During early engineering stage, one key issue surfaced right away — sealing the slanted joint between the metal housing and the plastic front panel. As I mentioned earlier, the enclosure meets IP21 requirements — technically speaking. That’s both true and not, for two reasons

First, the client insisted that a medical enclosure shouldn’t even have an IP rating — in their view, it should simply withstand having a beaker of disinfectant spilled over it (which I translated as IPX1). Second, the bent metal edges varied in thickness, which meant additional manual work during sealing and inspection. Still, that was the client’s deliberate choice, so yes — officially, the enclosure meets IP21

I suggested an alternative mechanical design — better suited for manual sealing and with a more refined rear section — but it was rejected on cost grounds. If you ask me, it all evens out: whatever we skip during mechanical design or production, we make up for later with extra care and manual work 

• This device — an electrosurgical coagulator — is an advanced electric scalpel that, in addition to cutting tissue (achieved by passing high-frequency current through it, causing heating and evaporation of cellular fluid — the function labeled Blend 1 on the device), also provides tissue coagulation, e.g., for bleeding control (Blend 2 — a combined cutting and coagulation mode).
• Coagulation can be soft — localized and gradual, with evenly distributed energy. This mode allows “sealing” blood vessels up to 1–2 mm in diameter, which is important when operating near delicate structures where minimal damage to surrounding tissue is critical. It can also be spark-type — with more intense heating, even causing tissue charring, but ensuring hemostasis of vessels up to 2–3 mm. Larger vessels require mechanical ligation or clips.
• The instrument held by the surgeon, connected to the device, can operate in monopolar mode (current flows through the patient’s body to a neutral electrode — NE) or in bipolar mode, where current flows only between two electrodes of the instrument.
• For convenience, the device provides two bipolar outputs to connect different tools (e.g., forceps and tweezers).
• To avoid manually switching the current on and off, the device features auto-start and auto-stop functions — energy delivery begins automatically upon tissue contact and stops when contact is lost.

• the foot pedal connector supports two pedals In monopolar mode: one for cutting and the other for coagulation — allowing the surgeon to select the function by foot;
• the single foot pedal connector is used in bipolar mode (for coagulation only);
• the connector for controlling an electromagnetic valve that supplies gas (e.g., argon) through the instrument. The gas improves the quality of coagulation and helps prevent tissue charring;
• Power input. The 500 Hz rating is not a mistake. A standard 50 Hz AC outlet cannot power this device directly. It must be connected through an inverter that raises the frequency to 500 Hz before feeding the electrosurgical unit. This design choice helps reduce the overall weight of medical equipment

The tall rubber feet provide enough airflow underneath (vent holes are only on the bottom due to sealing requirements). That’s doubly important if several units are stacked. The clearance also makes it possible to grab the device with your fingers — otherwise, we’d have to add handles

There’s plenty of fastening hardware — from plastic standoffs screwed to the bottom (to isolate the PCB from the metal body) 

When a client brings a finished PCB before a project starting (which happens more often than you’d think), it’s usually bad news for both the designer and the mechanical engineer — it limits the options for layout and ergonomics. Fortunately, in this project, the electronics and the enclosure were developed in parallel. Had the client come with pre-designed electronics, they would have ended up with a single front-panel PCB combining both the numeric indicators and the rotary encoder (the “knob”). That would have placed the displays of indicators deeper than the keypad — not great visually — and forced us to add glossy transparent windows above them, which means higher cost and glare issue 

Plus, those components wouldn’t have aligned with the designer’s intended ergonomics. The same goes for keypad connectors — they can’t be placed just anywhere. And it’s entirely possible the client would’ve picked the wrong connectors altogether — with a different pin pitch or no proper ribbon support — meaning the keyboard supplier’s interface wouldn’t fit 

If the client had come with a completely finished power-electronics module — yes, I’m still complaining — the main PCB would likely have awkward proportions and a rectangular outline. That would mean mounting the power supply above it, on the “ceiling” of the enclosure — a logical but space-inefficient choice. Then you’d need extra brackets, reducing service access and adding height. And of course, every extra part means extra cost

During the industrial design phase, the designer drew a cute, compact power button, perfectly aligned with the row of output connectors. But the real, working component we selected at the mechanical stage — the one that matched the electrical specs and looked decent — turned out twice as big. We had to adjust the keypad angle and enlarge the front section slightly just to make it fit. Acceptable. 

In total, the assembly consists of 11 unique parts. The project took eight months (including industrial design) and around 200 engineering hours. Could it have been done faster? In theory — yes. In practice — not really. From experience, these figures are pretty typical

If the client had responded instantly to every question and approved design proposals on the spot, the timeline could’ve been reduced to two or three months. But that only happens in emergencies — when they need it “yesterday”. That’s not healthy design; that’s just rushing a product out the door at the expense of quality

Under normal conditions, the client’s pace is… relaxed: a week or two to think, or sometimes disappearing for three. Meanwhile, they were developing the electronics and power sections in parallel — and also handled manufacturing themselves. So the first physical prototypes appeared about three months later

To reach a clean, balanced result that ties together the design, electronics, and every other technical requirement, it took dozens of iterations and a mountain of messages. If you add up just my written correspondence for this project, it comes to around 16,000 words. To put that into perspective: that’s roughly one full issue of National Geographic or Esquire, or a Wikipedia article the length of “History of the Byzantine Empire