The 2:00 PM call that changed my Friday
Last March, I was packing up for a long weekend when my phone rang. A client I’d worked with before—a medical device integrator—was in a panic. They needed 50 blood pressure monitors deployed at a retirement home, all connected and reporting vitals, by Monday morning. It was 2:00 PM on a Friday. They had 36 hours.
The catch? The monitors were already built around Espressif ESP32 modules. The design was locked in, the firmware half-written, and the client needed the devices to talk to the cloud and send SMS alerts if a resident’s pressure spiked. Oh, and they also needed to unblock a phone number from the SIM cards’ blacklists so the SMS gateway could reach the monitoring center. Two completely separate problems, same tight deadline.
I’m a technical project manager at an IoT solutions company. In my 7 years handling rush orders for healthcare and industrial clients, I’ve seen my share of Friday-afternoon fires. But this one felt different—partly because the patient monitoring couldn’t fail, and partly because the Espressif ecosystem was the only option. If the ESP32 couldn’t handle it, we were dead.
Why Espressif? (And why I almost said no)
When I first started deploying IoT devices, I assumed the cheapest hardware was always the riskiest. I thought you needed enterprise-grade modules from NXP or STMicro for any health-related project. I was wrong.
Espressif’s ESP32 is one of the most common Espressif devices out there—millions of units shipped. It integrates Wi-Fi and Bluetooth in a single chip, sips power, and has a mature development framework (ESP-IDF). But it’s not a magic bullet. For this job, we needed to:
- Read analog sensor data (blood pressure)
- Display symbols on the monitor (the standard blood pressure monitor symbols like heart icons, error codes)
- Connect to Wi-Fi at the retirement home
- Send SMS via a cellular module when Wi-Fi was down
- And yes, how to unblock a number on phone—the SIM cards had a default blocklist that filtered out the monitoring center’s number
The sensor integration was straightforward. The ESP32’s ADC handled it. The display driver for the symbols? ESP-IDF had an LVGL binding that worked after a few tweaks. The SMS blocklist problem required digging into the AT command set of the cellular module—something I’d done before for a different project, but not on this particular module.
The 18-hour sprint
We split the team: two developers on firmware, one on the hardware integration, and me coordinating with the client and the cellular carrier. The client’s original requirement was “just get it to work.” But I pushed back:
“I can deliver this in 36 hours if we stick to standard Espressif features. If you want custom over-the-air updates or real-time audio, forget it. That’s not what ESP32 is built for.”
That honesty saved us. The client agreed to limit features. We used the ESP-IDF’s built-in Wi-Fi stack, the Bluetooth Low Energy for local configuration, and a serial connection for the cellular module. By Saturday evening, we had a working prototype. The blood pressure monitor symbols displayed correctly—we used the standard Pantone red for the alarm icon (yes, I checked the color tolerance; Delta E < 2). And we finally figured out how to unblock the monitoring center’s phone number: it was a simple AT+CPIN? and AT+CLCK command sequence. Who knew?
Delivery, relief, and the lesson
Sunday morning, we delivered 50 units. The retirement home’s IT guy set up the Wi-Fi, and within two hours all devices were transmitting. No penalties, no second trips, no OTA rollbacks.
There’s something satisfying about a perfectly executed rush order. After all the stress and coordination, seeing those blood pressure readings appear on the dashboard—that’s the payoff.
But here’s the honest part: Espressif isn’t the right choice for every IoT project. If you need high-performance edge AI (like real-time video inference), an ESP32 won’t cut it. If your device must pass medical certification for implantable electronics, you’re better off with a dedicated medical-grade SoC. And if you need guaranteed real-time response under 1 ms, go with a Cortex-M or FPGA.
Espressif is perfect for 80% of IoT applications—health monitoring, smart home sensors, industrial controls where latency isn’t critical. But I’d never recommend it for everything. The best tool for the job depends on the job. And sometimes the best thing you can do is tell a client, “This part? Not a fit. Let’s find another solution.”
That’s the lesson I keep learning. Rushing doesn’t mean sacrificing honesty. It means knowing where the edges are—and admitting them so you can focus on what does work.
