Writing · Question
How long do battery e-paper signs last?
A battery e-paper sign lasts as long as its radio lets it: days if it checks for new content every few minutes, months at a few updates a day, and years at one update a day or less. The panel itself is almost never the problem. Here is the math we use when a customer asks, and the things that quietly drain cells in the field.
Published 2026-11-09. Numbers come from the ESP32-S3 + E Ink Spectra 6 stack we ship in 8frames and adapt for team projects. They are napkin math with stated assumptions, not a lab certificate — swap in your own figures.
The short answer
Count wakes per day. Every wake costs a fixed lump of charge: boot the microcontroller, join Wi-Fi, ask the server whether anything changed, download if it did, drive the panel if the image changed, then go back to deep sleep. Between wakes the device sleeps at microamps and the e-paper holds its image for free. So runtime is roughly:
Rule of thumb: Runtime (days) ≈ battery capacity (mAh) ÷ (wakes per day × charge per wake + sleep charge per day).
Everything interesting is in “charge per wake”. On our stack the panel refresh is the smaller part of it. In our ESP32-S3 + Spectra 6 field guide we budget about 0.11 mAh for a full Spectra 6 refresh (roughly 50 mA for 20–30 seconds), around 10 µA for the ESP32-S3 in deep sleep with the panel power cut, and a deliberately conservative 3 mAh for the Wi-Fi part of each wake.
Runtime by refresh cadence
Using those budgets on a 3,000 mAh Li-ion cell, and assuming every wake also redraws the panel (the worst case), here is what the formula gives. The right-hand column shows the same device if you get each Wi-Fi wake down to 1 mAh — which is why radio time is the first thing we optimize.
| Update cadence | Wakes / day | 3 mAh per Wi-Fi wake | 1 mAh per Wi-Fi wake |
|---|---|---|---|
| Every 5 minutes | 288 | ~3 days | ~9 days |
| Every 15 minutes | 96 | ~10 days | ~28 days |
| Hourly | 24 | ~40 days | ~3.7 months |
| Four times a day | 4 | ~7.8 months | ~1.8 years |
| Once a day | 1 | ~2.4 years* | ~6 years* |
| Once a week | 1/7 | self-discharge-limited* | self-discharge-limited* |
Sleep budget: 10 µA × 24 h ≈ 0.24 mAh/day. Panel refresh: 0.11 mAh per wake. *At this point the cell’s own self-discharge and calendar aging, not the display or the radio, set the real limit — treat multi-year figures as “the electronics won’t be what kills it”, not as a promise.
Two things fall out of the table. First, the jump from hourly to every five minutes is not 12× worse in practice — it is the difference between a sign you charge every quarter and a sign somebody has to plug in twice a week. Second, the single biggest lever is not a bigger battery; it is fewer, shorter radio wakes.
How this compares with vendor claims
Large-format vendors publish numbers in the same shape. Samsung says its 32-inch Spectra 6 EM32DX runs up to 199 days on a 4,600 mAh battery if content is updated once per day. Divide it out and that is roughly 23 mAh per day — a lot more than a 7.3″ or 13.3″ sign, which is what you would expect from a 2560×1440 panel and a bigger radio and processor budget. LG’s 32-inch Spectra 6 sign has a 72 Wh built-in battery and says runtime varies with content update frequency. At the other end of the size range, M5Stack’s 4-inch Spectra 6 module ships with a 1,250 mAh cell. Same technology, very different power envelopes. When you compare claims, ask three questions: what cadence, what radio, what temperature.
What actually drains e-paper sign batteries
1. Radio time per wake
Joining Wi-Fi, getting an address, and talking to the server is the expensive part of every wake. Our firmware sends the content tags (etags) it already holds on every sync, so a wake where nothing changed is a short status exchange rather than a download. Weak signal makes all of this slower: Visionect reports that very poor signal can cut e-paper sign battery life by as much as 70%, and that cellular connections drain faster than Wi-Fi.
