An idea of a potential Raspberry Pi Zero competitor chip

[technix]

This originates from the recent announcement of the PIC32MZ DA line. If a chip manufacturer can cook up a multi-chip module like this and sell it within 10 dollars, it would be able to pressure the Raspberry Pi folks. The chip contains:

* An 32-bit CPU core with MMU, be it MIPS Warrior (PUC32MZ DA) or Cortex-A7 MP1 (AllWinner V3s) or Cortex-A53 MP4 (BCM2837)
* A GPU that supports OpenGL ES and hopefully OpenCL if headless chips are used.
* At least 512MB of LPDDR2 SDRAM. Hopefully 1GB LPDDR3. Since the die is hard wired in the packages, a registerless hard wired memory controller can be used (like PSRAM.)
* At least 4GB of NAND storage, can be an eMMC die packaged in directly.
* Built in primary crystal oscillator and voltage regulators.
* Single 3.3V supply
* Packages: wide-body DIP28, DIP40, QFP44, QFP64, QFP100.

Features exposed by package:
* DIP28: 1x SPI with 3CS, 1x I2C, 1x I2S, 1x UART, 4x PWM, 1x USB OTG, 20x muxed GPIO
* DIP40: add 1x I2C, 1x SPI with 3CS, 1x SDIO, 12x muxed GPIO
* QFP44: add 1x USB Host and onboard Bluetooth/Wi-Fi combo chipset (and additional SDIO and UART on the main chip to support this RF chip.)
* QFP64: add Serial Digital Video Out complex (12 signals capable of being switched between MIPI DSI-2, HDMI and DisplayPort)
* QFP100: add two CSI-2 4-lane and one GMII interface.
* QFP144: add 24-bit parallel LCD interface, 3x more USB host ports (can be internal USB Hub) and 30x muxed GPIO

Package	GPIO (muxed)	SPI (w/ 3CS)	I2C	I2S	UART	PWM	SDIO	Parallel LCD	USB 2.0	CSI-2	SDVO	GMII	Wireless
DIP-28	20	1	1	1	1	4	0	0	OTG	0	0	0	No
DIP-40	32	2	2	1	1	4	1	0	OTG	0	0	0	No
QFP-44	32	2	2	1	1	4	1	0	OTG, 1x Host	0	0	0	Yes
QFP-64	32	2	2	1	1	4	1	0	OTG, 1x Host	0	1	0	Yes
QFP-100	32	2	2	1	1	4	1	0	OTG, 1x Host	2	1	1	Yes
QFP-144	62	2	2	1	1	4	1	1	OTG, 4x Host	2	1	1	Yes

[wraper]

This originates from the recent announcement of the PIC32MZ DA line. If a chip manufacturer can cook up a multi-chip module like this and sell it within 5 dollars, it would be able to pressure the Raspberry Pi folks.

Participate in the bloodbath of the race to the bottom? Does not seem to be viable if the goal is to earn money.

[JPortici]

This originates from the recent announcement of the PIC32MZ DA line. If a chip manufacturer can cook up a multi-chip module like this and sell it within 5 dollars, it would be able to pressure the Raspberry Pi folks.

crush your expectations, double it already. i see > 11 for a tape @mouser

budgetary pricing: http://www.microchip.com/ParamChartSearch/chart.aspx?branchID=30071

[legacy]

goddamn, electronics is going to throw up itself into a 2 penny coin.

[janekm]

The C.H.I.P. GR8 is pretty close to what you describe: https://getchip.com/pages/chippro (scroll down a bit for description and datasheet of the multi-die package). $6 each (I suppose that would almost be considered pricey these days...)

No on-board crystal or LDO or Flash, but that's not too hard to put on the carrier PCB...

[technix]

The C.H.I.P. GR8 is pretty close to what you describe: https://getchip.com/pages/chippro (scroll down a bit for description and datasheet of the multi-die package). $6 each (I suppose that would almost be considered pricey these days...)

No on-board crystal or LDO or Flash, but that's not too hard to put on the carrier PCB...

Too bad it is a BGA (thus not hand solderable.)

The chip I was talking above is a MCM packaged in DIP28/DIP40/QFP44/QFP64/QFP100/QFP144. Yes the two lowest pin count packages are DIP.

[krho]

The availability of GR8 is questionable. The next Co. Promises nothing.

[janekm]

The availability of GR8 is questionable. The next Co. Promises nothing.

Well, there is a clear promise on that product page... I suppose they haven't delivered on that promise yet?

[krho]

What i meant is that they are not promising long term availability. e.g 5, 10 years. When they decide that enough is enough they can stop selling.

[janekm]

What i meant is that they are not promising long term availability. e.g 5, 10 years. When they decide that enough is enough they can stop selling.

Got it, you're right that's not a safe solution in a case where one needs lone term availability (especially since as a startup they can "pivot" at any point).

[westfw]

You'd probably be better off thinking of the Raspberry Pi Zero as a marketing tool akin to free sample chips, rather than a realistic depiction of how cheap an independently manufactured SBC might be manufactured.  :-(

[wraper]

You'd probably be better off thinking of the Raspberry Pi Zero as a marketing tool akin to free sample chips, rather than a realistic depiction of how cheap an independently manufactured SBC might be manufactured.  :-(

Yep, dropping 3 connectors + a few very cheap parts compared to model A+, does not make it 4x cheaper. One sided part load likely is the biggest cost saver.