308 FAP/DPF fault -- Peugeot rip off??

R

robs

Active member
VIP Paid Subscriber
Fellow Frogger
Joined
Nov 1, 2010
Messages
550
Location
Blue Mountains, NSW
This video makes a pretty serious accusation against Peugeot (and other manufacturers, not named):


In brief it says that:
  1. at 120,000 miles a bogus fault comes up saying the DPF needs replacing
  2. the fault mode prevents further DPF regens
  3. car is quite driveable in this fault state so owners don't *rush* to the dealer
  4. by the time they get to the dealer the DPF is badly choked and needs replacement
  5. cost of repair is uneconomic. Car scrapped.
He claims to have dealt with this problem in "hundreds" of cars (surely an exaggeration).

I stick to RWD Pugs, so don't have a horse in the race, but if it's true, this is surely an order of magnitude worse (ethically and for the environment) than VW's "cheating" on emissions tests.

Have fun,

Rob.
 
I drive Japanese diesel trucks that have the DPF system and it does not work well in metro running. Often belching out smoke from a failed burn off When this idea was first introduced we were told that reconditioned DPF units would be available but that didnt happen.
 
P1445 is maximum ash in the particle filter. So clean it. Filters are cleanable and the transport industry has many service businesses available for this job.

It should never have reached that state if properly used. The system is designed to self regenerate at intervals, given the chance. (Filtered vehicles need periodic open road steady running. The motorways around outer suburban Sydney will do.)

The distance, converted, is 193,000 km. We have members here with HDIs that have travelled that distance.
 
Last edited:
The presenter was claiming that the warning was bogus and the filter was ok. He showed the live pressure readings across the filter and said they weren't excessive. I don't know the ins and outs of it. Maybe it was all set up to boost views, but it seemed like fair-enough reasoning and worth posting here. Owners of the model might be interested, and dealers/repairers might like to know the scurrilous claims getting around on the net.

Have fun,

Rob.
 
robs said:
stick to RWD Pugs, so don't have a horse in the race, but if it's true, this is surely an order of magnitude worse (ethically and for the environment) than VW's "cheating" on emissions tests
Click to expand...
You are on the pipe I guess that is so wrong educate yourself
 
I guess they do not last forever and eventually there will be a build up of material that can't ever be burnt off by a regeneration. P1445 would be in response to the amount of additive injected since the filter was new and advising it should be replaced. While it might appear, I don't think it actually causes any performance downgrade. Is it any different to some countries requiring periodic replacement of the oxygen sensor to manage emissions? What's the worst case here? Tell the car it has been replaced (it's a 'dumb' component) and reset everything and drive away and force a regeneration? You will find YT videos of people reverse flushing them with a pressure washer and that will eject ash that will not burn up. (Not a backyard DIY job though due to the wastewater.) However, one can't really do this with the later versions where the DPF is not a separate item but instead the catalyst and DPF are sandwiched together in a single assembly.
 
dmccurtayne,
Perhaps you need a blood-pressure check? In any case, please don't burst something on my account, it's really not that important.

David S,
The video was pretty clear in claiming it was a deliberate time-bomb in the firmware. As you say, the car remains driveable after the alert, but the DPF clearing stops from that point until you get the "dumb component replaced" box ticked. If you take too long about it, you end up with a completely choked DPF and (he says) scrapping the car due to uneconomic repair cost.

The time-bomb is an interesting claim. It could be proved if someone finds offending code lurking in the firmware, but it's pretty much impossible to disprove. As seasink points out, plenty of cars have got past that mileage. Ah, but the sneaky code only catches one car in ten, or it's only enabled in Ireland, or it happens if you hit the magic mileage on a Wednesday, or...

More interesting is thinking how such code might get there. It seems *extremely* unlikely Peugeot would specify it. News would soon leak out and, in any case, what benefit would it be to them? These ECUs are probably generic and Peugeot just gives their requirements to a standard supplier. Is there something lurking in a generic ECU library somewhere? That might explain the similar faults the guy reports in other manufacturers' cars. And is it there as a bug, or something deliberate by a disgruntled programmer, or... ?

Or is is just a guy on YouTube hoping for clicks? Beats me. It's a complicated world.

Have fun,

Rob.
 
A great deal has been done to get the PSA and engine ECU codes. The reading tools have existed for Linux for ages, and some are showing up for Windows.

The Car Hackers' Handbook (over 400 pp) was published roughly a decade ago, and is still a good starting read. There is a quick overview sample at

Tools such as these are commonplace in Linux boxes, even VW contributed, now installed by default in some cases:
github.com

GitHub - linux-can/can-utils: Linux-CAN / SocketCAN user space applications

Linux-CAN / SocketCAN user space applications. Contribute to linux-can/can-utils development by creating an account on GitHub.
github.com github.com
github.com

GitHub - barracuda-fsh/pyobd: open source obd2 car diagnostics program - reuploaded

open source obd2 car diagnostics program - reuploaded - GitHub - barracuda-fsh/pyobd: open source obd2 car diagnostics program - reuploaded
github.com github.com
vcan (a virtual CAN)
etc

This led to PSA-COM a commercial alternative to Diagbox (though not all the way yet).

