The case of the exhausting Crafter – By Gareth Davies CAE AMIMI, Euro Performance.
Life in the workshop can be quite the mixed bag sometimes, even when focusing on just a handful of brands. I applaud you technicians and business owners alike that welcome all brands through your door. Life at a specialist can be simpler at times, brand fault repetition and all that, and it can mean the bar has in some ways been set higher, because all manner of folk have been ‘at it’ before you. I had a surprisingly more common than not fault presented just recently in the workshop on a fleet vehicle we look after, but the direction of detection could have really taken any number of paths, one with the right conclusion, others leading to a seemingly positive, but short- lived conclusion.
My role within the workshop is very much a blend between workshop controller and senior master technician. I take an active interest (but know when I am not wanted) in most jobs that pass through the workshop. I attribute the interest to acquiring insights and data, but it can often be considered being a nosey bugger. I was recently approached by a fellow technician for some advice on a VW Crafter fleet vehicle we had in for various service and repair tasks, but also had a collection of warning lights on the dashboard. The vehicle presented with an engine management light, flashing glow plug light, and a particulate filter picture in the main driver information display.
The technician had checked fault codes in engine, only one was present, P2463 DPF excessive soot accumulation too high, and had rightfully posed questions on the repair strategy. I saw a prime opportunity to put two heads together and see what conclusions we could come up with between us. We were able to retrieve some key data PIDs in the first instance, using VCDS as our scan tool on this occasion. We ascertained that the measured and calculated soot loadings were 16g and 60g respectively. I saw this as a great opportunity for some on the job or ‘just in time’ training. We took a step back and pondered ‘plausibility’. I like this word a lot, for the sole reason that I find when attacking diagnostic work it really makes me stop and think for a minute. How complicated is this job really? Whether we are testing a simple circuit or sensor, or 26 pages of Can Bus comms faults, plausibility is a great point to stop and have a think.
Garbage in, garbage out…
I verbalised the question to the technician, how plausible is the overloaded/blocked DPF the vehicle reports? “I’m not sure, must be quite blocked I think given the calculated reading is so high?” It’s a reasonable assumption to make, but why not prove/disprove what the engine ECU thinks? After all, year 10 school IT taught me GIGO (Garbage In, Garbage Out) – Mr Griffiths I hope you are reading this! So our first test has to be data sampling and validation. We were able to see what the engine ECU was seeing, but this is fed from one of the myriad of sensors monitoring various states on the engine or associated sub sections.
In this instance, we needed to establish some data baselines and then create a test plan, which should lead to a hypothesis and, potentially, some conclusions. A little like Year 8 English (all good stories have a beginning, middle and end). So picking a selection of plausible data PIDs to examine were grouped under exhaust. We selected a wide array: exhaust temperatures, exhaust pressures and soot loading. We found some interesting detail when carrying out some tests in two different states. Key On Engine Off, KOEO, and Key On Engine Running, KOER.
Very quickly, without unnecessary testing, we highlighted an area of concern. DPF pressure had a significant positive ‘implausible’ reading across the exhaust when no combustion was taking place. The technician smirked, ‘that’s not right’. So how do we validate what’s right and wrong here then, is there a secret pressure in the exhaust from the particulate fairies, is the reading right, what could be causing this – the sensor, the wiring, the ecu?
Having established an area of concern, we now needed to establish how to test the sensor providing the data PID in question, that may then help eliminate said concern, or lead to further tests/conclusions.
For the purpose of the on-the-job training, we took to Alldata to do some light reading. This is a useful and pertinent step whatever data source you wish to use. Whether you’re fresh out of your apprenticeship armed with your ‘first multimeter’ thinking this is the excuse I needed to unwrap the packaging, or a seasoned diagnostician looking for your next electrical ‘wonder’ find, we don’t know it all. The reason we don’t, is the game keeps changing. Is it 12v supply, 5v supply, 3.3v supply etc.
Five minutes in front of the PC proved really useful for our next steps. The sensor is a 3 wire sensor and is made up of a 5v feed, a ground, and a signal wire. In component function/description it details the importance of the job of the sensor feeding back a signal voltage, that translates as a pressure. The pressure reading then equates to a significant factor of soot loading (how full the DPF may be of soot). Do we need to be DPF chiefs for the purpose of this test? Not really. We know what two values should be definitively, and the detail then explains that the signal voltage should be between 0.5v and 4.5v, which is indicative of pressure difference measurement.
Armed with all we needed to know and a multi meter we did some raw data acquisition. We were able to quickly conclude that the 5v feed was present and correct, the ground was good, see Figure 1, and we validated these readings via some ‘known goods’ elsewhere on the vehicle, just for good measure i.e., validating the ground to sensor via nominal battery voltage, and similarly for the 5v feed to alternative grounds.
