Coolant Loss With No Puddle
A 2015 40-foot bus with an 8.9L diesel, 297,000 miles, has used about 1 gallon of coolant every 450 miles for three weeks. There is no puddle under the bus and the pressure cap holds. The surge tank shows a dark oily film and the coolant smells of exhaust. Cold-start exhaust is white and sweet for roughly 90 seconds. Oil analysis shows potassium at 42 ppm and sodium at 58 ppm, up from 8 and 11 ppm on the previous sample. A chemical block test at the surge tank turns yellow-green in under two minutes. DPF regeneration frequency has climbed from once per 300 miles to once per 120 miles. The EGR cooler was replaced 40,000 miles ago with a remanufactured unit. The roster has 12 buses and two spares; one spare is already down for a transmission repair.
1. Interpret the evidence and rank the three most likely failure modes, explaining what pushes each up or down.
Read the worked answer
The evidence says combustion gas is entering the cooling system: positive chemical block test, exhaust smell in the surge tank, oily film, and rising potassium and sodium, which are coolant-marker elements, not normal wear metals.
Ranking:
- Internal EGR cooler leak - most likely. It explains coolant loss with no external puddle, gas in the coolant, and elevated soot loading (regen interval cut by 60 percent) as coolant dilutes combustion and fouls the aftertreatment. It was also a reman unit at 40,000 miles, which is young for a second failure and points at part quality or a coolant chemistry problem upstream.
- Head gasket or cracked head - plausible and must be excluded, because the block test cannot tell you which gas path is leaking. It would usually show more consistent white smoke rather than a 90-second cold-start puff.
- Cavitation-pitted wet liner - least likely but possible at 297,000 miles, especially if the coolant was run out of spec. It typically shows as coolant in the oil or rising oil level, which this bus has not shown.
The regen data is the tiebreaker: it is a downstream symptom that fits coolant in the intake charge better than a head gasket alone.
2. Should the bus stay in service while parts are ordered? What test confirms the diagnosis before teardown, and what is the operational risk of waiting?
Read the worked answer
Pull it from service. Continuing to run pushes coolant through the turbo and aftertreatment, and the regen frequency says that damage is already in progress. A gallon per 450 miles is aggressive loss, not seepage.
Confirmation before teardown, in order:
- Cold and hot cooling-system pressure test at cap rating, held 30 minutes, looking for a drop with no external trace.
- Isolate the EGR cooler: block its coolant supply and return, pressurize the gas side to a low regulated pressure, and watch for a drop or coolant displacement. This is the single test that separates cooler from head.
- Cylinder leak-down with the surge-tank cap off, one cylinder at a time, watching for bubbles. Bubbles on two adjacent cylinders point at the head gasket; bubbles on one cylinder suggest a crack or injector sleeve.
- Borescope the suspect cylinders and inspect liner walls for pitting and steam cleaning.
Operational risk of deferring: the run is covered today, but the failure mode degrades the DPF, the turbo, and possibly the block, turning an EGR cooler job into an engine replacement. That cost exceeds any savings from keeping the coach in the lineup.
3. What PM, inspection, or documentation practices failed here, and how would you correct them?
Read the worked answer
Three gaps stand out:
- Coolant chemistry was never trended. Nitrite, pH, glycol concentration, and conductivity should be tested at every PMI on a wet-liner engine. A coolant that has gone acidic or lost its inhibitor package is a direct path to liner pitting and to repeat EGR cooler failures.
- Oil analysis was collected but not read. Potassium and sodium were recorded every sample; the jump from 8/11 to 42/58 ppm is a five-fold change that should have triggered a cooling-system review at the previous PM, not three weeks into a known loss.
- Top-off volumes were not documented. Somebody added coolant for weeks without recording quantity or date. A simple fluid-consumption log with a threshold - for example any top-off over 1 quart between PMs requires a written note - turns a slow leak into a visible trend.
Also add regen frequency to the PMI trend sheet. It is a free early indicator of coolant, fuel, or oil entering the exhaust stream, and here it moved before the block test did. Finally, document the reman EGR cooler part number and supplier so a pattern of repeat failures can be found across the fleet.