Scaling up this h-bridge optoisolated parts help

Discussion in 'The Projects Forum' started by Steelmesh, May 14, 2012.

  1. Steelmesh

    Thread Starter Member

    Aug 25, 2009
    18
    0
    I found this circuit that includes a couple optoisolator's and no way to create a shoot-through condition:

    http://www.bristolwatch.com/ele/img/h-bridge.jpg

    Here are the MOSFETs I am looking at to build this circuit for higher amps:

    P-channel = URF5305PBF

    N-channel = HUF75321P3

    Opto = 4N25

    It will be used to drive this 24v 750w brushed DC motor via Arduino pwm controller.

    Will those components work? What size cap across the brushed DC motor? Along with that 22uf near the main source, should I add a bigger cap like a 1000uf?

    Thanks in Advance.
     
  2. praondevou

    AAC Fanatic!

    Jul 9, 2011
    2,936
    488
    I wouldn't be so sure that a shot-through condition is impossible with this circuit. Remember that even simultaneous conduction for a few tens or hundreds of nanoseconds is a shot-through.

    You'd be better off using separate gate resistor/diode combinations for each gate. Or even better RC combinations and then a schmitt trigger driver.

    The electrolytic capacitors should be at least 3300uF, mounted as near as possible to the bridge.

    Use short traces for gate and return signals.

    What frequency is the PWM?

    Charging and discharging the gates through a 10k resistor and the 4N25 is a bad way to do it if the frequency is high, your MOSFETs will heat up very much. The best is actually a dedicated gate driver like the HCPL-3180 (just an example)
     
  3. Steelmesh

    Thread Starter Member

    Aug 25, 2009
    18
    0
    I am thinking it will be very low frequency...100Hz?

    Let me verify your advice on the better path, get one of those pre-made gate drivers so I can just use the pwm signal straight from Arduino, no gate-resistors needed?

    10k resistors staying?
     
  4. mcasale

    Member

    Jul 18, 2011
    210
    12
    praondevou is correct. You do not need the R1 and R2 pull-up resistors if you are going to use a gate driver, which usually drives rail-to-rail. Gate resistors will slow down the edges of the gate pulses and cause the FETs to heat up.

    Drive each gate separately, and keep the runs as short as possible.

    Usually, a small resistance, like 10 to 50 Ohms is put in series with the gate to prevent big current spikes on your power supply. Go as low as your supply can handle.

    Let us know how you make out.
     
  5. Steelmesh

    Thread Starter Member

    Aug 25, 2009
    18
    0
    Is this what you mean? 4 digital inputs from micro? Or just splice each pair of optos for 2 digital inputs?

    [​IMG]
     
  6. Steelmesh

    Thread Starter Member

    Aug 25, 2009
    18
    0
    TTT

    Opps, CORRECTION, I think the Nfet gates need +20v through the opto, and keep the Pfet gates going to ground.
     
  7. praondevou

    AAC Fanatic!

    Jul 9, 2011
    2,936
    488
    The circuit from post #5 is missing pull/down resistors/voltage dividers.

    The p-FET gate voltage should be 24V to turn it off and 24V-15V (9V) to turn it ON. The n-FET gate voltage should be 15V or 0V to turn it on and off, respectively.

    If you use the DC-bus voltage (which is 24V) to drive the gate you need AT LEAST voltage dividers with two resistors.

    Don't drive the FETs with 24V as you may destroy them (depends on which you will be using)
     
  8. Steelmesh

    Thread Starter Member

    Aug 25, 2009
    18
    0
    - R1 is the Motor
    - Ignore Opto part# should be like a "HCPL-3180"
    - I put some anti-shoot-through relays in
    - Use a couple of these dc-switching regs for the 20v and 15v signals for the fets: http://search.digikey.com/us/en/products/L5973ADTR/497-5328-2-ND/1154573

    Revision 3:
    [​IMG]
     
  9. praondevou

    AAC Fanatic!

    Jul 9, 2011
    2,936
    488
    Anti-shot-through relays? Forget this, they bounce, they are slow, they simply have nothing to do in a PWM driven fullbridge circuit.

    1. How many separated 15V / 20V power supplies do you have?
    2. Does it have to be 2x PMOS, 2x NMOS or can it be 4x NMOS?

    The two lower NMOS can be driven by the same power supply for the gate signal.

    Did you have a look at the IR2110? Works quite well.
     
  10. Steelmesh

    Thread Starter Member

    Aug 25, 2009
    18
    0
    The relay idea was based on my assumption that they would be very low cycle, like 1 relay would be statically -off- during use and the other 1 would be statically -on- when you're in forward drive, I didn't know they would be cycled on and off a lot (I'm confused now). Not sure if I will ever use reverse drive, but cost $1 to do it so may as well via h-bridge. Another factor I was assuming that when I fire up the Arduino it may send out erroneous signals so I wanted a way ensure no screwups since the Pfets are normal closed; dont I want a battery disconnect to turn off main power when I initialize the arduino, then turn the battery on? *I think I was trying to accomplish a battery disconnect along with bonus feature to prevent shoot-through.

    NMOS vs. PMOS
    I thought I needed to use them as either sourcing or sinking...how can I use an NMOS on the positive side of the motor?

    15v and 20v power supplies...I was just thinking 1-each to be shared.
     
  11. praondevou

    AAC Fanatic!

    Jul 9, 2011
    2,936
    488
    You said PWM of 100Hz. That's low for FETs but high for mechanical relays.

