DIYEFI.org Forum

I found some interesting stuff about injectors. High impeadance injectors have an inductance of around 20 to 30 mH. Low impeadance injectors are around 3 to 10 mH.

EDIT: both below links where the same, I corrected the below this line.
http://sonic.net/~mikebr/ecm_555/inj_inductance.html
and here
http://sonic.net/~mikebr/ecm_555/injector_drvr.html
Which lead me to here
http://www.national.com/mpf/LM/LM1949.html#Datasheet

Note the date of 1995 and the peak and hold note backing down the hold current to produce faster on and off times, creating a more percise fuel consistancy. Hmmm, makes me wonder could you get a faster on time if you ran a small current through the injector. Not enough to turn it on, but enough to decrease the inrush current requirement. Hmmmm.

I also see fairchild has FDSS2407 a duel channel MOSFET. Not sure what the drain feedback is about. Hmmm what's the best injector drive circuit.

Thoughts?

IMO, and I believe others too, the best solution is a single resistor and single self protected FET per high imp injector. If low imp driving is required, there can either be a place for the driver chip jeans setup uses on the board for optional population, or an external box with that chip and it's own FETs could be used. The single self protected TTL drive FETs are an excellent choice for performance, reliability and simplicity. All things we are interested in. Esp for a first cut. I've got 24 or so of them here ready to solder up at some point, philips 8 amp ones i think i got. The other ones are pin compatible and up to 20 amps if you wanted some margin or to use them for P&H without the limiter kicking in and stuffing up your opening times etc.

Fred.

Perhaps you can scratch something on a napkin then scan it. Where is the resistor and what value. I'm picturing the resistor as a current sense resistor, so around .001 ohm.

Is peak and hold important? I was picturing perhaps two IGBT's side by side, on for massive current draw, the other with a series resistor to limit the current. Trade-offs are part cost vs functionallity. I'm assuming the PCB's are going to cost the same if it's layed-out either way, so why not plan for it, then we can populate parts if we want to.

I think I'd perferr two IGBT's vs one that attempts to current limit. Keeping them saturated should help keep them cool. However one could either PWM to current limit, or use an anlog resistor to keep the IGBT in the linear range.

Also the chips in hand, do you have part numbers handy? I'd like to plan for different form factors, features, ect so that other chips can be use if desired.

Arrrr, stupid work lack of spell check. Sorry for the Mrs.takes.

jharvey wrote:Perhaps you can scratch something on a napkin then scan it. Where is the resistor and what value.

The devices we will be wanting to use are internally current limited by measuring the potential across the silicone substrate. I don't think I need to draw it, it is simply a current limiting series resistor from the processor output to the gate input solely to limit current to the CPU pin and protect it. That really is all that is required for non peak and hold. Just select a FET that won't current limit with your particular injectors and it's good to go :-)

Is peak and hold important?

Not to me at the present time, but very certainly for all the big boost small displacement people (including me later). However, just because it's important, that doesn't mean we HAVE to cover it on the default board. The thing is, we could, but I personally would not want to anyway. I'd want it in a box in the engine bay such that the EMS doesn't suffer any noise from it. All it involves is placing a chip and a few other components around the FETs. I don't think we should include support on this default first single board though.

Also the chips in hand, do you have part numbers handy?

PIP3104 / PIP3104-P and they are 8 Amperes, I just checked. They have the same pin out as most of the other to220 autoprotected FET devices, so any of them should be OK. vnp20n07 and vnp5n07 and many others are just the same IIRC.

IGBT's are only for ignition duty, FETs for injector, fuel pump, and general heavy duty IO use.

I hope that irons it out :-)

Fred.

With the inductances you have stated there is quite a hefty inrush current that is used to build flux within the injector core, at this point the inrush current traveling around the circuit causes lots of power dissipation in all of the connecting paths, this causes further delay in the injector opening due to the momentary drop in voltage at the injector. this makes the matching of circuit paths and components a little more important, and obviously the strain on the driving FET/IGBT is more apparent.

As we know the fully open position usually takes about 1.0mS to achieve, most of this is due to acceleration of the needle valve but i can't help thinking that the priming of the injector using PWM will be beneficial. If we say for example the injector opens at 50% and closes at 20% Vcc then if we prime at say 40%, a full Vcc pulse to open, 60% PWM to keep open and a total shutoff to close would improve by maybe 10-30% on opening times.

