experimental H-bridge design mosfets

Discussion in 'The Projects Forum' started by hobbyist, Sep 14, 2009.

  1. hobbyist

    Thread Starter Distinguished Member

    Aug 10, 2008
    This circuit has major design flaws.

    robot motor controle H-bridge for small hobby motor..jpg

    After doing voltage measurements, I discovered that the top transistors are running linear while the bottom ones are saturated.

    top transistors Vds. around 4 to 5 volts.
    bottom transistors, Vds around 50mV.

    The reason why is the bottom transistors have there gate voltage refrenced to ground, so around 6 to 7 volts are sitting at the gate.

    While the top transistors have there gate voltages refrenced to the load (motor).

    The voltage at the source term. refrenced to ground is 3.34 volts. so 7 volts at the gate
    referenced to ground is only around 3 1/2 volts, refrenced to the source Vgs.

    Remedy attempt #1.

    First bias a mosfet into conduction and take voltage readings at the Vds, and Vgs, and gather data, and determine Vgs, needed for Vds to be around the low mV. region.

    Then take voltage readings across the motor to be used in this design, and gather data, concerning Volt. across motor under normal running conditions.

    Then work at biasing the mosfet on the supply side to have Vds. in the mV. region under motor running cond.

    This may require a higher battery voltage, well see.

    I will continue to add more posts to this thread as I experiment more in this design.
  2. hobbyist

    Thread Starter Distinguished Member

    Aug 10, 2008
    So far it looks like you can't saturate a mosfet on the HIGH side of a load.

    The source is not refrenced to ground so no matter how much the voltage supply goes up the Vds follows along, due to the load at the source.
  3. bountyhunter

    Well-Known Member

    Sep 7, 2009
    You can if you have floating drive. In the H bridge converters we used in our switchers, the gate drive circuitry was transformer coupled for that reason.

    In your case, you could just use P-FETs on the top side drivers and adjust their drive circuitry as required.

    BTW: any reason for the ridiculously huge ten MEG Ohm resistors? That looks like a simulation. If you had tried it you would realize that huge resistance won't drive the gate capacitance of the FETs well and they will switch slower than molasses in winter.
  4. SgtWookie


    Jul 17, 2007
    Here's a couple of hints for you.
    Since you won't typically be using the upper half of the bridge for PWM, it would make things a lot easier if you used P-channel power MOSFETs up there.
    You started out with IRF610 power MOSFETs, which are N-channel. In your library, you'll probably find IRFI9620, which are roughly P-channel complements of the IRF610.

    See the attached for a transistorized low-side driver. Replacing D1 with a Schottky diode will improve performance somewhat.

    The high-side driver can be made somewhat similar, but with fewer components and still be very high performance.

    You are only using 8v for your power source. If you are going to be using an 8v battery in your project, you will need to use logic level power MOSFETs. Your library likely does not have those, but there are equivalent parts you can substitute later for the real thing.
  5. SgtWookie


    Jul 17, 2007
    Sorry Bountyhunter, I didn't see your reply before I posted. I had the thread open for a couple of hours while I was fooling around with a simulator.

    He's using Circuitmaker Student or Circuitmaker 2000.
    He was trying to save battery power, as he explained in the thread where he first posted that circuit, here:
    His reply #17, mine #20, and he mentioned saving power on page 3.
    I told him something similar to what you said, only your response was a lot funnier. ;)
  6. DrNick

    Active Member

    Dec 13, 2006
    Another thing to consider is NOT driving the mosfets with a 1 Meg source resistance. They will turn on really really really slowly, causing massive switching losses...

    Senior Member

    May 26, 2009
    Could you share your voltage measurements, and show us how you got them?

    I'd appreciate it!
  8. Audioguru

    New Member

    Dec 20, 2007
    The datasheet for the Mosfets shows that they switch well when the Vgs is 10V, not 8V and not 4V.
    So why not feed them 10V??
  9. SgtWookie


    Jul 17, 2007
    All of the MOSFETs are N-channel, including the upper two.

