| ⭅ Previous (Flashing the Bootloader) | Next (6502 Available) ⭆ |
With all the electronics completed for the 6502 chiplab, it is time to look at adding a few other chips.
Many of the chips I would like to research with the chiplab are not commercially available, and can only be acquired by extracting them from existing systems.
I have acquired a few “for-parts” systems and intend to extract the key ICs for study. In this article I’ll share how I extracted the CPU and PPU chips from a NES, and the key SoC from an original Gameboy.
The first system I planned to disassemble was the original Gameboy, often called the DMG. The system has electronics split across two circuit boards. Most of the unique behavior comes from a single integrated circuit or SoC, so I only needed that board.
The chip of interest is the one labeled DMG-CPU. This is a surface mount package, meaning the chip sits atop the circuit board. We’ll see soon that this makes it easier to disassemble than some through-hole parts, which are installed through the board.
First I attempted to desolder these with my heat gun. By heating all the pins at the same time, the solder should become molten once more. I should then be able to extract the chip with a pair of tweezers.
The main chip is a QFP, or quad-flat package. Before trying on this main chip, I went to extract the two SRAM chips to the right.
Since the SRAM chips have pins only along the edges, I can use a trick when removing them. While heating one edge with the heat gun, I can use tweezers to lift that edge of the chip away from the board. This will prevent the solder from holding it against the board once it cools. Then I can heat the other side, and the chip should be easily removed from the board.
With this approach I was able to easily remove these two SRAM chips.
I tried applying this technique on the main chip, but with no success. Since there are leads on all 4 sides, it is harder to lift the chip after warming one side. And with my heat gun, I wasn’t able to keep the solder from all sides warm enough to extract the chip.
After attempting this for several minutes, it was time to try something else.
Another approach I have used successfully for other systems has been hot plate desoldering. As it sounds, this involves placing the circuit board atop a hotplate. This provides heat across the entire board at once, so all sides of a QFP should be easily heated at once. I also place a small bed of sand between the board and the plate, to help more evenly transfer heat.
The downside with this technique is that the plate isn’t temperature controlled, and so typically overheats the board. As soon as I am able, I need to remove the chip of interest and then remove the PCB from the hot plate. Otherwise various plastic components quickly begin to smoke.
With my hotplate I was able to successfully remove the chip. Here it is against a footprint I had printed for testing.
With the key component extracted, I’m ready to move on to the other system.
The NES is an older system (1983 vs 1989), and used an older assembly technique (through-hole vs SMT). Each of the pins run all the way through the board, and there is typically some solder on both sides of the board holding the chip in place.
This system has two chips I would like to extract for the Chiplab. The large chip on the left half is the “CPU” which has the sound functionality embedded. And the larger chip on the right is the PPU, which manages graphics functionality.
Since these chips have many pins, and have solder on both sides, I knew I would need to remove as much as possible. Otherwise the chip would be firmly held in place.
First I atttempted to use a soldering iron and desolder braid. This is a woven copper wire which tends to wick up molten metal held against it. Though I have used this technique successfully on other boards, the solder from this board was hardly wicking away from the board.
After doing a bit of research for other techniques, I settled on using a desoldering iron. Much like a soldering iron, this features a tip that heats up to melt the solder. The key difference is that it is combined with a “solder sucker”, or a handheld pump designed to suck away the solder while molten.
I turned the board over, and began desoldering with the iron. I was honestly amazed how well this worked. For nearly all the pins, it completely removed all the solder after a single round of heat + suction.
Here is the underside of the pcb after removing solder from the first four pins:
After repeating for the remaining pins, the CPU came right out. For the PPU, there were several pins which had significant solder on the other side (top) of the board. For these, I used the same technique from the DMG SRAM, heating one side at a time and lifting with tweezers.
Here the PPU is half way out:
And repeating for the other side:
Fully removed! Here are the two chips extracted from the system side-by-side:
In case you’re looking for tools for a similar project, here is what I used. These are affiliate links, so if you purchase through these links I’ll get a small share of profit at no extra cost to you.
With the key chips extracted, its time to get them added to the chiplab. I’ll share more about this, and hope to have the chips online soon.
| ⭅ Previous (Flashing the Bootloader) | Next (6502 Available) ⭆ |