A scan tool reports "no communication" with the ECM. A U0100 sets. The reflex — at a lot of shops — is to order a module. Half the time that module shows up, gets programmed, and the fault is still there. Now the customer is paying for a part that was never broken, and the actual fault is still sitting on the bus.
A no-comm code almost never means "this module is dead." It means something on the network stopped agreeing. Before anything gets condemned, you start at the physical layer.
Start at the wire, not the module
CAN is a differential pair — CAN-H and CAN-L — and it has a resting state you can measure. Key-on, engine-off, a healthy high-speed bus sits with both lines near 2.5V, splitting to roughly 3.5V / 1.5V as it talks. Across the two terminating resistors you should read about 60 ohms with the battery disconnected. If you read 120, you have lost a terminator. If you read near zero or open, you have a short or a break — and no amount of programming fixes copper.
That five-minute measurement rules out the most common no-comm causes before you have touched a single module:
- Corrosion and chafed wiring — especially in harnesses that live near water, heat, or rodents.
- A single shorted node dragging the whole bus down. One failed module can take every other module offline, which reads as "everything is dead" when only one thing is.
- Power and ground to the module that "is not communicating" — a module with no clean ground cannot talk, and that is not the module's fault.
When it is the module
Sometimes the network checks out and the module really is the problem — water intrusion, a cooked driver, a corrupted flash, a cracked solder joint under a connector. That is where bench-level work starts: we power the module on the bench, watch whether it wakes and arbitrates, and decide between repair, reflash, restoration, or a cloned replacement based on what the silicon is actually doing — not a guess.
That distinction matters financially. A reflash or a board-level repair is a fraction of a new-and-programmed module, and for a lot of discontinued units it is the only option left.
Why the generic scanner stops short
A $30 OBD reader talks to the gateway and reports what the gateway hands it. It does not see the sub-buses behind the gateway, it does not do security access, and it stops at the code. Signal-level diagnosis is a different job: oscilloscope traces on the actual pins, network topology mapping, and OEM-grade access to the modules the generic tool cannot reach. That is the gap between "no communication — replace module" and knowing which node, which wire, and why.
What to send us
We work mail-in, nationwide. If you have isolated a module and want it repaired, reflashed, restored, or cloned, send the unit with the vehicle's make, model, and year and a note on the symptom and any codes you have pulled. We log it in, diagnose at the signal level, do the work, bench-test it against spec across the bus, and ship it back with a signed report of what we found and how it verified.
Every job starts with a signal. Every job ends with a signed report.
Got a module that won't cooperate?
Mail-in diagnostics, reflashing, restoration, and custom engineering — nationwide, with a signed report on every job.
