The Day I Realized Our Team Had Been Misreading Battery Replacement Specs
It was a Tuesday morning in late March 2024. Our senior biomedical technician—call him Tom—walked into my office holding a Permobil F5 battery manual and a service log. He looked tired.
"We've been installing the wrong replacement batteries," he said.
I didn't believe him at first. We'd been working with Permobil wheelchairs for years. Our team prided itself on following manufacturer specs to the letter. But Tom flagged something that made me sit up: the torque specification for the battery terminal connection in the manual we were using differed by 15% from the spec listed on the Permobil technical portal.
Fifteen percent. That's not a rounding error. That's the difference between a secure connection and a fire risk.
I'm the quality compliance manager at a mid-sized rehabilitation equipment supplier. I review every product manual and service protocol before it reaches our technicians—roughly 200+ unique items each year. I've rejected about 18% of first deliveries in 2024 due to documentation errors. But this one slipped through.
Our Process: How We Ended Up with the Wrong Spec
Here's how it happened, step by step—painfully ordinary:
We received a batch of 12 Permobil wheelchair battery replacement kits in late February. Our inventory team checked the part numbers against the packing slip. Everything matched. The kits were logged into inventory. A week later, Tom pulled one for a routine service on an F5 Corpus.
The manual included with the kit had a torque value of 8 Nm for the terminal bolt. Tom followed it. On the second wheelchair he serviced that day, he noticed the connection felt loose compared to the first one. He checked the Permobil online portal—the spec there said 7 Nm. He checked the F5's original manual (the one the user received with the chair)—6.8 Nm.
Three different numbers for the same bolt.
I should mention: Tom has been a technician for 14 years. He knows that when you see variation like this, it's not usually a conspiracy—it's just that different versions of documentation get updated at different times. But in a quality system, consistency is everything.
Why This Matters Beyond Wheelchairs
Look, I'm not a mechanical engineer, so I can't speak to the physics of bolt torque in medical devices. What I can tell you from a quality perspective is this: if our documentation for a power wheelchair battery connection is inconsistent, how confident are we in our specs for more complex equipment?
Our facility also maintains mechanical ventilators and hospital beds. The same procurement process that brought in those battery kits also brings in ventilator circuits and bed components. If we can't trust a wheelchair battery manual, why should we trust the documentation for a life-support device?
That's not a rhetorical question. It's the question I had to answer for our executive team.
The Audit That Changed Our Approach
In Q1 2024, we conducted a focused audit of all technical documentation for our top 10 products by service volume. The results were sobering:
- Permobil F5 battery replacement documentation: 2 of 3 internal manuals had a version mismatch with the OEM portal.
- Mechanical ventilator maintenance guides: Correct, but cross-referencing took 30 minutes per device.
- Hospital bed assembly instructions: Outdated by one revision—the new version added a warning about a specific electrical component.
Why does this matter? Because in a medical facility, "close enough" isn't a standard.
We found that three different departments were using three different versions of the same manual for the same device model. The root cause wasn't malicious—it was just that updates happened, and no single person was responsible for ensuring everyone got the new version.
Simple. But it cost us $3,200 in rework on the two chairs Tom had already serviced.
The Fix: A Central Specification Registry
After the audit, I implemented what I call a "specification single source of truth"—a Google Sheet, honestly. Nothing fancy. It lists every device we service, the current documented spec for each critical parameter (torque, voltage, replacement intervals), and the source document with version and date.
When a new manual arrives—say, for a Permobil F5 battery replacement—the technician doesn't just file it. They compare the critical specs against the registry. If there's a discrepancy, the registry is flagged pending clarification from Permobil support.
It sounds obvious. But before this, each technician kept their own set of manuals. Some had PDFs from 2021. Some used printed copies from 2023. None of them were wrong on purpose—they just didn't know they were out of date.
What I Learned (And What You Can Steal)
Here's the part where I sound like a lecture, but I promise it's earned:
- Don't trust the manual that comes with the part. That manual might have been printed two years ago. The current spec might have changed. Always cross-reference with the OEM's current online portal.
- Specifications are not opinions. If two documents disagree, stop work until you know which one is current. The 20 minutes you spend verifying could save you a $3,200 redo.
- Small teams need spec registries more than big ones. Large facilities often have dedicated document control staff. Small facilities rely on the memory of their most experienced technician. When that person takes a vacation (or leaves), the institutional knowledge goes with them. A written registry protects against that.
This was accurate as of mid-2024. The medical device documentation landscape changes fast—manufacturers update manuals, regulations shift, and new versions of devices come out. Verify current specs before proceeding with any service work.
At our facility, the Permobil F5 battery spec discrepancy is resolved now. We use the Permobil portal as our single source of truth, cross-referenced by at least two technicians before any service. It's added about 15 minutes to each battery replacement job—but we haven't had a single torque-related rework since.
And yes, Tom still gives me a hard time about it. He earned that right.
