ayukawa

I was stuck in a similar situation for a day or two. Car would crank but not even try to start. Just spin and spin. I had not removed the injectors so I didn't have to worry about that part. I tried with very limited success to see if the injectors were being pulsed. You would think that measuring a 70V pulse would be easy (piezo-electric?) The closest I got to a signal was putting an old school multimeter on the connector. I got the needle to wiggle when cranking at about the correct period. I suspect they really should be tested when connected (i.e. under electrical load). If you don't have a signal to the injectors ... you then need to work through all the prerequisites: 1) crank sensor (very inaccessible; never checked mine) 2) virtual camshaft sensor (reason for doing the spray into the intake manifold) 3) fuel pressure (?) - I went as far as checking the signal at the fuel pressure sensor on the rail (far end from pump). You would have to have very shitty compression on all 3 cylinders to not have the engine sound like it is trying to start. Too much electrical junk in these modern diesels.

I agree. Doesn't sound like it is firing/injecting. Have you confirmed that you have fuel pressure in the common rail? Just crack one of the fittings, ideally using a flare nut wrench. Takes a bunch of possibilities out of the equation. As I mentioned, I was in a seemingly similar situation as you. After a bunch of messing around, including going down the path of bleeding out the fuel delivery pipes (this is from my historic diesel experience with old school diesel injection), it just started to fire. What I was next going to try was putting 1 tsp or so of diesel or light oil down the injector hole of #1 (the cylinder apparently that is critical for this damn virtual crankshaft timing system). But never had to go this far. Maybe worth a shot. Be very careful not to put too much down there (hydroloc the engine). Rough calculation for absolute max (in cc or ml): 800 cc displacement, 3 cylinders, 20:1 compression ratio 800/3/20=13cc . 1 tsp is 5ml.

BTW, I had a very similar experience trying to get my Smart Car to start after some repair work. It started, ran for a bit and then died. Would not start. I ended up testing a bunch of theories and learning about some of the inner workings of the system. Documented this on a wiki page on this site. In the end, the "WD40 into the intake hose" (in my case I used some penetrating spray) seemed to trigger the start. Suggested by Tolsen. Related to the virtual crankshaft sensor system. I'd give it a try. I was NOT a believer, but it did seem to work. Warning: Disconnecting the rubber hose to the intake manifold is tight and challenging the first time you do it. The hose will likely be stuck to the aluminum fitting and needs some gentle persuasion. You can get at the gear clamp but you will have to lever it off with some screwdrivers or something. No space for your hands. The other end closer to the air box is much easier. Do it first. Don't overstress the fuel lines that get in the way.

If that is the case I would: 1) get a can of Rust Check or whatever from Canadian Tire. Spray it on the alternator and move it manually to loosen it up. 2) charge up your battery with an external charger or boost it from a known good and charged battery. 3) make sure that the belt is tight (you loosened it in step 1). 4) turn off the fan and all the other electrics. If you are fortunate, it may start. There is a timer that cuts out all electrical load for about 30 seconds. You can definitely hear it kick in. Not sure if related to temperature but it doesn't seem to be. Perhaps if it gets a 'low voltage' signal, it shuts things down.

I replaced my alternator without disconnecting the X-brace. I believe I ended up taking the bolt out of the engine mount though. I don't have A/C option. Very tight to get out and in. Plus getting the fixings bolt back in was especially painful, if I recall correctly. But very doable in the driveway. The sequence to do the changes (different methods) are well documented on the web. I did the replacement since the belt squealing was getting unbearable (worn pulley problem). I now have a functional spare (but still with worn pulley ... still looking to source replacement. Apparently some Ford 5 groove pulley that can be modified to fit.)

I think I put one of my short old imperial sockets on the big nut and grabbed it with a pair of vice grips.

Belt slipping on the alternator pulley is not a good sign. Either the belt is very loose _or_ you have a seized alternator (stator to rotor). The latter could be the source of your starting problem. From the looks of your car (rust and corrosion), this increases the probability that you have this problem as well. Do a google on "smart car seized alternator".

I was lead to believe from the dealer that for 450 model, they needed to send VIN (or ECU code) and existing key code data to Germany to get a reprogramming code for both keys. Or something like that. Took an hour or so of waiting around. There is circumstantial evidence that something very complex is involved, based on the horror stories of those who lost their only key and were forced to get a new ECU. Chatting with the service manager in Kingston ON, he said that this vintage of Mercedes key system was the most painful he had come across to date but the newer system is much better. No calls to Germany.

Have you at any time measured the compression with a gauge? If you really want to diagnose this by swapping injectors, I would stick with the same set and return them back to the original position and see if you can replicate your first set of measurements. You essentially are trying to validate your theory (hope) that this is an injector problem. The first thing you need to do is make sure that your test procedure can be replicated.

Not sure what is easier ... but since I don't really understand HOW the STAR extrapolates to a compression value, I'd hook up an actual compression meter to cylinder 1 and measure it wet and dry. Plus I don't have a STAR This will give you indication of the raw compression values. If low: it will tell you if it is worn rings or a cracked/leaking head (including valves). If OK: it will tell you that it is the fuel injection Based on this information, you can decide what to do next. That said, swapping the injector would likely tell you if it is the injector on that bad cylinder. Other than the uncertainty around obtaining a decent virtual camshaft signal for the ECU to decide to go ahead and inject fuel. But maybe you can read that off your STAR system.

Very impressive that the STAR system can discern a weak cylinder. I wonder if it uses the angular velocity (acceleration) method similar to the virtual crankshaft? The correction factor could be derived by the ECU trying to even out any differences it sees from cylinder to cylinder. I have to get a STAR system or equivalent. I hope this Delphi system gives me a chance.

Not sure what temperature this was shot ... but doesn't look to be that cold. But certainly a rough idle when cold. Almost dies on you. Have you tried to isolate it to a specific cylinder? I have this mechanics stethoscope that I use for this purpose. I'll guess that you have a bad cylinder (injector, valve, wear). On a 3 cylinder engine, you don't have much room for a lazy one.

kdubya, I just noticed that your sig says Grande Prairie. Been there a few times to support our client in the oil and gas sector. Never saw a Smart Car up there. All big pick up trucks. The guys give me a very hard time about my Smart Car. Yup, you need to get the cold start under control for sure!

BTW, I could be mistaken but I don't think this injector style has a "cracking pressure". That is not to say that you need to meet a specific pressure to get a decent spray. But the signal the injector opens it up to the full common rail. This is pretty much analogous to petrol/gas multi-port fuel injection (indirect injection into intake manifold). Cracking pressure applies to the old school diesel with a sequential injection pump. 1 line per injector. In this mode the injection pump needs to overcome the "cracking pressure" to squirt diesel into the cylinder. Below this pressure the injector is not to leak diesel into the cylinder. BTW, you don't have any other suspicious observations? What is your oil consumption rate? This often correlates to worn rings. Does the piston sound change as it warms up? Hard to hear over the injection noise. But piston slap on one cylinder can be heard with some training. It goes away as the engine warms up. Google piston slap.

Wow. I only dream about having such visibility into the workings of this vehicle. As I understand it, 1) the problems with starting happen at about 2C. 2) the engine is cranking fine 3) if the interior of the car is warmed with a webasto heater, the car starts just fine. What is your current theory? BTW, how would these STAR read outs show the cylinder compression? Iron has about 1/2 the linear coefficient of thermal expansion compared to aluminum. Therefore, I'll hazard that the dry compression on a cold engine will be worse than that of a hot (the pistons will fit looser). So you want to measure the compression on a dry and cold engine (relatively speaking. Not that you have to do this at 0C).