    When the gate of an upper MOSFET is increased to 8v, the drain-source resistance drops, and the voltage to ground at the source terminal increases.

    This reduces Vgs (voltage from the gate to the source), so the MOSFET increases in resistance.

    The only way to turn the MOSFET fully ON is to get the gate 10v higher than the source terminal. He would have to use a different technique to get the upper MOSFETs to turn on properly.

    For this exercise, simply using P-channel MOSFETs for the upper half simplifies the problem a great deal.
  10. hobbyist

    Thread Starter Distinguished Member

    Aug 10, 2008
    Thankyou everyone with your replies.

    I will try to answer the questions, one at a time

    Designing mosfet circuits is fairly new to me, from what I've read in the past about biasing mosfets, was that since it is voltage driven and not current driven, then the Zin, can be very high, in the megohm regions,

    so I kind of ran with that, thinking if I want to save battery power, than make the gate volt. divider extrememly high resistance, because it's the voltage that works the gate.

    But now after sgt.wookie explained in good detail, about the gate capacitance, as well as others have said about , lately, I understand now that there is a lot more to designing with mosfets than just applying a gate voltage.

    After reading all these posts, I will gather as much information I can, and dsee if I can design a decent h-bridge from scratch taking into consideration, all that you guys have brought forth.

    Just a little background on my simulators.

    I have the multisim, and also the old old version of circuitmaker, as well as the downloadable student version.

    But when I seen that electronic simulation programs are so expensive nowadats, compared to when they first came out, I decided No more relying on simulator programs, to design circuits, I want to learn how to design simple transistor circuits using a calculator, design notebook, any needed data sheets, test equipment and a breadboard.

    Now I use my simulator for quick what -if scenarios, and just to draw up my schematic to study it.

    I try to prototype my circuits as much as possible and learn through trial and error, and I now keep notes on my computer of the whole procedures throughout the design.
    As well as I don't throw my scratch work out anymore I keep all scratch calculations in a seperate notebook, this realy helps in learning design work.

    So this circuit was breadboarded, and yes it runs slow, the top right transistor was hot to the touch.

    I wanted to use 8volts. so I can use a 9 volt battery.
    But if I need to then I will raise the supply voltage.

    One thing Ive learned so far just from the responses in this thread is I need to adjust the values of the pull down and up resistors, to avoid oerheating, due to delay switching.

    That seems to be the biggest issue to resolve first.

    Sgt wookie, thanks for that schem. I will look at it and learn from it, but I still need to learn through trial and error, by taking the concept of that circuit, and trying to come up with my own design, to achieve the same goal, I appreciate the work you did on that,
    but please understand I just can't take a ready made circuit and implement it into my project without first knowing why each component is used and the overall function of the circuit.

    Then when I understand it, than I like to try to come up with my own design that fits my way of designing a circuit.
    I'm sure you know what I mean.

    Also thankyou both of you guys about giving me some info about a P- channel mosfet, likee I said I will study this thread, and as I continue with this learning curve, I will post more results, to this thread,

    If theres more info that can be helpful please post it,
    because this thread can be helpful to other people who want to learn about mosfet circuit design, too.

    Thanks again guys.


    Sgt. wookie explained it best.
  11. hobbyist

    Thread Starter Distinguished Member

    Aug 10, 2008
    I probably will end up using p-channell, but this is the fun of this excersise, if it can be done with some technique, as stated above, that to me is what this hobby is all about to me, I will first put forth my effort in trying to come up with a way to use all n-channels, if it's possible, again through trial and error.

    please understand I am not trying to waste anybodies time by going this rout, but this is not a needed project but a challenge in a circuit design.

    So I'm going to direct this thread to seeing if I can possibly come up with a technique to use all n-channell mosfets, in the proper way, if I fail than I will at least learn why I can't do it, but this is what this hobby is to me.

