Slow Crank on a 24 V Transit Bus After an Overnight Sit
2017 40-foot transit bus, 24 V system: two 12 V Group 31 batteries in series, 24 V starter, 200 A alternator. Complaint: cranks slowly and occasionally fails to start after sitting nine hours overnight at 6 °C; starts fine warm or with a jump. Rest open-circuit voltage reads 12.62 V front battery, 12.58 V rear (25.20 V total). During a 10-second crank: 21.4 V at the battery posts, 620 A on the clamp, 16.9 V at the starter B+ stud, and a 4.4 V drop from battery positive post to starter B+ stud. Engine block to battery negative post drop: 0.15 V. Terminals are clean and tight, electrolyte at the split ring. One spare set of batteries is on the shelf and the bus is due out in three hours.
1. Is this a battery failure, a cable/connection failure, or both? State exactly what the numbers prove and what they do not.
Read the worked answer
Do not condemn the batteries yet.
- Cranking loop: 21.4 V at the posts minus 16.9 V at the starter leaves about 4.7 V unaccounted for. The 4.4 V positive-leg drop plus 0.15 V negative-leg drop accounts for essentially all of it, so the loss is in the positive cable path between the post and the starter stud.
- 4.4 V ÷ 620 A ≈ 0.0071 Ω, roughly 2.7 kW dissipated as heat in a cable that should drop well under 0.5 V total on a 24 V cranking circuit (about 0.2 V per connection). That is a hard fault, not a marginal one.
What the numbers do not prove: battery state of health, and starter condition. At 620 A the batteries are discharging through a resistor pack rather than a starter, and some of that 620 A may be excess starter draw caused by low terminal voltage. Neither the batteries nor the starter can be judged until the cable drop is corrected and the crank test is repeated with the same clamp position.
2. You have three hours and one spare set of batteries. What do you do first, and where do you put the meter leads?
Read the worked answer
Fix the cable first. Swapping batteries consumes the only spare set and leaves a 4.4 V series resistor in the cranking circuit.
- Record baseline state of charge and a crank sample before touching anything.
- Static resistance checks are screening only - a 0.007 Ω fault will not read reliably on a hand-held meter, so do not accept a "good" ohm reading as proof.
- Voltage-drop test under load, one segment at a time, meter probes on the cable ends rather than on the studs: battery positive post → master disconnect switch in → switch out → starter solenoid stud → starter B+ stud. Probe each crimp (lug to conductor) as well.
- Sweep the disconnect switch and any megafuse or breaker in that path. A pitted switch contact is a common culprit and is usually warm right after a crank.
- Replace or re-terminate the offending cable assembly; do not solder over corrosion.
- Repeat the 10-second crank test and record voltage at the posts and at the starter stud.
3. After repair, total crank drop is 0.4 V and battery voltage is 20.9 V at 540 A, but the engine still cranks noticeably slower than a sister bus. What do you evaluate next, and what data do you need?
Read the worked answer
The electrical path is now in spec (0.4 V total drop), so the remaining difference lies in the starter, the batteries' true condition, or mechanical drag.
Gather before wrenching:
- Starter specification: locked-rotor and free-run current, and cranking rpm.
- The same 10-second crank test performed on the sister bus, same temperature, same clamp position.
- Battery conductance test results and date codes for both batteries.
- Cranking rpm from the ECM (most transit ECMs broadcast engine speed at 1 rpm resolution during crank).
Then reason from the comparison, not from the parts catalogue: 540 A is lower than the earlier 620 A, which is consistent with less starter drag or a cooler engine, so the difference may be normal. If cranking rpm is genuinely low, check starter brushes and bearings, a bent or binding drive, starter case ground, cable gauge if a lighter replacement cable was installed, and mechanical causes such as accessory drag or cold ambient. Replace a component only when the comparison data points at it.