
Exposure to Absolute Maximum
Ratings for extended periods may affect device reliability.
It’s a shame the LS boards have no socket 3 clips. Check what I’ve done here on this Biostar UUD bending a paper clip around. Seems to hold it quite well but I think I’ll try it with thermal tape as well. I just need a desktop style AT case an an LS board would be fine with this cooler.
I have (had) 4 Am5x86 CPUs that can do 200MHz. One of them actually belongs to Anonymous Coward.
1. An ADW that can do 180MHz on air cooling at 3V but not fully stable with complex compute. With 5V Peltier fully stable. It can do 200MHz with 4V and 12V Peltier. Currently in the 180MHz rig.
2. One ADZ that can do 200MHz with 4V on air or 5V Peltier. But for some reason cannot do 180MHz at 3V. Needs 4V and 5V Peltier.
3. Two ADZ that can do 180MHz with 4V and 5V Peltier. Also, can do 200MHz with 5V and 12V Peltier.
When Chadti99 posted about running 200MHz at 5V on air cooling it clicked with me that CPU #2 can do the same.
Ran it at 4V with medium size aluminum heatsink + fan. It was fine in DOS interactive graphics. Not fine for Win95 and some more complex tasks.
Tried it at 5V - the chip didn't like that - no boot.
Back to 4V this time with 5V Peltier - all seemed good, but still inconclusive. Needed more time to confirm. Left it for the weekend.
Saturday morning (today).
I was ready to complete the testing (200MHz, 4V CPU, 5V peltier) and share a video with the results. Also wanted to see what was this 240MHz business I stumbled upon the other day.
I keep all CPUs in a bin - a pile, messy. Couldn't remember which one of them was #2 from the list above. Only one is marked with AC on the bottom for Anonymous Coward.
None of these in the bin worked. I was like - wtf. Then realized that the processor is still stuck on the bottom of the heatsink that is on the desk with few other components. Decided to give it one last try on air ... aaand that was the end of it.
@Feipoa
Glad I took your advise back then not to get rid of the excess junk.
@bitwrangler and @maxtherabbit
I was switching between 4 and 5 volts. Also this cpu saw a lot of action like that before, but under a Peltier.
While this is the first CPU I ever burned - page 58 got it right. I am believer 🙁
Now I have to source few more ADZs to find magical one like #2 ...
@Chadti99
This contraption is a disaster in disguise - but same sh!t appears on the test bench here all the time. : )
I find paper clips aren't able to produce that restoring force necessary to press the heatsink down. I'd recommend using thermal tape as well.
Has anyone witnessed adhesive thermal tape removing the silkscreening on CPUs upon removal? I've used such thermal tape on my SLC2-50MP recently and was hesitant to do so because of this uncertainty.
Plan your life wisely, you'll be dead before you know it.
Adhesive termal tape is sturdy, but massive pain in the ass to separate cpu from heatsink.
Still, it never affected markings on the cpu surface.
BTW, Feipoa, do you remember if you LSD recognized Gotek at all or just didn't boot from it ?
feipoa wrote on 2022-02-13, 04:26:I find paper clips aren't able to produce that restoring force necessary to press the heatsink down. I'd recommend using thermal tape as well.
Has anyone witnessed adhesive thermal tape removing the silkscreening on CPUs upon removal? I've used such thermal tape on my SLC2-50MP recently and was hesitant to do so because of this uncertainty.
I’ve only used it on a Matrox M3D and when I removed it the letters had transferred to the tape 🤣. Still lettering on the chip but not as vivid.
pshipkov wrote on 2022-02-13, 04:29:BTW, Feipoa, do you remember if you LSD recognized Gotek at all or just didn't boot from it ?
I'm pretty sure I only tested the Gotek at 66 MHz to see if it was bootable; it was not. Floppy at 66 MHz also not bootable. One needs to take the speed down to 60 Mhz, but that wasn't of interest to me, thus I did not test the Gotek at <=60 MHz. Are you having issues at 60 MHz with Gotek?
Chadti99 wrote on 2022-02-13, 04:50:I’ve only used it on a Matrox M3D and when I removed it the letters had transferred to the tape lol. Still lettering on the chip but not as vivid.
Darn. I wonder if it is sufficient to just use double-sided tape for these less intensive heating events. I'll have to source another Cyrix SLC2-50MP just in case.
Plan your life wisely, you'll be dead before you know it.
Lsd booting from floppy was always trouble, regardless of anything.
---
What I do in cases like your slc2 -xtwo very small drops of Loctite in the opposite corners.
Works great. Hold tight the heatsink, but with just the right amount of force can be separated from cpu without any side ffects.
This allows for use of regular termal paste.
I was considering that approach, although not with loctite. I only have the threadlock loctite, while it appears that you have some form of "glue" loctite.
I had tried the corner trick with paste in the middle before, but when it was inadvertently knocked, the heatsink came off. What I'd like is some kind of thermal tape that has the removable consistency of "permanent" double-sided Scotch Tape, even at the cost of reduced heat transfer characteristics. On the plus side, after 30 years, the adhesive in regular thermal tape should degrade to the point that there's no silkscreen destruction when removed.
Plan your life wisely, you'll be dead before you know it.
Where could I source some 10ns 28pin cache modules I could use in the Biostar UUD? Wanting to stick with 256k double banked for the moment for 66MHz FSB testing. My current 12ns modules can’t pull off a 3111 cache setting. Windows install stable at 3222 though.
Dip-28 packages go as low as 12ns.
Limited number available. Hurry up.
https://www.ebay.com/itm/QTY-15-IS61LV256-12N … redirect=mobile
Feipoa using tape in the corners will worsen heat transfer but worth trying, these are low temps, can be ok.
Thermal tape wasn’t quite enough to keep the AM5x86 running at 200MHz. Pulled it back off and here’s an example of what it pulls. Was hoping it hadn’t set long enough to do this.
That’s horrific. Can you show us the CPU?
I wonder if plastic topped CPUs fair any better.
Plan your life wisely, you'll be dead before you know it.
Ah should of snapped a pic before I put the thermal paste back on. Honestly it’s not that bad but the longer it sits the more it could pull. The alcohol pads I’ve used to clean paste will really do some damage so I’ve stopped and simply just remove paste with paper towels.
pshipkov wrote on 2022-02-12, 17:58:On-board IDE is stable at 66mhz and 1:1 for sure. It was never a problem with any of the LuckyStar boards here, afair. Trying to […]
On-board IDE is stable at 66mhz and 1:1 for sure.
It was never a problem with any of the LuckyStar boards here, afair.
Trying to think of why you are having this issue ...
Actually what exactly is the problem ? No boot or unstable after boot ?
It may just be an issue with this CPU already driven to the edge. I’ve tried on a few boards now. An LS-C2, Biostar UUD, and the LS-D and anything over 1:1/2 at 66MHz results in random freezes. Sometimes I can complete some benchmarks and sometimes not. Usually unable to boot into Windows. I’ve tried different GPU, external IDE controller, different memory. I should go back and try 1:1 at 60MHz.
For the benchmarks which did complete at 1:1 66 MHz, what was the improvement at 1:1 vs. 1:1/2 ?
Plan your life wisely, you'll be dead before you know it.
feipoa wrote on 2022-02-13, 23:45:For the benchmarks which did complete at 1:1 66 MHz, what was the improvement at 1:1 vs. 1:1/2 ?
I’ll retest, it makes a difference on most tests except Quake, same score.
