Hello all, this is my first post on this forum, so a little introduction first... I am Simon, I run a company (LPGC) that converts petrol vehicles to run on LPG (propane) in Yorkshire, England.
I am by no means an electronics expert compared to, I'm sure, many people on this forum but I may appear so to the layman - As a kid I used to fix amplifiers for bands / modify CB radios, modify computers to run from car batteries etc. More recently I have designed voltage buffering circuits to effect a repair on vehicle ECU's that have stopped reading temperature signals etc...
So I could probably fudge my way through and do what I would like to do next without asking for help, but I thought it better to ask the experts on this forum for advice!
Essentially, I want to build a circuit that: Takes a 0v to 5v voltage (this is a vehicle engine inlet manifold pressure signal) and turns that signal into a pulsing signal where 12v is off and 0v is the pulse. The pulse width should be between 0ms at very low manifold pressure (maybe close to 0v) and up to 16ms at atmospheric pressure (near 3v as opposed to 5v - more on this later). The frequency of the PWM signal needs to be clocked to pulse at the same frequency as engine rpm, so the system needs to take a clock signal from the engine ignition coil.
Having read the above, I am sure most people on the forum will think I am trying to build a standalone petrol or LPG engine fuel injection ECU, but this is not quite what I am trying to achieve...
On the market there already exists some very cost effective LPG system ECU's, but these are designed to work as a slave to the petrol fuel system, i.e. these ECU's cut the connection between petrol ECU and petrol injectors, monitor the pulse width of petrol injector signal, allow the installer to set a 'map' (which basically allows them to adjust multiplication of the petrol duration signal at many points, so a petrol signal of 3 ms might mean a LPG injector signal of 4ms and a petrol injector duration of 6ms might mean a LPG injector signal of 8ms, etc, to arrive at a pulse duration for LPG injectors). These LPG ECU's also compensate the final LPG injector signal according to temperature and LPG pressure readings (hot LPG is less dense than cold LPG, so pulse width is increased for hot and vice/versa, higher pressure LPG is obviously more dense)...
Slave type LPG ECU's are available cheaply, so I'm not trying to make this kind of ECU. But these LPG ECU's require the petrol injector signal to exist in the first place - they will only work with a petrol injector signal, they work as a slave to the petrol ECU. What I am trying to achieve is to get the LPG ECU to work on carb'd engines where no petrol injector pulses are available. So the circuit I design doesn't need to drive the (3ohm) LPG injectors directly, it only has to work as an input to an LPG ECU (50ohm input). It doesn't have to do compensation for LPG temperature or pressure, the 'slave' LPG ECU can do all that. It doesn't need to be capable of being 'mapped' to provide correct fuelling, the LPG ECU can do that.
I would be happy with the above for now, but I would hope to be able to modify the circuit in future to work as a 'closed loop' system, where PWM can be adjusted according to vehicle exhaust oxygen (lambda) sensors. Another future development might be to allow the system to fire each LPG injector sequentially, but this latter would be of less advantage than the closed loop ability.
I have been considering using a microcontroller, since it would be simple to adjust the program to work on turbo engines (where the 16ms would arrive at higher manifold pressures, higher prssure sensor voltage (up to 5v instead of up to 3v) and since the extent of lambda mixture adjustment could easily be adjusted in programming (we want mixture compensation to be disabled with a cold engine, slow at low rpm, quicker at higher rpm, disabled at high engine loads).
I welcome all sensible / informed replies.
Regards,
Simon
Lpgc
I am by no means an electronics expert compared to, I'm sure, many people on this forum but I may appear so to the layman - As a kid I used to fix amplifiers for bands / modify CB radios, modify computers to run from car batteries etc. More recently I have designed voltage buffering circuits to effect a repair on vehicle ECU's that have stopped reading temperature signals etc...
So I could probably fudge my way through and do what I would like to do next without asking for help, but I thought it better to ask the experts on this forum for advice!
Essentially, I want to build a circuit that: Takes a 0v to 5v voltage (this is a vehicle engine inlet manifold pressure signal) and turns that signal into a pulsing signal where 12v is off and 0v is the pulse. The pulse width should be between 0ms at very low manifold pressure (maybe close to 0v) and up to 16ms at atmospheric pressure (near 3v as opposed to 5v - more on this later). The frequency of the PWM signal needs to be clocked to pulse at the same frequency as engine rpm, so the system needs to take a clock signal from the engine ignition coil.
Having read the above, I am sure most people on the forum will think I am trying to build a standalone petrol or LPG engine fuel injection ECU, but this is not quite what I am trying to achieve...
On the market there already exists some very cost effective LPG system ECU's, but these are designed to work as a slave to the petrol fuel system, i.e. these ECU's cut the connection between petrol ECU and petrol injectors, monitor the pulse width of petrol injector signal, allow the installer to set a 'map' (which basically allows them to adjust multiplication of the petrol duration signal at many points, so a petrol signal of 3 ms might mean a LPG injector signal of 4ms and a petrol injector duration of 6ms might mean a LPG injector signal of 8ms, etc, to arrive at a pulse duration for LPG injectors). These LPG ECU's also compensate the final LPG injector signal according to temperature and LPG pressure readings (hot LPG is less dense than cold LPG, so pulse width is increased for hot and vice/versa, higher pressure LPG is obviously more dense)...
Slave type LPG ECU's are available cheaply, so I'm not trying to make this kind of ECU. But these LPG ECU's require the petrol injector signal to exist in the first place - they will only work with a petrol injector signal, they work as a slave to the petrol ECU. What I am trying to achieve is to get the LPG ECU to work on carb'd engines where no petrol injector pulses are available. So the circuit I design doesn't need to drive the (3ohm) LPG injectors directly, it only has to work as an input to an LPG ECU (50ohm input). It doesn't have to do compensation for LPG temperature or pressure, the 'slave' LPG ECU can do all that. It doesn't need to be capable of being 'mapped' to provide correct fuelling, the LPG ECU can do that.
I would be happy with the above for now, but I would hope to be able to modify the circuit in future to work as a 'closed loop' system, where PWM can be adjusted according to vehicle exhaust oxygen (lambda) sensors. Another future development might be to allow the system to fire each LPG injector sequentially, but this latter would be of less advantage than the closed loop ability.
I have been considering using a microcontroller, since it would be simple to adjust the program to work on turbo engines (where the 16ms would arrive at higher manifold pressures, higher prssure sensor voltage (up to 5v instead of up to 3v) and since the extent of lambda mixture adjustment could easily be adjusted in programming (we want mixture compensation to be disabled with a cold engine, slow at low rpm, quicker at higher rpm, disabled at high engine loads).
I welcome all sensible / informed replies.
Regards,
Simon
Lpgc
