Are you certain when you put the 5x86 back in, you had it positioned correctly in the socket? If you accidentally had the processor "clocked" incorrectly--meaning pin 1 of the CPU was 90/180/270 degrees off from pin 1 of the socket, most of the time it results in a fried 3v3 regulator.

I had the same problem with my LS486E mainboard, I have no experience with electronics so I just use 5V cpus or 3V with adapter.

gordon-creAtive.com wrote:

Disclaimer: I'm no electronics engineer.

I recently acquired a Lucky Star LS-486E C2 with a Am5x86-P75-133 CPU. After some weeks of usage the machine refused to boot. After exchanging various components I figured that it failed to post with CPUs other than those who require 5V. Because of this thread I decided to exchange the TIP107 transistor. I soldered in it and first booted again with a Intel 486DX, it ran fine. Then I put in the Am5x86-P75-133, adjusted the jumpers, triple-checked the jumpers and turn the machine on. It didn't post, instead smoke escaped from the new TIP107. I turned it off immediately. I desoldered the transistor, put the DX back in, adjusted the jumpers: It still works.

Now my question is: Do you have any idea what when wrong? Could you point me into some direction to solve this issue?

If you have followed the manual, then most likely the manual are wrong on the FSB settings.
On my board, it is "reversed" from what is explained in the manual. I have had help on Vogons before.

Re: Problems getting Lucky Star ls486e Rev.C to post.

Anyway....
Regarding the TIP107, are there any heatsink on your's? I think I had some issues with my Transistor as well.
For some reason it was DOA, yet I changed it. I could not run CPU's under 5v with the original one.
And I even think that I mounted a bigger heatsink as well. Just to be shure that it would be cooled well.

Thank you for your responses.

FesterBlatz wrote:

Are you certain when you put the 5x86 back in, you had it positioned correctly in the socket?

Well, I took extra care this time, but in retrospective it's hard to tell. Now given I did put it in the wrong way, the thing broken should be the TIP107, right? (And eventually the CPU I guess.) Given that I didn't put it in the wrong way, what else could have caused the TIP107 to overheat like that?

dirkmirk wrote:

I had the same problem with my LS486E mainboard, I have no experience with electronics so I just use 5V cpus or 3V with adapter.

I thought about that too, unfortunately these boards are quite rare and really expensive, so I figured I'd go with soldering a cheap replacement transistor. Well, lesson learned, kinda.

brostenen wrote:

If you have followed the manual, then most likely the manual are wrong on the FSB settings.
On my board, it is "reversed" from what is explained in the manual. I have had help on Vogons before.

But that should only apply to the FSB right? Which shouldn't cause the TIP107 to overheat, just general CPU malfunction of some kind, I presume.

brostenen wrote:

Anyway....
Regarding the TIP107, are there any heatsink on your's?

I put the heatsink of the original TIP107 on there. I instantly burned my fingers when I touched it and the TIP107 had bubbles on its surface when I removed the heatsink. So I don't think that was a problem of the heatsink, it definitely became much too hot for 4 or 5 seconds of normal operation.

Well I do have the impression that in case I did not put the CPU in the wrong way there might be something else that's faulty. I'm thinking about measuring the voltages at the TIP107 mounting holes, I feel like there's something wrong. Do you got any other ideas?

Measure the voltage getting to the CPU. Some CPUs wont run with 3.3V and require something a little closer to 3.45V.

Can you post a close-up photo of the VRM area of your board? I have one LS486E board, and had some problems with the VRM at first. There were signs that the transistor (TIP127A) ran very hot (oxidized, broken solder joints), and since the terminals were extremely short and i couldn't re-solder them properly i decided to put a new transistor in its place. So i went to buy transistors, returned home with TIP127, figuring the lack of "A" in the name wouldn't make such a difference. Checked the datasheets, and to my surprise there was a big difference: TIP127A is a standard PNP transistor whereas TIP127 that i had bought is a Darlington.
Not being in the mood to go again after transistors i searched around my stuff and found a MJE15031 (not a perfect match for the TIP127A, but close enough), soldered it, attached a heatsink bigger than the original and...no more problems.
I'm not very experienced in electronics, but a Darlington used as a series regulator for a VRM sounds a bit strange. So that's why i'm asking for a pic of your board, especially VRM area. If it's identical to mine, i'm pretty sure that you need a regular PNP power transistor in there, not a Darlington.

quicknick wrote:

i'm pretty sure that you need a regular PNP power transistor in there, not a Darlington.

No, please be careful when giving this kind of advice-because in this case it may prove harmful.

Sometimes a darlington is necessary because a regular PNP doesn't provide enough current gain. While it's possible a standard PNP may survive for a little while, it will likely cause poor regulation and could eventually damage both the VRM circuit's variable zener (TL431 or similar) and/or the non-darlington pass transistor due to sky-high Ib (base current) as a result of insufficient Hfe.

I've repaired many Socket 3 boards over the years and many of them used a darlington for the output pass transistor. This helps reduce the complexity and cost of the VRM circuit and is generally necessary because on most Socket 3 boards, the output voltage (3.3v-4v) is so close to the input voltage of 5v. If a darlington is what his board was equipped with, that's what needs to be used.

The user can also use a VRM interposer board and set his motherboard's CPU jumpers for 5 V to skip over the board's VRM circuit.

FesterBlatz wrote on 2017-12-15, 19:01:

Are you certain when you put the 5x86 back in, you had it positioned correctly in the socket? If you accidentally had the processor "clocked" incorrectly--meaning pin 1 of the CPU was 90/180/270 degrees off from pin 1 of the socket, most of the time it results in a fried 3v3 regulator.

Have MB with same issue.
How often it goes further than just frying TIP127 (or whatever power BJT installed)?
I mean is LP2951CM also under suspect?

UPD: I've fixed the issue by replacing dead H2585 that I found in my Rev:D to TIP127 that is know to be used in Rev:C. After that problem with Am5x86-P75-133 is gone.