There's a Win tool collection at http://obdtester.com/downloads

So what comes from this? Just that if any of the many PSA investigators hits something as scandalous as described in the YT, it would be all over the internet, and not just coding sites. It isn't. So I'm sceptical.
 
robs said:
... Or is is just a guy on YouTube hoping for clicks? Beats me. It's a complicated world.
Click to expand...
It certainly gives him something to talk about and maybe brings some clicks and work his way. No idea about his motivation really, but it's making it out to be a bigger problem than it is.

If the fault appears, and it's not an Ireland-only code, it's because the manufacturer predicts that the particle filter has reached the end of it's useful life. They have an obligation to ensure ongoing emissions performance. This is a high mileage issue, but it appears to relate to the cumulative amount of Eolys injected. It's rather like the predictive oil wear counters and inferred remaining Eolys quantity (there is no level sensor). I would guess it will show up at a lower mileage on a vehicle used for many short trips and with a part filled fuel tank. As I suggested, I can't imagine it's a problem for any competent tech to work around if the filter is not actually containing a large amount of permanently stuck material that cannot be burnt in a regeneration. Unlike the particle filter systems in some other makes, the PSA system actually works quite well except perhaps for owners who only ever drive their car very short distances at low speed - e.g. station duty. Keep in mind that AdBlue is a separate and additional technology used for diesels and some people will confuse the two or think that an AdBlue car doesn't have the particle filter.
 
Last edited:
seasink,
Thanks for the link to the Hackers' Handbook sample. Not planning on reversing any CAN packets, but it was an interesting skim. I'd go a lot deeper if I was younger and had brain-space to spare.

Have done a small amount of CAN programming to get my LCD dashboard working with the Megasquirt CAN protocol. What I learnt in that effort was that the CAN packet structure is pretty "flexible" (i.e. not standardised) and the bit rate is also "flexible" -- Megasquirt fixed on 500kHz. The HH'book only mentioned the packet flexibility. Do hackers not need to worry about speed?

Anyhow, back to the DPF fault: I'm sceptical too. All the same I know it's technically possible and I don't think it would necessarily be spottable in CAN packets.

A somewhat related story: my brother's Subaru Forester had a fault code "failed secondary air pump". This pump only runs for a few seconds after startup, I presume it's to help burn unburnt fuel in the exhaust. Code was correct: the pump had seized up. To encourage you to get this *meh* problem fixed. Subaru's firmware disabled cruise control as long as the fault persisted. My suspicion is that Subaru added this "punishment" behaviour to satisfy a regulator. So that's another way strange behaviour might lurk in the code.


David S,
I would read the moral of his story that if (I suppose his claim is *when*) this fault code comes up, *don't dally*. It's no biggy if you get right onto it, get the thing cleaned and the filter marked as replaced. For people who follow this path, he's overstating the problem. However, people who opted to sit on it because "the car's going fine", might have a different view when it has become too expensive to repair.

Have fun,

Rob.
 
The whole book is worth a read although it's getting dated. You have to start somewhere. I own a copy. Eons ago we used to pore through core dumps when things went belly up. Hacking cars brings it all back! can-utils has 30 or so utilities in the package which handle low level CAN stuff. They are listed at the page linked.

I know of one regulator enforced punishment. If Adblue runs out on newer HDI models your choices are fix it soon or no go. The EU knows very well that motorists don't want the complications.
 
On the PSA models, the Adblue 'countdown' usually starts at over 2,000km, but the documentation mentions the distance to no start is actually decreased by units of 50km with each vehicle network awakening. So the actual distance to the repair could be much less than expected from when the problem begins if you make a series of short trips. Just something to keep in mind if you enter the countdown.
 
seasink,
On top of my dwindling faculties, a lack of a modern car to tinker with makes CAN hacking a bit academic for me. Fully with you on the nostalgia trip, though what we have in today's microcontrollers is really a lot better than the processors we worked with in the late-'80s: similar performance, but with facilities like hardware breakpoints, flash memory, etc. Fun for programming. Fun with the soldering iron too. Golden age, etc.

On the original topic, it must make for interesting meetings when they're deciding what to sabotage. Subaru asks the question "secondary air pump failure" and comes back with the answer "disable cruise control". Of course! And for the 308, the meeting where they asked "120000 mile mandatory DPF check" (or whatever), "passive alert at startup" seems to have been a mistake. What *should* they have gone for? Disable higher gears? Constant irritating squeal? Move driver's seat to full forward position? Douglas Adams might have some ideas (if he weren't dead).

Have fun,

Rob.
 
Adams would have loved cars that start without a metal key. Zaphod would have a theft device ready in minutes.
 
You must log in or register to reply here.
Top