Well...I mean we did sell the products so there was a market for them. And yeah, before we released them they were not commercially available...so you got me there ;)
Functioning RF power harvesting MCUs have been demonstrated for a number of years now. Obviously there are challenges when the state-of-the-art is pushed forward, but the claims don't seem far fetched.
There are a lot of interns working at the company! IDK much about Norway - is that the name for a graduate?
I assume the micro itself is not developed by them given the team size - even with 1 guy looking after this, we're looking at a factor of 30+ in terms of effort to develop, test and certify this type of IC
The combination of a mask rom and reprogrammable flash on the same product surprised me. Why would you want a customizable mask rom if you’ve already got 8x or more of flash? Is it a power consideration: are mask roms lower-power to read from?
One possible explanation could be the desire to have a guaranteed-never-modified bootloader that is rock solid, which allows for firmware updates on the flash (in addition to booting the system in normal use). This is how Apple's iOS devices are designed, for example.
NOR flash is also slow. That is why a lot of microcontrollers that run at hundreds of megahertz have very strange flash accelerators that read slowly in parallel (STM uses a 128-bit-wide bus for example), and then dispense data/code to the CPU quickly in 16 or 32-bit pieces. Most ROMs can be read very quickly.
They are ex-Nordic cadres. You can note, they keep their own key tech cadres incommunicado :). Unlike business people, they have not a single bit of their contact information listed on their page.
Would love to know if this is going to capitalise on the 5G rollout .. I seem to remember a few years ago someone making claims that once 5G was deployed, ULP sensors would be more feasible ..
Seems like a great time to join this company and help make this product come alive after years of work. I would guess they are an efficient and driven team where everyone knows everyone else.
I am not affiliated with them, but working at companies when they start to gain traction and to feel how proud everyone is at getting a product out with a relatively small team is great fun!
Nope. They collect contact info but send nothing. At least not automatically. No email sent to my trash email account when I filled out the form as Johny Smith
If this comes out and is affordable, it'll have some really cool applications. As per datasheet, it also has UWB support so you could make a "where's my keys????" (or inventory tracker) without having to worry about dead batteries or limited range of RFID.
> 1KB of mask ROM (stdlib, math etc). 2KB RAM. 8/16/32KB ULP Flash. 100k Writes. Read supported down to 850mV.
Even 32kB of rom is going to be very tight to fit a wireless stack like BTLE. It’s not clear if their BTLE peripheral includes the stack or just the radio. I would hope it’s the latter.
The harvester(s) are of course built in. There are some pasives - one resistor for a master bias and a handful caps. The antenna is taken directly in - or solar cell type.
How low power is it - it all depends is the correct answer. The chip is tuned for a few different application scenario:
1. DVFS mode: Here you select a clock speed/voltage your application require - and you will get the associated voltage/clock speed in return. You can doctor your code with calls to have various performance levels.
2. Deadline based: You select a deadline for the work to complete within - and the device will throttle by setting clock speed to match this requirement. The voltage is matched for the selected clock speed.
IO voltage is maintained from a seperate regulator - so internal logic voltage can vary with stable IO voltage.
Mask rom + flash...why? Maskrom for some secure boot stuff, stdlib and math - it is less energy per fetch from mask rom vs. flash. Ideally everything should be in ROM - but that makes it hard for the end user to add his/her application
26 comments
[ 3.3 ms ] story [ 66.0 ms ] threadI assume the micro itself is not developed by them given the team size - even with 1 guy looking after this, we're looking at a factor of 30+ in terms of effort to develop, test and certify this type of IC
They have open positions on their career page: https://www.onio.com/career.html
Seems like a great time to join this company and help make this product come alive after years of work. I would guess they are an efficient and driven team where everyone knows everyone else.
I am not affiliated with them, but working at companies when they start to gain traction and to feel how proud everyone is at getting a product out with a relatively small team is great fun!
Does it have the RF harvesting bits built in? or is it a bunch of passives that I have to add after.
How low power is it? are we talking microamps per mhz(I know we're not...) How much useful time do you have per charge cycle?
the articles section seems to be full of generic health mumbojumbo, which isnt a good start... https://www.onio.com/article-list.html
The initial versions of the datasheet will be read by NOBODY if this is kept up! If you make it hard to see what your product does, who will care?
Learn from STM32 - public datasheet!
edit: my bad, that's just for the "key features", not the full datasheet
Even 32kB of rom is going to be very tight to fit a wireless stack like BTLE. It’s not clear if their BTLE peripheral includes the stack or just the radio. I would hope it’s the latter.
How low power is it - it all depends is the correct answer. The chip is tuned for a few different application scenario:
1. DVFS mode: Here you select a clock speed/voltage your application require - and you will get the associated voltage/clock speed in return. You can doctor your code with calls to have various performance levels.
2. Deadline based: You select a deadline for the work to complete within - and the device will throttle by setting clock speed to match this requirement. The voltage is matched for the selected clock speed.
IO voltage is maintained from a seperate regulator - so internal logic voltage can vary with stable IO voltage.
Mask rom + flash...why? Maskrom for some secure boot stuff, stdlib and math - it is less energy per fetch from mask rom vs. flash. Ideally everything should be in ROM - but that makes it hard for the end user to add his/her application