2. A schedule nobody tuned
Our firmware supports two schedule styles: a list of wall-clock times (up to 24 per day, in the device’s own time zone), or a fixed interval of every N minutes up to once per 24 hours. If a device has never received a schedule it falls back to waking hourly. For most retail and workplace content, a handful of wall-clock wakes (before opening, after lunch, at close) beats a short interval by an order of magnitude on the table above — with no visible difference to shoppers.
3. Power-management chips running on defaults
This one cost us a real field failure. The 13.3″ driver board we build on uses an IP5306 power-bank PMIC. Left on its power-on defaults, a chip like that can cut its boost output when the load gets light — which is exactly what a deep-sleeping e-paper device looks like. A frame went dark overnight. The fix was to configure the chip at every boot and read the registers back. If your board has a power-bank or charger IC, check its light-load behavior before you trust any sleep-current number.
4. Cold, heat, and where the sign is mounted
Battery chemistry is temperature-sensitive. Visionect cites 20–25 °C as the ideal range and notes indoor signs generally outlast outdoor ones. A freezer-aisle header or a loading-dock label will not get the runtime of the same hardware in a meeting room.
5. Extras that never sleep
Front lights, touch controllers and always-listening radios all add a constant draw that the e-paper itself does not have. Visionect also lists touch as a battery cost. If the product does not need them, leave them off the BOM, or switch them off hard in sleep.
How we check battery life in the field
A sign on a shelf has no serial cable attached, so “the battery died early” usually arrives as “the picture stopped changing”. Our firmware keeps a black-box log for exactly this:
- Battery voltage at boot and its minimum during the panel refresh, which is when sag shows up first.
- Why the device woke (timer or button) and why it last reset (including brownouts).
- A count of panel BUSY timeouts, which separates power problems from display problems.
- The last few log lines ride along in every cloud sync, so the history from an unplugged stretch lands as soon as any sync works.
For a fleet, that is the difference between guessing and knowing which stores need a charging round. It is the same telemetry the multi-location fleet dashboard is built around.
Typical cadences by use case
- Brand shelf and endcap signs — weekly promo swaps, so primary-cell or long-interval rechargeable designs are realistic. See retail shelves.
- Meeting-room and desk signs — booking changes are event-driven; cadence is the main design decision. See workplace signage.
- Kanban and bin labels — as often as the ERP or MES pushes state, throttled to the panel’s refresh. See industrial e-labels.
If you are still choosing between e-paper technologies, the refresh and power trade-offs differ by film — our Spectra 6 vs Kaleido 3 vs Gallery 3 vs BWRY comparison covers that.
FAQ
How long does a battery-powered e-paper sign last on one charge?
Anywhere from a few days to several years. The deciding variable is how many times per day the sign wakes up, joins Wi-Fi and fetches content. On a 3,000 mAh cell with our conservative per-wake budget, once a day works out to roughly 2.4 years on paper, hourly to about 40 days, and every five minutes to about 3 days.
Does an e-paper display use power when the image is not changing?
No. Spectra 6 is a reflective, bistable film: the pigments hold their position with zero power, so a static image costs nothing. What still draws current between updates is the microcontroller in deep sleep and any regulator or power-management chip you leave on.
Why do vendor battery claims for e-paper signs vary so much?
They assume different update cadences, panel sizes, radios and test temperatures. Samsung quotes up to 199 days for its 32-inch EM32DX at one update per day; a 4-inch module on a 1,250 mAh cell is a different device entirely. Always ask for the refresh cadence and connectivity behind a battery claim.
Is Wi-Fi or cellular better for battery e-paper signage?
Wi-Fi, where it is available and strong. Visionect reports that poor signal can cut e-paper sign battery life by as much as 70%, and cellular radios are hungrier than Wi-Fi in general. Survey signal at the mounting spot, not at the store office.
Should a shelf or wall sign be rechargeable or use primary cells?
If it updates daily or more, rechargeable Li-ion with USB-C is usually right. If it updates weekly or less, the math often lets a primary cell outlast the campaign. We scope this per project on the scoping call, because mounting, temperature and who will swap cells matter as much as the math.