The race to the bottom could be the end of real OSHW. Real meaning HW that really has an impact on the world, not just something cheap that does what a thousand boards already do, like blink LEDS, but is really cheap. There were a lot of MSP430 boards sold for $4.30. I suspect that most of them are now in a desk or some other place. Impact was not lasting.
There is a number out there that says the price should be 2.6x the cost of the BOM. I will say that the X-15 will be less than 2.6x the BOM, but it is not in the pitiful range of the BBB margins. There is some margin there to allow people to support it and invest in it, making it better. And hopefully use it for something more than a fast Linux build machine or an LED flasher.
At some point, the race to the bottom has to stop. I suspect we will see $1 boards soon with $35 shipping and handling. The reality is the price you see is not the price you pay. I will say, the price you see for the X-15 will be the price you pay.
Gerald
There were a lot of MSP430 boards sold for $4.30
Two of those are sitting in a box here, right next to 4 LM4F120 LP boards. We actually made a reflow oven controller with 1 ( of 4 ). The LM4F Launchpads however were completely free from TI. Didn’t even have to pay shipping.
I think an important thing to note however would be that the X15’s processor is not some garbage bottom of the barrel processor either. 4 PRU’s on die is worth much to many out there, and I will just about guarantee there will probably be thousands sold to people who use them in commercial products. All of which will care less about the piss poor graphics IP situation.
I would also not consider the beaglebone black as garbage. However I was recently reminded of how slow it really is. When creating / using an i386 CANBus development system, in a virtual machine, on an i7 quad ( 8 threads ) laptop. So, in the context of performance, I’m sure the X15’s will be more than welcomed by many.
The race to the bottom could be the end of real OSHW. Real meaning HW that
really has an impact on the world, not just something cheap that does what
a thousand boards already do, like blink LEDS, but is really cheap. There
were a lot of MSP430 boards sold for $4.30. I suspect that most of them are
now in a desk or some other place. Impact was not lasting.
Well, my $4.30 board (and trip to the initial 4/30 day) turned into a
few 10's of thousand parts shipped for my day job. We migrated from
Atmel AVR parts to the MSP430 when the AVRs became "unobtanium" due to
a mismanaged fab migration on Atmel's part (pro tip: don't shut down
one fab before the other one is _really_ online!).
I'm just glad I was aware of the MSP430 family, so I didn't have to
seriously consider porting to a PIC or (god forbid) an 8051! 
There is a number out there that says the price should be 2.6x the cost of
the BOM. I will say that the X-15 will be less than 2.6x the BOM, but it is
not in the pitiful range of the BBB margins. There is some margin there to
allow people to support it and invest in it, making it better. And
hopefully use it for something more than a fast Linux build machine or an
LED flasher.
I'd personally prefer to see a viable product that has enough margin
to support continued development than rock-bottom prices. I like the
Arduino model, where official quality boards are affordable but still
have a good profit margin. There's room at the bottom for Chinese
clones if you only care about price, but I am willing to pay more for
boards I know are built well, quality tested, and which give back to
the community by funding further development.
Keep up the great design work, Gerald!
“Once you stop learning, you start dying”
A.E.
Gerald
“Once you stop learning, you start dying”
That sounds depressingly like an obituary, because the main lessons of the last several years in this space have been that:
- Price is king.
- The sales/price curve is an exponential over price, and even a poor design that’s cheap can sell millions.
- A SoC slapped onto a PCB with SDRAM and some connectors cannot have a high price unless someone is price gouging or their manufacturing systems are hopelessly uncompetitive.
- There is no shortage of manufacturers of ARM SoCs, so a good engineer has a smorgasbord from which to choose.
- There is no shortage of good ARM boards, so a high price merely ensures that your audience is small or niche, or zero.
- Embedded is not a niche market so competition is intense, and ignoring the competition is a business deathwish.
These are all very important lessons, and they represent the business environment of today in ARM boards. Someone once said “Once you stop learning, you start dying”, and in a business sense that’s very true. Companies either learn and fit into the business environment of the day or they don’t survive.
I very much appreciate my many Beaglebones of both colours, all made by CircuitCo and not clones, and I was quite inspired along with thousands of other people by this very effective 2-member series of boards. But time doesn’t stand still, and it’s the right time now for another younger and even more nimble member to join the family and impress a generation eager for more IoT devices.
I’ve read some depressing “There’s nowhere to go” comments on this list resulting from TI’s lack of a newer $5 volume SoC to continue the AM335x family line, but that’s just narrow thinking. There’s plenty of room for new products within the Beaglebone form factor, and it was proven just very recently by Seeed Studio’s newly announced Beaglebone Green.
There’s nowhere to go only if your mind is chained to an existing spec. All change that improves the ability to meet some embedded requirement that wasn’t well served before should be on the cards for consideration. Lowering the price definitely fits into that category because it opens up so many new embedding opportunities, and it’s definitely the single factor which most affects mass adoption.
In summary, I think reports of the end of the BBB line reflect more a lack of energy for innovation than a lack of possibilities. For example, a new and cheaper “BBE” (Embedded-only) could drop HDMI as Seeed did very sensibly, as well as lose the eMMC. This would undoubtedly drop the price significantly since the change of eMMC size from 2GB to 4GB was blamed for the rise in BBB prices. And with the drop in price will come a rise in demand (lesson #2) and hence in sales too, if CircuitCo can cope with that.
The end isn’t necessarily nigh. It’s a very good form factor.
Morgaine.