[Phambili Tech] creates a battery powered mountable display, called “the Newt”, that can be used to display information about the time, calendar, weather or a host of other customizable items.
The Newt tries to strike a balance between providing long operating periods while still maintaining high refresh rates and having extensive features. Many of the battery powered devices of this sort use E-Ink displays which offer long operating windows but poor refresh rates. The Newt uses an LCD screen that, while not being as low power as an E-Ink display, offers extended battery operation while still being daylight readable and providing high refresh rates.
The display itself is a 2.7 inch 240×400 SHARP “Memory In Pixel” LCD that provides the peppy display at low power. The Newt is WiFi capable through its ESP32-S2-WROVER module with a RV-3028-C7 Real Time Clock, a buzzer for sound feedback and capacitive touch sensors for input and interaction. A 1.85Wh LiPo battery (3.7V, 500mAh) is claimed to last for 1-2 months, with the possibility of using a larger battery for longer life.
The different pieces of functionality that take in outside data are brokered through MQTT API services coded into the device. Everything is open source software and open source hardware, including certification, so the enterprising hacker could point the services to anything desired or extend the Newt with their own communication protocol and functionality.
We’ve seen many other great wall mounted projects that use E-Ink displays as well as many projects that take advantage of the high refresh rate and low power consumption of LCD screens. We hope to see more applications that can combine the best ideas of both worlds of the e-paper application devices and devices using LCDs, while still being open source and hacker friendly.
How much longer would these units last if part of the wall panel was an amorphous thin film solar panel? Indefinitely even with a moderately sized panel?
Based on the battery life numbers, power usage is about 2 mW. Amorphous thin film solar cells have efficiency of about 6%. Indoor illumination of 100 lx is about 0.5W of light energy per square meter. If room is lit 12 hours per day, this would give 40 x 40 cm square as the needed panel size. Probably easier to run a cable :)
Ugh… it uses WiFi which is a HUGE power drain. Change that out for an RF receiver that turns on at timed intervals and run it using an AVRtiny and you could run it off a small amorphous solar cell and a super capacitor. Of course this means you need a separate transmitter that grabs the info. Lots of approaches to do that but in theory you could run a service on your router that would grab and transmit the info in intervals that include the correct time (allowing the omission of an accurate RTC), updated weather info (EEPROM could hold an entire week of data), and include a signature if you want it to validate the source. Timing accuracy only needs to be accurate to the minute until the next update interval.
Hey – this is Darian – the creator of the Newt. Thanks for your comments.
@Gravis is right – Wifi is a major impact on power/battery needs. I considered using RF as a data transmission channel, but that would require a second device (with an RF transmitter and a WiFi chip) to obtain information (NEWT currently connects every 30 mins to get the latest weather information; and hourly to get updated weather/UV/AQI forecasts, weather alerts, etc).
I have been looking into Matter (which is built into Alexa, Google Home, Apple Homekit, etc)… that might be a better option once the Matter receiver technology (or rather, documentation about it) is widely available.
I like the design. Available now at Mouser for USD $92.00 each plus shipping.[1] I know that Sharp/Kyocera 2.7″ Memory-In-Pixel (MIP) display is expensive. For example Adafruit sells it for $44.49.[2] But $92.00 – really? I have a couple more nit-picks: A. It uses a Micro Crystal AG (SWATCH Group) RV-3028-C7 RTC module, which is run-of-the mill stability-wise (±1ppm @ 25°C).[3][4] But that’s probably OK if it’s indoors and sets itself via SNTP a couple of times a day. I would prefer it had a Dallas/Maxim/ADI DS3231S Extremely Accurate I²C-Integrated TCXO RTC instead.[4] Unfortunately the DS3231S costs around ten bucks a-pop and are very difficult to find these days.[5] B. I do not think there is any backlight.
1. NEWT-01 @ Mouser $92.00.
https://www.mouser.com/ProductDetail/Phambili/NEWT-01?qs=Li%252BoUPsLEnueTLbjz6GTXQ%3D%3D
2. Adafruit SHARP Memory Display Breakout – 2.7″ 400×240 Monochrome $44.95
https://www.adafruit.com/product/4694
3. RV-3028-C7 SMT RTC Module
https://www.microcrystal.com/en/products/real-time-clock-rtc-modules/rv-3028-c7/
4. RV-3028-C7 32.768kHz 1ppm Qty.-1 $3.06
https://www.mouser.com/ProductDetail/Micro-Crystal/RV-3028-C7-32.768kHz-1ppm-TA-QC?qs=u16ybLDytRaU7ZFJl2AxfQ%3D%3D
5. DS3231X Extremely Accurate I²C-Integrated TCXO RTC
https://www.maximintegrated.com/en/products/analog/real-time-clocks/DS3231.html
6. DS3231S @ Mouser
https://www.mouser.com/c/?q=DS3231S
+1 Thanks for the links.
