Peak-Hold Fuel Injector Voltage Clamp

[CaptainCarp]

View attachment OUTPUT (1).pdfI am designing a peak-hold fuel injector driver and need some suggestions about modernizing a sub-circuit.I need to have a means of "clamping" the injector to BATT during PWM of the "hold" current (a controlled "clamp").
I incorporated a "clamp" circuit that I found in an old injector driver datasheet but I am not satisfied that I couldn't do better with a power MOSFET topology. I don't like the amount of power being dissipated in Q4 during pwm (1A x ~2V) and was hoping that someone here could suggest a differnt topology. Thanks

 

[Genomerics]

View attachment 52573I am designing a peak-hold fuel injector driver and need some suggestions about modernizing a sub-circuit.I need to have a means of "clamping" the injector to BATT during PWM of the "hold" current (a controlled "clamp").
I incorporated a "clamp" circuit that I found in an old injector driver datasheet but I am not satisfied that I couldn't do better with a power MOSFET topology. I don't like the amount of power being dissipated in Q4 during pwm (1A x ~2V) and was hoping that someone here could suggest a differnt topology. Thanks

That's the sort of circuit that can hurt your head. I assume Q1 is one of those all singing all dancing 'special' protected Mosfets that includes limiting functions one of which will be a clamp to limit the Drain voltage and you are using that during injector reset..

It looks strange because the PNP transistor appears to be inserted the wrong way up but during reset its Emitter is taken positive so turning on Q3 will turn on Q4 and clamp to Vbatt as you expect.

I would think that it is probably feasible to substitute Q4 with a P-channel Mosfet. You will still need D2 in place to block the Body-Source diode but otherwise that should not affect the circuits operation.

Some problems you might run into relate to protecting the Mosfet, switching speeds and managing to effectively turn off the device... Try this,



Q1/Q2 and Q3/Q4 are current limited level shifters. Q1/Q2 will deliver about 50mA with Q3/Q4 delivering about 6mA. Adjust the Emitter resistors to adjust the current levels.

The first set will turn on the upper Mosfet, M2. The second pair will turn on Q5 turning off the upper Mosfet. Its gate is protected by the 15V zener diode. R6 is, nominally, there to damp local oscillation in M1.

Having drawn it I could put some 'real' components in and Spice it to see what happens but it seems reasonable to me..

Genome..

 

[CaptainCarp]

Thank you Genomerics

It will take me some time to understand the circuit details and hopefully get back to you with some resonable questions.

 

[CaptainCarp]

Genome -
I think I am ready to ask you some questions (please forgive my ignorance).

1. Why turn on the M2 mosfet with a current limited source (why can't I get rid of Q2 and connect the Q1 emitter to gnd?)?

2. Under what scenario is the zener protecting the gate of M2?

Thank you

p.s. I have just installed LTspice and was wondering where you obtained the battery component. I can't seem to find it under components.

 

[Genomerics]

Genome -
I think I am ready to ask you some questions (please forgive my ignorance).

I did not notice you had any. It's probably me that has little knowledge of your subject.

1. Why turn on the M2 mosfet with a current limited source (why can't I get rid of Q2 and connect the Q1 emitter to gnd?)?

Because you are dealing with an automotive environment then effectively voltage levels are not guaranteed. That might raise a question about what happens when your battery is heavily loaded.. This will be my inexperience.

Otherwise..

Depending on the device you choose then Mosfets need a certain gate voltage to fully turn them on. There is also a maximum voltage you can apply to the gate before you blow them up. That's why the Zener Diode is there.

A 12 Volt battery is probably good to go but in order to fully enhance the Mosfet you will need all of those volts which means you can't place a series resistor in the collector of Q1. It would reduce the available drive current to what might be effectively zero.

When 'nasty' things happen, load dump or something, you may well end up with 'Vbatt', at your circuit location well above the expected 12V at which point/time you would apply it across the gate of the Mosfet and kill it. Either that or if the driver was not current limited it would get stuffed via the Zener Diode.