I think clever use of PWM with an active flyback circuit could provide serious benefits to opening times, but how do we test and prove this is actually working before attaching to an engine?
I have an idea which involves 2 reasonably matched injectors on a common rail supplied with an inflammable fuel substitute spraying into 2 graduated beakers both squirting at the same time for the same duration, but one is PWM primed and the other standard 0 and 1. The pre primed should open sooner and over a period of time fill the beaker quicker. During this experiment it may be worth measuring the delta T in the 2 injectors and collected liquid to make sure the temperature rise is not going to cause heating of the fuel or intake manifold which could cause variation in Intake air density and fuel atomization rate.

furthermore how do we control it? Personally I think this has to be software controlled to allow serious tweakability. But does the 9S12C512 have enough high speed PWM outputs to control all 12 injectors at the PWM frequency required? or will it have to be mutiplexed externally?

thats enough thinking for tonight, I have had enough C language learning today and my head hurts! i'm off to bed to think about circuit ideas for this,

davebmw wrote:With the inductances you have stated there is quite a hefty inrush current that is used to build flux within the injector core, at this point the inrush current traveling around the circuit causes lots of power dissipation in all of the connecting paths, this causes further delay in the injector opening due to the momentary drop in voltage at the injector. this makes the matching of circuit paths and components a little more important, and obviously the strain on the driving FET/IGBT is more apparent.

Agreed, some hefty caps at the drive, or some place along the line will come in quite handy. I feel that this parameter is a good one to measure and customize. To much capacitance causes a low pass filter, while the injector acts as a high pass filter. So a balance might be quite handy. Let the RF energy through, but don't block the DC. A balance is in order, it will keep the noise down.

To measure, I'm picturing a circuit that you can use to get a ball park figure for you injectors. X = 6.28FL so assuming .025H and 60hz (US) 50hz (EU), we get an impedances of 9.42 (US) and 7.85 (EU), lets assume 5 ohms. Now with the use of a power supply, a multimeter, and a small bucket of water, you can make a resistor divider and measure the reactant. Once you know the frequency and impedance, you can calculate the inductance. The meter is probably fairly good for a 10V scale, so I'll first use that scale with an expected 2 amp draw.

I don't see why you can't do the same with a wall wart, mV scale, and resistor. I chose the higher power because I can, and it's fits the expected power requirements a pinch better. With luck I'll get to try this out this weekend.

davebmw wrote:As we know the fully open position usually takes about 1.0mS to achieve, most of this is due to acceleration of the needle valve but i can't help thinking that the priming of the injector using PWM will be beneficial. If we say for example the injector opens at 50% and closes at 20% Vcc then if we prime at say 40%, a full Vcc pulse to open, 60% PWM to keep open and a total shutoff to close would improve by maybe 10-30% on opening times.

I don't see why we couldn't do that. Perhaps good for version 2 software. Should be easy to PWM prime and hold. This is a low cost method, I'd kind of like for that to be done in hardware however. The prime and hold currents are low compared to the fat inrush IGBT. Having perhaps as many as three IGBT's might be a good idea. It would reduce the software stress and heat stress quite a lot if we can keep them saturated, not in the transient zone.

davebmw wrote:I think clever use of PWM with an active flyback circuit could provide serious benefits to opening times, but how do we test and prove this is actually working before attaching to an engine?

I don't like the term flyback. I've only seen that in MS land, and I feel it shows electrical ignorance. This circuit shunts current, so I recommend we call it ISC for Inductive (or Injector) Shunt Circuit. I think that will help separate the Engineers from the mechanics. Thoughts?

davebmw wrote:I have an idea which involves 2 reasonably matched injectors on a common rail supplied with an inflammable fuel substitute spraying into 2 graduated beakers both squirting at the same time for the same duration, but one is PWM primed and the other standard 0 and 1. The pre primed should open sooner and over a period of time fill the beaker quicker. During this experiment it may be worth measuring the delta T in the 2 injectors and collected liquid to make sure the temperature rise is not going to cause heating of the fuel or intake manifold which could cause variation in Intake air density and fuel atomization rate.

I think that sounds like a great jig, is it feasible for you to build it? I don't know your soft background, but I'm sure you could get plenty of help making that kind of a jig.

Fred wrote:The devices we will be wanting to use are internally current limited by measuring the potential across the silicone substrate. I don't think I need to draw it, it is simply a current limiting series resistor from the processor output to the gate input solely to limit current to the CPU pin and protect it. That really is all that is required for non peak and hold. Just select a FET that won't current limit with your particular injectors and it's good to go

I think you mean figure 17 from http://www.fairchildsemi.com/ds/IS%2FISL9V2540S3ST.pdf it shows .01 ohm as a shunt resistor. The problem with trying to sense this way, is that the signal is very small. 8 amps produces 80mV. That enough signal to measure reliably, but it requires a conditioning circuit. This is why I'm pro PSoC. PSoC can condition on it's own, no extra crap. If we're going to hope for current sense in the first freeEMS board, then I should probably recommend we add space for a hall sensor. Much less complicated hardware wise and costs about the same dollar wise.

