injectors - [56k beware page 2]

[gearhead]

Delta,

There are quite a number of 'logic level' fets which are 'on' at significantly below 5V.

The IRL series from International Rectifier are all logic level with turn on voltages about 2V. Examples are IRLU110 and IRL540. The protected 'Omnifets' from ST are also 'nice' in this way. The Vth for a VND5N07 is 3V max. Very nice that we do not have to use a FET driver to turn them on.

Gearhead

[Delta]

Having a logic level fet will infact fix the driving problem and hence put it in switching mode, but none of the other problems will be fixed by this. I don't have a 540 in LTspice so I can't check, so I'll just assume it works. Looking at the datasheet they basically have a rail to rail opamp built in, so you could do without the first op amp in my circuits. I'm going to keep having a look, so far i can see that the Zener will definitely burn. Also in the graphs above, the solenoid current isn't shown. The sharp rise in voltage which is being clamped is actually during the off time not the one time, and is not a good thing. (EDIT - its always going to happen but we want to shunt it a quickly as possible in the safest way possible - sorry should have made that clear) Solenoid opening is a function of current through it. You will see if you check currents in various parts of the circuit that all is not well. (EDIT - On that note, check what happens if you short circuit the injector)

[Fred]

The other thing you have neglected is that you are driving your fets with only 5V while their Vds is at least 12v, this will put it in the triode region and hence it will act as a voltage controlled resistor, and hence disipate a fair amount of power.

That is fine with the auto protected FETs, they pretty much all have TTL inputs on them instead of the standard gate arrangement. That's why we want to use them over ordinary FETs. That and they are indestructable in almost every way. It makes for a very simple setup if you choose them over normal ones.

EDIT : missed those posts... They cover all bases really :

• TTL input
• Output clamp @ 50V+ to replace the zeners and other items
• Thermal overload protection
• Short circuit protection

In summary, they are awesome and you almost can't kill them and don't need much/any circuitry around them to have them work well.

Fred.

[Delta]

They are most definitely NOT indestructable, yes if they overheat they will shutdown - what will happen is that you won't have you injectors being driven anymore. What you want is something that stays in the safe opperating area no matter what. The other oversight (which is easily corrected) is that you don't use a zener for controling an inductive spike, a freewheel diode is a much better method as the device is in its forward mode rather than controlled breakdown. The zener in your circuit has to conduct 600mA at 17.5-18V volts, which is a lot of wattage. If it fails then you will have over 1kV across the Mosfet, which will put it in its avalanche region, and protected or not, it will eventually fail under those conditions.

[Fred]

Can you do some calculations that show that they will even come close to dissipating enough heat driving simple High Z injectors without external parts? I just can't see it really happening.

There is no Zener involved either, the voltage clamp protection works internally by switching the FET ON to hold the Voltage at an acceptable level. I believe this is independent of the overheat functionality which just disconnects the drive from the gate. Thus, once the temp gets up it removes all means for temperature to become present.

PS, none of those circuits are mine. My circuit is one resistor, one CPU, and one auto FET.

Fred.

[Delta]

Can you do some calculations that show that they will even come close to dissipating enough heat driving simple High Z injectors without external parts? I just can't see it really happening.

There is no Zener involved either, the voltage clamp protection works internally by switching the FET ON to hold the Voltage at an acceptable level. I believe this is independent of the overheat functionality which just disconnects the drive from the gate. Thus, once the temp gets up it removes all means for temperature to become present.

PS, none of those circuits are mine. My circuit is one resistor, one CPU, and one auto FET.

Fred.

