Smart questions
Smart answers
Smart people
Join Eng-Tips Forums

Member Login

Remember Me
Forgot Password?
Join Us!

Come Join Us!

Are you an
Engineering professional?
Join Eng-Tips now!
  • Talk With Other Members
  • Be Notified Of Responses
    To Your Posts
  • Keyword Search
  • One-Click Access To Your
    Favorite Forums
  • Automated Signatures
    On Your Posts
  • Best Of All, It's Free!

Join Eng-Tips
*Eng-Tips's functionality depends on members receiving e-mail. By joining you are opting in to receive e-mail.

Posting Guidelines

Promoting, selling, recruiting, coursework and thesis posting is forbidden.
Jobs from Indeed

Link To This Forum!

Partner Button
Add Stickiness To Your Site By Linking To This Professionally Managed Technical Forum.
Just copy and paste the
code below into your site.

Helpful Member!  CaptainCarp (Automotive) (OP)
17 Jan 11 21:18
I 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

dgallup (Automotive)
18 Jan 11 14:42
Easiest thing to do is just use a National Semiconductor LM1949 injector driver as long as you are happy with a 4:1 peak:hold current ratio.
CaptainCarp (Automotive) (OP)
18 Jan 11 22:09
Thank you dgallup for your reply, but I am aware of the LM1949.  It is a very old part (not even available in surface mount) and really is intended for linear operation and not pwm.  The transistor that it controls will act like a variable resistor and dissipate about 12W during the injector "hold" period (remember, I am trying to keep dissipation down).  Thanks again.
murpia (Mechanical)
19 Jan 11 8:04
If you do not want to dissipate any power linearly, I would suggest implementing a half-bridge with the injector connected between VBatt and the bridge.

If you turn on the high side FET whenever the low side is off (with shoot-through protection) the injector current can be easily adjusted.

Regards, Ian
dgallup (Automotive)
19 Jan 11 10:52
Pages 9 & 10 of the LM1949 data sheet show how to use it as a switcher.  It may be old but it works very well.  Be sure to EMI test your home grown circuit, I've seem some of them go absolutely haywire in a noisy environment.  Never had a problem with the LM1949.

IRstuff (Aerospace)
19 Jan 11 11:47
Why not a programmable buck converter that switches between 8V and 2V instead of running directly to the battery, which has quite a bit of variance, anyway.  Then you can run your transistors into saturation and not burn excess power there?

The way you appear to have it wired, you are burning either 48W, or at least 12W, just to get 32W, or 2W.  With a decent converter, you'd be looking at around 90% efficiency.


FAQ731-376: Forum Policies
Chinese prisoner wins Nobel Peace Prize

CaptainCarp (Automotive) (OP)
19 Jan 11 19:20
murpia and IRstuff,
I am looking to update a design that works fine with only a couple of transistors and a coupe of resistors, the only issue is that the Darlington will be have about 2V across it when it is on (only used in saturated mode).  I am looking to update the design by using a P-MOSFET based circuit to replace the Darlington based circuit.  Your proposals don't seem to by me anything other than more parts.  The Darlington circuit's function is to provide a re-circulation current path for the injector during the PWM "off" time - that's all.  Note that the suggestions are appreciated.

While the data sheet may show the LM1949 being used in a PWM mode, it is not as efficient than more modern designs using power MOSFETs (the Darlington has a saturation voltage of about 2V).  Plus the circuit shown will always clamp the flyback voltage at a high voltage, which decays current flow in the injector faster (great if you're trying to turn the injector off, not so great if you're trying to PWM it.  Thanks

murpia (Mechanical)
20 Jan 11 5:47
I count one less transistor in a half-bridge compared to your scheme.

Most modern PWM outputs from microcontrollers can do centre-synchronised PWM (useful for motor control). If you can do that too you don't need your recirc diode D9 either.

Regards, Ian
CaptainCarp (Automotive) (OP)
20 Jan 11 8:58

Perhaps I was too quick to discount your suggestion.

