- Site Admin
- Posts: 5039
- Joined: June 18th, 2006, 9:31 am
- Location: Bedford, UK
Old style ignition systems we all know, use bob weights and calibrated springs and stops to control the movement of the baseplate (on which the contact breaker points or inductive pickup are mounted) relative to the distributor spindle, to achieve advance, while mapped systems vary the advance electronically.
The old type of distributor without vacuum is insensitive to anything except speed. Vacuum advance by means of a diaphragm sensor - a sort of crude engine load sensor working from inlet manifold vacuum is merely a device to advance the ignition on part throttle - to improve torque and economy. Perfectly reasonable, the part-throttle requirement for advance does tend to be higher than full throttle anyway. When there is no vacuum in the inlet manifold - ie: at full throttle or thereabouts, it doesn't add any advance.
Because it is only speed sensitive the old style bob-weight distr (distributor) cannot be setup to accommodate part-throttle advance and to 'optimise' it requires the bob weights to be locked up engine to be run at full throttle at a range of speeds, from idle to max (800 - whatever-u-like) and the distributor manually adjusted to give peak torque without inducing knock - detonation. Detonation is a very hard thing to pick up, even with modern sensors, so doing this MUST go hand in hand with some knowledge of what the demand for ignition is likely to be! It is also a very demanding thing to do because someone has to physically turn the distr to and fro to preset marks on the distr-engine to calibrate it, while the engine is on full throttle, responding to the callout from the guy on the brake dyno control as to whether the torque is going up or down.
Of course you could say, let's make a vernier to do it remotely. Torque going down can mean it needs more advance or it's detonating, so you can see it's not a job for the faint-hearted. To keep and engine on full throttle at set speeds needs very good load control at the control console and in the cell where you're moving the distributor, the heat coming off the exhaust at high load is tremendous. Plus if she blows up you're right next to it. I've done it by hand around 7000-7500 between 30 and 36 deg on the National Hot Rod 2 liter. It was a nasty, noisy, finger-burning exp and I would NOT like to have to do it again! Having the distr just loose enough to turn means it is quite likely to spin on its own as soon as you take your hand off it.
Having gained the knowledge of what the advance should be you can build your own distributor, by messing around with weights, springs and stops and if you have a distributor testing machine to calibrate it. No-one does this for their own one-off engine, trust me!! A run of 50 racing distributors, maybe. For a race engine you have to remember that it is probably racing at 5000 plus rpm all the time anyhow and if it's an 8v unit TC the advance it needs is somewhere between 33 and 35 deg, a 16v probably abut 30-32 deg and that's that! No fancy accuracy needed in the low rpm centrifugal range, none at all. Think about it. In a lot of engines, the baseplate is locked up at maximum advance and the engine started with 24 volts.
Mapping up an electronic ecu (electronic control unit) system is altogether different and much more fun. Here you'll be dealing with a unit that measures throttle plate angle via a potentiometer on the spindle, and is often setup using a laptop display with appropriate software. There are systems that run the fuel injection or igntion or both, usually sold with the hardware - sensors, throttle bodies, pickups, trigger wheel, harness.
Most programmable systems can be adjusted quite readily by turning a dial. You can calibrate them real time to suit each load site - in other words combination of throttle position and engine speed. For example 20 deg throttle and 3000 rpm. That would be a load site. You need at least 220 load sites to make the job worth doing, using say, 500 rpm steps from 1000 to 8000 and 4 or 5 deg throttle increments from zero deg -
near closed (on the idle adjuster screw) up to 90 deg - full throttle. Of
course the dyno load needs constant adjustment to keep the speed stable. There will be, for each load site - a commensurate setting for the fuelling, if you're on carbs you can leave them to themselves to an extent, but for injected systems you need to check and adjust the mixture constantly, because one affects the other, obviously.
The optimum ignition setting is that which generates the maximum torque at that speed/throttle combination, you have to adjust the ignition advance to find out where it is. As you approach the best setting torque will start to peak - just before the engine starts to knock, if you have good ears it is worth getting to know what LIGHT knock sounds like at low speed and light throttle - you'll hear a crackling or splitting noise. Not a good idea to try this at high load, you can hole a piston in 2 engine revolutions. The warning signal for knock is a sudden drop off in torque for no good reason, then you have to very swiftly dial out a few deg of advance and run thru that load site again to check. You need keen
eyes and lightning reactions to do this well.
Are knock sensors much use? Well, yes if they were original equipment for that engine, no if they were not. There is still no bullet-proof way of detecting knock, so again exp of the type of curve you're expecting generate is a big help. Mapping turbo systems without real-time knock sensing (as opposed to not real-time, ie: examining holed pistons and blown gaskets) is highly risky to say the least. As a rule the less advance for given boost the better! If the ignition settings at the numerous load sites were predictable that'd be fine but it's the opposite of predictable! That is, dare I say it, the reason mapping is so good, you CAN go looking for the optimum. There is some useful data on p 134 and 135 of my TC book. My table illustrates 'graphically' how the ignition demand varies wildly. Look at the curves - if they get uploaded - from graph p 135 and you'll see how unpredictable the demand is. It goes up and down randomly. I've studied these kind of maps over and over and I CANNOT tell you it needs more or less advance because the cylinder is filling well or because the rampipes are too short or too long etc!
Yes, that GC engine was of such-and-such a type and tune and yes it was fuel injected but that is a VERY good base model for anyone attempting his own mapped system. That's why I published it. If your own peak advance was capped anywhere at 38 deg compared with that map (the true demand on that Case History 2 liter motor went as high as 44 deg in places) you'd be pretty safe.
Do you get 'holes' in the torque curve from missed load sites? Well probably. The ignition system can't self teach - it's not like the fuel system where Lambda feedback from the exhaust gas controls the next injection cycle. Is it possible to have some odd incorrect interim setting somewhere which you didn't pick up where the ignition is way over advanced, yes, possibly. But in the main you're pretty safe provided you've spent time over mapping, because you'd as likely as not just drive thru the incorrect setting without even noticing, you'd be very unlikely to hold the absolute precise setting of throttle and miss the conditions necessary to generate knock. What happens if atmos temp changes compared with test? Well a good mappable system should incorporate an intake temp sensor that can retard the ignition progressively for increases in intake air temp.
And lastly What happens if you don't use the same octane gas as used on test? You could be in trouble, that's what!
- This is trigger wheel, Weber as it happens, with sensor. The teeth and the one 'missing tooth tell the ec the rpm and where tdc is. The inductive pickup must be accurately and robustly mounted or it will cause misfire.
- Weber ign 1.JPG (11.16 KiB) Viewed 20099 times
- Weber mappable igntion system - ecu, pickup, trigger wheel, temperature sensor, coil pack and harness. Race proven.
- Weber ign 5.JPG (15.18 KiB) Viewed 20099 times
Users browsing this forum: No registered users and 2 guests