Thats right the voltage clamp works by turning on the gate - but not fully - it puts it in the triode region functioning as a resistor to shunt the current required to lower the voltage. This will seriously heat the device. Also, since this requires time to turn on (internal capacitance) the device momentarily goes into avalanche which is less damaging in a MOSFET than a transistor but still over time does stuff the device up. The resistor + cpu + autofet will function basically the same as the circuit I have here - but you MUST use a freewheel diode across the inductive load. Its even in the datasheet. My circuit pre-dates auto fets, and basically has all the same circuitry as one, except externally, so you can set the current limit etc however you like.....I suppose it all depends on pricebreak and how much space you want to save, the circuit I use has 3 op amps in it and 2 mosfets. The MOSFET's take up space, the opamps are in a quad package. The MOSFETS are like .90c and the quad opamp is around the same from pretty much anywhere in any number, so its only a few dollars. Autofets - depends where you get them from and how many, but most are over $5. So its a trade off between cost, adaptability, and space. cheap autofets won't have the current capabilities to run Peak and Hold injectors in future, but save space now.....

[Fred]

I understand all of the above, but "seriously heat the device" seems to me to be an exaggeration. Can you qualify it some how to make it stick in my mind? We are talking about a single high Z injector per FET. During operation the injector itself will not heat the device significantly (pretty much at all). Which leaves the inductive kick back having to heat a whole FET from cold to "overheated" while attached to some form of heat sink.

I just checked and neither the philips nor the vnp series show any diode in use in the datasheets.

Obviously I want the design to be right (even this preliminary one), but at the same time, I don't see a problem that needs solving.

Fred.

[jharvey]

Hmm, heat requirement. How much are we talking about. From page 3, I see Vcap is at 12V, and the other side of the injector is about 17.5V for about 2.5ms. The current goes from 1 amp to about 0 amps in this time. So a voltage difference of about 5.5V and an average current of .5 amps. So a power of about 2.75 watts for about 2.5 ms in a period of around 100ms, so an average power of about .07w... Less at a 200ms period... That can't be right. I'm sure the inductor(s) must hold more energy than that.

Fred can you make a quick bread board test circuit with the PIP, hold the PIP and measure how long it takes until you think it's hot? Try a 10ms on duty, and 100ms period for starters. Careful 50V is considered low voltage, but if something goes wrong it can bite ya. Should be much less then a spark plug however.

Can you also do a current and voltage measurement across an injector? Perhaps the spice is wrong in the discharge time, voltage, current.

I expect most of the heat will be from inductor discharge, and small amount from the on time transition. If the power is really around .07 watts, I don't think heat sinking will be much of an issue. My power guess has to be wrong.

[Delta]

I suppose being an electronic engineer I tend to look at the worst case scenario hahha, I'm thinking along these lines - what happens when you have all your fets driving injectors on a common heatsink, in a case with no airflow, in a kick panel of a car on a 50 deg day (I'm an aussie), and its being thrashed around the track. There is already the switching events to cause heating (although they aren't really a big deal - but more revs = more heat) however the inductive spike is actually quite large. In the simulator I see 9W ramping down as the EM field collapses in a low Z injector and just under 2W with high Z. This is through a freewheel diode where the current is dumped to positive and only a few volts are across the diode. If you don't have the diode the auto clamping of the fet will allow the voltage to rise to 40-50V and it will dump across the fet to ground.

Lets put it this way....the done thing is to put a freewheel diode across any inductive load that you don't want a spike on. If one is missing, then the fets must be designed to acommodate the extra heating and loading for all circumstances. I'de prefer to install a 10c part rather than buy a $1 more expensive fet, and risk its failure.

[jharvey]

Have you noticed the discharge times for the spice on pg 3 of this thread? The autoFET drops like a rock, because it drops so quickly, the off time tolerance is tighter, so more accurate control.

I know that freewheel (snubber, fly-back, suppressor, catch, TVS, ect) component is very common. However it really rolls over from the relay days, when there were no other options. I think the autoFET shows newer tech, increasing quality. If the dissipated power for the autoFET is really a fraction of a watt, that would be something. I think there is a good chance it really is that low. Most FET's get hot because this spike can put you in a tranient zone, where you're simply a resistor across the battery. However, this isn't a normal FET, it appears it will make that transient period much shorter then normal.

I'll be eagerly waiting for Fred's reply.