Can you please detail your suggestion a bit more so that I can "see the light" (one less transistor is always better, unless it costs me $$$ to get rid of it).

I would appreciate it if you could identify an application note (or schematic) to help me understand your suggestion and if you had a part recommendation, that would be great.
Thank you
LionelHutz (Electrical)
20 Jan 11 9:46
I really don't get the point of Q4 at all for an injector driver.

Why not dump the Q4 circuit completely and just rely on the D9 diode when Q1 is off. Use a Schottky diode to achieve a low voltage drop, the lower the better.

I don't see any point in using a traditional 1/2 bridge for this application because the diode across the upper FET will be carrying the current, not the FET itself.
CaptainCarp (Automotive) (OP)
20 Jan 11 11:14
LionelHutz -
The reason for having Q4 (and the overall reason for this posting) is that I have a couple of conflicting requirements for driving the injector during the "injection" event.

During the PWM, 1A current "hold" time, I want the current to recirculate through the injector for as long as possible so that the PWM frequency and duty cycle can be minimized.  Since the current decay time during recirculation is dependant on the voltage across the inductor (the smaller the voltage, the longer the decay), having a diode across the injector is great during current "hold" time.

Unfortunately, when I need to turn the injector "off", I need the voltage across the inductor to be large so that the current decays "now" as injector timing is critical.  So I need to be able to turn off my "diode" and let the flyback voltage build up to a higher clamp voltage.  I am doing this by using a "protected" Q1 mosfet, which has a built in voltage clamping function (I hope to use one that clamps at around 60V).

Good question, hope I provided a reasonable answer.  
murpia (Mechanical)
22 Jan 11 16:59
OK, this is basic stuff and I'm not going to hold your hand....

Typing 'half bridge' into Google images provides 21,000,000 hits of which numbers 2 and 5 immediately show what I'm on about.

If you drive both MOSFET gates from your microcontroller you have eliminated Q3 and D9 from your schematic.

Regards, Ian
CaptainCarp (Automotive) (OP)
23 Jan 11 22:58
murpia -

This is really not "basic" stuff" and I don't want you to hold my hand and lead me down the path of ignorance.

You have not read (or do not understand) what I am trying to accomplish.  AGAIN! - I want to be able to turn on and off, a voltage clamp to BATT, with a minimal amount of parts, and in an elegant / efficient manner.

As LionelHutz already pointed out, an N-MOSFET has an intrinsic Source to Drain diode which means it will always be in circuit and clamp the inductive flyback to BATT - ALWAYS, which is what I don't want.

murpia (Mechanical)
24 Jan 11 8:05
Use an IGBT?

Regards, Ian
LionelHutz (Electrical)
24 Jan 11 9:18
CaptainCarp - now you throw in the desire for a clamp to quickly decay the current. Your first circuit doesn't even have such. Turn off Q4 and there is no current path. Maybe you turn it partially off? At any rate, is the p-mos not working like you want?

murpia - The 4qd circuit description is crap. The diode across the MOSFET is carrying the current they are drawing.
CaptainCarp (Automotive) (OP)
24 Jan 11 10:04
LionelHutz - my first circuit does have this higher voltage clamp function, it was just my poor initial description of the circuit (in an attempt to limit discussion to the active clamp to BATT sub-circuit) that has caused confusion.

My symbol for Q1 shows a device with a PROT section connected to the D, G, S, of MOSFET.  This was my attempt to show a "PROTected" part such as the NCV8405 that has a ~45V clamping function for inductive loads built in.

