Killing a Diesel - a detailed how-to (much info on commonrail technology)
 

Bbjsw10's Kill switch how-to thread got a little hijacked by details, so I'm starting a new thread to cover diesel engines and how to make them stop.

I recently found some detailed info on the workings of my engine and hopefully it will be enough to make a safe kill switch. I also hope that this will help others in their quest for understanding the Zen of Diesel.

Quick intro:
My engine is a commonrail turbodiesel. It is a second generation version of PSA's (Citroen/Peugeot) HDi direct injection technology. The first generation were the 1.4, 2.0 and 2.2 liter 8v engines, the second are the 1.6 and 2.0 liter 16v. The 1.4 and 1.6 engines can also be found in Fords as the TCDi Duratorq, and also in Volvos, Suzukis, Renaults, Fiats, etc.
Older diesels had a pump for each injector, while in commonrail technology there is one pump responsible for creating very high fuel pressure in a reservoir used by all four injectors. This, along with direct injection, gives a much better air-fuel mixture, which increases engine efficiency, power by around 25% and low end torque by 50%, reduces fuel consumption by around 20%, and reduces emissions. Of course, this is made possible by heavy use of electronics (calculators with myriads of sensors). This may help in finding a safe way to stop an running diesel engine, but may be very easy to break something (like something very small and very expensive).

After finding out that neither disconnecting the cam position sensor, nor the crankshaft sensor stops the engine (doesn't even throw a code), dcb proposed unpowering the fuel pump. HDi engines have two fuel pumps, the first pumps low pressure fuel from the tank, the second creates high pressure (200-1350 bar) for the injectors. Turning the first one off may just starve the second one, causing the pressure at the injectors to get lower and lower, possibly messing up emissions. On the other hand, the high pressure pump is powered off the camshaft, so it can't be easily turned off. It does, however, have a switch for disabling one of its three cylinders. This is used to reduce the pressure when there is no load on the engine (in normal operaton the high pressure fuel pump uses up to 3.5kW of power!). The pump uses the third cylinder either when load is between 66%-100% of max or when idling. I guess that deactivating this third cylinder during idle won't turn the engine off, but would it help use less fuel? I need that SGII!

So, next I tried to find out when the ECU cuts the injectors. This happens either when engine speed is above 5300rpm, or when the gas pedal is in "0" position when in gear. In the second case, fuel is sent to injectors when engine speed drops below 2200rpm. Hmmm, could I trick the engine into cutting the fuel supply to the injectors by sending it a sygnal to make it think that engine speed is above 5300rpm? Somehow, this doesn't seem like a good idea. But maybe I could tweek it so that fuel is cut off down to about 1200rpm, instead of 2200rpm (I hardly ever go that high, even when engine braking)? I really need that SGII!

Here is what happens when ignition is turned off (key turned):

1. Fuel pressure regulator voltage = 0 = minimum RCO (open cycle ratio?)
2. Fuel pump (low pressure) power off
3. Stop engine thru injector cut off
4. Turn off injector calculator

Every time the engine is stopped, these four events are done in a different order to allow the ECU to run a diagnostic.

There is also an emergency engine shut down procedure, which happens when the ECU finds an error in one of the following:

• EPROM in the injector calculator
• Engine speed sensor
• Cam position sensor
• Fuel pressure sensor
• Injector error (1 to 4)

So, all of this information, and what to do with it? I have a few ideas on how to kill the engine, but I'm not sure whether they would harm the engine, harm the ECU, or at least make it throw a code and go into limp mode until I pay $$$ at a service station to turn it off. How about:

1. Kill the lower pressure fuel pump and see what happens
2. Hack the fuel pressure sensor to tell the ECU the pressure is too low/high
3. Trick the ECU into thinking the engine speed is such that the fuel should be cut to the injectors
4. Cut sygnal to one of the injectors and see if the ECU sends an emergency shut down (rebuild engine if not)

Any safer ideas? As I mentioned earlier, the engine doesn't react to unplugging neither the cam, nor the crankshaft (engine speed) sensors, though I didn't pull both at the same time, so maybe the ECU is able to get its info from only the one that is available? Or maybe instead of unplugging it, the sygnal lead should be grounded to mass? I'd like input from anyone who's tried, even those dumb hall sensors cost a fortune.

Good info there, useful for people with newer TDIs in NA.

I'd try 1 and 4 at idle to see what happens. 2 might get the ECU to tell the pump to adjust pressure and bad things could happen if there's not a second feedback loop.

Even older diesels, 1997-2003 TDIs in NA, had a single rotary pump, in which case you can just short the wires going to the fuel quantity adjuster to have the ECU shut the pump off.

Ok, I see the confusion, I did not mean the fuel pump, I was trying to reiterate my previous post about the fuel cutoff valve, it shuts down the fuel leaving the distributor nearly instantly (I think, sorry haven't tested it yet), it just happens to be housed in the mechanical pump:
http://ecomodder.com/forum/showthrea...tml#post143293

"my diesel injector pump has a solenoid on it to turn on the fuel, I was planning on interrupting that for a diesel kill switch"

Why do you want to kill the engine?
 

