I just received a 1400 MHz Tualatin-S in the mail and found that pin AN34 (TRST) had broken clean off the CPU body . The CPU will not boot without the pin, but when I put the pin in the proper socket (hoping it would make contact with the CPU), I found that the CPU boots and works fine .
I would contact the seller and get either a replacement or refund. Nicely the seller offers 30 day money back so he/she would be easier to work things out with.
I would contact the seller and get either a replacement or refund. Nicely the seller offers 30 day money back so he/she would be easier to work things out with.
A replacement would have been great, but considering the fact that the auction was cancelled because of stock quality issues, my hopes are not very high for that .
A refund, partial or otherwise, that lets me keep the item still means I need to live with it or fix it.
A complete refund and return would leave me without the item while pretty much all other avenues for such a CPU (Tualatin-S PIII @1400MHz) are currently at least twice as expensive.
And to further complicate this, I just realized the shipping packaging was accidentally discarded (doh!) , so I can not provide any pictures of it upon request .
At this point, I feel I am pretty much "stuck" with the CPU . Incidentally, the it runs great at 1050 MHz in my PowerLeap adapter (8 hours+ Prime95 stable).
Does anybody know what the TRST (Test ReSet ?) pin is for on a socket 370 CPU ? Is it only used during POST or some kind of CPU self-test, or is it something else entirely ?
it is an inverted signal, so when it's a logic 0, it's actually in logic reset mode. (The line above the signal name is what states that it's inverted)
the good news is that it will be a VERY low current pin, though I'd try to get something out of the seller if you could.. if not, it's not a current carrying pin that's going to start a fire. But your PC won't likely be reliable.
it is an inverted signal, so when it's a logic 0, it's actually in logic reset mode. (The line above the signal name is what states that it's inverted)
the good news is that it will be a VERY low current pin, though I'd try to get something out of the seller if you could.. if not, it's not a current carrying pin that's going to start a fire. But your PC won't likely be reliable.
Thank you very much for the information. I think I will at least let the seller know about the issue . As for improving reliability, would dabbing the pin top with something like conductive epoxy, be a worthwhile endeavor ? If not, if the thing keeps on working as well as it has so far, I doubt I would ever remove the CPU and pin from the PowerLeap adapter .
Sounds fairly likely to work but I'm not certain. Epoxy can be tricky to work with. A bit of added resistance from it would probably be okay, but be really careful that the pin ends up reasonably flush while also not getting any adhesive into the socket or accidentally gluing the chip or pin to the socket. Also make sure you don't get any adhesive on the pin itself which when dried will prevent it from inserting into the socket, and it could be a pain to clean off. If you could somehow hold the pin flush in time for the epoxy to set without using the motherboard to do it, it might be worth it (if the seller doesn't help you and you don't want to fight it, which would probably involve shipping it back).
I contacted the seller and got no response after 7 days and I don't feel up for a fight, so I think I will just try to repair it . I have some parts from a broken slocket that should allow me to keep the pin steady while the epoxy dries . In the meantime, I ordered another CPU, because prices seem to be on the rise for this model .
You really should have an electrically AND mechanically sound connection, simply epoxying in place probably won't cut it for very long. After heat-cycling enough times, it's likely the pin's electrical contact to the CPU will eventually weaken and fail.
I came up with a method for replacing broken CPU pins that worked excellent, but I've only done this with ceramic CPUs. Perhaps it's possible on your CPU too, but I suspect you'd need to be VERY quick with the soldering iron to prevent damaging whatever substrate that CPU package is made of.
This method assumes there is still a metal base or stub of the broken pin remaining for soldering onto. You'll also need a very fine soldering iron tip...I used a Metcal STTC-144 30° bent sharp tip for this, but any fine tip should do...
Get yourself a new precision DIP socket similar to this one. Ideally you'll want to find one with the thinnest pins possible...
Carefully remove a pin, and with it clamped gently in a small vice, use a small flat file to file the opening at the top of the pin down until it's almost completely flush with the base of the metal socket opening leaving only a concave shaped surface that just barely cups over the broken pin's base on the CPU.
Place a dab of flux on the "cupped" base of the new pin, and on the "stub" of the broken pin as well. QUICKLY tin the stub with solder. You don't need much solder left on there, a pinhead sized ball is plenty.
Using a .5mm metal tipped mechanical pencil, insert the new pin inside the pencil which you can then use as a tool to hold the new pin's cupped end perfectly centered on the tinned "stub" on the CPU.
A quick touch with the fine-tipped soldering iron should melt and flow the solder up onto the cupped surface of your new pin. Be quick with this! I used silver bearing solder for it's slightly higher strength, but the benefit is probably insignificant.
If done carefully, after gentle cleaning with a toothbrush wetted in isopropyl or acetone, it should look something like this. Two pins were replaced on this CPU:
It's very important to clean any remaining flux from the pins. Flux can be mildly corrosive, especially when heated...so you don't want that getting down into your CPU socket.
Last edited by FesterBlatz on 2017-03-21, 16:32. Edited 3 times in total.
The "broken" pin is actually in one piece, I imagine I can solder it back into place . I should probably take some photos when I get a chance and post them before I try anything .
What really surprises me, though, is how reliable the CPU has been so far with the pin simply set in place .
Thanks, now if my understanding is correct and the TAP is only used during hardware-level debugging, and needs to be driven low (ONLY?) during power on Reset, reliability should still be high with my current setup .
The only question that remains is if I could simply bridge the connection to a ground pin to keep it asserted at all times and expect the CPU to work normally .
Thanks, now if my understanding is correct and the TAP is only used during hardware-level debugging, and needs to be driven low (ONLY?) during power on Reset, reliability should still be high with my current setup .
The only question that remains is if I could simply bridge the connection to a ground pin to keep it asserted at all times and expect the CPU to work normally .
try it, put a 1Kohm resistor between the pin and the ground.
I just got a new socket 370 board and put in my damaged CPU and "loose" pin .
Guess what ? It still works fine .
At this point, I think I am just going to leave in there as is until (if) I have an issue with it and only try to fix it then . I mean, it's not like it is ever going to leave that socket unless it has an issue, so I can't really see it getting any more physical wear and tear .
Well others have done the same, broken pin/ pins on an newer AMD cpu like AM3, AM4, just find where the broken pin needs to be on the socket, drop it there and you are good to go...
The whole idea is to get contact, the cpu itself is fine, as the CPU die is the important thing, aslong as the die is not broken the cpu will "just works" ...
Mah systems retro, old, newer (Radical stuff)
W3680 4.5/ GA-x58 UD7/ R9 280x
K7 2.6/ NF7-S/ HD3850
IBM x2 P3 933/ GA-6VXD7/ Voodoo V 5.5K
Cmq P2 450/ GA-BX2000/ V2 SLI
IBM PC365
Cmq DeskPRO 486/33
IBM PS/2 Model 56
SPS IntelleXT 8088
