Very excite. Many horror. Much word. 1300 it would seem. Let me know if I should have had someone other then myself proofread it.
Ancient Chinese Secret
As experienced by Oelmuvun.
One day I acquired an untested Macintosh PowerBook 180. This is a nice little machine from 1992 with a 33MHz 68030, 68882 FPU, 4MB onboard RAM expandable to 14MB. For storage and expansion you have a 1.4MB floppy and 2.5” SCSI HDD, internal modem slot, ADB(keyboard& mouse), two serial ports, HDI-30 SCSI, video out(8-bit colour @ 832x624), line level in & out with built-in mic and speaker. The 180 in particular has 128KB VRAM for the internal display that is a 10” 640x400 greyscale Active Matrix LCD. Not bad.
When I tested this machine I was seeing video corruption, so I swapped in a compatible but slower PowerBook 160 daughterboard (25MHz 030, no FPU) to find the issue gone. I had no luck cleaning and fiddling with various connectors now that the problem was isolated to the 180’s daughterboard.
What next? Let’s try the video RAM. I would rather not sacrifice the PB160 unless I really had to. I cannot remember if it used a different chip, or if I just could not find a datasheet for what it had, or if I just did not want to sacrifice it but I started looking elsewhere. A PB145 proved incompatible. What I did notice was that my two PowerBook 150s shared the same VRAM chip as the 180. The 150 is a machine I have mixed feelings about. On the one hand it uses IDE HDDs, 640x480 screen res, 40MB RAM ceiling, and has a 33MHz 68030. On the other hand that LCD is an absolutely terribad passive matrix panel. It has no ports except a single serial, optional modem, and HDI-30 SCSI. No audio in or out, no ADB, no video out. It has no mic either. As a budget machine of that time period I am certain it made people happy, but if I am going to play with low end machines I can think of better ones.
One of my PB150s is not in good shape and I did not mind donating it’s bits. The VRAM chip in question is a 40-pin SOJ package, easy peasy. Unfortunately the video corruption remained after the swap! Bugger; How about the main video chip? It is a 100-pin QFP and I can do that, no worries. However this comes at the cost of my PB160 since the pair of 150s and 145 available to me do not have the part I need.
At this point I get out the hot air and whip off the chips from both daughterboards, all is well.
Now there are three ways I can put this sucker back on.
#1: Clean the pads and pins, flux up, tack down a couple corner pins and solder each individual pin with a teeny tiny iron tip and thin solder. I had no tiny solder on me at the time.
#2: Tin the pads, more flux, then fart some hot air onto the whole mess. The IC should more or less align itself with the pads and settle in nicely.
#3: Tack corners of the IC, goober some solder all over the rest of the pins, and drag some desoldering braid along the pins to pick up the excess solder. This is a popular method, but one I am not all that experienced with.
There are times when I should be off in a corner wearing a dunce cap. This was one of them. Can you guess what option I went with? Yep, I figured I should build some experience with option #3. Normally it would have gone fine I am sure. However this time I was a bit short on my usual solder and for some stupid reason which was probably being too lazy to order more, would have preferred conservation… for projects that matter more and really need a lot. MEGA-FACEPALM(in default Unreal Tournament announcer voice even)
But that was fine right? I had a whole spool which came with an “Ancient Chinese Secret” iron. Not even a second thought; for once it was a good thing that a cheaply made item had lead in it! 63/37 Solder is 63/37 Solder is 63/37 Solder, right? Hold on a second, I need to facepalm some more as some of you may already have determined.
I blob some flux and tack opposing corners before proceeding to slop the new solder on. Seems okay I guess, not quite as shiny though… Oh well, grab the desoldering braid……. It is not wicking why is this not wicking? More flux? Still not wicking!? Will a vacuum pump be my saviour? Oh god it just sort of sits there and does nothing! Dilute it with good solder, anything to get this garbage off!
It would have been a really good idea to test this solder on something dead first. At this point mixing in good solder then using desoldering braid sort of works, but is extremely wasteful and is dumping extreme heat into the IC, delicate traces, and PCB itself. This is a lost cause; it is time to retreat and remove the IC with hot air.
The situation so far is that I have ripped a pad off the PCB while the IC and pads are covered in un-removable solder. I toss the IC into some helping-hands and spend way too much time eventually managing to clean it off. The PB180 daughterboard is a parts-board now; perhaps someday I will stick the FPU and CPU onto a future PB160 board and bump the clockspeed up? Who knows?
Let’s put the PB160’s video chip back on the PB160 daughterboard and see if it still works… Nope, I killed it with heat. Time for more hot air to remove the dead chip; I am going to toss the PB180’s faulty chip on the PB160 daughterboard just for the sake of making sure the PB160 daughterboard still works and make sure I at least had isolated the right chip. Yep, video problems; at least I was on the right track.
Typically you would fix a damaged pad or trace by taking some tiny gauge wire and routing it from the pin of the IC to the trace’s nearest VIA or other part. The problem with this is that the trace went under the video chip and to a VIA that came out under another IC on the underside of the board. Not happening. I suppose I could at least try bending all the pins of this QFP package down to gain a bit of clearance for some 30 AWG wire wrapping wire to reach it’s VIA… But no, not happening. There is still some of that nasty solder on the rest of the pads anyway and trying to remove it would likely take the pads along too. I did not bother cleaning the flux off the PB180 daughterboard either, by now the board is likely toast even if I did bypass the missing pad.
Ultimately I need a new daughterboard. I do not remember what models exactly are compatible but the list goes something like 160, 165, 180, I think. I am not sure about 165c or 180c. I am pretty sure the 170, and 140 are out. I flat out know the 190, 150, 145, and 100 are not happening. If I get another video IC from a parts board I can still snag the CPU and FPU from the PB180 board I killed and bump up the clockspeed of the PB160 board. If I get a PB165 daughterboard it will just need a FPU since it is already 33MHz like the PB180. No crystal oscillator swap required.
And after all that I get to go spend $100 on a CF or SD to 2.5” SCSI adapter. Yay. I actually really dislike old 2.5” SCSI drives and that is a big reason why I have not pursued this project further.
Someday I will come back to it. I may even learn another lesson. Haha.