What is that kill for?

Just a matter of fact...

I am sure you have a reason in mind ... security, may be?

OldBeaver

Ok, de-energizing the fuel shutoff valve works fine for me (2001 TDI). That's my plan.

http://ecomodder.com/forum/attachmen...1&d=1260125605

I turned off the "ignition", and clipped the lead to the solenoid (don't ground anything out!!) so I could see what happens when I disconnect it cleanly while the car is running.

http://ecomodder.com/forum/attachmen...1&d=1260125605

The engine shuts right down, as expected.

The solenoid is a somewhat hefty circuit though so I will rig up a microswitch on the shifter and control a NC relay with it.

Stopping a diesel with a "classic" distributor pump is quite straightforward as DCB shows. But the common-rail engines are quite different as Pivoslaw shows. Why not use the ignition switch connections?!

PS: I take it you've seen the description of this simple engine at http://www.christiantena.pwp.blueyon...operation.html

:)

I forget commonrail is a completely different animal, is there a better schematic available?

Bouncing the whole cpu sounds like it might reduce the life expectancy of the cpu, so interrupting the injectors is probably the most thorough, least likely to break something else approach.

we mess w/cam or crank sensors on the gas cars because it is an easy way to kill both injection and ignition, with a switch that only needs to handle signal currents.

I don't know if the ign switch would simply kill power to the ECU? Seems a bit brutal!

Although the hdi may seem complicated it's mechanically very simple and akin to petrol injection. Albeit with very high pressures and direct injection from 4 injectors. The link above gives about the best information I've found, with some useful schematics. (If you navigate back to the parent pages there's also some VW TDI stuff).

If the ECU "thinks" rail pressure is below 120 bar then it won't allow injection to start. Try grounding that signal and see if it also forces a stop? Might result in a fault code (but in my experience the light goes out after 200 miles!) :)

I believe that I've found something very similar, just translated into Polish. The figures and tables look the same. Lots of info. Add to that electrical diagrams and I'd be set, except that my engine is too new: All the materials I've found are for the older 2.0 and 2.2 liter HDi diesels, nothing so far for the 1.6. Of course, most of the info is the same, it's just that some of the sensors aren't in the same places, etc.

I looked into my car's manual and found the circuit breaker that is responsible for the fuel pump. So to test fuel cutoff, I pulled the breaker and tried to start the engine. It shouldn't have, but it started. This means that either my engine not only doesn't need the cam and crankshaft sensors for operation, but it doesn't even need fuel!! (How's that for hypermiling?) The second hypothesis is that the breaker is not for the fuel pump, or anything vital. In fact, I'm starting to believe that the box of circuit breakers is just a dummy, since a month ago I found out that pulling the ignition switch breaker doesn't disable it at all. I can the turn the key and all is as before. Kind of makes me wonder...

I'd look for the cable going to the fuel pump in the tank, but everything in my car is so well hidden, and so hard to get to even if I happen to find it, plus it's getting cold outside. Brrr. Too bad this didn't spring up in the summer, since in September I removed all the seats and carpets for a week as the car was drying. Fooling around with the cables would have been so much easier then.

I'm don't want to do an ignition switch cutoff, since then the computer, radio, etc. all turn off and take 5-10 seconds to restart. I want to keep the CPU on when the engine dies. I'll look for the fuel pressure sensor when it gets warmer. I might have to wait until April:rolleyes:

I navigated up to a different chapter, High pressure pump. There is a table there labeling all 88 channels of the ECU, and I noticed that channel 36 is called "Engine immobiliser serial line". I wonder what that is about? Maybe a signal on that line will kill the engine?
There is also channel 66 "Input:wake up for injection ECU, wake up for ADC ". Maybe grounding this would stop the injection calculator? On the other hand, I'd rather not mess with that, as the injector signals are 50-80V and screwing something up would be very bad.

All diesels, even the old ones, are direct injection.

Although a better fuel/air mix does help with common-rail injection systems, the main advantage comes from ability to fine tune the injection event. Old diesels just injected a slug of diesel into the cylinder at the appropriate time (giving that traditional diesel bang). With high pressure common-rail systems the injection period can be spread out over time - slowly at the beginning of the stroke and then pouring it on during the peak power stroke and then tapering off at the end. This wrings every bit of available power out of the stroke, resulting in more power, more torque, more efficiency, and most notably - less noise.

According to Wikipedia (not that it's the only source of info...)

Common Rail to an old diesel .... ¿Possible?
 

Very interesting...

What if one purchase a Bosch Common Rail system complete with high presure pump, electronic control and install it in an old diesel (taking out the mechanic injection pump)?

