Skoda Rally Blog

.. for most people means taking an extra day off.  This time of year for me it means working on the car a lot!

The plan was fairly simple – the car ran OK the other day, so spend the day driving it, slowly improving the way the engine runs.  This, in principle, is fairly simple.  Last year I invested in a couple of piece of kit that allow me to map the way the engine runs – they measure the engine’s speed and how far open the throttle is, and measure the air/fuel mixture at that point, and record all of this information.  After driving the car for a bit (to average out little errors and to drive in as many different speed/throttle positions as possible) this log can be downloaded and then analysed using the LogWorks software which can also take the engine’s current settings and calculate new ones.  That’s how it should work, and within a few sessions you get a fairly driveable engine.  Then the car can be taken to a rolling road to fine-tune all of these settings and also to set the ignition timing (which is critical for maximum performance).

However, things aren’t quite that simple.  The first issue is the software that comes with the ECU that I’m using.  In short, it’s awful.  When I got the car mapped before Rally GB 2008, the dyno operator hated it, and I can’t argue with that at all.  Anyone who knows me knows I’m a bit of a techie/software geek, and I’m quite happy using things on computers that would freak most people out, so it’s not just a case of it being complex or even a bit geeky.  It’s just awful.  The kind of software that constantly leaves you saying “who did this?”, every time you change something.  Now, this wibbling might seem irrelevant, but it’s not;  it actually creates a real problem when setting the car up.  The reason?  You can’t copy or paste from/to the fuelling map.  No really, copy and paste, something which has been in the Windows World since, oooh, 1988.  No, you can’t do that.  Which makes for a far more complex setup.  I will spare you the full horror, but let’s just say it’s good that I’m game to use any means to get something done, in this case working out the file format that the software stores the engine settings in, and then converting that to a spreadsheet, and then putting the newly-generated values into that.  Then it’s the reverse – taking those settings and converting them back into the format the ECU will understand.  It’s about a 5-stage process, but at least I’ve done it enough times now to have it fairly well down.  And the big bonus?  You can go for a drive for 30 minutes and change maybe 75% of the fuelling map.

OK, so that’s the real plan.  And if it had gone that simply, then I would have been writing this 24 hours ago….

Saturday started out simply enough – got the car running, and went for a first little drive; Tammie was in the passenger seat, and it all seemed OK.  However, the car wasn’t keen to re-start, and as it was running so lean I thought it best to add some more fuel before driving back.  Two things happened – one was an amazed passenger that you can transform the way the car behaves without having to “do something under the bonnet” and two, the car drove better.  But the laptop’s battery was dead and I don’t like going out without it as it can diagnose an electronic issue instantly, so it was time for lunch, and for me to get on with it alone in the afternoon.

The problem came that the car just didn’t want to start up in the afternoon.  No amount of coaxing would get it to fire.  And it soon looked like the problem was the battery – when cranking, the ECU would disconnect from the laptop, I assumed from low voltage.  So time for a charge, and a few more hours wasted, but I took the opportunity to go buy some more spark plugs.

Even after that, when the car did run, it ran like crap.  Lumpy, unhappy, and really just plain odd.  Usually when a car is missing on one cylinder it’s fairly easy to diagnose which one, but removing an injector connector on each cylinder just made an incremental kind of worse, rather than changing anything massively (or as it should do, not making any difference if that was the bad cylinder).  Really weird and frustrating; sometimes it would seem to clear for a few seconds, and then run rough again.  A change of plugs didn’t do much for it.  I’d checked things over and over again, and didn’t seem to be getting anywhere.  And when I tried the ECU’s “test” mode (where it fires the injectors and spark plugs), nothing happened.  I knew it would fire normally (otherwise it wouldn’t run), but there was clearly something odd.  In addition the laptop kept having communication errors with the ECU and with the logging hardware (both use serial ports, and I use an old laptop solely for this purpose). It was getting dark and I’d had enough, so I gave up.  At this point, any chance of going to RallyDay seemed to have vanished.

A good night’s sleep often does wonders.  Admittedly I did wake up with that “oh no…” feeling, but thought about checking all the little things that could be a problem, starting from scratch.  And found a few things – firstly, the reason for the ECU’s test mode not working – I’d followed Skoda’s practice of powering the injectors and coil from the fuel pump output, which isn’t triggered in test mode (for obvious reasons, you’d flood the engine and set fire to it at the same time!), so I rewired to power them from the ECU power, which enabled test mode (which plays a lovely out-of-tune tune) and showed that coils and injectors were all working (I primed the fuel pressure first and then collected fuel into a jar, one injector at a time).  So, two hours down, and no actual progress.  To eradicate the laptop problem, I had to go and buy a USB-Serial adapter (so I could use my ‘proper’ laptop), which meant I could download a log, but it still doesn’t work with the ECU software as it refuses to see the port in question.  GRRR!  Then something struck me.  I’d adjusted the valve clearances when building the engine, but not since it had been hot and then cooled, and some settling in may have occurred.  And it had – although there was clearance on each valve, on some it was much less than the spec 0.25mm, so I re-did them all. I also realised I’d not reset the ECU to the throttle body (having noticed when it was running that closed throttle was registered as 20%), and then fired up.  It didn’t sound great, but it did run.  And well enough for me to strap everything back in place, and try to go for a drive.

It drove OK.  Not amazingly, obviously, but above 3000rpm it was clean and clear, and went pretty well.  Well enough to do a 20 minute log drive, and to come home and look at the data, and realise I’d not reconnected the RPM trigger to the logger, so that data was all null as the engine was apparently stopped!

Another drive was done after it was connected, and the engine behaved much the same – quite a misfire below 3000, but above seemed clean and clear, despite running quite rich after my hamfisted “richen the map everywhere” adjustment.  I was careful not to stress the engine too much, and it seemed happy enough, so I did about 20 minutes and then went through the process of recalculation, and repeated this three times over the course of the day – each time with the car driving better, and being able to take it to more “extreme” areas of the map, either low speed with big throttle or lots of load, etc.  The main problem, though, was that when you got to 5500rpm in 4th or 5th gear, then the engine would keep stuttering; it woudl just die and kick back in and so on.  Didn’t do it in any other gear (it accelerates quite quickly), which was odd.  I had a think about it and then went for a final run with the laptop connected to the ECU; the software of horror has one screen which is really useful, which shows you a running count of things you’d not like to see, including crank trigger problems.  And lo and behold, every time the stuttering occurred, the crank trigger was at fault – the counter would jump up by 10 or 20 each time, and never change at any other time in the half hour I ran it like this.  It seems to me that the bracket that it’s on is vibrating at that speed (as it works fine above it), and when you’re in a high gear it’s at that speed long enough to let it get going.  A strengthened one is on the list of things to do this week.

So, having had a horrific Saturday, Sunday has actually been good and productive.  OK, it’s taken me all day (aside from slacking off to watch the F1 and have lunch), but it’s now fairly driveable, and with any luck the stronger bracket will kill the misfire for good (it also does it sometimes at about half that speed, which makes sense and adds credence to the ‘resonance’ theory).  As Gene Krantz said, “Work the problem.  Don’t make things worse by guessing.”  Hopefully this approach will see the car at RallyDay!