#RaspberryPiPico silicon teardown time!
Board as received optical + x-ray images
Inspecting package. X-ray shows bond wires easier after removing thermal slug
Thermal slug removed w/ HCl + H2O2
Decapped to show copper bond wires connecting die to package pins. Initially used only nitric acid which destroyed copper bond wires (last image). Fixed by switching to mixed nitric + sulfuric acid
Zoomable top metal die image: siliconpr0n.org/map/raspberry-…
@Raspberry_Pi marking it's turf
B0 => chip revision. On every layer? A chip I worked on did this and unfortunately when we fixed a bug on a *different* metal layer we didn't want to remake the masks with the text. Thus "B1" chips were mislabeled as B0!
M8 => eight metal layer chip. Higher metal layers are coarser to carry more power and reduce manufacturing costs. You can also see M7 and M6 marked
Copper bond wires were eaten by nitric acid leading little dimples where they used to be attached to the bond pads
Several unused power pads. Possibly breaks out internal core voltage for debug / test
Time lapse etching chip layers
This area wasn't 100% etched, leading to vibrant blue/purple from SiO2 thin film interference
1) 20x air vs 2) 60x oil objective: 60x has higher resolution but lower depth of field, making the image appear different. Ex: although the lower left splotch is large, it's only visible in the 60x image. I might take SEM images later for better resolution
Transistors arranged into standard cells form the majority of the digital logic. This includes both the CPU and the digital portion of most peripherals
A few different I/O showing logic, ESD protection, etc between the top metal and transistor layer. A number of different I/O IP blocks are used. Would be interesting to match them up to the datasheet for analog functionality like ADC, USB, etc
Memory, probably the boot ROM
SRAM is scattered throughout the die. By analyzing the memory sizes and comparing to the datasheet it might be possible to guess what some of these are for
CMP fill is added to unused areas to keep polishing more even when preparing for the next layer. As such, it looks like it was omitted on M8 (top metal) but is found on other layers
That's all for now. Hope you enjoyed and let me know if you want more information on something!

• • •

Missing some Tweet in this thread? You can try to force a refresh

Keep Current with John McMaster

John McMaster Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!


Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @johndmcmaster

26 Jan
Motorola MC14500 "Industrial Control Unit (ICU)," an early programmable logic device controlled via serial commands

Zoomable image: siliconpr0n.org/map/motorola/m… ImageImageImageImage
Decapping this chip I believe was inspired by @UsagiElectric who has some videos on it (ex: ), and I believe also provided the chip through a series of exchanges
@kenshirriff and others have already started doing some analysis on the chip, so I'll let them push that out when they see fit. I'll try to follow up with delayering in the near future
Read 4 tweets

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!

Follow Us on Twitter!