    Thanks again I appreciate all the effort you all put forth to help me with this, I just hope I don't seem like I'm wasting your time now that I'm going this rout with this project design.

    Thanks again.
    You all were a great help.
  12. SgtWookie


    Jul 17, 2007
    MOSFETS do have very high gate impedance. Once they have been charged/discharged to a certain voltage, it requires practically no current to keep them there. This property also makes them very vulnerable to static electricity.

    Low power enhanced FETs like the 2N7000 have a very small total gate charge, but they're also limited in current carrying capacity and voltage. The 2N7000 is very handy, as it can be controlled by logic level voltages at a reduced drain current (75mA), and is rated for up to 500mA, 60v.

    In the schematic I attached above, the key components are D1, R4, and Q4, along with (obviously) Q1. D1 provides a ready low-resistance charge path for the gate of Q1, if a 10v source were applied to the junction of D1 and R4.

    When that 10v source is removed, R4 pulls the base of Q4 to ground, turning it on, causing the gate to discharge through Q4. This happens quite quickly down to about 0.7v, which is well below the threshold voltage of most power MOSFETs.

    Q2 and Q3 merely convert a high-impedance logic-level control input (0v-5v) to 0v-12v at low impedance.
  13. hobbyist

    Thread Starter Distinguished Member

    Aug 10, 2008
    Thanks, that pretty much confirms I do understand that circuit pretty well.

    When I first seen it, I realized that Q3, was acting as a pull up voltage at the gate, and Q4 was the pull down voltage.

    The diode can be used to isolate R4 from the gate, R4 is needed to give Q3 it's proper bias at the collector, and Q2 deals with input signal.

    Whoops I just read your post again, R4 gives base bias to Q4.
    Q3 pulls positive at the base of Q4 also which turns off Q4.
    I see that clearly.

    Of course not as technical as you explained it, but basic point of view.

    see I'm gettin there,, slowly but surely.

    I have quite a few of the 2N7000, that I'm breadboarding to be used as switches across my capacitors for the timing circuits used to drive my H-bridge.

    Thanks again..
    Last edited: Sep 15, 2009
  14. hobbyist

    Thread Starter Distinguished Member

    Aug 10, 2008
    Hi guys.
    You guys are great
    thanks for all the help.

    please hold off on any suggestions or schematics, at the moment,
    I'm very close to coming up with a scheme. using only a single power supply to be able to utilize all N-chanl. mosfets in a motor driving bridge.

    This is just for the challenge (learning) of it.

    When I build and test it I'll post results .

  15. hobbyist

    Thread Starter Distinguished Member

    Aug 10, 2008
    Ok the couple of methods I came up with will use a lot of battery power, simulation wise they work find, but not practical for a circuit to design and build, so the other alternative, which is the way I originally planned to build it as, is to use 2 seperate supplies, one for the logic, and the other just for the motor and h-bridge.

    The reason I ran into problems with this in the first place is what was already mentioned about to high of pull resistors, and such, and the scheme. sgt. wookie gave will take care of that, now that I have a good grasp on it's function,
    The other major problem with the original BREADBORDED circuit was I have six AA batteries for the motor and bridge, (high current) and a rechargeable 9 volt battery for the logic.

    The 9volt battery was reading 7.7V. and the 9volt AA batteries were reading well over 8 volts close to 9v.

    So already there is a mismatch on the voltage sources.

    Now my simulation shows using 2 voltage sources that a double in voltage at the gate, with respect to the drain voltage source will result at around 54mV ,Vds when a grounded load is present at its source term. of 10 ohms.

    This is just a 1 mosfet simulation, (No bridge), if put into the bridge I would have to take into account all volt. drops from ground through bottom mosfet and load.

    However I will look into p-chanl. or use one voltage source and build 2 regulators one at 5V. (motor) and another at 10V. for the logic.

    That's all for now.

    Thanks again for everyones help..