Please view this spice file.. (fuel injector driver amp)

Compare the current flow and power losses, as presented. Why are the top drive components even present?

You can't model a electromotive element as a simple inductor or transformer. During the initial or final periods of energisation the magnetic structure changes.

If you've got to dump energy somewhere, use a resistor if you can't clamp and recycle it. They're cheaper and can be more reliable.

RL

snip

where is it going to be used? it is far too slow to work in a car engine

do you have the option of driving both ends of the injector or are you limited to the low side, with the high side connected to the battery?

-Lasse

Slow? how so? I was doing 1khz with no problems, I 'll try a faster test later.

I am not limited to how the coil is connected, I am going to connect the other end to a bridge so that I can switch the poles on coil collapse.

jamie

The usual problem driving fast solenoids is dumping the stored energy; for that, you need a high reverse/flyback voltage. One complication is that most solenoids change their magnetic path when the armature is seated, making the closed solenoid have a lot more inductance than the un-energized one. You could test for that. For serious speed, you need to take that into account.

You can dump the stored energy into a dissipative element, a resistor or a zener, or you can recover it somehow, like using a full bridge driver.

Why not just use a power fet and a zener clamp, or buy a full-bridge driver chip?

John

On 11/25/2011 6:32 PM, Jamie wrote:

Here ya go:

Version 4 SHEET 1 880 680 WIRE 320 -192 -272 -192 WIRE 592 -192 320 -192 WIRE 608 -192 592 -192 WIRE -272 -144 -272 -192 WIRE 320 -96 320 -192 WIRE -272 -48 -272 -64 WIRE 32 16 -144 16 WIRE 320 16 320 -32 WIRE -144 32 -144 16 WIRE -144 32 -176 32 WIRE 592 32 592 -192 WIRE 320 192 320 80 WIRE 352 192 320 192 WIRE 592 192 592 112 WIRE 592 192 352 192 WIRE 656 288 528 288 WIRE 528 320 528 288 WIRE 656 336 656 288 WIRE 32 384 32 16 WIRE 160 384 112 384 WIRE 320 384 320 192 WIRE -176 432 -176 32 WIRE -176 432 -304 432 WIRE 160 432 160 384 WIRE 256 432 160 432 WIRE 528 432 528 400 WIRE 656 432 656 416 WIRE 320 512 320 480 FLAG 320 512 0 FLAG -272 -48 0 FLAG -304 512 0 FLAG 352 192 MIN FLAG 528 432 0 FLAG 656 432 0 FLAG 672 16 Injector_coil SYMBOL voltage -272 -160 R0 WINDOW 123 0 0 Left 2 WINDOW 39 24 44 Left 2 SYMATTR InstName V1 SYMATTR Value 12 SYMBOL Misc\\signal -304 416 R0 WINDOW 123 0 0 Left 2 WINDOW 39 0 0 Left 2 SYMATTR InstName V2 SYMATTR Value PULSE(0 10 0 100u 100u 8ms 10ms 1000) SYMBOL npn 256 384 R0 SYMATTR InstName Q3 SYMATTR Value FZT849 SYMBOL res 16 400 R270 WINDOW 0 32 56 VTop 2 WINDOW 3 0 56 VBottom 2 SYMATTR InstName R1 SYMATTR Value 500 SYMBOL ind2 576 16 R0 SYMATTR InstName L1 SYMATTR Value 10mh SYMATTR Type ind SYMATTR SpiceLine Rser=2 SYMBOL ind2 512 304 R0 SYMATTR InstName L2 SYMATTR Value 100mh SYMATTR Type ind SYMBOL res 640 320 R0 SYMATTR InstName R2 SYMATTR Value 1k SYMBOL zener 304 -96 R0 SYMATTR InstName D1 SYMATTR Value 1N5378B SYMATTR Description Diode SYMATTR Type diode SYMBOL diode 336 80 R180 WINDOW 0 24 64 Left 2 WINDOW 3 24 0 Left 2 SYMATTR InstName D2 SYMATTR Value MUR460 TEXT -328 -216 Left 2 !.tran 0 .100 0 startup TEXT -408 -176 Left 2 !K1 L1 L2 .9

The energy is being clamped through the amp output when attempts to exceed input bias reference exist.

I have both poles designed in the circuit so the coil can be connected either way and also by doing so, any unexpected voltage on the output is going to force either the top or bottom to clamp as required to maintain it.

For example. If I have a reference of 5 volts on the input and lets say the output device or anything attached attempts to be higher than 5 volts it'll force the bottom pole on to drain it off. If the output was lower than 5 Volts, the top pole(rail) components will then turn on to make up for loss.

So as far as magnetic collapse energy coming back, if it does not match with the input reference it will be corrected.

