ESP32 Forum

Hello, i was wondering if it was possible to suggest a new product type. I wanted to see if it were possible to make a new version of the ESP32-C3 chipset with a smaller SoC package. This SoC package could possibly be a DIP-8 chip or SMD variant designed for very small devices that do not need much GPIO, but require good processing power and WiFi capabilities.

The reason i am asking this is because this idea came from a dream i had one night, and i wish it was a real thing. The DIP-8 chip can contain 4 to 5 GPIO and a reset pin, the pins used for both versions are the serial pins 20 and 21, and the boot pin pin 9. The 4 pin version can have one extra GPIO pin with an external antenna pin, while the 5 pin version will have 2 extra GPIO pins but no antenna (unless it’s possible to embed an antenna inside the package).

This idea may not be practical at all but i felt curious to see if anyone would take a look into it. If it would be possible to make both an SMD version and a through hole version for different use cases and scenarios. The through hole version could be used as a bare breadboard chip using an FTDI adapter for power and programming, while the SMD version could be soldered onto a custom PCB for a project that’s meant to fit in a small space.

I look forward to your consideration about this idea and i hope that you will be able to produce the 8 pin chips. Thank you for your time.

I asked about something similar once, and from what I understand, DIP packages for ESP32 chips are infeasible because the WiFi antenna pins would make for a too long path, with too many impedance mismatches. The SMD version would possibly be feasible, but given the fact that something like a SOIC-8 is larger than the 4x4mm package the ESP8685 form of the ESP32-C3 comes in, I don't think there'd be much demand for it.

True, it may be a bad idea. Have you already emailed the company? I would think it would be easier to make the 8 pin chips by just adding the existing dye chip into a different package and connecting only a few of the pins.

XanCraft21 wrote: ↑

Sun Jun 04, 2023 4:27 pm

True, it may be a bad idea. Have you already emailed the company? I would think it would be easier to make the 8 pin chips by just adding the existing dye chip into a different package and connecting only a few of the pins.

I don't need to email the company; I work for Espressif A simple repackaging would not be enough to get the ESP into an 8-pin package, by the way; you'd loose so many pins for the crystal, power supply pins etc that you would have nothing left. As such, it'd need to be a SoM, like an ESP32-Pico-V3, which internally is a substrate (think tiny PCB) with both the ESP32 chip as well as a bunch of components like the crystal mounted, and the entire thing is encapsulated in the black material of the package. Adapting something like that to some 8-pin package would be non-trivial, though, and as I said the market for it likely is miniscule.

Pico-V3's are way to go IMO - everything you really need is in the package while still providing some GPIO pins to connect other peripherals (like RMII ethernet, which is quite pin-hungry). It would only be better if the solder pads would not just be on the bottom side but rather also extended to the sides of the package (like the core-chips are).

Difference is that currently you can only really solder them with hot air and solder paste. But chips with pads extended to sides can quite easily be soldered by hand without paste (or rather hotair with possiblity to repair contacts by hand). That could significantly improve prototyping/development or servicing operations.

Fyi, I think the production process makes that hard. From what I know, 'real' chips are built using a leadframe, that is, the pads are actual 'slabs' of metal, they have a certain thickness so they can 'stick out' on the sides. SoMs like the Pico are a substrate (which is more-or-less a PCB) with the pads more akin to a foil; as they don't need structural integrity as the substrate provides that, they're made very thin and don't 'stick out'.

Ahh, it was worth a try. Thank you for taking the time to look at this.