A fleet manager running 80 Class 8 tractors told me last year that his Samsara deployment caught every DTC his drivers ignored — but missed the SPN 3216/FMI 18 NOx sensor drift pattern that put three aftertreatment systems into forced regen lockout within the same month. The repairs ran $4,100–$5,600 per unit. The data was in the ECM. The platform just never surfaced it in a usable form.
That's the real question in the Geotab vs Samsara fault code debate. It's not which platform has a cleaner dashboard or better driver scoring. It's which platform reads deeper into the J1939 data bus, captures more diagnostic parameter IDs, and structures that data in a way a director of maintenance can actually act on before a roadside failure turns a $200 sensor swap into a $14,000 regen system overhaul.
What J1939 Diagnostic Depth Actually Means
Every modern heavy-duty diesel communicates on SAE J1939 — a CAN-based protocol that carries Suspect Parameter Numbers (SPNs), Failure Mode Identifiers (FMIs), and a library of Parameter IDs (PIDs) covering everything from engine coolant temperature to DEF quality sensor voltage. Passenger vehicles use OBD-II PIDs under SAE J1979. The architectures are different, the data density is different, and the failure consequence gap between a missed J1939 fault and a missed OBD-II code is enormous.
A basic telematics device reads active DTCs — SAE J1939 DM1 messages. That's table stakes. The diagnostic depth question is whether the platform also captures:
- DM2 (previously active fault codes with occurrence counts)
- DM4 (freeze frame data at fault activation)
- DM25 and DM26 (expanded freeze frame, emission-related test results)
- Continuous PID streaming for parameters like SPN 190 (engine RPM), SPN 110 (engine coolant temp), SPN 94 (fuel delivery pressure), SPN 1600 (EGR valve position), and SPN 3480–3483 (SCR system efficiency)
- Proprietary OEM SPNs that fall outside the base J1939 spec
That last category is where the platforms diverge most sharply.
Geotab's Architecture: Deep OEM Integration at Scale
Geotab's hardware — primarily the GO9 and the newer GO9+ with BLE — connects via the J1939/OBD-II port and reads standard diagnostic messages plus an expanding library of OEM-specific PIDs through what Geotab calls its "Custom Data" framework and the MyGeotab SDK. The platform's ProPlus subscription tier enables multi-PID logging at configurable intervals, with some fleets reporting active monitoring of 80–120 discrete SPNs per asset on supported OEM integrations.
For Cummins X15 and ISX engines specifically, Geotab's integration captures SPN 5246 (aftertreatment 1 SCR efficiency), SPN 1569 (engine protection torque derate active), and SPN 3251 (DPF differential pressure) alongside standard fault data. On Paccar MX-13s, the platform can pull SPN 3936 (NOx sensor upstream) and SPN 3719 (DPF soot load percentage) — two of the most diagnostically valuable predictive indicators for emission system failures.
The DPF soot load percentage figure deserves specific attention. When SPN 3719 climbs past 85% of modeled capacity without a successful active regen completing, you're typically 40–60 engine hours from a forced parked regen or a DPF pull. That window is everything. The true cost of a deferred DPF cleaning in a 25-truck diesel fleet makes the financial case in granular detail — but the short version is that a $350 cleaning becomes a $2,200–$4,800 replacement when the window closes.
Geotab also maintains an Engine Fault Code database with over 100,000 entries cross-referenced by SPN/FMI combination and vehicle year/make/model. This metadata layer distinguishes a SPN 3597 FMI 1 (NOx sensor signal below normal) from SPN 3597 FMI 3 (voltage above normal), which have entirely different diagnostic pathways and repair cost profiles.
Samsara's Architecture: Breadth Over Depth, With Trade-offs
Samsara's VG34 gateway and the CM32/CM34 vehicle gateways also read J1939 DM1 and DM2 messages and stream a solid set of standard PIDs. For fleets running mixed assets — light commercial, medium-duty, Class 8 — Samsara's unified dashboard and its broad OBD-II coverage for non-commercial vehicles is a genuine operational advantage. Its fault alerting is fast, its mobile interface is well-designed, and for fleets that primarily need driver behavior and basic fault notifications, it does the job.
The gap opens at OEM-specific PID depth. Samsara's proprietary SPN library is narrower than Geotab's for heavy-duty diesel applications. In independent testing documented by TMC fleet working groups and echoed consistently in fleet manager forums, Samsara tends to capture 60–70% of the diagnostic PID set that Geotab surfaces on the same Cummins or Detroit DD15 application. For standard J1939 parameters, the gap is minimal. For OEM-extended SPNs — the ones that actually drive predictive maintenance decisions — it's material.
Samsara's freeze frame data (analogous to J1939 DM4) is available but less granularly accessible via API than Geotab's. For fleets building custom integrations into maintenance management systems like Dossier, TMT, or Decisiv, this matters. The depth of what you can extract programmatically determines whether your predictive maintenance model runs on rich fault context or just fault codes and timestamps.
A Concrete Failure Scenario: 2019 Freightliner Cascadia, 487,000 Miles
A private carrier operating temperature-controlled LTL routes ran a 2019 Freightliner Cascadia DD15 through a Samsara fleet. At 487,000 miles, the unit generated a SPN 3719 DPF soot load alert — but the alert triggered at a generic threshold, not calibrated to the DD15's specific regen cycle profile. The driver completed a highway regen. Fault cleared. No work order generated.