    If you want to be able change the motor speed through PWM and reverse the direction you will need the fullbridge. If you don't reverse the direction a simple transistor could also do it.

    something like this with 2x IR2110 would be more suitable:

    [​IMG]

    You can isolate the signals "input 1" and "input 2" if you want.
    There is only one power supply 15V needed for this IC. The upper NFET's drivers receive their power supply voltage from the 15V as well (bootstrap).

    The reason why I asked what PWM frequency you want to use is that you where trying to charge/discharge the FETs gates through high resistances. This means you increase switching losses in the transistors. It may be possible to use simple optocouplers but then the switching frequency would have to be quite low. This however doesn't work well if you want to smoothly change the motor speed through PWM modulation.
     
    • ir.JPG
      ir.JPG
      File size:
      21.3 KB
      Views:
      414
  12. shortbus

    AAC Fanatic!

    Sep 30, 2009
    4,000
    1,512
    Also, since the ir2110 (or similar drivers) are "isolated" in the driver itself, an opto is not necessary.
     
    Steelmesh likes this.
  13. Steelmesh

    Thread Starter Member

    Aug 25, 2009
    18
    0
    Since I don't have Private Messaging unlocked in this forum...THANK YOU for helping me with this.

    I am still lost in the thought that I have to cycle the mechanical relays; in my design I am using a full-bridge, however I am simply disconnecting it one-way so the motor will only run one-way...it would be like me actually disconnecting a wire so that there is no path to ground, only through the motor. I am not trying to argue to put it in the design, I am just trying to turn the light bulb on in my head.
     
  14. Steelmesh

    Thread Starter Member

    Aug 25, 2009
    18
    0
    Parts I should get:
    1N4007-E3/54GICT-NDDIODE GP 1A 1000V DO41 $1.72
    IR2110-2PBF-NDIC DRIVER HIGH/LOW SIDE 14-DIP $14.54
    1N4148VSCT-NDDIODE SWITCH SS 75V 300MA DO35 $1.38
    22DVCT-NDRES FUSE METAL 22 OHM 1/2W 5% $0.64
    445-2903-NDCAP CER 10UF 50V RADIAL $4.38

    I guessing I don't need Pfets, so I can still use the HUF75321P3 Nfets, I have 4 of those already =)
     
  15. praondevou

    AAC Fanatic!

    Jul 9, 2011
    2,936
    488
    $14 for a 2110?

    try futureelectronics.com, even their shipping is cheaper:
    http://www.futureelectronics.com/en/CSRichProduct.aspx?ItemNumber=9962770&IM=0&Language=en-US

    I suggest using the UF400x instead of the 1N400x... It doesn't need to be the 4007, you are working with 24V.

    If you can use a power supply with a current limit, otherwise you will probably burn some of your MOSFETs in your first tests.

    Read and apply the informations from this app note, you will avoid many possible errors regarding layout and component selection: http://www.irf.com/technical-info/appnotes/an-978.pdf
     
  16. Steelmesh

    Thread Starter Member

    Aug 25, 2009
    18
    0
    Last edited: Jun 12, 2012
  17. shortbus

    AAC Fanatic!

    Sep 30, 2009
    4,000
    1,512
    No, the Hin and Lin is Vss-0.3 and Vdd+0.3 volts. then you look to see what Vss and Vdd values are.
     
  18. Steelmesh

    Thread Starter Member

    Aug 25, 2009
    18
    0
    I tore apart google last night, and there were people using ir2110 running 170vdc through fets, 12vdv at VCC, and 5v / -5v logic for inputs.

    But what datasheet tells us, is...if VDD is +15vdc then HIN logic is +15.3vdc, and if VSS is -15vdc, then LIN logic is -15.3vdc. And how about VCC =( Something is not clicking, I am no EE.

    Do I need another circuit to amplify the pwm from mcu? Also, do I need an inverter for the LIN signal?

    Am what I doing very similar to an inverter build? Seems like thats where all the google action is...

    I am just trying to get this pretty close before I hook any power to it. I also added 1k pull-down resistors to each fet, g to s.

    Posts I dug deep for:

    http://www.edaboard.com/thread219997-4.html

    http://www.edaboard.com/thread239471.html#post1024529

    http://4hv.org/e107_plugins/forum/forum_viewtopic.php?131378

    http://www.circuit-projects.com/fp/motor-driver/schematic.jpg

    http://electronics.stackexchange.com/questions/18423/bootstrap-capacitor-selection-with-ir2110-3

    http://www.diyaudio.com/forums/class-d/113532-ir2110-output-problems.html

    http://www.physicsforums.com/showthread.php?t=564463

    http://forum.allaboutcircuits.com/showthread.php?t=36971

    http://www.electro-tech-online.com/...sign-ideas-reviews/97093-problems-ir2110.html
     
    Last edited: Jun 13, 2012
  19. Steelmesh

    Thread Starter Member

    Aug 25, 2009
    18
    0
    You mentioned I don't need a whole lot of stuff, just a simple transistor for 1-way motor. I wanted to learn full-bridge so that I can then integrate regeneration, but controlled by mcu. I would imagine additional circuit after low fets to redirect generated power to another load or battery. Have you ever imagined an H-bridge with regeneration?
     
  20. shortbus

    AAC Fanatic!

    Sep 30, 2009
    4,000
    1,512
    Quote "But what data sheet tells us, is...if VDD is +15vdc then HIN logic is +15.3vdc, and if VSS is -15vdc, then LIN logic is -15.3vdc. And how about VCC =( Something is not clicking, I am no EE."

    No, what that means is the MOST you can put on Hin is +0.3 V over the voltage of Vdd. Or for Lin -0.3V of Vss.


    Don't worry about the charts in the data sheet. All you really need is listed in the heading called "features" on the first page. Items at dots number- 2-4-5.
     
Loading...