Holy mad Solder skills!

timiskander

Well-known member

Donor 3 years: 2011-2012, 2017
Joined
May 3, 2006
Messages
2,667
Reaction score
61
Location
Fayetteville, New York
Not specifically arcade related, but...

I know many guys here and on RGP are always bemoaning having to work with surface mount parts. I guy I work with pulled this ball grid (.8mm pitch) DRAM part off, flipped it around, and rewired it (with magnet wire) to the board to correct a trace issue (time was of the essence!). The bus is running at 100MHz and the whole thing is working great!

Not recommended practice for sure, but I was impressed as hell that he could pull this off.. it took him 4 hours to do! He even made a twisted pair for the clock lines!

attachment.php
 

Attachments

  • rework.jpg
    rework.jpg
    100.5 KB · Views: 210
amazing work!! I do work like this all day long to ECU's in cars and trucks ,what a PITA but it feels good when you've accomplished something of this caliber. Are those wires bare copper, and if so how do they not short since they are touching?
 
amazing work!! I do work like this all day long to ECU's in cars and trucks ,what a PITA but it feels good when you've accomplished something of this caliber. Are those wires bare copper, and if so how do they not short since they are touching?

That's magnet wire... enamel coated
 
Not specifically arcade related, but...

I know many guys here and on RGP are always bemoaning having to work with surface mount parts. I guy I work with pulled this ball grid (.8mm pitch) DRAM part off, flipped it around, and rewired it (with magnet wire) to the board to correct a trace issue (time was of the essence!). The bus is running at 100MHz and the whole thing is working great!

I'd hate to see the signals travelling down those wires... what a signal integrity nightmare -- impedance mismatches everywhere. Good thing it was only 100M :)
 
I'd hate to see the signals travelling down those wires... what a signal integrity nightmare -- impedance mismatches everywhere. Good thing it was only 100M :)

Yup... we were all a bit concerned about that, but it turns out that "only 100M" was good enough to get it running. Tested the entire address range multiple times and its holding up... just needs to last long enough to get the reworked board in :)
 
I'd say he was damn lucky it's working! See all those snake traces on the BGA footprint on the board? Those are matched-delay lines. i.e. the signal flight times to/from the BGA pins are finely-tuned. Either he was incredibly careful to make all of the magwires exactly the same length (unlikely), or he was just damn lucky!
 
I'd say he was damn lucky it's working! See all those snake traces on the BGA footprint on the board? Those are matched-delay lines. i.e. the signal flight times to/from the BGA pins are finely-tuned. Either he was incredibly careful to make all of the magwires exactly the same length (unlikely), or he was just damn lucky!

We were lucky that it worked, but the guy who did the work is no slouch. It was a bit of a gamble, but he knows what he's doing.

But like Mark said... no-one wants to look at the signals on that! (well, I kind of do, but only 'cause I'm twisted :) )
 
We were lucky that it worked, but the guy who did the work is no slouch. It was a bit of a gamble, but he knows what he's doing.

But like Mark said... no-one wants to look at the signals on that! (well, I kind of do, but only 'cause I'm twisted :) )
Gotta give your buddy props. Not only for the soldering but for the plain fact that he was willing to take the time to take on such a task. Most would have said something like this wouldn't work and would have given up without even trying. You have to respect that "can do" attitude.
 
I'd say he was damn lucky it's working! See all those snake traces on the BGA footprint on the board? Those are matched-delay lines. i.e. the signal flight times to/from the BGA pins are finely-tuned. Either he was incredibly careful to make all of the magwires exactly the same length (unlikely), or he was just damn lucky!

I suspect a little bit of both. Things that should be matched tend to be very near each other on the BGA footprint (that is, the balls are physically close). That means if he keeps the wires of similar length, he's got a good chance of pulling it off.

That's some DAMN fine work. Did you make whoever screwed up the traces on the board buy beer for the guy who did that? Because he deserves it.
 
He's also lucky it was a nice "big".8mm BGA.

I fucked up a board a few years back and had to solder 32ga wire wrap wires to pins on a .4mm LQFP, and that was *exremely* un-fun.
 
Back
Top Bottom