As for the p-mosfet, I am not a power guy and have never designed a circuit using a p-mosfet.  I do have an idea as to how I would connect it, but my "idea" circuit still needs D9 because of the intrinsic S-D diode of the mosfet, and with no experience, considerations like making sure the G-S voltage limit isn't exceeded (during high voltage clamping) come late to me (I currently am considering clamping this voltage with a zener).  So, I basically was requesting a complete robust, efficient, and elegant active clamp circuit design with my initial question.  Thanks
murpia (Mechanical)
24 Jan 11 16:15


murpia - The 4qd circuit description is crap. The diode across the MOSFET is carrying the current they are drawing.
I agree the description could be better written, but my understanding is the top MOSFET is turned on using complementary PWM to the bottom one, thus recirculating the current through the low resistance of a turned on MOSFET as opposed to using the intrinsic diode.

I'm satisfied the OP's requirement can be fulfilled with a half-bridge of bottom N-channel MOSFET (avalanche rated) and a top P-channel IGBT (avalanche rated), driven by complementary PWM.

Regards, Ian
LionelHutz (Electrical)
25 Jan 11 14:22
The circuit was not clear. I believed that prot block was to protect the gate circuit. At any rate, a n-channel and p-channel should be able to work together to do what you want.

murpia - I'm just saying the 4qd page is crap. It shows 2 n-channel devices and then claims the current will go backwards through the upper device.

murpia (Mechanical)
25 Jan 11 15:36


murpia - I'm just saying the 4qd page is crap. It shows 2 n-channel devices and then claims the current will go backwards through the upper device.
There are many application notes on PWM switching of half-bridges or full-bridges which show current flowing that way through N-channel devices. E.g. the LMD18200 AN694.

The 4QD circuit would need a charge-pump voltage booster to generate the gate voltage to turn on the top N-channel device. Dedicated gate driver chips are available to do that job (and others, like provide shoot-through protection). For this application a P-channel device might be a better bet.

Regards, Ian
murpia (Mechanical)
25 Jan 11 15:39
Also, N-channel MOSFETs conduct 'backwards' in synchronous rectification applications.

Regards, Ian
LionelHutz (Electrical)
26 Jan 11 8:45
Sure, conducting through the diode, not the device...
murpia (Mechanical)
26 Jan 11 10:23
Er, no...

See figure 4.

Regards, Ian
IRstuff (Aerospace)
26 Jan 11 12:42
Not exactly: "Figure 4. If you use the complementary gate-drive approach and the output is loaded lightly, the inductor current reverses during the synchronous rectifier's on time and the next half cycle begins with current flowing backward through the high-side MOSFET (MOSFETs are bidirectional). [bold]During the switching dead time, current flows through the parasitic diode.[/bold]"

Current can flow through BOTH.  However, the substrate is usually quite lightly doped, and therefore presents a relatively high series resistance for any current through the substrate.  There would also be some additional parasitics, depending on the number of substrate contacts and where they're located.  

The channel, while rather thin, is usually inverted to a much higher conduction state than the substrate, and for a power transistor, the contact and other parasitic series resistances are intentionally minimized.  Therefore, when the transistor is biased ON, the channel will present the least resistance path for current flow.  When the transistor is OFF, the substrate will carry the current, albeit at a much higher voltage drop.


FAQ731-376: Forum Policies
Chinese prisoner wins Nobel Peace Prize

murpia (Mechanical)
2 Feb 11 4:14

Did you ever take a decision on trying any new circuit ideas?

Regards, Ian
CaptainCarp (Automotive) (OP)
2 Feb 11 19:09

I have got a little further with how I think I want to proceed.  I have selected the main components (I think) but still have to select some resistor values.  I have installed LTspice and will try to simulate the circuit, but it may take a while as although I know how to "run" the simulation, creating / including components that are not part of the supplied libraries is new to me so I have to learn how to do this.


Reply To This Thread

Posting in the Eng-Tips forums is a member-only feature.

Click Here to join Eng-Tips and talk with other members!

Back To Forum

Close Box

Join Eng-Tips® Today!

Join your peers on the Internet's largest technical engineering professional community.
It's easy to join and it's free.

Here's Why Members Love Eng-Tips Forums:

Register now while it's still free!

Already a member? Close this window and log in.

Join Us             Close