Four weeks later: SPN 3251 FMI 18 (DPF differential pressure below normal — indicating a cracked substrate, not a loaded filter) combined with SPN 5246 FMI 18 (SCR efficiency below threshold). Parked regen attempted, failed. Unit towed. DPF replaced: $3,100 parts and labor. SCR catalyst inspection found thermal damage from repeated incomplete regens: add $2,800. Total event cost: $5,900 plus 31 hours of asset downtime on a temperature-sensitive lane.
The post-mortem showed SPN 1209 (exhaust back pressure) had been trending upward for six weeks before the failure event. That PID was in the ECM data stream. It was not being captured or alerted on by the platform in use. On a Geotab deployment with ProPlus and the active DD15 PID set configured, SPN 1209 monitoring with a configurable threshold alert is standard. The intervention window was there. The data capture wasn't.
This is the practical cost of telematics diagnostic depth — or the lack of it. For a model of how that downtime compounds against revenue by vehicle class and route type, the fleet downtime cost model provides a structured framework for converting shop hours into financial exposure.
Where the PID Gap Shows Up in Predictive Maintenance
Predictive maintenance isn't pattern recognition on fault codes. It's pattern recognition on parameter trends that precede fault codes. The difference is roughly a 2–6 week intervention window versus a roadside event.
The PIDs most diagnostic for heavy-duty diesel predictive maintenance — and where platform depth diverges most — include:
| SPN | Parameter | Predictive Use Case | |-----|-----------|--------------------| | SPN 94 | Fuel delivery pressure | Lift pump and injector wear trending | | SPN 1600 | EGR valve position | EGR cooler fouling, valve sticking | | SPN 3719 | DPF soot load % | Regen frequency drift, substrate life | | SPN 1209 | Exhaust back pressure | DPF loading independent of modeled estimate | | SPN 3251 | DPF differential pressure | Substrate crack vs. loading differentiation | | SPN 5246 | SCR system efficiency | DEF quality, injector degradation | | SPN 3480 | DEF dosing unit temp | Dosing line freeze risk, injector coking | | SPN 84 | Vehicle speed (wheel-based) | Drivetrain anomaly detection vs. GPS |
Geotab captures and alerts on all eight of these in a properly configured ProPlus deployment on supported OEMs. Samsara captures SPNs 94, 3719, and 84 reliably across most Class 8 applications. Coverage on SPNs 1600, 1209, 3251, and 3480 varies by OEM integration version and firmware — and that variability itself is an operational risk for multi-OEM fleets.
Fault Code Triage: Why Raw Data Isn't Enough
Capturing more PIDs creates its own problem: fault alert fatigue. A 100-truck fleet running ProPlus on Geotab can generate 300–500 fault events per week across the fleet. Without a structured triage framework, that volume overwhelms service writers and dispatchers until they start ignoring alerts entirely.
This is where platform capability and operational process have to work together. Geotab's rules engine allows threshold-based alerting with SPN/FMI-specific escalation logic. Samsara's alert framework is less granular at the SPN level but offers faster mobile dispatch notification. Neither platform solves the triage problem out of the box — that requires a severity tiering framework mapped to your fleet's specific asset types, duty cycles, and SLA requirements. The fault code triage matrix for 50–200 vehicle fleets covers the operational architecture for building that layer, regardless of which telematics platform you're running underneath it.
The data depth question and the triage question are connected. If your platform doesn't surface SPN 1600 EGR valve position drift, you never need to triage it. You just pay for the cooler replacement 40,000 miles earlier than you should have.
Which Platform for Which Fleet Profile
For fleets running primarily Class 6–8 diesel assets from Cummins, Detroit, PACCAR, or Volvo, Geotab's diagnostic depth advantage is real and consequential. The OEM PID library, the configurable multi-PID logging, and the API flexibility for CMMS integration make it the better choice for a predictive maintenance program built on leading indicators rather than fault events.
For mixed fleets with significant light commercial or medium-duty asset populations, or fleets where driver behavior, ELD compliance, and operational visibility are the primary use cases, Samsara's broader OBD-II coverage, faster alerting, and cleaner mobile experience provide genuine operational value that partially offsets the diagnostic depth gap.
For fleets where the maintenance cost driver is emergency versus planned repair ratios — and if your shop is seeing more than 35% emergency events, it is — the diagnostic depth difference translates directly to budget. The real cost gap between emergency and planned repairs by failure type quantifies that delta across drivetrain, emission, and electrical failure categories. The numbers make the platform selection conversation look like a procurement decision, not a technology preference.
The Bottom Line
Geotab captures more diagnostic PIDs for predictive maintenance in heavy-duty diesel applications — specifically on OEM-extended J1939 SPNs covering aftertreatment, EGR, and fuel system parameters that provide 2–6 week leading indicators before active fault events. Samsara is a capable platform with real strengths in mixed-fleet visibility and alert speed, but its OEM-specific PID coverage on Class 8 diesels is materially thinner than Geotab's, and that gap shows up in repair bills, not benchmark reports.
Rouutiq gives fleet maintenance teams exactly this kind of fault pattern visibility — cross-referencing SPN/FMI combinations against asset history, mileage bands, and repair cost data to flag the trends that precede failures, not just report them after the fact. Start a free trial at Rooutiq and see what your current telematics data is and isn't telling you.