I've already examine some type EV6 and throttle body injectors with a scope to monitor current and kick back using a basic MOSFET low side drive with and with out diode clamp. What I saw with both was a slight current curve ripple with initial first cycle pulse and I can only assume this is due to the inner components interacting with the magnetic core material as it moved in place. The only useful information that lead to this was the possibility of detecting pintle movements which for now is not at the top of the list.

What I saw with the scope and a basic equivalent inductor was close enough to be used as a model in a basic sketch.

We will be metering Xylene through the injectors. It looks like a throttle body type may work the best. With that I can simple do the initial open and then meter it with PWM to maintain it. The test we did with Kerosene seems to confirm that we should be doing a first full open and then following pulses are current monitor and can be regulated via PWM.

The final circuit is going to be monitoring current for max magnetic current peak. This is going to allow a automatic parameter to indicate what the base frequency of the inject should be and once that is calculated the PWM will be a window of that range for a variable maintainable control. This will be done on the start of each initial use of the injector. I did this to help in the event when the inject gets warm the variation of core material and pull in time will auto adjust, hopefully.

I most likely will have a small uC behind this all.

Jamie

The problem is current, while a chip driver can handle the small ones we have a throttle body type that seems to require a lot more current and the energy store is massive compared to the Bosch EV6 manifold types

Our first test model on the bench does use a MOSFET on the low side with a wheeling diode on the coil. It works fine but I found the diode getting rather hot with the throttle body type, due to the very inductive core it has..

With the design of that amp I did, it does drain off the energy very nicely. I picked an inductor that matches closely what the injector does. If you watch the analysis of the circuit and view the current, you'll see the amp soaking up the back current and I don't need a wheeling diode.. The reason I went this way is I may decide to have a long decay on the pulse so to let the inductor collapse slowly.

We'll be metering xylene and the reference for each cycle of spray will be coming from an encoder pulse as a one shot, in case the machine stops on a tooth and from there, it's going to be a time period of injection depending on the LBS/Ph demand.

I am at home for the time being and I didn't bring any injectors home with me for electrical testing so I am relying on the measurements I took of the coil at work and using a inductor with the series R that matches it.

Jamie

I think you better go back to the drawing board on that one? did you see the analysis on that?

The back energy is up to 120V, the low side sink transistor is only rated for 30V.

If I go in there with the proper clamp, the coil then looks like my amp analysis except its being handled in the amp and not a wheeling diode.

Jamie

Is that supposed to be a trick question?

I know. I was not trying to do your design for you. Just give you some direction.

So, are you telling me that you want the *average* current through the coil to be 1.5 amps with the peak at 2.4 amps and the valley to be .95 amps? If so, then you're there.

Note that Q1, Q2, and Q5 are doing absolutely nothing. You can take them out and not change the circuit operation. I think RL pointed this out earlier.

Note that you are dissipating 13W in Q3.

That make you happy?

I think you better go back to the books.

Energy is in joules (or ft-lbs, or some such) but not in volts.

Some transistors are rated for more than 30V. You don't have any that can take more than that?

I have no idea what you are trying to convey in your last sentence.

Are you trying to do the saturation or the peak-and-hold? Either way, your circuit sucks big time.

Sure, but I wasn't planning on using those any way. Just the first thing in the list that look good.. I still may end up going the MOSFET and large clamp diode way, but we'll see. We do have a large count of TIP3055 that will do the job just fine.

We have been thinking of generating a half trapezoid wave on the decay side for the input reference, one that we can adjust.

Jamie

if you say so, it does work. I manage to find some parts at home base here to load test it using a DC solenoid coil. The coil isn't an injector but a spring loaded unit actuator so I can experiment a bit. I am getting ready to set up the AB-GEN for a experimental wave form I have in mind.

Have a good day. As always.

Jamie

Oh, okay. So it works. Then why have you asked for a review? Did you just want approval for your crap? Don't expect it here. Go away.

On 26 Nov., 18:51, Jamie wrote:

es

the point is that with that very slow rate of decay you have no dynamic range on the pulswidth, and on top of that it will be variable with pressure and everything else at exactly what current it closes

try something like this: (the highside driver is a bit slow but it gets the point across)