Does a new diesel engine have special characteristics as to tolerate more pressure, more efforts, etc., from a CRDI, compared to an old diesel?

Would like to hear your comment...

Tks,

OldBeaver

Wikipedia often leaves a lot to be desired in accuracy. If you look up direct injection diesel on Wiki you get one article that says engines after 1980 were not direct injection, while the Wiki article you found says that engines before 1980 were not direct injection. The fact is, from the very first engine Rudolph Diesel built, diesels have been direct injection. Direct injection of fuel into hot compressed air in the cylinder is the main principle on which diesel engines are based.

Wrong, I believe you are mixing up what direct and indirect mean in terms of a diesel versus a gasser for example.

In a diesel Indirect injection occurs in a prechamber and was common practice on most smaller auto motors up until very recently, my 89 6.2 diesel suburban is not true direct injection for example.

On Gas motors indirect injection normally meant throttle body injection which is impossible to implement on any automotive size diesel motors that I know of and likely that is what you are thinking of, it simply can't work that way on a diesel only gas. But never the less a diesel can be indirect injection and most automotive diesels up until very recently where indirect as in having a prechamber.

Cheers
Ryan

Absolutely right. Indirect refers to the use of a precombustion chamber. This was needed to start the combustion process in a mild manner, to avoid what would have been intolerable NVH levels in a passenger car. The first VW and Peugeot/Citroen engines from the late 1970s onwards were good examples. These engines had distributor pumps and around 175bar injection pressures.

When higher injection pressures, then better injection control, became possible, true direct diesel injection was feasible and the VW TDi was born. A bowl in the top of the piston helps swirl the mix. Early TDi still used the ditributor pump. Peugeot/Citroen and Bosch/Siemens put their efforts into the HDI, commonrail, system, which took a long time to develop. Commonrail originally had up to 1300 bar available and a small number of injection events but I think may have up to 11 shots per cycle now. HDi was released in 1999 and enabled an incredible advance in every area- as featured in Pivoslaw's car, the 307. VW meanwhile went the PD route to get high pressures - for a while. That got them 2000 bar. Now VW group have ditched PD for commonrail - for lower emissions, I understand.

How do I know all this? Been driving diesels since 1990 and just swapped a HDi engined car for a noisier PD-engined one!

Please excuse my ignorance of lots of other diesel car makers - my slant on it is that VW and Peugeot/Citroen (or Bosch, Siemens and Delphi) led the way and others followed. For example, Ford have teamed up with Peugeot/Citroen for their commonrail diesels. (You really didn't want to drive a Ford-engined Ford diesel in the 1990s!)

Simon

=

Quick update: Earlier I wrote that I unplugged the crankshaft position sensor and nothing happened. Well, it turned out that that wasn't the sensor I thought it was. (In fact, I still don't know what I was unplugging earlier.) After going over hard to find engine wiring diagrams I finally found the crankcase sensor, and it does kill the engine (and throws a code if I disconnect it while the car is moving). Unplugging only the camshaft sensor still doesn't kill the engine, only prevents it from starting - the ECU synchronizes it with the crankshaft sensor at start up, so if it gets unplugged later the crankshaft sensor takes over.

On my gasser the injectors have a common +12v-line and all their separate ground wires go to the ECU that connects each wire to ground when it wants fuel to be injected through that injector. If your injectors have a similar setup, you could kill the engine by simply interrupting the common +12v-line.

In certain engines (HDi's for example) both leads of each injector go to the ECU. This is because the injectors double as induction coils for increasing voltage to the required 50V-80V.

My (extremely old skool) Isuzu engined diesel has an engine stop solenoid which is how it stops the engine when you turn the key off. I was going to energise (or de, Ive got to check the wiring) this via a switch. Should be easier for people with older technology like me :)

Do not kill your lift pump. Starving your injection pump, especially such a high pressure pump, will cause cavitation. I work in an industry that uses high pressure pumps and I’ve seen the damage of cavitation. It essentially pulls the metal right off the inside of the pump.

Older diesels used a valve to cut air in and emergency. No diesel will run without air. Have you considered an actuated swing-out, or butterfly valve before turbo?

I’m a newbie, so I’m a little naive. What’s the purpose of cutting the engine like you are trying to do; to keep rolling at speed with the engine killed? Why not just turn the ignition off? It works for me.

If you can do something like have a microswitch on your shifter to kill the engine, then it is fluidly becomes part of your driving and does not kill the speedo and everything else (or risk locking the steering column, though not a huge risk). Also some of us have very old ignition switches that don't need the aggravation.

Don't kill the lift pump !
Having a lift pump go out tends to wipe out the main pump. That's $2000 last time I checked.
I would cut power to all 4 injectors. Its safer than the other options.

Gotchya!

What about a high pressure 3-way valve in-line with the common rail? Your switch would send supply fuel to return AND kill the injectors. That way you don't create a high temp condition from the IP running and damage to the injectors from dry-firing.