I'm thinking digi's hall 480-2002-ND would work nice, it would require some calibration, but that shouldn't be a huge deal. Should be as simple as running a known current through the lead, then taking a calibrating reading in software. The offset should be constant, so once you have the first reading, the rest should be simple math. We won't all get it in the exact same place, so this calibration will be needed. And if it ever moves for any reason, it will need recalibrating. There are some that are designed for this, I've got one from allegro if I recall correctly. It straddles the trace, and doesn't need calibration because it the parameter don't change from install to install. Cost quite a bit more.

jharvey wrote:

davebmw wrote:

davebmw wrote:I think clever use of PWM with an active flyback circuit could provide serious benefits to opening times, but how do we test and prove this is actually working before attaching to an engine?

I don't like the term flyback. I've only seen that in MS land, and I feel it shows electrical ignorance. This circuit shunts current, so I recommend we call it ISC for Inductive (or Injector) Shunt Circuit. I think that will help separate the Engineers from the mechanics. Thoughts?

Sorry to all of you who I have offended by using the term Flyback, I didn't realise that the term Flyback was rude in any way or that Flyback as used in the MS design was patented.

In my experience this flyback circuit should actually be called active power factor control, when it is used on large industrial motor inverter drives it recycles the back EMF and damps the reverse currents from the load back into the smoothing caps.
The same design is used in most switching PSU's these days as it makes a huge difference to the current Vs Voltage relationship in the supply cables to the installation.

The other term which is more applicable i think due to its wide spread use in DC circuit designs is Synchronous Rectification, this is used (usually on chip or driver controlled) to re use the Back EMF from stepper motors.

Flyback is a term that should only be used (I feel) when you are dealing with:
a: An inverting power supply that uses the back EMF from a charged inductance and a charge pump in the form of diode and cap to produce higher or negative voltages as a result
or
b: the term given when the scan line in a CRT hits the end and flys back to the start position of the next line on the phosphorus screen to begin another scan.

OK so we have a couple more choices:

Active power-factor control APC
or
Synchronous rectification SR

davebmw wrote:I have an idea which involves 2 reasonably matched injectors on a common rail supplied with an inflammable fuel substitute spraying into 2 graduated beakers both squirting at the same time for the same duration, but one is PWM primed and the other standard 0 and 1. The pre primed should open sooner and over a period of time fill the beaker quicker. During this experiment it may be worth measuring the delta T in the 2 injectors and collected liquid to make sure the temperature rise is not going to cause heating of the fuel or intake manifold which could cause variation in Intake air density and fuel atomization rate.

I think that sounds like a great jig, is it feasible for you to build it? I don't know your soft background, but I'm sure you could get plenty of help making that kind of a jig.

I have got a couple of injector rails around here although i don't know whether they are low or High Imp, I also have a spare fuel pump, fuel pressure regulator and a gallon of paraffin which whist still flammable is a bit less explosive than petrol/gas!
I only have an MS2 which will probably require more mods than its worth, to drive as required i may as well knock up something on some strip board.
I will start playing with some ideas today.

Sorry all I deleted one of the HTML quote markers in the above post.

Some really good v2.0 ideas there guys! :-)

jharvey wrote:I think you mean figure 17 from http://www.fairchildsemi.com/ds/IS%2FISL9V2540S3ST.pdf

No, those are test circuits for measuring component behaviour with your load.

That part number is an ignition driver, NOT an injector driver, these are less suitable and more expensive than the appropriate FET parts.

it shows .01 ohm as a shunt resistor.

Nope, no such part in fig 17. No point looking at the other diagrams either as it's an ignition driver and they are test setups anyway.

If we're going to hope for current sense in the first freeEMS board

I'm hoping that you don't try and do that :-) For non P&H injectors we don't want or need the extra complication.

then I should probably recommend we add space for a hall sensor. Much less complicated hardware wise and costs about the same dollar wise.

Less complicated than what? Hall sensors interface directly with the CPU (via basic diode/resistor/capacitor protection and filtering) and live in the engine bay, not on the board. For Hall, not a lot needs to be done, for VR we need a conditioner, but that doesn't belong in this thread.

The appropriate FET parts that I posted part numbers for have INTERNAL current sensing for protection purposes only. If the board builder uses FETs that are sized appropriately, (5A per channel for a single 8 ohm injector should be heaps) then there is nothing you need to do for it to work close to perfectly.

Good call on the circuit naming BTW.

As dave said, the pintle movement is the vast majority of the opening time, so very little is to be gained by increasing complexity of software an order of magnitude to do pwm before and after with a peak in the middle in a 4 phase system. That is only applicable to P&H injectors anyway which I don't think we should be trying to support with this bread and butter v0.1 hardware.

I think what I need to do (as this sort of discussion comes up quite often) is create an advanced section for discussion of features for future versions to help keep this section focused.

Cheers for your input :-)

Fred.