Version 4 SHEET 1 2440 680 WIRE 1536 -496 1120 -496 WIRE 1296 -224 1040 -224 WIRE 1536 -224 1536 -496 WIRE 1536 -224 1376 -224 WIRE 1536 -144 1536 -224 WIRE 1808 -144 1536 -144 WIRE 1536 -128 1536 -144 WIRE 1536 -128 1440 -128 WIRE 1040 -48 1040 -224 WIRE 1184 -48 1040 -48 WIRE 1808 -48 1808 -64 WIRE 1808 -48 1440 -48 WIRE 1808 -32 1808 -48 WIRE 1040 32 1040 -48 WIRE 1184 32 1104 32 WIRE 1584 32 1440 32 WIRE 1584 48 1584 32 WIRE 1760 48 1584 48 WIRE 1808 112 1808 64 WIRE 2048 112 1808 112 WIRE 1040 128 1040 96 WIRE 1104 128 1104 96 WIRE 1312 128 1312 80 WIRE 1808 208 1808 112 WIRE 1808 400 1808 288 WIRE 1456 480 1248 480 WIRE 1760 480 1536 480 WIRE 1248 512 1248 480 WIRE 1248 608 1248 592 WIRE 1808 608 1808 496 FLAG 1120 -416 0 FLAG 1312 128 0 FLAG 1808 608 0 FLAG 1248 608 0 FLAG 2048 176 0 FLAG 1104 128 0 FLAG 1040 128 0 SYMBOL PowerProducts\\LT1910 1312 -48 R0 SYMATTR InstName U1 SYMBOL voltage 1120 -512 R0 WINDOW 123 0 0 Left 0 WINDOW 39 24 44 Left 0 SYMATTR InstName V3 SYMATTR Value 12 SYMBOL nmos 1760 -32 R0 SYMATTR InstName M2 SYMATTR Value IRL3915 SYMBOL res 1792 -160 R0 SYMATTR InstName R3 SYMATTR Value .025 SYMBOL voltage 1248 496 R0 WINDOW 123 0 0 Left 0 WINDOW 39 24 44 Left 0 SYMATTR InstName V4 SYMATTR Value PULSE(0 12 1m 1u 1u 50m 200m 1) SYMBOL ind2 1792 192 R0 SYMATTR InstName L3 SYMATTR Value 10mh SYMATTR Type ind SYMATTR SpiceLine Rser=3D2 SYMBOL schottky 2064 176 R180 WINDOW 0 24 72 Left 0 WINDOW 3 24 0 Left 0 SYMATTR InstName D1 SYMATTR Value MBRS360 SYMATTR Description Diode SYMATTR Type diode SYMBOL nmos 1760 400 R0 SYMATTR InstName M1 SYMATTR Value irl3915mod SYMBOL cap 1088 32 R0 SYMATTR InstName C1 SYMATTR Value 5n SYMBOL res 1392 -240 R90 WINDOW 0 0 56 VBottom 0 WINDOW 3 32 56 VTop 0 SYMATTR InstName R4 SYMATTR Value 10k SYMBOL zener 1056 96 R180 WINDOW 0 24 72 Left 0 WINDOW 3 24 0 Left 0 SYMATTR InstName D2 SYMATTR Value BZX84C6V2L SYMATTR Description Diode SYMATTR Type diode SYMBOL res 1552 464 R90 WINDOW 0 0 56 VBottom 0 WINDOW 3 32 56 VTop 0 SYMATTR InstName R1 SYMATTR Value 1k TEXT 1632 -560 Left 0 !.tran 0 .100 0 startup TEXT 1984 504 Left 0 !.model irl3915mod ako: irl3915 bv=3D55

-Lasse

me

oes

e.

the stroke on an in jector is very small so I don't think it changes that much, though on some you can see see a small wiggle in the waveforms when it opens and closes

to get any kind of accurate metering of fuel it needs to dump the current fast so the closing time it repeatable

that how all the car manufacturers do it, a temperature/current protected fet with a clamp voltage at something like 36/42/55 volts

-Lasse

You're so funny, we've already been to whole crap load of sites on this subject, but thanks any way.

Btw, if you read that page you so posted, you'll find that I've been telling you about what we had to do with our project as far as doing the Peak+Hold or just high-Z units, also if you remember, I did talk about the affects being shown in the wave forms as the mechanics inside move into place.

We are most like going to use a throttle body type which is a Peak and hold, this means I either need to use that amp circuit so I can drop the output voltage at the initial open or use PWM and have a long start pulse followed by short ones to maintain hold, that I could do with the MOSFET if I go PWM. Either way, dealing with the energy is hump I am getting over with.

Have a good day.

Jamie

That is an interesting idea, I can do that with antique parts using a dual timer. one for the holding pulse stream and the other a one-shot for volume measuring that needs to be adjusted of course.

We will be using an encoder that can ramp the volume as the pulse rate increases. I think I can then adjust the one shot as the ratio mix.

I will be fully clamping the back voltage. It will actually be an aid to hold the valve in a steady position at lower current rather than a problem. Injector to be used needs a high current pulse then it idles back to about 1.5 amps or so. That is where I need the pulse if I go the full switching way :)

Lots interesting. Thanks.

Thanks,.

I was only interested in the opinion of how I decided to handle the back energy which I can meter that way. For example, Lets say I will allow it to ramp back up to like 30 volts. I can use a voltage divider network in the off stroke to set the clamp level with that circuit..

Have a nice day,.