Since the driver is current limited then under over voltage conditions Q1 will 'take up the slack' or difference between the 15V Zener and any over voltage.

Otherwise it is a 'nice' thing because it guarantees you that drive current to the Mosfet. I'll admit it is not large but it should be sufficient to switch the device quickly enough. That will be my inexperience again.

2. Under what scenario is the zener protecting the gate of M2?

As suggested, given an automotive environment, then voltages are not guaranteed and there is the real chance that you may end up exceeding the Vgs, Gate to Source, rating of the device. The Zener diode, in conjunction with the current limited drive, will... hopefully prevent that from happening.

Thank you

p.s. I have just installed LTspice and was wondering where you obtained the battery component. I can't seem to find it under components.

As far as I know the Battery Component is nothing special. It is just a graphical representation of a voltage source. Otherwise it lives under 'misc'.

Genome

 

[Genomerics]

Since you have grabbed LTSpice here you go,

* C:\inj\injhold.asc
Q1 COLL1 N004 N006 0 NPN
Q2 N004 N006 0 0 NPN
R2 N006 0 12R
R1 N004 P001 1K
Q3 COLL3 N005 N007 0 NPN
Q4 N005 N007 0 0 NPN
R4 N007 0 1K
R3 N005 P002 1K
A2 PTO 0 0 0 0 0 P002 0 BUF Vhigh=5V Vlow=0V
A1 PTO 0 0 0 0 P001 0 0 BUF Vhigh=5V Vlow=0V
M1 N002 N003 VCLMP VCLMP FDS4885C_P
D§Z1 COLL1 VCLMP ZID
D1 N002 N001 DID
R5 N003 COLL1 22R
I1 N001 VCLMP 1A
Q5 COLL1 COLL3 VCLMP 0 PNP
R6 COLL3 VCLMP 4K7
VBATT N001 0 24V
D2 0 VCLMP ZCL
V§PTO PTO 0 PULSE(0 5 0 10n 10n 5m 10m)
.model D D
.lib C:\Program Files\LTC\LTspiceIV\lib\cmp\standard.dio
.model NPN NPN
.model PNP PNP
.lib C:\Program Files\LTC\LTspiceIV\lib\cmp\standard.bjt
.model NMOS NMOS
.model PMOS PMOS
.lib C:\Program Files\LTC\LTspiceIV\lib\cmp\standard.mos
.MODEL ZID D(RON=100m ROFF=100K VREV=15V)
.MODEL ZCL D(RON=10m ROFF=100K VREV=45V)
.MODEL DID D(RON=10m ROFF=1E9)
.tran 0 50m 1u uic
.backanno
.end

If you copy and paste the above into a text editor then save it as an .asc file then, hopefully, LTSpice will open it for you.

Being me I have used 'ideal' components as a starting point apart from the Mosfet and I just 'grabbed' one of those from the libraries. I1 is your injector hold current and D2/ZCL clamps it to 45V during 'reset'.

It seems to work but it does show I have a bit of a misconception about the Q3/Q4 level shifter. Apart from transients all it needs to do is supply the base current for Q5 and current through that devices Base-Emitter resistor...

Otherwise

With VBATT set to 12V



The clamp occurs at 12V, that will be plus RDSon of the Mosfet and current through it, with reset at 45 volts as set by D2/ZCL. After the event, since the gate Zener is effectively open circuit, Q1 collector current drops to zero.

If I raise VBATT to 24V then,



Now Q1 is supporting the difference in voltage between VBATT and the Gate Zener, 9V, and its current is limited to about 65mA which was the purpose of implementing that driver in the manner shown.

Q3/Q4 behaves, as suggested, differently. A guess but perhaps in this case a mistake.

Genome

Edit

I might still be bothered about Q3/Q4/Q5..... Or maybe I worry too much.

 

[CaptainCarp]

Hi Genomerics,
For some reason I was not notified of your latest replies. I appreciate all the assistance that you have provided.
I will load your circuit into LTspice and "play" with it and let you know what I come up with.
Thank you

 

[Genomerics]

No worries. Thanks for saying thanks.

Genome.