Page 1 of 1

Assorted small repairs

PostPosted: January 25th, 2025, 7:29 pm
by LongRunner
First off, the Kingston A400 SSD (I always regarded Kingston as the Maxtor of SSDs, although even the former consumer-SSD strongholds Samsung and Crucial have now fallen too) faded out after not-quite 3 years of use, so (after temporarily reusing an old Samsung HD250HJ, which was at least just moderately slow all the time rather than the Kingston's regular severe, several-second lagging) I ordered in a proper Exascend SE3 enterprise SSD (from Mouser together with my usual electronic parts, exact model in my signature) to replace it, and now that it's installed, my PC runs lightning-fast again :clap:

I've also replaced my ST380817AS's +5V TVS diode (its lone STMicroelectronics part, which I never fully trust especially after >20 years) with an ON one from an ST310014ACE. Then I swapped the +12V TVS diodes between an ST360021A (STMicro TVS, otherwise good electronics) and an ST340014A (Fairchild TVS, but STMicro MCU so who cares there :P; anyway the HDA is heavily corroded and its head has suffered, although it still has no bad sectors!)

Barracuda reunion.jpg
Strange how the nominal currents changed, but only IBM/Hitachi was known for accurate power ratings (and weights)…
Barracuda reunion.jpg (388.04 KiB) Viewed 1135 times
Barracudas from beneath.jpg
Notice the shinier solder joints where I replaced my ST380817AS's +5V TVS. Their SATA headers also have much thicker gold than even most enterprise hardware nowadays (including the Exascend SE3 SSDs), and seem stronger‑than‑average too.
Barracudas from beneath.jpg (431.08 KiB) Viewed 1135 times
Meanwhile I've reunited the ST380817AS with an ST3160827AS from eBay – only 4 days newer, lightly-used in good health again, although some of the silver PCB areas are very tarnished (including the motor contact pads, so again time to refinish). Both thankfully have the most-reliable Agere+SH6950 combination, although this particular ST3160827AS does have an STMicroelectronics serial flash (but at least that's an off-the-shelf part, rather than custom-made for the specific drives as the MCU and motor driver are) whereas my ST380817AS has an Atmel; but since both have firmware 3.42, their boards swap just fine.
Therefore my S-tier backups can expand from 80GB to 160GB, and the lounge PC gets the ST380817AS (was using an STMicro-MCU ST380013AS)…

ST3160827AS PCB before.jpg
ST3160827AS PCB before.jpg (913.32 KiB) Viewed 1135 times
ST3160827AS PCB refinished.jpg
Like Seagate's own leaded soldering on the Barracuda ATA IV & V (and still in the 7200.7 HDA during 2003/2004), I've got to make NASA proud.
ST3160827AS PCB refinished.jpg (884.08 KiB) Viewed 1135 times
I've refinished the HDA contact pads (by flowing on leaded solder, then wicking it off and thoroughly cleaning the flux) on my Barracuda 7200.7 PCBs to ensure continuing reliable connection (except my youngest ST340014A from mid-2005, apparently fully RoHS-compliant observing its dull motor pins :-/ I'll cross the bridge when I come to it there…); thankfully Seagate always used high‑quality PCBs back then (unlike on even relatively high‑end WD drives now).
(And I imagine Seagate would have held out closer to the 2006-07-01 deadline in the 7200.7 HDA, if not for pressure from their competitors complying…)
Interestingly the 7200.8/9/10 HDAs (even the single-platter slimlines) actually have gold‑plated pins, implying that Seagate's engineers wanted ENIG PCBs (or separately gold‑plated contact pads and the usual silver or tin elsewhere) but accounting made them stay with immersion silver :-/

FW100 before.jpg
FW100 before.jpg (324.92 KiB) Viewed 1135 times
FW100 with PPTCs.jpg
FW100 with PPTCs.jpg (482.31 KiB) Viewed 1135 times
This FireWire card, while solidly-built overall, cut a corner in replacing the self-resetting PPTC fuses with 0Ω resistors; I've restored the originally-designed safety.
Although FireWire devices draw up to 1.5A (on +12V in normal PCs), you need a 2A PPTC (to allow for ambient temperature) rated for ≥12.6V (16V being standard).
I've already recapped it with ceramics, as >20-year-old TREC electrolytics (even general-purpose) aren't worth keeping…

LaCie PTI–205N.jpg
LaCie PTI–205N.jpg (839.92 KiB) Viewed 1135 times
This LaCie PTI–205N USB 2.0 card was stored so poorly that even the gold-plated PCI connector tarnished enough to lose contact; yet it still works fine after buffing the connector with Jif (just as a one‑off, thankfully the gold thickness was generous), a testament to the reliability of NEC chips (unlike VIA :rapidfire:).
I've ordered nice Panasonic FC non-aqueous electrolytics (nothing greater than +5V here, so 6.3V or 10V is fine even though the original Chocons are 16V) and will look for the Amphenol 87520‑3010 ports (rated for 5000 insertions rather than the standard 1500) later…

U304S before.jpg
U304S before.jpg (636.75 KiB) Viewed 1135 times
U304S braced.jpg
U304S braced.jpg (510.58 KiB) Viewed 1135 times
And I almost broke the power input off this USB 3.0 card; but thankfully the PCB provided through‑holes for a brace, and 3 strands of 0.5mm² (0.8mm diameter) wire fit nicely and ensure it will never break again :cool: Since it's only a cheapo, I'll worry about installing proper PPTCs later.
(It's used, of course; I would go for better-quality if I was buying new, but this is just for a lounge PC so it's no big deal.)