Ask HN: Why can't the Raspberry Pi be properly cloned? | Hacker News Reader

Ask HN: Why can't the Raspberry Pi be properly cloned?

8 pointsandrewstuart27 comments

Since the earliest days of the PC. companies having been creating clone computers.

There's big money for whoever can identically clone the functionality of a popular computer - witness Compaq.

Is it "hard but doable by appropriately skilled modern engineers", or "impossible"?

Surely the rewards are there for whichever company can do it properly.

So why is there no proper clone of the Raspberry Pi?

One aspect to keep in mind. In those "earliest days" the target that was being cloned was the IBM PC, which Wikipedia [1] quotes as having an introductory price of $4,455 in 2020 dollars.

Wikipedia also quotes a price range of $4.00 to $70.00 [2] for various Pi model introductory prices (they don't quote as "2020 dollars").

There is much more room for profit available to someone who is cloning a $4,455 product than there is to someone cloning a $4 to $70 product. The comparative very low price of a Pi vs. an IBM PC from 1981 likely makes it not profitable to "clone" and so no one is attempting to clone the Pi's.

[1] https://en.wikipedia.org/wiki/IBM_PC

[2] https://en.wikipedia.org/wiki/Raspberry_Pi

This isn't really a space where anyone is attempting to beat the already competitive price point of the Raspberry Pi.

Most similar boards are created with their own vision such as faster storage or more networking but at different price points. These include the PINE A64-LTS, Orange Pi 4B, or BeagleBone.

This market also includes many carrier boards for the Raspberry Pi compute module (3 or 4). For example the Turing Pi V2 mini-ITX cluster board with four Raspberry Pi CM4 modules!

Because Broadcom won't sell anybody else the processor, and nobody who can make chips is going to waste their time and incur the legal risk of cloning some weird Broadcom SoCs when they could instead do something more sensible. And the only reason people would buy a Pi clone would be if it was better value for money, so now they also have to beat a quite cheap product in price.

I was under the impression that RPI chose "stock" Broadcom SoCs for their boards that, but the main problem with Broadcom is that they won't speak to you unless you have a substantial enough order?

As far as I know the current SoCs are special-made for RPi, but even when it was a stock part they didn't sell to other companies that wanted to make SBCs with them - some tried.

The BCM2835 was a chip designed for set-top video boxes but turned out to be a loser. RPi was started by a Broadcom engineer that found a use for the chip.

What magical thing is in Raspberry Pi that cloning it would be an advantage over other Linux SBCs/SOMs or EVKs like Beagle?

Oh wait I know the answer - the price. You want it cheaper. But it's already subsidized by Broadcom so how do you think a cloner would make any money off of it?

Or are you trying to use it in a product and you need a second source? Because that way leads to doom as well.

The software.

Okay, let's continue that line of thought.

So there must be some special software/firmware that only runs on RPi and a cloner could take advantage of that work when selling their board. What software would that be?

Linux distributions with large enough ecosystems to smooth out the rough edges.

Why does Broadcom subsidize Raspberry Pi?

It provides Broadcom with value by making Linux distros stable on their chips.

There is no one answer for this.

Looking at job ads, the industrial uptake of Raspberry Pi has been amazing.

Some key things that Raspberry Pi Trading has done to be competitive include:

- long product lifetimes

- good documentation

- listening to their customers (and adapting their products to meet their needs)

It isn't really "just price", as evidenced by the price of the compute modules (mainly used by industry).

But the point being made is that the only reason to clone it would be price. Because a true clone would benefit from all those things you listed, same as an original rpi.

No, the reason to clone it is software.

So maybe resubmit your question and take the hardware out of the equation.

"Why does RPi have such a robust ARM Linux ecosystem and why isn't that portable to other modules?"

No - that's not my question.

It was. By odroid.

However Broadcom didn't appear to like where/how they sourced the components and refused to sell them any more chips.

https://hackaday.com/2014/07/30/a-real-raspberry-pi-clone-no...

Presumably, Any company could clone the Pi, but they'd never be able to compete on price.

It would be great for industrial uses if someone made a pricier, but more readily available clone.

It would also be great if they could cut deals with "pro" electronics distributors here in Europe. None of their current distributors are able to ship very fast.

There are a bunch of clones out there.

None of them are compatible.

If you want compatibility buy another board or embedded system. The BCM SoC on it can’t be bought so you’re screwed for second sourcing anything.

The software support for all other SBC's is universally terrible. They're Linux support is always questionable, incomplete, out of date, poorly documented etc etc.

If software matters then the only real choice is Raspberry Pi.

There are good ones. They just cost real money. Which is the issue.

What do you mean by that? Isn't the Pine64 stuff supported by mainline Linux?

Much of it is mainline, yes. So I’m thinking they expect x86 like compatibly.

Reply on news.ycombinator.com