Movement from a two-piece to a one-piece board:
Once the Kickstarter was fully funded (that's how the CELLv1.0 first started out for those who don't know), we asked the backers whether or not the originally designed 2-pc board was preferred or the 1-pc and why.
To give you a little back-story on the original CELLv1.0 2-pc design (top "cellular board" and bottom "jumper board"), it was created with the thought that if we could fully certify the top board as a PTCRB "final product" that we would have a fully modular unit which could be used in any situation including:
- Prototyping - with the 2nd piece "jumper board" the exposed header pins could be used
- Shield - with the Raspberry Pi, Arduino, or any custom shield, use both top and bottom boards + shield to directly connect
- Final product - when trying to fit into a small package like a tracking device just use the top "cellular board"
The planned downsides were increases in costs due to the two PCBs, assembly, and extra connectors.Why then move to the one piece?
1. The vast majority wanted a one piece.
Most people thought it would be best to have a 1-pc, and those that wanted a 2-pc wanted it potentially for future projects. But there was nobody that had a real use for the board and in reality, those that wanted the 2-pc should ideally take the schematics and parts we provide to add it to their own projects.
If someone truly wants to resell this device in their own system, then there would need to be certification of that system anyway, so why not save extra space and just include it into your own design?
There are loads of benefits and a minimal number of drawbacks.Benefits:
- One piece versus two
- Easier circuit traceability - There will be one set of pinouts versus two for the CELLv1.0 itself
- Less expensive sale price - minimizing PCB, parts, and assembly will help bring this to more people by making it more affordable
- Less cumbersome - shields are only 2 pieces now versus 3 high
- 1mm bigger: Went from 34mm x 40mm to 35mm x 41mm (1.38" x 1.61")
There were no losses in functionality, voltage conversion, mounting holes, or anything else!
3. We over-engineered it
Sometimes as engineers we over-think things, and by sometimes I mean all the time. In an effort to maximize modularity we lost sight of the best part of this board and what most people wanted it for:
- it's small size
- Any Dev Environment - ability to connect to any development environment, even all those the future may bring b/c of header pins
- 3G speeds and connectivity - AT&T is discontinuing 2G in the US
- Lower Cost than any other shield on the market
Circling back to these main ideas allowed us to make a beautiful one piece design which will lower the cost of the unit and make it more widely available.
4. The rules changed.
It used to be possible to certify non-standard interfaces (like header pins), now you're limited to only having USB, PCMCIA, Compact Flash, MMC, RS-232, or IEEE-1394 interfaces for data and standard ports for power - that's it! Not much room for those who want something small or integrated. In fact, anything out there that isn't a standard interface, isn't truly PTCRB certified.
We had a chance to talk more thoroughly with the PTCRB board members and PTCRB testing facilities and realized there was no real gain with the 2-pc design. If one were to productize around the top "cellular board" and create a PTCRB "final product" they would need to go through PTCRB with that design regardless of the certification level of the top "cellular board" itself. The only way to bypass PTCRB testing is to use an "end product" which means the "end product" would only have the previously specified standard interfaces. And even then, if you change anything in the power path, RF path, or SIM card path, you must re-certify and go back through PTCRB.I think that's about all she wrote on that story...