Hey @drone – thanks for your comments. I thought a lot about the components that went into the NEWT. Some I’d keep – others I’d change. I also designed and built this during the chip shortage, which impacted certain decisions.
a few callouts
1. I chose the RV-3028-C& RTC because I needed
– an RTC that provided a separate clock out pin and interrupt pin
– clock out pin is used to send a 1HZ pulse to the display (so there is no screen burn in)
– Interrupt pin supports three alarms that can be used at the same time – Periodic (changes the time every min, countdown timer – used for the regular and Pomodoro timer, and alarm)
2. You’re right – the Sharp displays are expensive. Adafruit provides only the display for $45 (which I purchased and used for early prototyping). The NEWT includes the display plus:
– ESP32S2
– RTC
– QWIIC connector
– Speaker and Amplifier
– 10 cap touch pans
– Battery charging circuitry (with a low batt indicator)
– Works “out of the box” software that includes a bunch of stuff
That – plus the need to use Crowd Supply as a selling channel – is why the units are $92 (they were cheaper at the launch date). [no shade against Crowd Supply – I love working with them and would work with them again]
That being said… $92 is a lot of money… so I’m all for people building their own – or better yet, building a better version. I’ll add a comment below with links to all the software (device and server side) and hardware designs.
And if I had to do it differently:
A. I might use a NE555 to send a 1 HZ pulse to the display, and use a different RTC- as long as it was low cost, low power, and supported multiple alarms/timers. Or maybe I’d add a crystal to the ESP32 and use internal RTC (which is super inaccurate w/o an RTC).
B. Maybe I’d use a combination of a low power LCD 7 segment display (for time) with a 4.5 inch eInk screen.
C. I think I’d add a legit battery fuel monitor (I use a voltage monitoring chip right now, that goes HIGH when the batt voltage falls below 3.5V). There were few to no LiPO fuel gauge chips in stock when I launched NEWT
–
As alternative to the RV-3028-C7, you can use the RV-3032-C7 TCXO RTC from Micro Crystal.
With +/-2.5 ppm time accuracy over -40 to 85°C, it complies with your selection criteria. Moreover, it has a 12-bit embedded temperature sensor – adding one more data (room temperature) to the system.
RV-3032-C7 SMT TCXO RTC Module
https://www.microcrystal.com/en/products/real-time-clock-rtc-modules/rv-3032-c7/
Sweet
A while ago I wanted the same thing for home automation: some wireless screens with 1+ years of battery life. I could not find a display that could work, the consumption was too high for getting that much battery life from AA/18650 cells. If i would use many batteries/bigger size the thing around look quite ridiculous: huge box with tiny screen.
I have up and went for esp32 LCD with a USB connection.
There are a few chips that now promise almost an (almost) always wifi connection with low battery drain (enough to last a year – depending on the battery). The DA16200and the ESP32S6 (I think). Played around with the DA16200, but paused the project (not enough time to tool around with it)
My point is that the LCD will be the limiting factor in power consumption. I for one use RFM69, not WiFi which is way more power efficient.
But yeah, I could not find a LCD in the 2-3 inch range with low enough consumption.
I stand corrected looking at my notes:
-Nokia LCD could run close to a year on 2-3 AA batteries but it was too small
-some small LCDs on aliexpress also existed that claim low power, but were also small
-a plethora of quite cheap, quite large LCDs exist, but had consumption in the mA range. Finally:
-some MIP screens could also be bigger and run for one year on 2-3 AA batteries, but the cost was unjustifiably high compared to just running something plugged in.
So i went for plugged in color LCDs
I wish the interface could be like this https://hackaday.io/project/187369-esp32-lumia
Thanks for the link… I’ll look into it. The display is monochrome, so I don’t have many (any) options for color… but maybe I can pull some inspiration from the lumia device for the next sofwtare update
Hey all – this is Darian (the creator of NEWT). Thanks for your comments; really appreciate your thoughts and suggestions.
NEWT is fully open source, and the software, hardware specs, eagle files, etc. are available on Github
Software: https://github.com/Phambili-Tech/Newt_Display
Hardware: https://github.com/Phambili-Tech/Newt_Display_Hardware
In addition, we have a pretty active community of users. I think we’d all welcome your feedback on how to make NEWT better. Right now, we’re focused on improving the software (v 1.1.0-RC-1 is out for Beta and will be released GA in the next few weeks).
Our discussion board, wiki, and issue log are all up to date:
Discussions: https://github.com/Phambili-Tech/Newt_Display/discussions
Issues: https://github.com/Phambili-Tech/Newt_Display/issues
Wiki: https://github.com/Phambili-Tech/Newt_Display/wiki
A happy owner here. I am supporting you and your approach through purchasing, not just buying a gadget. Please showcase that! It is what pushed me past the “high price” hurdle and I am sure others as well.
Nicks’ comments are spot on about your NEWT work representing how I want to see products released in the future.
I recognize that the price represents a significant hurdle in “comparison shopping”, but it is an investment in the community and your “track record” to build trust. Playing the long game, I can see the possibility of scale as the community grows and others seen the track record of delivering a product “in the open”.
Most people are concerned about sharing $ numbers and there isn’t much precedent for it, but I encourage you there as well. I imagine 95% of the time, there isn’t much profit left over (if you include the value of your time) and people should see the level of investment (in time and up front costs) you are making.
Thanks for making NEWT a reality!
Thanks [Darian] for the helpful comments and conversations on here. I would love to see more products designed and sold this way – “here’s something cool that I built, here’s a link to buy one. Here’s a link to all the source files if you want to make your own. Here are some sensible answers to questions about my component choices, links to discussion forums, and comments about future development. ”
Great stuff!