For the signal wire evaluation we checked, and with KOEO, we had just under 1.8v, see Figure 2. Without knowing scales (or needing to know at this point), how plausible is this reading?
If we think of the scaling of the signal thresholds, not very. On this application we knew that voltage increase was dictated by pressure difference increase. There was definitely no pressure present here, the engine wasn’t running.
We discussed at this point what to do next, and my suggestion was to measure the pressure ourselves. Thinking ahead of other tests that may be required with this line of inquiry, we got the Mityvac ready (vacuum/pressure tester). The pressure gauge separates from the vacuum tester on this particular model, so with two pipes pulled from the pressure sensor we decided to measure the pressure ourselves without the pipes connected to the sensor (we have become the sensor). This sensor is a pressure difference type with two pipes, at two points of measurement on the exhaust, one before the main particulate monolith and one post monolith. Surprise surprise, with KOEO we found zero pressure registered at both points (thankfully no fairies this time!) The next test proved very useful. KOER we measure pressure pre DPF and post. Virtually nothing at all (0.5PSI). We need no scale to realise that this DPF is unrestricted. Looking positive now.
Using known ‘goods’ as reference…
The last test before condemnation of the sensor is substitution values, or simulation. I think this is a good test that will usually give you as the technician great reassurance for your evidence and conviction you are carrying out the right repair. In this instance we used a voltage simulator (Ditex) and proceeded to take control of the signal wire. We did this by using some break out leads from the sensor to the harness connector, leaving the signal wire disconnected from the sensor and to be controlled by us. Now you can also do this by using a decade box, either will give you similar results.
A known good for this sensor is around 1v at rest. We proceeded to give the sensor 1v and had a look at our live data whilst doing so with KOEO. What we noticed is we now had virtually no pressure (1 hPa) across the filter. Of further interest, when we started the engine and kept the signal at 1V (simulating the verified low pressure difference we had just measured ourselves) and gently picked the engine idle up to around 1100rpm, is that the measured soot mass began to fall quite quickly. This test proved that if we replace and adapt the sensor correctly, we have no wiring defect, no engine control processing defect, and that accurate measurement of the load across the particulate filter reflects the system can and will begin to work properly again.
We subsequently recommended a replacement particulate filter pressure sensor and service regeneration of the particulate filter for repair. Duly authorised, the sensor was replaced and taught in (adapted). This is quite a key step, as the adaption value stored for the vehicle was 8 hPa for pressure difference. This changed to 2 hPa after adaption. Had we left this un-adapted, it may have further skewed our repair process and potentially, the longevity of the repair. Interestingly, this sensor has been superseded for one reason or another to an updated part number.
In the pursuit of getting a little more understanding and data acquisition, before finalising the repair we carried out a further quick test to understand the behaviour of the sensors, new and old, for future reference. If working on brand volume this can be particularly useful for you for future cases or suspicions. We connected the old sensor to the vehicle harness, and with ignition on and our meter connected to the signal wire (back probing), we looked at what happened to the voltage when we applied a vacuum and a pressure using our vacuum tester. We could see a positive voltage rise when applying a pressure, also mimicked by the new. But the baseline readings of voltage from old to new was 1.7v and 1.0v respectively, see Figure 3. We could now be happy this was our issue.
The last job to carry out was the DPF regeneration. Now this was thwarted somewhat by the stored calculated soot mass of 60g, with the threshold of 45g maximum value for regeneration well surpassed. On this occasion, given the low pressure across the DPF being a reasonable indication of no residual blockage or fault, we adapted this down to 40g for the purpose of carrying out a regeneration. The key here is we did not reset it (as though the filter had been replaced) and this would have reset both soot, ash and a whole load of other metrics that need to remain as they were. It allowed us to carry out the service regeneration, which took around 40 minutes, and continuing data evaluation showed good temperature change across the sections of exhaust, no super hot or cold spots. The van drove well post repair, and happy to report to date no return of the warning lights.
Troubleshooting faults can seem complicated and a minefield at times, but I very often see case studies from SimplyDiag.net members, and Steve Scott himself, that stress the importance of breaking everything down into as simple terms and steps as possible.
In writing this article, it may seem like a long process but in reality the whole process of diagnosis took some 45 minutes in total. This included doing our ‘reading up’ and bouncing ideas about the various theories and outcomes the diagnostic process can lead you to. Some background information is key when carrying out DPF analysis.
This vehicle is what we would consider a known quantity. It does 30k miles a year, its serviced regularly using quality parts and lubricants, and spends most of its time on the motorway, never running around with just fumes leftin the tank. Furthermore, we know this vehicle has not previously been looked at or tampered with elsewhere.
All of this is conducive (when no faults are present) to a long and healthy self-maintaining particulate filter lifespan.