Gregory's Serverthis is fascinating: the power board has several laser trimmed resistors (sure, pretty common) but these have been laser trimmed AFTER being soldered onto the board!
|
161 comments
[DATA EXPUNGED]
 @tubetime Huh? On a power supply board? What sub-circuit are these resistors part of?  @attilakinali there are 7 of them. they trim the output voltages and a few other parameters @tubetime interesting. I would have thought that 1% resistors would do the job for the voltage feedback of something like a computer. Any idea why they need this level of accuracy? @attilakinali 1% is indeed fine for this. but they probably thought doing it this way was some sort of competitive advantage lol  ok i have a theory about why this was failing. symptoms: randomly the computer would shut off and pull a ton of current... 1) crowbar circuit fires. it can fire due to a rail going overvoltage... 2) a buck converter's MOSFET wasn't switching properly... 3) oh look this diode wasn't connected anymore due to a bad via! want to stare at the schematic yourself? i've put it on GitHub. https://github.com/schlae/Thinkpad700CPower
[DATA EXPUNGED]
 I just checked and (thankfully) CHM doesn't have one. They actually have almost no ThinkPads in the collection. I have nightmares about how many early 90s machines in the collection have pcb rot. @tubetime I had totally forgotten about active vs passive matrix displays! The minute I saw an IBM active matrix LCD, I knew the end was nigh for the CRT. It was amazing for its time. @tubetime Active matrix, man. A million psychedelic colors... I bet it looks crispy in the dark. @tubetime The single strip of Scotch tape in the middle is a nice touch.  @tubetime Did you have to make your own contacts of could you reuse something else? @tubetime Perhaps the Hirose DXM connector? https://www.mouser.com/new/hirose/hirose-dxm-connectors/ @maehem yeah i sorta think its a Hirose product but it's not DXM, which has only 2 rows of contacts, not 4.  @tubetime Here's the 0.8mm pitch version.  @tubetime and apparently that specific one is only found on 700 and 720 series. Probably this one is ISA only, whereas the following series have PCI and/or CardBus.  @tubetime looks like some kind of mezzanine connector  @tubetime Now that would be a truly useful AI - upload a picture of a connector (perhaps with a ruler for scale), and it tells you what the connector is and where to get it. Yours was a tough one, but its hard even for more common connectors. Next time I go to Anchor I’m going to take some stuff with me to see if I can find matching connectors. @tubetime Others have speculated in the past that IBM used Nextel Suede coating. https://www.nextel-coating.com/produkt/nextel-suede-coating-3101-90fh-11-black/?lang=en&v=3a52f3c22ed6. I am not sure if that info is correct. Difficult to get in the USA and $$$. I know folks who may have very limited quantities (used on aircraft instrument panels), may be the wrong color or variant.  @jonhendry by shining a flashlight through the board and looking at the shadow of the trace. if it is more distant or fuzzy, then it's on the opposite side of the aluminum core  @tubetime I'm guessing they trimmed to tune the system as a whole, not just the single resistor value? @tubetime My first job out of school was doing fpga stuff at a company that did laser trimming. It was wild. They used analog circuits to drive the galvos which had the mirrors. If the circuitry was off by a bit, you could see the “ringing” of the circuits in the right angle cuts (instead of an angle you’d see a little degrading sine wave). I’m still convinced analog EE’s are wizards. |
