Audi 5 cylinder

[Crasher]

I need to equip a 1986 Audi 5 cylinder 10v K-Jet engine with wideband self learning Lambda control using an LSU4.9 and to control the Lambda by operating a genuine VAG retrofit frequency valve that intercepts the control pressure; basically it is an injector with a hose and coupling on it and can be used to adjust the control pressure +&- within a small window, enough to keep lambda 1 for a cat. I used these years ago with the Racelogic V-Sam so I know they are low impedance and require a ballast resistor; I found out the hard way! Additionally I would add an air flow sensor plate position sensor, MAP, IAT and CTS  sensors plus the throttle body has shut and WOT switches; the addition of  knock detection would be very nice. The Bosch dizzy is vacuum and centrifugal timing control and hall sender. What I don't know is if I can intercept the dizzy output and tweak it or do I have to use a non advance dizzy (technically possible but awkward) and scratch buo a 3D timing curve and use say an Atom ECU to control the frequency valve using the signals provided AND critically intercept and modify (or not) the G40 signal. This would be almost like a classic "piggyback" installation like I used to do with the V-SAM, does anyone have any idea if this would work? The reason it needs to be kept simple is the owner hates complexity and electronics so the engine bay aspect needs to look quite OE plus when finished, the car is being shipped to NZ as the owner is returning home so he would have some support available in the ECU's homeland. The actual reason for the unit is the engine is being taken from 2.2 to 2.5 litres with a big valve head and mild cam, the metering head has been professionally rebuilt KMI and modified for suitable extra fuel flow but due to the unavailability of 99 RON fuel in NZ, I need to set it up for 95 RON and he needs to be able to deal with any knock.

 

[Adamw]

Whats the resistance of the injector?  From memory the K-jet runs around 6bar pressure and with a ballast resistor added I would have some doubt that the old style ev1 which is designed for P&H drivers would open reliably or predictably at that pressure. 

For the ignition you would really want to lock the distributor and have all advance done by the ecu.  A spot of MIG on the advance mechanism locks them pretty quick.   

[Crasher]

A fixed dizzy is no problem. The system runs at 5.5 Bar and I have used a ballast resistor on these valves for decades but I don't have one to hand at the moment.

[Adamw]

Is there still a mechanical regulator in these systems or does this injector do all the press control?  Do you know if they usually only vary duty cycle or frequency as well?

So electrically it will work with a ballast as you have found, but with a ballast those will be very slow to open and that opening time will vary a lot based on voltage on the other side of the injector, but the closing time will be similar or faster than it would be with a P&H driver.   Deadtime is opening time - closing time, so with a ballast the deadtime will be very long and quite variable.  I dont know what duty cycle these would typically operate at, but as duty cycles get small or frequency gets big they are going to give poor control.  So for example your fuel pressure will change a lot when electrical load changes such as a fan or headlights turn on. 

My initial thought was it would be best to run this device using the closed loop fuel pressure control function (rather than as an injector) so you then have a 3D feedfoward table, a PID control loop for very quick and stable response, and a 3d fuel pressure target table.  You could have CLL correction on one axis of the target table so you have lambda feedback.  This function uses a fixed frequency, so that means you could choose a relatively low frequency so that the deadtime was a smaller proportion of the pulse width required therefore having less influence on the result.  

But if you needed to vary the frequency for some reason (I dont know if these were normally timed to any mechanical event) then you would probably have to run it as an injector (using the normal fuel table etc), which would effectively only give proportional only control and im less confident that would give what I consider acceptable control.  

[Crasher]

Sorry, only just had a reply notification. It has a Control Pressure Regulator which is both electrically and engine block transfer heated, as the temperature increases the pressure increases and this forces the plunger in the fuel distributor down, covering the fuel slots and reducing the flow of fuel. The frequency valve bleeds off the control pressure so you can increase fuelling on demand. VAG used them with the control pressure set higher so the valve could +/- fuel.