Low impedance injector driver - HELP!

[1insuare]

Hi everybody, sorry not to reply earlier, I've been away this weekend. Regarding to the circuit #ronv did it looks great, I more or less what I was thinking but using the timer. I have to analyze both but what the difference between them? Is the second one using a comparator with hysteresis to open and close the mosfet?


I have no more specifications apart from the injector impedance (around 2ohms) and the injectors must be driven by MOSFET. As this drive circuit is part of a ECU, the power supply must be more or less controlled in terms of power consumption. I have a question: is possible to decrease the time from zero to the peak level?

 

[1insuare]

Sorry #conv, I can't see the circuit properly, is it possible to attached the image with other format as jpg ?

 

[alec_t]

is possible to decrease the time from zero to the peak level?

The only way to do that is to use an injector with a lower inductance and/or increase the drive voltage above 28V.

 

[1insuare]

#conv, can you tell me which comparator are you using ? I am trying to replicate the circuit in LTspice but I don't distinguish the model.

Thank you

 

[1insuare]

I have a new question, I hope easy for all of you: Because I will need to build this circuit in a PCB, the 0.1ohms with 5W power dissipation, the resistors I found was quite big (around 17mm length) for my design. Is any other way to do for example using more smaller resistors?

Thank you in advance

 

[ronv]

PWM Injector

Here is a better picture. I made a couple of changes to improve the turn on turn off of the FET. Use a low gate charge FET like this one:

https://www.electro-tech-online.com/custompdfs/2012/01/FQP20N06.pdf

Since the LM119 comparitor doesn't drive much current it will switch faster with low gate charge.

The rise time of the current is determined by the inductance of the injector (typically 2 Mh for your 2 ohm one), and the voltage (28 volts) along with any series resistance (.1 ohm plus .06 ohm for the FET). This is probably about the best you are going to do. Is it a very high RPM engine?

 

[ronv]

Depending on the RPM it will run at you might get by with as small as 1 watt maybe 2 would be better.
If you have to order parts take a look here:

https://www.mouser.com/Passive-Comp...s/_/N-7h7z5?P=1z0vocvZ1z0wo8tZ1z0vo2aZ1z0x780

 

[alec_t]

If you want to use the LM1949 with a simple MOSFET drive adapter here's a revised circuit. The components in the dotted rectangle model the action of the LM1949 (including the output clamp action at pin 2 and the peak time limiter at pins 6,7,8).
An LM324 op-amp (or any other having a common-mode input which includes ground) is used as a combined level-shifter and low-impedance driver. A zener diode D8 drops its supply voltage down from 28V (too close to the op-amp limit) to a safer 18V. A zener (not shown) could also be used to supply 5V to the LM1949.
The FET dissipates most power during the hold period. More power is dissipated in the 33V zener at the FET drain than in the injector.

 

[1insuare]

Its near 5000rpm maximum, do you think I can reduce the power rating of the 0.1ohms? I will try to simulate it with other comparitor since my software doesn't have the LM119 installed. Thank you so much #ronv and everybody.

 

[alec_t]

Its near 5000rpm maximum

Average power in the circuit depends on several factors apart from speed. Is this for a common-rail system (only 1 injector), or is there 1 injector per cylinder? How many cylinders? 4-stroke or 2-stroke?

 

[1insuare]

its 2 injector two stroke engine. I will use an air injector as well. Both must be driven at the same time, I mean: 1air injector/1fuel injector every pulse.

 

[alec_t]

As the injectors will be operating under different pressures and with different fluids they will open/close at different rates so each will need its own control circuit. For a 2-stroke system at 5000 rpm the pulse repetition period is 12mS. Simulation shows that for a 2mS pulse duration the average dissipated power when driving a 2 Ohm 2mH injector is:-
Injector 1.4W, FET 4.0W, Spike suppression zener 3.0W, Current sense resistor 40mW.
So a 0.5W rated sense resistor should be adequate.

Be aware that during the hold part of the cycle the FET tends to hunt about the hold level, i.e. it oscillates, so will generate considerable RFI which will need to be suppressed. A darlington instead of the FET would be less likely to oscillate (lower parasitic capacitance).

 

[1insuare]

The engine I need to run, should turn the injector on 5 times per revolution. So for 4000 rpm maximum (sorry I said 5000rpm) should run every 3ms.... A part from that, the pulse timeon goes from 1.5ms to 5ms. Does it make sense?

So I will need to build 4 different circuits to run every injector I guess.... Thanks!

 

[alec_t]

should run every 3ms.... A part from that, the pulse timeon goes from 1.5ms to 5ms

Not sure how an on time of 5mS can be squeezed into 3mS repetition period ?

Edit: Simulation shows the injector turn-off time is ~ 0.4mS, so the max pulse on time at 4000rpm would be ~ 2.6mS.
With a 2.5mS pulse every 3mS the average power consumption leaps up:-
Injector 6.1W, FET 20.5W, Diode 12W, Sense resistor 175mW.
Some serious heatsinks will be needed!

 

[1insuare]

sorry, I was completely wrong. Yes, there is one injection per revolution, so injection every 12 ms at 5000 rpm. For the worst case, the injection time pulse is 4.8 ms. Is a way to calculate the average power dissipation in LTspice? Thank you again

 

[ronv]

If you set up the time in LTSpice for 1 complete cycle (12 ms) you can then left click on the waveform you want to average (lets say IL1). This will give you the average and RMS current thru the injector. Multiply times the voltage and you will know the power. The thing I noticed while doing this is that the current in the zener running in PWM mode is quite high. You might need 4 9 volt 5 w zeners not the 1 5w. Funny National missed it to.

 

[1insuare]

I noticed the same yesterday. For 5000rpm (12 ms of period) and a 4.8 ms of Ton, the average power dissipation in the zener is 10W. So which zener I need to chose?

#ronv, you said 49V 5W . Will be enough with this? Thank you

 

[1insuare]

In order to reduce the current through the zener, can I place other in parallel to distribute the current? I did it and it seems to be OK....but Im not sure...

 

[alec_t]

In order to reduce the current through the zener, can I place other in parallel to distribute the current?

Unless the zeners are closely matched they will not share the current equally. Better, IMHO, to place zeners in series. I think Ronv was suggesting to use 4 zeners, each of 9V 5W rating, in series.

 

[ronv]

Right you are Alec. Poor choice of nomenclature on my part.
Seems like there ought to be a better way but I haven't thought of one yet.